JOUR. r—+J>—i R C 280 U8 N38 1996 PUBL THE NATIONAL CANCER INSTITUTE — National Institutes of Health Consensus 1996 Conference on Cervical Cancer Number 21 Contents Consensus Statement vii National Institutes of Health Consensus Development Panel Epidemiology of Cervical Cancer 1 Rolando Herrero Screening Patterns for Cervical Cancer: How Best to Reach the Unscreened Population 7 Carol L. Brown The Bethesda System, the Pathology of Preinvasive Lesions, and Screening Technology 13 Michael Henry Cervical Human Papillomavirus Infection and lntraepithelial Neoplasia: a Review 17 Michele Follen Mitchell, Guillermo Tortolero-Luna. Thomas Wright, Asis Sarkar, Rebecca Richards-Kortum, Waun K. Hong, David Schottenfeld Histopathologic Prognostic Factors in Early Stage Cervical Carcinoma 27 Jo Ann Benda Integration of Diagnostic Imaging in the Clinical Management of Cervical Cancer 35 Anthony H. Russell What Is the Appropriate Management of Early Stage Cervical Cancer (International 43 Federation of Gynecology and Obstetrics Stages I and HA), Surgical Assessment of Lymph Nodes, and Role of Therapeutic Resection of Lymph Nodes Involved With Cancer? David H. Moore, Frederick B. Stehman Management of Stage IA Cervical Carcinoma 47 Mitchell Morris Primary Surgery for Stage IB-IIA Cervical Cancer, Including Short-Term and Long-Term 53 Morbidity and Treatment in Pregnancy Javier F. Magrina Primary Radiotherapy for Stage 13 or IIA Cervical Cancer 6| Perry W. Grigsby Stress and Quality of Life Following Cervical Cancer 65 Barbara L. Andersen New Surgical Approaches to Treatment of Cervical Cancer 71 Kenneth D. Hatch. Alton V. Hallum III. Magdy Nour Adjuvant Therapy After Primary Surgery for Stage I-IIA Carcinoma of the Cervix 77 Gillian M. Thomas Adjuvant Surgery After Radiotherapy 85 Patricia J. Eifel Optimal Management of Locally Advanced Cervical Carcinoma 89 Henry Keys. Susan K. Gibbons Neoadjuvant Chemotherapy Before Surgery in Cervical Cancer 93 Gary Lee Eddy. Sr. Concomitant and Neoadjuvant Chemotherapy in Conjunction With Radiotherapy in the 10! Management of Locally Advanced Cervical Cancer Martin H. N. Tattersall ( C omen/s continued (m back cover) Become a National Cancer Institute Information Associate and receive one copy of this and each upcoming Monograph at no additional charge! As an Information Associate, you customize your own package of benefits to meet your individual needs. In addition to Journal Monographs, you can receive the distinguished Journal of the National Cancer Institute (available by subscription exclusively to members) . . . communicate with your peers in the oncology community through our dial-up Bulletin Board System . . . or search PDQ, the National Cancer Institute’s comprehensive cancer information database, via Internet. Best of all, the cost of becoming an Information Associate is surprisingly modest —just $100 for US. members and $150 for non—US. members per year. And if you’re not completely satisfied, we’ll refund your money. To find out more about NCl’s Information Associates Program, simply complete the form below and return it to NCI today. Or, tojoin today, call l-800-NCl-7890 within the United States or 301-496-7600 outside the United States. i: ngUSC EiKE YES! I want to learn how I can access National Cancer Institute ’5 network of oncology information resources. ASSOCIATES Name []Dr.[ ]Mr.[ ]Ms.[ ]Mrs. Organization Phone Fax 13-mail Return this form to NCI Information Associates Program, 9030 Old Georgetown Road, Bethesda, MD 20814—1519, USA, or fax to 301—231-6041. Number 21 ISSN1052—6773 Richard D. Klausner MONO GRAPHS JOURNAL OF THE NATIONAL CANCER INSTITUTE Director, National Cancer Institute Susan Molloy Hubbard Director, International Cancer Information Center Julianne Chappell Chief, Scientific Publications Branch EDITORIAL BOARD Barnett S. Kramer Editor-in-Chief J. Gordon McVie European Editor Eric J. Seifter Book Reviews Editor J. Paul Van Nevel News Editor Frederic J. Kaye Douglas L. Weed Reviews Editors Martin L. Brown Economics Editor ASSOCIATE EDITORS Susan G. Arbuck Frank M. Balis William J. Blot Peter M. Blumberg John D. Boice. Jr. Louise A. Brinton Bruce A. Chabner Ross C. Donehower Susan S. Ellenberg Suzanne W. Fletcher Michael A. Friedman John K. Gohagan Frank J. Gonzalez Michael M. Gottesman Peter Greenwald Donald E. Henson Susan M. Hubbard Frederic J. Kaye Hynda K. Kleinman EDITORIAL ADVISORY BOARD James O. Annitage Bruce C. Baguley Laurence H. Baker William T. Beck Clara D Bloomfield Benjamin Bonavida George J. Bosl C. Noman Coleman 0. Michael Colvin Thomas H. Corbett Pelayo Correa Stephen P. Creekmorc Johanna T. Dwyer Merrill J. Egorin Soldano Ferronc lsaiahJ Fidler Richard 1. Fisher David FitzGerald Dystein Fodstad Antonio Fojo Lori J. Goldstein Harvey M. Golomb Elieser Gorelik Jean L. Grem Amold H. Greenberg Maureen M. Henderson Gloria H. Heppner Ronald B. Herberrnan Waun Ki Hong William J. Hoskins Alan N. Houghton James N, Ingle David H. Johnson V. Craig Jordan John S. Kovach Margaret L. Kripke Theodore S. Lawrence W. Marston Linehan Marc E. Lippman Scott M. Lippman Dan L. Longo Douglas R. Lowy Susan G. Nayfield David L. Nelson Kenneth Olden Drew M. Pardoll Ross L. Prentice Alan S. Rabson Robert H. Shoemaker Richard M. Simon Michael B. Spom Maryalice Stetler—Stevenson J. Paul Van Nevel Douglas L. Weed Noel S. Weiss Mark G. Kris Donald W. Kute Bernard Levin Brian R. Leyland—Jones Allen S. Lichter Guy McClung Frank L. Meyskens Anthony B. Miller Malcolm S. Mitchell James J. Mule' C. Kent Osbome John J. O’Shea David M. Ota David F Paulson Henry C. Pitot Igor B. Roninson Edward A. Sausville ThomasJ Sayers 1996 STATISTICAL EDITORS Janet W. Andersen Donald A. Berry Barry W. Brown Bernard F. Cole Susan S. Ellenberg Scott S. Emerson Eric Feuer Laurence S. Freedman Edmund A. Gehan Sylvan B. Green Susan G. Groshen Richard A. Olshen Philip C. l’rorok Philip S. Rosenberg Harland N. Sather Daniel .I. Schaid Richard M. Simon Donald M. Stablein Robert E. 'l‘arone Peter F. Thall David Sehottenlbld Herman A. J. Schut Richard K. Severson William R Shapiro Roy E. Shore Paul M Sondel Patricia S Steeg Herman D. Suit Sandra M. Swain Mario Sznol Raymond 'l‘aetle Ian Tannock Joel E. Tepper J. 'l‘atc ’l'higpen Peter R Twentyman Larry M. Weisenthal EDITORIAL STAFF Scientific Editor: Edward Trimble. MD. Monograph Coordinator: Elaine Price Beck Manuscript Editors: Elaine Price Beck Joan O’Brien Rodriguez Editorial Assistant: Jamia V. Capehan MARKETING STAFF Marketing Director: Jean Griffin Baum Press Contact: Linda Anderson EDITORIAL POLICY: Manuscripts from key conferences dealing with cancer and close— ly related research fields. or a related group of papers on specific subjects of importance to cancer re- search. are considered for publication. with the understanding that they have not been published previously and are submitted exclusively to the Journal oft/w National Cancer Institute Monographs. All material submitted for consideration will be subject to review. when appropriate. by at least one outside reviewer and one member of the Editorial Board of the Journal oft/1e National Cancer Institute. Opinions expressed by the authors are not necessarily those of the publisher or the editors. Proposals for monographs should be submitted to the Editor-in-Chief. Journal oft/1e National Can- car Institute. National Cancer Institute. 9030 Old Georgetown Rd.. Bethesda. MD 20814. Journal of the National Cancer Institute Monographs are available on request to members of the National Cancer Institute Information Associates Program (one copy ofeach monograph per mem- ber). Monographs are also available through the US. Government Printing Office. To request a copy ofa monograph. or for more information about the Information Associates Program. call 1-800-624- 7890 (301-496-7600 outside the United States). National Institutes of Health Consensus Conference on Cervical Cancer Proceedings of a Conference Held at the National Institutes of Health Bethesda, Maryland April 1—3, 1996 Conference Sponsors Office of Medical Applications of Research. NIH John H. Ferguson. MD. Director National Cancer Institute Richard D. Klausner. MD. Director Conference Cosponsors National Institute of Nursing Research Patricia A. Brady. RN. Ph.D. Director National Institute of Allergy and Infectious Diseases Anthony S. Fauci. MD. Director Office of Research on Minority Health, NIH John Ruffin. Ph.D. Associate Director Office of Research on Women’s Health, NIH Vivian W. Pinn. MD. Director Centers for Disease Control and Prevention David Satcher. M.D.. Ph.D. Director para, National Institutes of Health Consensus I 996 Conference on Cervical Cancer Number 21 Contents Consensus Statement National Institutes of Health Consensus Development Panel vii Epidemiology of Cervical Cancer 4 ~, Rolando Herrero I Screening Patterns for Cervical Cancer: How Best to Reach the Unscreened Population Carol L. Brown / I. The Bethesda System, the Pathology of Preinvasive Lesions, and Screening Technology at ‘ ' ,, Michael Henry Cervical Human Papillomavirus Infection and Intraepithelial Neoplasia: a Review Michele Follen Mitchell, Guillermo Tortolero-Luna, Thomas Wright. Asis Sarkar, Rebecca Richards—Kortum, Waun K. Hong, David Schottenfeld Histopathologic Prognostic Factors in Early Stage Cervical Carcinoma Jo Ann Benda Integration of Diagnostic Imaging in the Clinical Management of Cervical Cancer Anthony H. Russell What Is the Appropriate Management of Early Stage Cervical Cancer (International Federation of Gynecology and Obstetrics Stages I and HA), Surgical Assessment of Lymph Nodes, and Role of Therapeutic Resection of Lymph Nodes Involved With Cancer? David H. Moore, Frederick B. Stehman Management of Stage IA Cervical Carcinoma Mitchell Morris Primary Surgery for Stage IB-IIA Cervical Cancer, Including Short-Term and Long-Term Morbidity and Treatment in Pregnancy Javier F. Magrina Primary Radiotherapy for Stage IE or IIA Cervical Cancer Perry W. Grigsby Stress and Quality of Life Following Cervical Cancer Barbara L. Andersen New Surgical Approaches to Treatment of Cervical Cancer Kenneth D. Hatch, Alton V. Hallum III, Magdy Nour Adjuvant Therapy After Primary Surgery for Stage I-IIA Carcinoma of the Cervix Gillian M. Thomas Adjuvant Surgery After Radiotherapy Patricia J. Eifel Optimal Management of Locally Advanced Cervical Carcinoma Henry Keys, Susan K. Gibbons Neoadjuvant Chemotherapy Before Surgery in Cervical Cancer Gary Lee Eddy, Sr. Concomitant and Neoadjuvant Chemotherapy in Conjunction With Radiotherapy in the Management of Locally Advanced Cervical Cancer Martin H. N. Tattersall Is There a Radiobiologic Basis for Improving the Treatment of Advanced Stage Cervical Cancer? Jacob C. Lindegaard, Jens Overgaard, Soren M. Bentzen, Dorte Pedersen 27 35 43 47 53 61 65 71 77 85 89 93 101 105 Radiotherapy for the Treatment of Locally Recurrent Cervical Cancer Rachelle Laneiano Surgery for the Treatment of Locally Recurrent Disease Manuel A. Penalver. Giselle Barreau. Bernd—Uwe Sevin. Hervy E. Avereue Chemotherapy for Stage lVB or Recurrent Cancer of the Uterine Cervix George A. Omura Radiation l’alliation of Cervical Cancer William J. Spanos, Jr.. Thomas J. Pajak. Buhman Emami. Philip Rubin. Jay S. Cooper. Anthony H. Russell. James D. Cox New Directions for Radiation Biology Research in Cancer of the Uterine Cervix Daniel S. Kapp. Amalo J. Giaccia Human Papillomavirus Biology Douglas R. Lowy. John T. Schiller Human l’apillomavirus Immunology and Vaccine Prospects Michael A. Steller. John T. Schiller ll7 l23 [27 131 141 145 National Institutes of Health Consensus Development Conference Statement: Cervical Cancer, April 1-3, 1996 National Institutes ()fHealth Consensus Development Panel Objective: The objective was to provide physicians and the general public with a responsible assessment of current screening, prevention, and treatment approaches to cervical cancer. Participants: A non-Federal, nonadvocate, l3-mem- her panel representing the fields of obstetrics and gynecol- ogy, gynecologic oncology, radiation oncology, and epidemiology participated. In addition, 28 experts in obstetrics and gynecology, gynecologic oncology, radiation oncology, gynecologic surgery, and psychology presented data to the panel and a conference audience of 500. Evidence: The literature was searched through Medline, and an extensive bibliography of references was provided to the panel and the conference audience. Experts prepared abstracts with relevant citations from the literature. Scien- tific evidence was given precedence over clinical anecdotal experience. Consensus process: The panel, answering predefined questions, developed their conclusions based on the scientific evidence presented in open forum and the scientific literature. The panel composed a draft statement that was read in its entirety and circulated to the experts and the audience for comment. Thereafter, the panel resolved conflicting recommendations and released a revised statement at the end of the conference. The panel finalized the revisions within a few weeks after the conference. Con- clusions: Carcinoma of the cervix is causally related to infec- tion with the human papillomavirus (HPV). Reducing the rate of HPV infection by changes in sexual behaviors in young people and/or through the development of an effective HPV vaccine would reduce the incidence of this disease. Papanicolaou smear screening remains the best available method of reducing the incidence of and mortality from in- vasive cervical cancer. Persons with stage [A] disease have a high cure rate with either simple hysterectomy or, where fertility preservation is an issue, by cone biopsy with clear margins. For patients with other stage I or stage IIA disease, radical surgery or radiation treatment is equally effective. These patients should be carefully selected to receive one treatment or the other but not both, because their combined use substantially increases the cost of and morbidity from treatment. Women with more advanced, nonmetastatic dis- ease should be treated with radiation. Recurrent cervical cancer confined to the pelvis should be treated with the modality not previously received. Radiation therapy is recommended to palliate symptoms in patients with meta- static disease. [Monogr Natl Cancer Inst l996;21:vii-xix] Journal of the National Cancer Institute Monographs No. 21. 1996 Introduction Carcinoma of the cervix is one of the most common cancers in women and accounts for 15 700 new cases (6% of all can- cers) and 4900 deaths in the United States each year. World— wide. cervical cancer is second only to breast cancer as the most common cancer in both incidence and mortality. More than 471 ()00 new cases are diagnosed each year. predominantly among the economically disadvantaged. in both developing and industrialized nations. During the last 50 years in the United States. the utilization of screening programs based on the Papanicolaou (Pap) smear and pelvic examination has led to a steep decline in incidence ofand deaths from cervical cancer. Both invasive cervical cancers and precursor lesions have been firmly associated with the presence of human papil— lomavirus (HPV) DNA. It has also been well established that the majority of squamous cell cancers of the cervix progress through a series of well—defined preinvasive lesions and that, during this usually lengthy process. the disease can be easily detected by Pap smear screening. During this preinvasive stage. cervical squamous intraepithelial lesions (SlLs) can be con- trolled with nearly uniform success and with the retention of fer- tility. Many treatment and quality-of—life issues remain unresolved for women with cervical cancer. For women with early stage disease. key issues include determining guidelines for the extent of treatment. the pathologic and clinical indicators for the inten— sity of therapy. and the selection of a treatment modality among several competing options. For women with advanced stage dis- ease. critical issues include optimal radiotherapy techniques, whether chemotherapy or combined modality regimens improve outcome, the morbidity and benefit of salvage therapy for recur- rent disease. and palliative treatment. Additional topics include advances in screening technology, the implementation of the Bethesda System for Pap smears. the role of HPV testing and subtyping. treatment selection for patients with preinvasive dis— ease. advances in laparoscopic surgical staging and therapy techniques, and the application of newer imaging techniques such as magnetic resonance imaging (MRI). Prospects for both prophylactic and therapeutic vaccines against HPV offer hope for fundamental alterations in the prevention and management of this disease. See “Note" section following “References." vii To address these and related issues. the National Cancer In— stitute and the National Institutes of Health (NIH) Office of Medical Applications of Research convened a Consensus Development Conference on Cervical Cancer.I The conference was cosponsored by the National Institute of Nursing Research, the National Institute of Allergy and Infectious Diseases, the Of- fice of Research on Minority Health, the Office of Research on Women's Health of NIH, and the Centers for Disease Control and Prevention. After 11/2 days of presentations and audience discussion, an independent, non-Federal consensus panel weighed the scien— tific evidence and developed a draft statement that addressed the following key questions: 0 How can we strengthen efforts to prevent cervical cancer? 0 What is the appropriate management of low stage cervical cancer (FIGO [International Federation of Gynecology and Obstetrics] stages I-IIA)? - What is the appropriate management of advanced stage and recurrent cervical cancer? ' What are new directions for research in cervical cancer? 1) How Can We Strengthen Efforts to Prevent Cervical Cancer? A strong causal relationship has been established between HPV and cervical cancer and its precursors. The evidence for this statement is as follows: . HPV DNA is present in virtually all cases (93%) of cervical cancer and its precursor lesions. 0 Multiple epidemiologic studies indicate that HPV infection is the major risk factor for SILs and in- vasive cervical carcinoma. 0 Studies have demonstrated that the HPV genes E6 and E7 are integrated into the host genome and that the transforming proteins encoded by these genes are tumorigenic. More than 70 types of HPV have been identified. However, only 23 of these infect the uterine cervix; of these, only one half are associated with SILs or invasive cervical cancer. These are further classified into low—risk types, HPV types 6 and 11, and high—risk types, most commonly HPV types l6, 18, 31, and 45, which account for more than 80% of all invasive cervical can— cers. An unknown percentage of women infected with HPV will develop either low—grade SIL or high-grade SIL. One third of all grades of SILs will regress, whereas 41% persist and 25% progress. Of lesions that progress, 10% progress to carcinoma in situ and 1% to invasive cancer. Three quarters of all grades of SlLs will not progress. This virus is transmitted through sexual intercourse, with a peak prevalence of infection in women in the age group 22-25 years. The prevalence of infection decreases with increasing age, suggesting that most infections in women and men resolve over time through host immune responses. Epidemiologic studies are now focusing on cofactors and host factors that may explain the natural history of HPV infections viii and their associated lesions. Factors under investigation include smoking, use of hormonal contraceptives, number of live births, young age at first sexual intercourse, use of vitamins such as carotenoids, vitamin C, and folic acid, co-infection with other sexually transmitted diseases (e.g., herpes simplex, human im- munodeficiency virus [HIV], and chlamydia), growth factors, cytokines, and humoral and cellular immunity. Screening Squamous cell cervical cancer is an ideal disease for screen— ing because of the typically long preclinical phase, which per- mits early detection. Use of the Pap smear is effective in reducing morbidity and mortality from cervical cancer. Despite the recognized benefits of Pap smear screening. substantial sub- groups of US. women have not been screened or are not screened at regular intervals. One half of the women with newly diagnosed invasive cervical carcinoma have never had a Pap smear, and another 10% have not had a smear in the past 5 years. The unscreened populations include older women, the unin- sured, ethnic minorities (especially Hispanics and elderly blacks), and poor women (particularly those in rural areas). One fourth of the cases of cervical cancer and 41% of the deaths occur in women aged 65 years and older. Data from the 1992 National Health Interview Survey indicate that one half of all women aged 60 years and older have not had a Pap smear in the past 3 years. Although older women are screened less frequent- ly, they have the same number of recent physician visits as younger women, which indicates the need to educate older women and their health care providers about the importance of Pap smear screening. For patients who are not involved in routine screening programs, any health care encounter should be an opportunity to obtain a Pap smear and offer other screening modalities. On the other hand, recent evidence demonstrates that the gap in the incidence of cervical cancer between black and white women under age 50 years is disappearing, suggesting that the rate of screening has increased among young black women. To improve outreach to unscreened populations, reasons for nonparticipation in screening must be determined and addressed with appropriate interventions. Community-based approaches to reaching diverse ethnic populations are recommended and should include using community leaders and members to assess attitudes and concerns prior to instituting screening programs and as part of the process of education and awareness. Cultural- ly sensitive and linguistically compatible staffing for outreach and screening is a key component. Logistical problems associated with screening in both metro- politan and rural settings should be addressed during outreach planning (e.g., transportation, child care, duration of appoint- ments, multiple site referrals, and accessible screening sites). Options such as mobile screening services and incentives should be considered. A concerted effort to standardize Pap smear terminology resulted in the Bethesda System (Table 1). This system evaluates the specimen for adequacy, uses diagnostic terminol- ogy, and makes recommendations pertaining to the smear when necessary. Determining the adequacy of the specimen is a major Journal of the National Cancer Institute Monographs No. 21, 1996 Table l. The 1991 Bethesda System* Adequacy ofthe specimen Satisfactory for evaluation Satisfactory for evaluation but limited by (specify reason) Unsatisfactory for evaluation (specify reason) General categorization (optional) Within normal limits Benign cellular changes; see “Descriptive diagnoses" Epithelial cell abnormality; see “Descriptive diagnoses“ Descriptive diagnoses Benign cellular changes Infection Trichmmmas vaginalis Fungal organisms morphologically consistent with Candida sp. Predominance of coccobacilli consistent with shift in vaginal flora Bacteria morphologically consistent with Animmrvrm sp. Cellular changes associated with herpes simplex virus Other Reactive changes Reactive cellular changes associated with Inflammation (includes typical repair) Atrophy with inflammation (“atrophic vaginitis") Radiation Intrauterine contraceptive device Other Epithelial cell abnormalities Squamous cell Atypical squamous cells of undetermined significance: qualify)“ Low-grade squamous intraepithelial lesion encompassing HPV: mild dysplasia/CIN l High—grade squamous intraepithelial lesion encompassing moderate and severe dysplasia, CIS/CIN 2, and CIN 3 Squamous cell carcinoma Glandular cell Endometrial cells, cytologically benign, in a postmenopausal woman Atypical glandular cells of undetermined significance: qualifyt Endocervical adenocarcinoma Endometrial adenocarcinoma Extrauterine adenocarcinoma Adenocarcinoma, not otherwise specified Other malignant neoplasms (specify) Hormonal evaluation (applies to vaginal smears only) Hormonal pattern compatible with age and history Hormonal pattern incompatible with age and history; specify Hormonal evaluation not possible because of. . . (specify) *CIS = carcinoma in situ; CIN = cervical intraepithelial neoplasm. ’rAtypical squamous or glandular cells of undetermined significance should be further qualified as to whether a reactive or a premalignant/malignant process is favored. iCellular changes of human papillomavirus (HPV) (previously termed “koilocytotic atypia“ or “condylomatous atypia") are included in the category of low-grade squamous intraepithelial lesion. contribution, because retrospective reviews of smears from women with cervical cancer have shown that many were un- satisfactory. Smears may be unsatisfactory for a variety of reasons, the most common of which are obscuring blood or inflammation. Evaluation of others may be less than optimal be- cause of factors such as absence of sampling from the trans- formation zone. Among the diagnostic terminologies are low—grade SIL and high-grade SIL. Another category of abnormal squamous cells is atypical squamous cells of undetermined significance (ASCUS). Management modalities for high—grade SIL are established and include colposcopy—directed biopsy and endocervical curettage, followed by conization with scalpel, cautery, laser, or loop electrocautery excision procedure. Management modalities for ASCUS and low-grade SIL are not as uniform. A large clinical Journal of the National Cancer Institute Monographs No. 21, 1996 trial is currently under way to determine whether HPV testing can effectively triage these patients, to develop clinical manage- ment guidelines and provide prognostic information. and to identify areas for cost reduction in screening and treatment. The glandular cell abnormalities are divided into two categories. atypical glandular cells of undetermined significance and adenocarcinoma. Methods of specimen acquisition, preparation. and evaluation of the Pap smear have changed little since its introduction in the 1940s. Although the Pap smear is highly effective in screening for preinvasive lesions of the cervix, a single test has a false- negative rate estimated to be 20%. One half of the false—nega— tives are due to inadequate specimen sampling. and the other half are attributed to a failure to identify the abnormal cells or to interpret them accurately. Pap smears should be obtained in con- junction with a pelvic examination. If a gross lesion is visual- ized, a biopsy of the lesion should be done, inasmuch as a Pap smear alone is inadequate in this situation. To improve the adequacy of the cervical smear specimen, a variety of sampling devices is available (e.g., spatula, endocer- vical brush, broom, and cotton swab). Liquid—based specimen collection methods are currently being evaluated to improve sampling and cell preservation and presentation. In the fall of 1995, the Food and Drug Administration ap- proved two automated instruments for rescreening smears evaluated as negative on the initial screen. Data from clinical tri— als suggest that these instruments could reduce the rate of false- negative smears. Neither the efficacy in routine practice nor the cost—benefit of these devices has been determined. In addition, these and other devices are being evaluated for use as primary screening instruments. In 1988, a group of experts recommended that annual Pap smears and pelvic examinations begin at onset of sexual activity or at age 18. After three consecutive examinations in women with normal findings, the interval between screenings may be increased at the discretion of the physician and the patient. In 1995, the American College of Obstetricians and Gynecologists (ACOG) recommended that patients with one or more risk fac- tors for cervical cancer (e.g., HIV or HPV infection, a history of low-grade SIL, high-risk behavior) be screened annually. Women older than the age of 65 years should continue to be screened. Prevention Primary prevention of HPV infection will require 1) directing education efforts toward adolescents and health care providers regarding the strong causal link between acquisition of HPV as a sexually transmitted disease and development of cervical can- cer and its precursors, 2) encouraging delayed onset of sexual intercourse, 3) developing an effective prophylactic vaccine, and 4) developing effective vaginal microbicides. The data on the use of barrier methods of contraception to prevent the spread of HPV are controversial but do not support this as an effective method of prevention. Secondary prevention efforts must focus on 1) developing ef- fective antiviral agents to treat HPV and/or prevent transforma- tion by E6/E7, 2) developing therapeutic vaccines to prevent HPV progression, 3) improving the sensitivity and specificity of screening for the precursors of cervical cancer. and 4) expand- ing education and screening programs to target underreached populations. 2) What Is the Appropriate Management of Low Stage Cervical Cancer (FIGO Stages I-IIA)? Table 2 lists the staging criteria based on the FIGO staging system for cervical cancer. The diagnosis of stage IAl cervical squamous cell carcinoma should be based on cone biopsy, not punch cervical biopsy, preferably by using a technique that does not result in cauterized margins. Systematic pathologic evaluation of the cone specimen is necessary. Where early invasion is identified. serial sections may be necessary to determine the extent of maximal depth of invasion. lateral involvement. and the presence of lymph vas— cular invasion. In patients with stage lAl cervical squamous cell carcinoma. simple hysterectomy or cone biopsy (with negative margins) is virtually 100% curative. The choice of therapy should be influenced by the patient‘s desire to preserve fertility. Although lymph vascular involvement is generally considered to be an adverse prognostic factor in cervical cancer. the prog- nostic significance of lymph vascular involvement in stage 1A1 cervical squamous cell carcinoma is uncertain. Because of this uncertainty. some clinicians have suggested that patients with lymph vascular involvement in stage 1A1 disease might be more appropriately treated with either radical hysterectomy or radia— tion therapy. At our present state of knowledge. a category of cervical adenocarcinoma that could be treated conservatively in order to preserve fertility cannot be identified. This needs further inves- tigation. Patients with stage IA2 lesions can be treated with primary radical or modified radical hysterectomy or primary radiation therapy with equivalent results. The choice of therapy should be influenced by such factors as ovarian preservation, comorbid Table 2. Criteria based on International Federation ol'Gynecology conditions, and potential late side effects. The availability of physicians with appropriate experience and training in gyne— cologic oncology procedures and physicians who are experts in radiation therapy should also influence the decision. Estimates of lymph node involvement in patients with stage IA2 disease range from 4% to 10%. Whether these patients should have lymph nodes treated with lyrnphadenectorny or radiation therapy should be determined on a case-by-case basis. Radiation therapy for stage 1A2 disease consists of intracavitary brachytherapy. Patients with stage IE or IIA cervical cancer are appropriately treated with either radical hysterectomy with pelvic lym- phadenectorny or radiation therapy. with equivalent results. The choice of therapy should be influenced by the same factors as described in patients with stage IA2 disease. To minimize mor- bidity. primary therapy should avoid the routine use of both radical surgery and radiation therapy. The combined use of radi- cal surgery and radiation therapy results in high morbidity and cost. Radiation therapy for stage 1B or 11A disease should con- sist of external—bearn therapy and brachyther'apy. For patients with stage [B or 11A disease. factors such as lymph node in- volvement. increasing lesion size, deeper stromal invasion. un— favorable histopathologic type. and lymph vascular involvement adversely affect prognosis. In patients with positive pelvic lymph nodes documented at radical hysterectomy. postoperative radiation therapy reduces pelvic recurrence but has not been proven to affect survival. Pel- vic radiation therapy may also be useful in reducing subsequent pelvic relapse in those patients with either close or involved sur- gical margins. The value of reducing pelvic recurrence is impor— tant to ameliorate or prevent pelvic pain and bleeding. even in those patients who may experience subsequent distant recur— rence. Patients who undergo simple hysterectomy for presumed benign disease and who are found to have invasive cervical can— cer (greater than stage 1A1) are considered candidates for post- operative radiation therapy or radical parametrectomy and lymphadenectomy. and Obstetrics (HGO) staging system for cervical cancer. 199-1 Stage Criteria 0 Carcinoma in situ. intraepithelial carcinoma. Cases of stage 0 should not be included in any therapeutic statistics for invasive carcinoma. 1 The carcinoma is strictly confined to the cervix (extension to the corpus should be disregarded). 1A Invasive cancer identified only microscopically. All gross lesions. even with superficial invasion. are stage IB cancers. Invasion is limited to measured stromal invasion with a maximum depth of 5 mm and no wider than 7 mm. (The depth of invasion should not he more than 5 mm taken from the base of the epithelium. either surface or glandular. from which it originates. Vascular space involvement. either venous or lymphatic. should not alter the staging. 1A1 Measured invasion of stroma no greater than 3 mm in depth and no wider than 7 mm. 1A2 Measured invasion of stroma greater than 3 mm and no greater than 5 mm in depth and no wider than 7 mm. [B Clinical lesions confined to the cervix or preclinical lesions greater than IA. 1B1 Clinical lesions no greater than 4 cm in si/.e, 1B2 Clinical lesions greater than 4 cm in size. 11 The carcinoma extends beyond the cervix. but it has not extended onto the pelvic wall; the carcinoma involves the vagina. but not as far as the lower third. 11A No obvious parametrial involvement. ”8 Obvious parametrial involvement. III The carcinoma has extended onto the pelvic wall; upon rectal examination. there is no cancer»free space between the tutnor and the pelvic vi all: the tumor involves the lower third of the vagina; all cases with a hydronephrosis or nonfunctioning kidney should be included. unless they are known to be due to other cause, IIIA No extension onto the pelvic wall. but involvement of the lower third of the vagina. IIIB Extension onto the pelvic wall or hydronephrosis or nonfunctioning kidney. IV The carcinoma has extended beyond the true pelvis or has clinically involved the mucosa of the bladder or rectum. IVA Spread of the growth to adjacent organs. IVB Spread to distant organs. x Journal ofthe National Cancer Institute Monographs No. 21. 1996 The optimal role for imaging studies to define the extent of disease at presentation as well as to plan radiation therapy needs further investigation. Modalities requiring further study include MRI. ultrasound. computed tomography (CT), and positron emission tomography (PET). Lymphangiography with fine— needle aspiration cytology has been demonstrated to be an ef- fective means of assessing lymph node status. Although lymphangiography is not widely available, it appears to be more effective in assessing pelvic and para-aortic lymph node status than both MRI and CT scanning. However, MRI and CT are useful in some patients to assess extent of disease and to select and plan optimal therapy. lf grossly involved pelvic lymph nodes are detected at the time of radical hysterectomy. data suggest that excision of the grossly involved lymph nodes results in improved local control. However, these findings should be confirmed in additional studies. Conflicting data exist on whether to remove the uterus or to leave it in place to assist brachytherapy. A subset of patients may exist who benefit from prophylactic or therapeutic para-aortic lymph node radiation. The greatest benefit is likely to be noted in patients who have bulky stage [B. ”A, or [1B carcinoma of the cervix with a high probability of control of disease in the pelvis and who have either no evidence of gross para—aortic disease or resected micrometastatic disease. One randomized study assessing the value of prophylactic para- aortic lymph node radiation showed survival benefit. In patients who receive primary radiation therapy for bulky stage IB cervical cancer. published results do not justify routine performance of postradiation hysterectomy. In patients who can receive optimal brachytherapy. combining hysterectomy with primary radiation therapy for bulky stage 18 tumors increases cost and morbidity without clear improvement in local tumor control. Stage for stage. the outcome of the treatment of pregnant patients with cervical cancer is similar to that of nonpregnant patients. Although data are limited. information suggests that patients with stage [A or small stage IB disease may have limited delays in therapy to allow fetal viability without serious- ly compromising patient survival. The safely of therapy delay for patients with bulky stage I lesions or more advanced stages has not yet been established. Neoadjuvant chemotherapy before planned radiation therapy in bulky stage IE or II disease has been investigated in several studies, but it has not been demonstrated to provide a benefit. Several prospective randomized trials evaluating neoadjuvant chemotherapy before planned radiation therapy were recently completed. and the results of these trials will be of interest. The potential usefulness of neoadjuvant chemotherapy before planned surgery is also under investigation. The measurement of serum tumor markers in patients with in— vasive cervical cancer has not been found to be of clear benefit but is worthy of further study. Estrogen replacement therapy may be safely prescribed to patients with invasive cervical can- cer regardless of the histology and stage of their cancer. Journal of the National Cancer Institute Monographs No. 2 l. I996 3) What Is the Appropriate Management of Advanced Stage and Recurrent Cervical Cancer? For advanced stage disease, FIGO stage MB or greater. the standard of care is primary radiation therapy using external- beam radiation and brachytherapy. With improved irradiation techniques. especially the increased use of brachytherapy, im— provement in tumor control and long-term survival has been noted in studies examining the patterns of practice in the United States since I973. Increased risk for therapeutic failure is most closely associated with patients who have a large volume of pri- mary tumor. bilateral parametrial disease. lymph node metas- tases, poor performance status, and low hemoglobin values. Retrospective studies have demonstrated increased risk for pel- vic recurrence when lower total doses of radiation and prolonged treatment schedules are employed. In patients with advanced disease. the prognostic effect of different histologic types is not clearly an independent prognostic variable. Optimal management utilizes megavoltage radiation energies and brachytherapy. Multiple field arrangements need to provide adequate coverage for the tumor volume. especially when lateral fields are used. The use of low—dose-rate (LDR) brachytherapy has been shown to significantly reduce the rate of local recur- rence in patients with advanced stage disease. LDR brachy- therapy is the most commonly used and most extensively defined technique in the United States, The use of high-dose— rate (HDR) brachytherapy has been increasing. although more studies are needed to define optimal fractionation schemes as well as long—term complications of this method. Pending further study, the use of HDR brachytherapy fractionation schemes that demonstrate rates of local control equivalent to LDR brachy- therapy techniques and use smaller dose per fraction is recom- mended to decrease the probability of long-term complications with HDR brachytherapy. Interstitial therapy has been used for unusual or difficult tumor geometry. Primary surgical therapy for advanced cervical cancer is limited to the management of some patients with stage lVA cancers who present with vesico— vaginal and/or rectovaginal fistulas. The utility of cytotoxic chemotherapy in patients with more advanced or recurrent disease has been investigated in clinical trials with a variety of study designs. These include neoadjuvant chemotherapy. administered prior to surgical resection or radia— tion therapy; concomitant chemotherapy and radiation therapy. in which both modalities are administered together; adjuvant chemotherapy, in which surgery or radiation therapy is followed by chemotherapy: and chemotherapy as sole treatment for patients with widely disseminated tumor not suitable for pallia- tive radiation. Cisplatin is the drug with the best documented single—agent activity in cervical cancer. with response rates of l8%—31% in multiple trials conducted in more than 900 patients. Addition of other drugs to cisplatin has not been associated with survival ad- vantages. although the objective response rate is sometimes in- creased. One large randomized trial found both diminished survival and increased pelvic failure rates with neoadjuvant chemotherapy prior to radiation therapy. While results of ongo- ing randomized trials of neoadjuvant chemotherapy are awaited. there is little to be gained from additional phase II studies. xi Early controlled trials of concomitant hydroxyurea with radiotherapy were suggestive of survival benefit, but they were flawed methodologically. Radiation techniques have matured, necessitating reconsideration of a radiation therapy—alone arm in such studies. Presently. there is no evidence that hydroxyurea or any other concomitant chemotherapy agent should be incor— porated into standard practice. Several controlled trials of adjuvant chemotherapy admin— istered after completion of radiation therapy are under way or in the planning stage. but at present no data support this practice. Recurrent Disease After Primary Surgical Therapy Patients with locally recurrent disease after hysterectomy should receive pelvic radiation therapy because radiation can provide long-term pelvic control and prolong survival in ap— proximately 40% of patients. The role of concomitant chemo— therapy for recurrent disease awaits definition. The results of ongoing trials for locally advanced disease may help to define this issue. Surgery for Recurrence After Radiation Therapy Therapy for recurrence of cervical cancer after maximal radiation therapy depends on site of recurrence and extent of disease. The clearest role for surgical therapy is for centrally recurrent disease, and the surgery chosen is usually a form of pelvic exenteration. More limited surgical techniques such as radical hysterectomy have been explored for recurrent small lesions located on the cervix. Although there has been a sugges- tion of benefit in patients with lesions of less than 2 cm, this ap- proach was associated with a high incidence of urinary tract complications. For the majority of patients, tailored pelvic ex- enteration remains the standard surgical approach. Candidates must be chosen carefully on the basis of their psychological and medical status. The procedure-related mortality from pelvic exenteration when performed by experienced surgeons is less than 10% and continues to decline. In addition. advances in continent urinary reservoirs, vaginal reconstruction, and low rectal anastomoses allow resumption of a more normal lifestyle following the pro— cedure. The overall 5-year survival after exenteration varies be- tween 30% and 60%. For pelvic sidewall recurrences or other limited but not central pelvic recurrences, investigational therapy has employed combinations of surgery and intraoperative electron beam or brachytherapy. Aggressive therapies for recurrent disease after radiation therapy are emotionally, physically. and economically costly for women. All these factors should be considered in making treatment decisions. Palliative Therapy Palliative treatment is appropriate for patients with symp- tomatic disease. The goals of the intervention should be defined with the patient before therapy is initiated. Palliation of pelvic symptoms can be achieved in most patients by radiation therapy. Short courses of radiation therapy are well suited to this popula— tion. However. large. single—dose fractions of radiation have been associated with higher late complication rates in some xii studies. Symptomatic extrapelvic sites such as bone are also ef- fectively palliated with short courses of radiation therapy. Sys- temic chemotherapy may also have a role in palliation of symptoms. although the benefit is usually of short duration. On- cologists should assure patients that psychological support and adequate treatment of all symptoms. including pain, are part of the treatment plan. 4) What Are New Directions for Research in Cervical Cancer? Although much is known about the incidence, etiology. and treatment ofcervical cancer, many issues remain unresolved. It is well recognized that invasive carcinoma of the cervix is, in theory. a preventable disease. Modification of high—risk be- havior in young people could change the pattern of HPV infec- tion, and research in this area is warranted. Additional research is needed to determine the optimal methods of evaluating and treating HIV—positive women who have preinvasive and in- vasive lesions of the cervix. Improving screening in populations who are typically underscreened, such as the elderly. ethnic minorities. and the poor. will require research directed toward overcoming the barriers specific to each group. Support should be given to research on provider behaviors that influence patient and clinician compliance with Pap smear screening. Additional research into methods of improving the accuracy and interpreta— tion of cytologic sampling techniques. including liquid-based systems and computer automation, should be encouraged. Currently, there are large numbers of women with Pap smears showing SIL each year. The disease in a minority of these women will progress to invasive cancer. and it would be ad- vantageous to develop predictive markers to identify those women. This would allow low-risk women to avoid costly and potentially morbid diagnostic and therapeutic procedures. Addi- tional research to identify molecular, pathologic, and im— munologic markers that would assist in this triage is needed. Clinical trials, including the ASCUS/Iow-grade SIL trial, deserve support. In women with invasive cancer of the cervix, a number of key issues merit further investigation. Research into the role of modern radiologic imaging (CT, MRI, ultrasound, and PET) for determination of tumor volume and extent of disease is war- ranted. This imaging information could also be incorporated into radiotherapeutic treatment planning. For those patients with microinvasive adenocarcinoma. an acceptable definition and guidelines for selecting patients for fertility-preserving treatment are needed. Additional information to identify those patients with microinvasive squamous cell cancer who can be treated safely with conservative, fertility—sparing surgery is also needed. Methods to improve the traditional modalities of surgery and radiation therapy in the treatment of cervical cancer represent an important area of research. The advent of newer laparoscopic procedures offers the potential for a relatively nonmorbid his- topathologic staging technique of pelvic and para-aortic lymph nodes. The status of these lymph nodes. determined either laparoscopically or by other methods, is one of the most impor— tant prognostic indicators, and pretreatment knowledge of lymph node status could allow for a more rational allocation of Journal of the National Cancer Institute Monographs No. 21. 1996 patients into therapeutic groups. Prospective trials evaluating the usefulness of histopathologic lymph node staging are indicated. Studies are needed to assess quality-of—life issues in patients undergoing therapy for both preinvasive and invasive lesions of the cervix. In patients with low—grade SIL, studies on the impact of follow-up only versus active intervention are needed. The im- pact of frequent follow—up visits and the uncertainty of receiving no treatment for a prelnvasive lesion with an unknown natural history may be significant and should be studied. Data are also needed regarding quality-of—Iife issues related to the selection of radical hysterectomy versus definitive radiation treatment in patients with early invasive cervical cancer. The optimal role for chemotherapy in the treatment of early or advanced invasive cervical cancer is unknown. In the area of concomitant chemotherapy and radiation therapy, a number of clinical trials have been completed or are under way. At this time, however, there is no proven benefit to combining chemo— therapy with radiation therapy. Additional studies, including quality—of-life studies, are warranted. Several issues related to radiation therapy for cervical cancer need to be addressed. The impact of p53 status and HPV sub- types on radiation responsiveness is a promising area of research that may allow optimization of treatment strategies. Dose— response relationship, time—dose relationship, improvements in technical instrumentation, and optimization of brachytherapy techniques need further study. Predictive assays for tumor and/or normal tissue radiation sensitivity would allow for individualization of radiation pre— scriptions, while addressing the influence of hypoxia and anemia could improve the radiation responsiveness of the tumor. Studies in these areas are ongoing and should be supported. Support should also be given to research to develop topical microbicides designed to prevent HPV infection as well as HIV infection and other sexually transmitted diseases. One of the most exciting areas of research in the prevention and treatment of cervical cancer is the development and testing of prophylactic and therapeutic vaccines against HPV. The firm- ly established causal relationship between HPV infection and cervical neoplasia makes vaccine strategies uniquely appealing as a prophylactic and therapeutic approach. Research efforts in this area should be given the highest priority. Conclusions Carcinoma of the cervix is a substantial public health issue worldwide and remains an important issue for women‘s health in the United States, especially when one considers the totality of invasive disease and its precursor lesions. The evidence presented at this Consensus Development Conference has led to the following conclusions: 0 Carcinoma of the cervix is causally related to in- fection with HPV. Reducing the rate of HPV in— fection by encouraging changes in the sexual behavior of young people and/or through develop- ing an effective HPV vaccine would reduce the in— cidence of this disease. Journal of the National Cancer Institute Monographs No. 2 l, I996 Screening with the Pap smear remains the best currently available method of reducing the in- cidence of and mortality from invasive cervical cancer. Inability of women to adhere to screening guidelines and failure of many health care pro- viders to recommend screening to their patients are issues that need to be studied further and reme- died. Specific attention should be paid to popula— tions known to be underscreened, including the elderly, the uninsured. ethnic minorities (especial- ly Hispanics and blacks), and poor women (par- ticularly those in rural areas). Women with cervical cancer should have access to appropriate specialists and clinical trials. Microinvasive squamous cell carcinoma with 3 mm or less of invasion and 7 mm or less of lateral spread (stage IAl) is highly curable either with simple hysterectomy or, in cases where preserva— tion of fertility is an issue, by cone biopsy with clear margins. Of patients with other stage I or stage IIA disease, 70%—85% are cured. Treatment with radical surgery or radiation therapy is equally effective. Selection criteria for treatment with a particular modality should be established to ensure treatment with one modality or the other, but not both. The combined use of radical surgery followed by radia- tion therapy increases the cost and morbidity of treatment substantially. Patients with more advanced, nonmetastatic dis- ease are treated with radiation. Although 5-year survival rates in the range of 40%-60% are re- ported, there is clearly room for improvement. The addition of systemic chemotherapy in this subset of patients is an active area of investigation. Recurrent cervical cancer confined to the pelvis is treated with the modality that the patient has not previously received; i.e., if the patient has received radiation therapy, she is treated with surgery and vice versa. Radiation therapy may be useful in the palliation of symptoms due to metastases. The ef— ficacy of chemotherapy in those with metastatic disease is best evaluated in the context of a clinical trial. Additional research efforts are needed to improve detection, staging, treatment, and quality of life for cervical cancer patients. Included among these ef— forts are investigations into optimal pretreatment and post—treatment imaging, improved screening compliance and technical evaluation of Pap smears, prognostic markers to improve treatment selection, laparoscopic surgical techniques, radio- biology, and systemic chemotherapy. Cervical cancer can, in theory, be prevented and treated by HPV vaccine therapy; this research holds promise for a profound impact on this disease. xiii Consensus Development Panel Patricia S. Braly. MD. Conference and Panel Cochairperson Professor and Chief Section of Gynecologic Oncology Department of Obstetrics/Gynecology Louisiana State University Medical Center New Orleans. LA Allen S. Lichter. M.D. Conference and Panel Cochairperson Isadore Lampe Professor and Chair Department of Radiation Oncology University of Michigan Medical Center Ann Arbor. MI Deborah K. Ash Director of Programs and Planning National Coalition for Cancer Survivorship Silver Spring. MD Judith L. Bader. MD. Chairman Department of Radiation Oncology Suburban Hospital Bethesda, MD Ross S. Berkowitz. M.D. William H. Baker Professor ofGynecology Division of Gynecologic Oncology Department of Obstetrics and Gynecology Harvard Medical School Director ofGynecoIogy and Gynecologic Oncology Brigham and Women‘s Hospital and Dana—Farber Cancer Institute Boston. MA Joanna M. Cain. MD. Professor and Chair Department of Obstetrics and Gynecology Pennsylvania State University Hershey, PA Amy M. Fremgen. PhD. Associate Manager Clinical Information Commission on Cancer American College of Surgeons Chicago. IL Irene Gage. MD. Instructor Division of Radiation Oncology The Johns Hopkins Oncology Center The Johns Hopkins Hospital Baltimore. MD Jennifer Stubblefield Hutchison. MD, Birth and Women's Health Center Tucson, AZ Daniel C. Ihde, MD. Professor of Medicine Division of Medical Oncology Department of Internal Medicine Washington University School of Medicine St. Louis. MO Audrey F. Saftlas. PhD. MPH. Assistant Professor Department of Epidemiology and Public Health Yale University School of Medicine New Haven. CT xiv Patricia E. Saigo. MD. Chief. Cytology Service Department of Pathology Memorial Sloan-Kettering Cancer Center New York. NY Richard L, Sweet. MD. Professor and Chair Department of Obstetrics. Gynecology. and Reproductive Sciences University of Pittsburgh/Magee Women‘s Hospital Pittsburgh. PA Speakers Barbara L. Andersen. PhD. Professor Department of Psychology Ohio State University Columbus. OH Jo Ann Benda. MD. Associate Professor of Pathology and Obstetrics and Gynecology University of Iowa Hospitals and Clinics Iowa City. IA Carol L. Brown. MD. Attending Surgeon Gynecology Service Department of Surgery Memorial Sloan—Kettering Cancer Center New York. NY Gary Lee Eddy. Sr.. MD. Associate Professor Director. Division of Gynecologic Oncology Department of Obstetrics and Gynecology State University of New York Health Science Center at Syracuse College of Medicine Syracuse. NY Patricia J. Eifel. MD. Associate Professor Department of Radiotherapy The University of Texas M. D. Anderson Cancer Center Houston. TX Perry W. Grigsby. M.D.. M.B.A.. F.A.C.R. Professor Division of Radiology Department of Radiation Oncology Mallinckrodt Institute of Radiology Washington University School of Medicine St. Louis. MO Kenneth D. Hatch. MD. Director. Gynecologic Oncology Department ofObstetrics and Gynecology University of Arizona Tucson. AZ Michael R. Henry. MD. Department of Pathology National Naval Medical Center Bethesda. MD Rolando Herrero. MD. Unit of Field and Intervention Studies International Agency for Research on Cancer. Lyon, France Journal of the National Cancer Institute Monographs No. 21. 1996 Anthony E. Howes. M.D.. PhD. Associate Professor Harvard Joint Center for Radiation Therapy Harvard Medical School Boston. MA Daniel Kapp. M.D.. PhD. Professor of Radiation Oncology Department of Radiation Oncology Stanford University Medical Center Stanford. CA Henry M. Keys. MD. Professor and Chairman Depanment of Radiation Oncology Director. Cancer Center Albany Medical College Albany. NY Rachelle M. Lanciano. MD. Radiation Oncologist Department of Radiation Oncology Fox Chase Cancer Center Philadelphia. PA Douglas Lowy. MD. Chief. Laboratory of Cellular Oncology Division of Basic Sciences National Cancer Institute National Institutes of Health Bethesda. MD Javier F, Magrina. M.D. Chair. Department ofGynecologic Surgery and Oncology Mayo Clinic Scottsdale Scottsdale. AZ Michele Follen Mitchell. M.D.. MS. Associate Professor of Gynecologic Oncology Division of Surgery and Anesthesiology Department of Gynecologic Oncology The University of Texas M. D, Anderson Cancer Center Houston. TX Mitchell Morris. MD. Associate Professor of Gynecologic Oncology Department of Gynecologic Oncology The University of Texas M. D. Anderson Cancer Center Houston. TX George A. Omura, MD. Professor Emeritus. Medicine and Gynecology University of Alabama at Birmingham Medical Director BioCryst Pharmaceuticals. Inc. Birmingham. AL Jens Overgaard. M.D.. D.M.SC., F.R.C.R. Professor. Head of Department Department of Experimental Clinical Oncology Danish Cancer Society Norrebrogade. Denmark Manuel A. Penalver. MD. Professor and Associate Chairman of Gynecology Division ofGynecologic Oncology and Gynecology Department of Obstetrics and Gynecology University of Miami School of Medicine Jackson Memorial Hospital Miami. FL Journal of the National Cancer Institute Monographs No. 21. 1996 Daniel G. Petereit. MD. Assistant Professor of Human Oncology Division of Radiation Oncology Departments of Human Oncology and Obstetrics and Gynecology University of Wisconsin School of Medicine Madison.WI Marcus E. Randall. MD. (Discussant) The William A. Mitchell Professor and Chair Department of Radiation Oncology Division of Medicine Indiana University Medical Center Indianapolis. IN Anthony H. Russell. M.D.. F.A.C.R. Radiation Oncologist Radiation Oncology Center Mercy General Hospital Sacramento. CA William J, Spanos. Jr.. MD. Professor and Acting Chairman Department of Radiation Oncology James G. Brown Cancer Center University of Louisville School of Medicine Louisville. KY Frederick B. Stehman. MD. Professor and Chair Department ofObstetrics and Gynecology Indiana University School of Medicine University Hospital Indianapolis. IN Michael A, Stellcr. M.D. Head. Section of Gynecologic Oncology Surgery Branch Division of Cancer Treatment National Cancer Institute National Institutes of Health Bethesda. MD Martin H. N. Tattersall. M.D.. F.R.A,C.P, Professor of Cancer Medicine Department of Cancer Medicine University of Sydney Sydney. New South Wales. Australia Gillian M. Thomas. M.D.. F.R.C.P.C. Head. Division of Radiation Oncology Toronto—Sunnybrook Regional Cancer Centre Professor of Radiation Oncology. Obstetrics and Gynecology University of Toronto Toronto. ON, Canada Planning Committee Patricia S. Braly. M.D. Conference and Panel Cochairperson Professor and Chief Section of Gynecologic Oncology Department of Obstetrics/Gynecology Louisiana State University Medical Center New Orleans. LA Jerry M. Elliott Program Analyst Office of Medical Applications of Research National Institutes of Health Bethesda. MD XV Alan H. Epstein. MD. Assistant Professor of Radiology and Nuclear Medicine Uniformed Services University of the Health Sciences Bethesda, MD John H. Ferguson. MD. Director Office of Medical Applications of Research National Institutes of Health Bethesda, MD Perry W. Grigsby. M.D.. M.B.A., F.A.C.R, Professor Division of Radiology Department of Radiation Oncology Mallinckrodt Institute of Radiology Washington University School of Medicine St. Louis, MO William H. Hall Director of Communications Office of Medical Applications of Research National Institutes of Health Bethesda. MD Penelope J. Hitchcock, D.V.M.. M.S. Chief. Sexually Transmitted Diseases Branch Division of Microbiology and Infectious Diseases National Institute of Allergy and Infectious Diseases National Institutes of Health Bethesda. MD Carrie P. Hunter, M.D. Special Assistant to the Director Office of Research on Women‘s Health Office of the Director National Institutes of Health Bethesda. MD Barnett S. Kramer. M.D., M.P.H. Associate Director Early Detection and Community Oncology Program Division of Cancer Control and Prevention National Cancer Institute National Institutes of Health Bethesda, MD Herschel W. Lawson. MD. Medical Epidemiologist Division of Cancer Prevention and Control Centers for Disease Control and Prevention Atlanta. GA Allen S. Lichter. M.D. Conference and Panel Cochairperson Isadore Lampe Professor and Chair Department of Radiation Oncology University of Michigan Medical School Ann Arbor, MI June R. Lunney. Ph.D.. RN. Chief Health Promotion/Disease Prevention Branch National Institute of Nursing Research National Institutes of Health Bethesda, MD Caroline McNeil Senior Science Writer Press Office National Cancer Institute National Institutes of Health Bethesda, MD xvi Nubia Munol. Chief, Unit of Field and Intervention Studies International Agency for Research on Cancer Lyon, France Cherie Nichols. M.B.At Chief. Planning, Evaluation. and Analysis Branch Office of the Director National Cancer Institute National Institutes of Health Bethesda. MD Robert C. Park. MD. Distinguished Professor Division ofGynecologic Oncology Department of Obstetrics and Gynecology Uniformed Services University of the Health Sciences Walter Reed Army Medical Center Washington, DC Vivian W. Pinn. MD. Associate Director for Research on Women's Health Office of Research on Women‘s Health National Institutes of Health Bethesda. MD John Ruffin. Ph.D. Associate Director for Research on Minority Health Office of Research on Minority Health National Institutes of Health Bethesda. MD Susan Scherr Deputy Director National Coalition for Cancer Survivorship Silver Spring, MD Mark Schiffman. MD. Medical Epidemiologist Epidemiology and Biostatistics Program Division ofCancer Etiology National Cancer Institute National Institutes of Health Bethesda. MD Charles Sherman. PhD. Deputy Director Office of Medical Applications of Research National Institutes of Health Bethesda. MD Hugh Shingleton, M.D, Vice President American Cancer Society Atlanta, GA Deborah M. Smith, M.D.. M.P.H. Medical Advisor Office of Women‘s Health Food and Drug Administration Rockville. MD Diane Solomon. M.D. Office of the Deputy Director Division ofCancer Prevention and Control National Cancer Institute National Institutes of Health Bethesda, MD Journal ofthe National Cancer Institute Monographs No. 2]. 1996 Michael A. Steller. M.D. Head. Section of Gynecologic Oncology Surgery Branch Division of Cancer Treatment National Cancer Institute National Institutes of Health Bethesda. MD Gillian M. Thomas. M.D.. F.R.C.P.C. Head. Division of Radiation Oncology Toronto—Sunnybrook Regional Cancer Centre Professor of Radiation. Obstetrics and Gynecology University of Toronto Toronto. ON. Canada Edward L. Trimble. M.D.. M.P.H. Head. Surgery Section Clinical Investigations Branch Cancer Therapy Evaluation Program Division of Cancer Treatment National Cancer Institute National Institutes of Health Bethesda. MD Conference Sponsors Office of Medical Applications of Research. National Institutes of Health John H. Ferguson. M.D. Director National Cancer Institute. National Institutes of Health Richard Kluuxner. M. D. Director Conference Cosponsors National Institute of Nursing Research Patricia A. Grady. R.N., Ph.l). Director National Institute of Allergy and Infectious Diseases Anthony S. Fum‘i. M.D. Director Office of Research on Minority Health. National Institutes of Health John Ruflin, Ph.D. Associate Director Office of Research on Women's Health. National Institutes of Health Vivian W. Pinn. M.D. Director Centers for Disease Control and Prevention David Satcher. M.D.. Ph.D. Director Bibliography The following references were provided by the speakers“ listed above and were neither reviewed nor approved by the panel. Anderson BL. Kiecolt—Glaser JK. Glaser R. A biobehavioral model of cancer stress and disease course. Am Psychol 1994;49:389-404. Andersen BL. Psychological aspects of cervical cancer. In: Rubin SC. Hoskins WJ. editors. Cervical cancer and preinvasive neoplasia. Raven Press. In press. Andersen BL. Surviving cancer. Cancer 1994;74:1484—95. Andersen BL, van der Does J. Sexual morbidity following gynecologic cancer: an international problem. IntJ Gynecol Cancer 1994;4z224—40. Andersen E. Husth M. Joerjenson A. Nielsen K. Laser conization from microin— vasive carcinoma of the cervix: short—term results. Int J Gynecol Cancer 1993:3zl283-5. Journal ofthe National Cancer Institute Monographs No. 21. I996 Averette HE. Lichtinger M. Sevin BU. Girtanner RE. Pelvic exenteration: a 15— year experience in a general metropolitan hospital. Am J Obstet Gynecol 1984;150:179-84. Berek JS. Hacker NF. Lagassc LD. Rectosigmoid colectomy and rcanastomosis to facilitate resection of primary and recurrent gynecologic cancer. Obstet Gynecol 1984;64:715-20. Birdsong GG. Automated rescreening of Papanicolaou smears: what are the im- plications [editorial]? Diagn Cytopathol 1995:13:283«6. Bloss JD. Lucci JA 3d. DiSaia PJ. Manetta A. Schiano MA. Ramsinghani N. et al. A phase II trial of neoadjuvant chemotherapy prior to radical hysterectomy and/or radiation therapy in the management of advanced carcinoma of the uterine cervix. Gynecol Oncol 1995159: 105-10. Bosch FX. Manos MM. Munoz N. Sherman M. Jansen AM. Peto J. et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspec- tive. International Biological Study on Cervical Cancer (IBSCC) Study Group [see comment citation in Medline]. J Natl Cancer Inst 1995;87:796-802. Breitburd F. Kirnbauer R. Hubbcn NL. Nonnenmacher B. Trin—Dinh»Desmar- quet C. Onh G. et al. Immunization with viruslike particles from cottontail rab- bit papillomavirus (CRPV) can protect against experimental CRPV infection. J Virol 1995;69:3959-63. Brown JM. SR 4233 (tirapazamine): a new anticancer drug exploiting hypoxia in solid tumours. BrJ Cancer 1993;67:1163-70. Childers JM. Hatch K. Surwit EA. The role of laparoscopic lymphadenectomy in the management ofcervical carcinoma. Gynecol Oncol 1992;47:38~43. Crutn CP. Nuovo GJ. Genital papillomaviruses and related neoplasms. 1st ed. New York (NY): Raven Press. 1991. Delgado G. Bundy BN. Fowler WC Jr. Stehman FB. Sevin B. Creasman WT. et al. A prospective surgical pathological study of stage I squamous carcinoma of the cervix: a Gynecologic Oncology Group study. Gynecol Oncol 1989;35:314— 20. Delgado G. Bundy B. Zaino R. Sevin BU. Creasman WT. Major F. Prospective surgical-pathological study of disease»free interval in patients with stage IB squamous cell carcinoma of the cervix: a Gynecologic Oncology Group Study. Gynecol Oncol 1990;38:352—7. Eddy DM. Screening for cervical cancer [see comment citations in Medline]. Ann Intern Med 1990;113:214—26. Eddy GL. Manetta A. Alvarez RD. Williams L. Creasman WT. Neoadjuvant chemotherapy with vincristine and cisplatin followed by radical hysterectomy and pelvic lytnphadenectomy for FIGO stage IB bulky cervical cancer: a Gynecologic Oncology Group pilot study. Gynecol Oncol 1995;57:412-6. Eifel PJ. Burke TW. Morris M. Smith TL. Adenocarcinoma as an independent risk factor for disease recurrence in patients with stage IB cervical carcinoma. Gynecol Oncol 1995;59:38-44. Eifel PJ. Morris M. Wharton JT. Oswald M]. The influence of tumor size and morphology on the outcome of patients with FIGO stage IB squamous cell car— cinoma of the uterine cervix. IntJ Radiat Oncol Biol Phys 1994;29:9—16. Fowler JM. Carter JR. Carlson JW. Maslonkowski R. Byers LJ. Carson LF. et al. Lymph node yield from Iaparoscopic lymphadenectomy in cervical cancer: a comparative study. Gynecol Oncol 1993;51:187—92. Fu KK. Phillips TL. High—dose»rate versus low-dose—rate intracavitary brachytherapy for carcinoma of the cervix [see comment citation in Medline]. Int J Radiat Oncol Biol Phys 1990;19:791—6. Fuller AF Jr. Elliott N. KosloffC. Hoskins WJ. Lewis JL Jr. Determinants of in— creased risk for recurrence in patients undergoing radical hysterectomy for stage IB and IIA carcinoma of the cervix. Gynecol Oncol 1989;33:34—9. Girinsky T. Rey A. Roche B. Haie C. Gerbaulet A. Randrianarivello H. et al. Overall treatment time in advanced cervical carcinomas: a critical parameter in treatment outcome. Int J Radiat Oncol Biol Phys 1993127: 1051-6. Greer BE. Koh WJ. Figge DC. Russell AH. Cain JM. Tamimi HK. Gynecologic radiotherapy fields defined by intraoperative measurements. Gynecol Oncol 1990;38:421-4. Grigsby PW. Perez CA. Radiotherapy alone for medically inoperable carcinoma of the cervix: stage IA and carcinoma in situ. Int J Radiat Oncol Biol Phys 1991 :21:375—8. Grigsby PW. Russell A. Bruner D. Eifel P. Koh WJ. Spanos W. et al. Late injury of cancer therapy on the female reproductive tract. Int J Radiat Oncol Biol Phys 1995;31:1281—99. xvii Harlan LC. Bernstein AB. Kessler LG. Cervical cancer screening: who is not screened and why? Am J Public Health 1991181 :885—90. Hayward RA. Shapiro MF, Freeman HE. Corey CR. Who gets screened for cer- vical and breast cancer? Results from a new national survey. Arch Intern Med 1988;148:1117-81. Heller PB. Maletano JH. Bundy BN. Barnhill DR. Okagaki T. Clinical- pathologic study of stage 118. Ill. and IVA carcinoma of the cervix: extended diagnostic evaluation for paraaortic node metastasis—a Gynecologic Oncology Group study. Gynecol Oncol 1990;38:425-311. Hockel M. Baussinann E. Mitze M. Knapstein PG. Are pelvic side»wall recur- rences of cervical cancer biologically different from central relapses? Cancer 1994;74:648-55. Hreshchyshyn MM. Aron BS. Boronow RC. Franklin EW 3d. Shingleton HM. Blessing JA. Hydroxyurea or placebo combined with radiation to treat stages IIIB and IV cervical cancer confined to the pelvis. Int J Radiat ()ncol Biol Phys 1979;5:317-22. Hutchinson ML. Isenstein LM. Goodman A. Hurley AA. Douglass KL. Mui KK. et al. Homogeneous sampling accounts for the increased diagnostic ac- curacy using the ThinPrep Processor [sec comment citation in Medline]. Am J Clin Pathol 1994;1(11 :215-19. Kapp DS. Lawrence R. Temperature elevation during brachytherapy for car- cinoma of the uterine cervix: adverse effect on survival and enhancement of dis- tant metastasis. lntJ Radiat Oncol Biol Phys 1984;10:2281-92. Kirnbauer R. Booy F. Cheng N. Lowy DR. Schiller JT. Papillomavirus L1 major capsid protein self—assembles into virus-like particles that are highly im— munogenic. Proc Natl Acad Sci U S A 1992;89:121804. Kinney WK. Alvarez RD. Reid GC. Schray MF. Soong SJ. Morley GW. et al. Value of adjuvant whole-pelvis irradiation after Wenheim hysterectomy for early—stage squamous carcinoma of the cervix with pelvic nodal metastasis: a matched—control study. Gynecol ()ncol 1989;34:258-622. Kjorstad KE. Kolbenstvedt A. Stricken T. The value of complete lymphadenec- totny in radical treatment of cancer of the cervix. Stage 18. Cancer 1984; 54:2215-9. Kolstad P. Follow—up study of 232 patients with stage Ia] and 41 1 patients with stage Ia2 squamous cell carcinoma of the cervix (microinvasive carcinoma). Gynecol Oncol 1988;33:265-72. Kurman RJ. Solomon D. The Bethesda System for reporting cervical/vaginal cytologic diagnoses: definitions. criteria and explanatory notes for terminology and specimen adequacy New York (NY): Springer—Verlag. 1994. Lagasse LD, Smith ML. Moore JG. Morton DG. Jacobs M, Johnson GH. et al. The effect of radiation therapy on pelvic lymph node involvement in stage I car— cinoma ofthe cervix. Ant J ()hstet Gynecol [974:1 19:328-34. Lockhart JL. Remodeled right colon: an alternative urinary reservoir. J Urol 1987;138:730. Lowy DR. Human papillomaviruses and epithelial cancer. In: Mukhtar H. editor. Skin cancer: mechanisms and human relevance. Boca Raton (FL): CRC Press. 1995. Makuc DM. Freid VM. Kleinman JC. National trends in the use of preventive health care by wotnen. Ath Public Health 1989;79:21-6. Massi G. Savino L. Susini T. Schauta-Amreich vaginal hysterectomy and Wert- heim—Meigs ahdotninal hysterectomy in the treatment of cervical cancer: a retrospective analysis [see comment citation in Medline]. Am J ()hstet Gynecol 1993;168:928-34. Mendenhall WM. McCarty PJ. Morgan LS. Chafe WE. Million RR. Stage IB or IIA-B carcinoma of the intact uterine cervix greater than or equal to 6 cm in diameter: is adjuvant extrafascial hysterectomy beneficial? Int J Radiat Oncol Biol Phys 1991;21:899—904. Morris M. Mitchell MF. Silva EG. Copeland LJ. Gershenson DM. Cervical con- ization as definitive therapy for early invasive squamous carcinoma of the cer- vix. Gynecol ()ncol 1993;51:193-6. Munoz N. Bosch FX. de Sanjosc S. Shah KV. The role of HPV in the etiology of cervical cancer. Mutat Res 1994;305:293—301. ()mura GA. Blessing J. Vaccarcllo L. Berman M. Match D. Clarke-Pearson D. A randomized trial of cisplatin versus cisplatin plus mitolaetol versus cisplatin and ifosfamide in advanced squamous carcinoma of the cervix by the Gynecologic Oncology Group (abstract). Gynecol ()ncol 1996160: 120. ()mura GA. Chemotherapy for cervix cancer. Sentin ()ncol 1994;21:54.62. xviii Park RC. Thigpen JT. Chemotherapy in advanced and recurrent cervical cancer: a review. Cancer 1993z7 l (4 Suppl): 1446—50. Patel FD. Shanna SC. Negi PS. Ghoshal S. Gupta BI). Low dose rate vs. high dose rate brachytherapy in the treatment of carcinoma of the uterine cervix: a clinical trial. Int J Radiat Oncol Biol Phys 199412833541. Perez CA. Fox S Lockett MA. Grigsby PW. Camel HM. Galakatos A. et al. Im— pact of dose in outcome of irradiation alone in carcinoma of the uterine cervix: analysis oftwo different methods. Int J Radiat Oncol Biol Phys 1991;21:885»98. Perez CA. Grigsby PW. Casro»Vita H. Lockett MA. Carcinoma of the uterine cervix. II. Lack of impact of prolongation of overall treatment time on morbidity of radiation therapy. Int J Radiat Oncol Biol Phys 1996;34:3-1 l. Perez CA. Kao MS. Radiation therapy alone or combined with surgery in the treatment of barrel—shaped carcinoma of the uterine cervix (stages 18. 11A. IIB). lntJ Radiat Ottcol Biol Phys 1985;] 1:1903—9. Petereit DG. Sarkaria JN. Chappell R. Fowler JF. Hartmann TJ. Kinsella TJ. Stitt JA. et al. The adverse effect of treatment prolongation in cervical carcinoma [we comment citation in Medlinel. Int J Radiat Oncol Biol Phys 1995;32:1301- 7. Petereit DG. Sarkaria JN. Schink JC. Fowler JF, Kinsella TJ. Buchler DA. HDR brachytherapy for carcinoma of the uterine cervix: the University of Wisconsin experience. Proceedings of the Xth International Brachytherapy Meeting. 95-98. Nice. France: November 25-28. 1995. Piver MS. Barlow JJ. Vongtama V. Blumenson L. Hydroxyurea as a radiation sensitizer in women with carcinoma of the uterine cervix. Am J Obstet Gynecol 1977;129:379-83. Platz CE. Benda JA. Female genital tract cancer. Cancer 1995;75:2711-94. Potish RA. Downey GO. Adcock LL. Prem KA. Twiggs LB. The role of surgi- cal debulking in cancer of the uterine cervix. Int J Radiat Oncol Biol Phys 1989: 17:979-84. Querleu D. LeBIanc E. Castelain B. Laparoscopic pelvic lymphadenectomy in the staging of early carcinoma of the cervix. Am J Obstet Gynecol 1991:1642 579—81. Ressing ME. Sette A. Brandt RM. Ruppert J. Wentwonh PA. Hanman M. et al. Human CTL epitopes encoded by human papillomavirus type 16 E6 and E7 identified through in vivo and in vitro immunogenieity studies of HLA—A*()2()1- binding peptides.J Itnmunol 1995;154:5934-43. Russell AH. Contemporary radiation treatment planning for patients with cancer of the uterine cervix. Semin Oncol 1994;21:3(1-41. Russell AH. Walter JP. Anderson MW. Zukowski CL. Sagittal magnetic resonance imaging in the design of lateral radiation treatment portals for patients with locally advanced squamous cancer of the cervix. Int J Radiat ()ncol Biol Phys 1992;23:449-55. Sardi J. Sananes C. Giaroli A. Bayo J. Rueda NG. Vighi S. et 211. Results of a prospective randomized trial with neoadjuvant chemotherapy in stage IB. bulky. squatnous carcinoma of the cervix [we comment citations in Medlinel. Gynecol ()ncol 1993;49:156-615. Schiffman MH. Schatzkin A. Test reliability is critically important to molecular epidemiology: an example front studies of human papillomavirus infection and cervical neoplasia. Presented at the 4th International Conference on Anticar- cinogenesis and Radiation Protection. April 18-23. 1993. Baltimore. MD. Can- cer Res (Suppl) 1994;54:1944—7s. Stehinan FB. Bundy BN. Thomas G. Keys HM. d‘Ablaing G 3d. Fowler WC Jr. et al. Hydroxyurea versus misonidazole with radiation in cervical carcinoma: long-term follow—up ofa Gynecologic Oncology Group trial. J Clin ()ncol 1993: 1 11523-8. Stelzer KJ. Koh WJ. Greer BE. Cain JM. Tatniini HK. Figge DC. et al. The use of intraoperative radiation therapy in radical salvage for recurrent cervical can» cer: outcome and toxicity. Am J Obstet Gynecol 1995;172:1881—8. Stitt JA. Fowler JF. Thomadsen BR. Buchler DA. Paliwal BP. Kinsella TJ. High dose rate intracavitary brachytherapy for carcinoma of the cervix: the Madison system: I. Clinical and radiobiological considerations [we comment citation in Medlinel. lntJ Radiat Oncol Biol Phys 1992;24:335-48. Suzich JA. Ghim SJ. Palmer-Hill F]. White WI. Tamura JK. Bell JA. et al. Sys— temic immunization with papillotnavirus L1 protein completely prevents the development of viral mucosal papillomas. Proc Natl Acad Sci U S A 1995: 92:11553-7. Tattersall MH. Lorvidhaya V. Vootiprux V. Cheirsilpa A. Wong F. Azhar T. et al. Randomized trial of epirubicin and cisplatin chemotherapy followed by pel» vic radiation in locally advanced cervical cancer. Cervical Cancer Study Group Journal of the National Cancer Institute Monographs No. 21. 1996 of the Asian ()ceanian Clinical Oncology Association [we comment citation in Medline]. J Clin ()ncol 1995;13:444-51. Thomas GM. Demho AJ, Myhr T. Black B. Pringle JF. Rawling G. Long—term results of concurrent radiation and chemotherapy for carcinoma of the cervix recurrent after surgery. Int J Gynecol Cancer 1993:32193-8. Thomas GM. Dembo AJ. Is there a role for adjuvant pelvic radiotherapy after radical hysterectomy in early stage cervical cancer? Int J Gynecol Cancer 1991:1:1-X. Thomas WW Jr. Eifel PJ. Smith TL. Morris M. Delclos L. Wharton JT. et a1. Bulky endoccrvical carcinomas: a 23-year experience [we comment citation in Medline]. IntJ Radiat Oncol Biol Phys 1992;23:491-9. van Bommel PF. van Linden AC. Kock HC. Leers WH. Neijt JP. A review of prognostic factors in early-stage carcinoma of the cervix tFIGO 1 B and II A) and implications for treatment strategy. EurJ Obstet Gynecol Reprod Biol 1987; 26:69-84. van Nagell JR Jr. Greenwell N. Powell DF. Donaldson ES. Hanson MB. Gay E. Microinvasive carcinoma of the cervix. Am J Obstet Gynecol 1983;145:981-91. Vermorken JB. The role of chemotherapy and squamous cell carcinoma of the uterine cervix: a review. Int J Gynecol Cancer 1993;}:129—42. West CM. Invited review: intrinsic radiosensitivity as a predictor of patient response to radiotherapy. BrJ Radiol 19951682827517. Wheeless CR Jr. Hempling RE. Rectal J pouch reservoir to decrease the frequen- cy of tenesmus and defecation in low coloproctostomy. Gynecol Oncol 1989: 35:136-8. Zaino RJ. Ward S. Delgado G. Bundy B, Gore H. Fetter G. et a1. Histopatho— logic predictors of the behavior of surgically treated stage IB squamous cell car— cinoma of the cervix: a Gynecologic Oncology Group study. Cancer 1992: 69:1750—8. /.ur Hausen H. Human papillomaviruses in the pathogenesis of anogenital can— cer. Virology 1991 :184:9-13. Notes lAbout the National Institutes of Health (NIH) Consensus Development Pro- gram: NIH Consensus Development Conferences are convened to evaluate available scientific information and resolve safety and efficacy issues related to a biomedical technology. The resultant NIH Consensus Statements are intended Journal of the National Cancer Institute Monographs No. 21. 1996 to advance understanding of the technology or issue in question and to be useful to health professionals and the public. NIH Consensus Statements are prepared by a nonadvocate. non-Federal panel ofexperts. based on 11 presentations by investigators working in areas relevant to the consensus questions during a 2—day public session. 2) questions and state- ments frotn conference attendees during open discussion periods that are part of the public session. and 3) closed deliberations by the panel during the remainder of the second day and morning of the third. This statement is an independent report of the panel and is not a policy statement of the NIH or the Federal Government. Statement availability: Preparation and distribution of this statement are the responsibility of the Office of Medical Applications of Research of NIH. Free copies of this statement and bibliographies prepared by the National Library of Medicine are available from the Office of Medical Applications of Research. NIH. or from the NIH Consensus Program Information Center by 24—hour voice mail. In addition. free copies of all other available NIH Consensus Statements and NIH Technology Assessment Statements may be obtained from the follow— ing resources: NIH Consensus Program Information Center P. 0. Box 2577 Kensington. MD 20891 Telephone: l—RXX—NlH-CONSENSUS (888—644—2667) Fax: (301) 816-2494 NIH Office of Medical Applications of Research Federal Building. Rm. 618 7551) Wisconsin Ave. MSC 9120 Bethesda. MD 20892—9120 Internet World Wirlc Web http://text.n1m.nih.gov/nih/nih.html F'l‘l’ ftp://pub1ic.nlm.nih.gov/hstat/nihedes Gopher gopher://gopher.nih.go\'/HEALTH AND CLINICAL INFORMATION xix Epidemiology of Cervical Cancer Rolando Herrera * Cervical cancer is a major public health problem through- out the world, and despite important declines in incidence and mortality observed in developed countries in the last 20 years, those indicators remain almost unchanged in develop- ing countries. In addition, a recent increase in cervical can- cer among women under age 50 years is being observed in some areas, particularly for adenocarcinoma, without a clear explanation. Marked socioeconomic and ethnic dif- ferences are evident in incidence, mortality, and survival from the disease, with the less affluent groups having a much higher impact. Epidemiologic research has concluded that certain types of human papillomaviruses are the central cause of cervical cancer and its precursors, and this has opened the door to the possibility of developing vaccines. However, many aspects of the epidemiology still need to be elucidated, and the development of clinically applicable vac- cines will require a number of years to be achieved. There- fore, renewed efforts at improving screening programs are necessary, particularly among women in the lower socio- economic regions and groups. In this paper, we present the descriptive epidemiology of invasive cervical cancer, with particular emphasis on the situation in the United States, based on the latest data from the National Cancer Institute’s Surveillance, Epidemiology, and End Results Program (up to 1992). In addition, a brief overview is made of current ideas on the role of human papillomaviruses and other risk factors for the disease. [Monogr Natl Cancer Inst 1996; 21:1-6] Descriptive Epidemiology Cervical cancer is the second most common cancer among women worldwide, with an estimated 471 000 new cases diag— nosed and 213 000 deaths occurring in 1990 (Pisani P: unpub— lished results). In many developing countries. where 80% of the cases occur. it is the most common female cancer and the most common cancer-related cause of death (1). Incidence of and mortality from cervical cancer vary widely between countries. with up to [0—fold differences between high— and low-incidence areas (Fig. I). The areas with the highest in— cidence are Latin America. sub—Saharan Africa. and southern and southeastern Asia. while those with the lowest are Western Europe, North America, the Middle East, and China. In addi— tion. regional variations within countries are very marked. with the rural and less developed areas usually reporting the highest rates (2). Incidence and mortality have declined over the last 40 years in Western Europe. the United States, Canada. Australia. New Zealand, Japan. Hong Kong. and Singapore (3). It is generally accepted that screening has played a major role in the reduction Journal ofthe National Cancer Institute Monographs No. 21. 1996 of cervical cancer, and the areas where screening efforts have been better organized have obtained the most significant reduc— tions in incidence and mortality (4). However. the reduction in cervical cancer rates started in most countries before the initia- tion of screening. probably in relation to reductions in parity and other risk factors. better access to treatment. and detection at earlier clinical stages (5). Correspondingly, in eastern Europe and in most developing countries. these rates have experienced very little change. partly because in those areas the prevalence of risk factors for the dis— ease remains high and partly because the complex process in— volved in screening for cervical cancer requires important investment and organization that are often not feasible in the less affluent societies. In addition. a reversal of the decreasing trend has been ob- served in some developed areas in the last two decades, and an important increase in incidence and mortality has been reported among women under age 50 years. particularly in England and Wales. Scotland. Ireland. New Zealand. and Australia. and also in several Nordic countries and Eastern Europe. This is probably related to preceding changes in sexual behavior and concomitant increases in the incidence of sexually transmitted diseases (STDs) that have been well documented in some of those countries. This increase in STDs probably included infection with certain types of human papillomavirus (HPV). which are now considered the central cause of cervical cancer (see below). An additional factor that may have played a role in the UK. is the fact that. until recently. screening for cervical cancer in— cluded only women age 35 years and over (3). The incidence of adenocarcinoma of the cervix. which usual- ly accounts for about 5%-l()’/r of the cases (6), is increasing in most developed countries. particularly among younger women. In Canada. for example. the cumulative risk of developing adenocarcinoma of the cervix by age 50 years among women born after 1955 is about three times higher than the similar risk for women born in 1925 (Vizcaino: unpublished results). It has been hypothesized that this increasing trend may be related to the exposure to oral contraceptives in these age groups. which appears, in some studies. to be associated with risk of adenocar- cinoma (7,8). Alternatively. it may be related to changes in the prevalence of HPV infection in these younger groups or with in- creased reporting of this histologic subtype. In the United States. according to the National Cancer Institute’s Surveillance. Epidemiology. and End Results (SEER) Program] (9). the incidence of invasive cervical cancer (age—ad— :l:C(H‘I‘t’A/HIII(/(’Ilt‘1‘ m: Rolando Herrero. M.D., International Agency for Re- search on Cancer. World Health Organization. 150 Cours Albert Thomas, 60372 Lyon Cedex ()8. France. See “Notes" section following “References." I <9] <14] I<288 I <38.0 I <77.4 Fig. 1. Estimated worldwide. age-standardized incidence rates of invasive cervical cancer. Pisani P. Parkin DM. Munoz N. Ferlay J. Cancer and infection: estimates of the attributable fraction for 1990: International Agency for Research on Cancer. 1996 (submitted for publication). justed for the standard world population) for the period 1988 through 1992 was 8.1 cases per 100000 women. on the basis of information from the nine areas covered by the Program, and the age-adjusted mortality for the entire United States was 2.6 deaths per 100000 women. Corresponding figures for white women are 7.6 (incidence) and 2.2 (mortality) and for black women 12.0 and 5.7. respectively. Cervical cancer is the seventh most common cancer among women in the United States, and an estimated 15 800 invasive cancers and 4800 deaths occurred in 1995 (/0). The median age at diagnosis of cervical cancer is 47 years and. excluding childhood cancers and lymphomas. patients with cervical cancer die younger than women with any other type of cancer (average, 60 years). Thus. the disease has a high social impact and is responsible for the loss of approxi- mately 118000 years of life every year (SEER estimate for 1992). Since 1950. incidence and mortality rates in the United States have declined 76.6% and 74.5%. respectively. and in the period from 1973 through 1992. incidence declined 37.0% among white women and 55.6% among black women. Corresponding figures for mortality are 43.0% (white women) and 46.9% (black women). Fig. 2 presents age-adjusted incidence rates by race per 100 000 women during the period. It is evident that white women have reached a plateau in the decline of their incidence while black women have not. and therefore. the “gap" between races is decreasing. Until recently. this was more evident for women under age 50 years. but the black/white ratio of in— cidence in 1992 was 1.3 for that age group compared with a ratio of 1.5 for women aged 50 years and over. With respect to mortality, however. the gap is not closing at the same rate (Fig. 3). and the black/white ratio of mortality was still 2.3 for women under age 50 years and 3.0 for women aged 50 years and over. to The age-specific incidence rates also reveal important racial differences (Fig. 4). with rates among white women reaching a plateau at around 15 per 100000 after age 35 years. On the other hand, among black women. the rates continue to increase consistently with age in a pattern similar to that observed in some high-incidence areas. Different prevalences of risk factors by race and age group are likely to explain this pattern (ll). Other ethnic groups in the United States also have higher in- cidence rates of invasive cervical cancer than white women (12). For example, Hispanic, Korean, and Alaskan native women have about twice the incidence rates of white women. Notably. Vietnamese women have five times the incidence of white women. On the other hand. Japanese and Chinese women Incidence rate 30 25 20 15 10 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 Year of diagnosis Fig. 2. SEER age—adjusted incidence by race and year of diagnosis. 1973 through 1992 (cases per 100 000 women). Journal ofthe National Cancer Institute Monographs No. 21. 1996 Mortality rate 14 12 10 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 Year of death Fig. 3. US. age-adjusted mortality by race and year of diagnosis. 1973 through 1992 (deaths per 100 000 women). residing in the United States have a slightly lower incidence than their white counteiparts. There are also important differences in incidence and mor- tality rates for cervical cancer between different states (9). The five states with the highest mortality during the period 1988 through 1992 are the District of Columbia (4.9 deaths per 100 000. age-adjusted for the 1970 US. standard population). Kentucky (4.1). Alabama (4.0). South Carolina (3.9). and Delaware (3.7). These states have in common either important proportions of black population. rural poverty. or both. The five states with the lowest rates are North Dakota (2.1). Nebraska (2.1). Hawaii (2.0). Idaho (1.8). and Minnesota (1.6). With respect to the recent increase observed among young women in several developed countries. a very slight tendency to increase was observed from 1987 to 1990 in the United States. However. as shown in Fig. 5. it is restricted to white women and declined again after 1990. Fig. 6 presents age—specific incidence rates of adenocarcinoma among white women for three different periods from 1975 through 1989 and shows an increase among women under age 55 years for the period 1985 through 1989 (3). A recent analysis ([3) of the SEER data demonstrates that Age-specific rate 60 50 40, o . t . 1 . . ,a,4 15-19 25-29 35-39 45-49 55-59 65-69 75-79 85+ Age at diagnosis Fig. 4. SEER age-specific incidence by race. 1988 through 1992 (cases per 100 000 women). Journal ofthe National Cancer Institute Monographs No. 21. 1996 Incidence rate 18 16 14 12 10 4 2 L 0 t 1 1 . . 1 . 1 . i . . . 1 1 4*, 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 Year of diagnosis Fig. 5. SEER age-adjusted incidence under age 50 years by year of diagnosis. 1973 through 1992 (cases per 100 000 women). invasive adenocarcinoma has been increasing in both white and black women. but the increase is statistically significant only among whites. with an increase of 4.2% per year for those born since 1935. Notably. the rate for white women has surpassed that for black women during the last study period (1988—1990). The stage distribution of cervical cancers diagnosed in the United States in the period 1986 through 1991 was as follows: 51% localized. 33% regional. 8% metastatic. and 7% unstaged. according to SEER data. Table 1 presents stage distribution by age and race for the period 1986 through 1991 and demonstrates that cases are diagnosed at more advanced stages among older women than among younger women and among black women than among white women. Overall. the 5-year survival rates are 71.1% for white women and 56.2% for black women. These figures have remained con- stant since 1974. without evidence of improvement in survival in any race. On the contrary. a statistically significant deteriora— tion is observed for the period 1986 through 1991 among black women. Table 2 shows the 5-year survival by stage and race and demonstrates that white women under the age of 50 years tend to have better survival. even after stage of diagnosis is taken into account. The reasons for this difference in survival by age and race could be explained by blacks having higher real stage within diagnosed stage. Alternatively. differential access to medical care could also explain such a difference in survival. Of 17 678 invasive cervical cancers reported to the nine registries included in the SEER Program from 1973 through 1987 in the United States (14). 77.1% were squamous cell car- cinomas. 2.5% were adenosquamous or adenocarcinomas with squamous metaplasia. 10.9% were adenocarcinomas. and 8.4% corresponded to other carcinomas or carcinomas not otherwise specified. Rare histologic types. corresponding to 1.1% of carcinoma. and tumors. include transitional cell mixed miillerian tumors. SilI'COIUiIS, In summary. squamous cell cervical cancer is decreasing in the United States. but important racial and ethnic differences persist. The increase in adenocarcinoma requires further epi— demiologic studies to elucidate its origin and take the corre— sponding measures for control. Incidence rate 3.5 3.0 Age 2'5 ___, — —30-34 2‘0 ______________________________ 40'44 Fig. 6. SEER age-specific incidence rates of adeno~ 1.5 _____ . _ . 50.54 carcinoma, white women, 1975 through I989. Mod- __________________ ified from (3) (cases er 100 000 women). 1.0 a 60-64 P 0.5 0.0 - . 75-79 80-84 85-89 Year of diagnosis Risk Factors tive risks in the order of up to several hundreds, both for cervi- The association of cervical cancer with sexual behavior was recognized early on because of the increased occurrence of the disease among women reporting multiple sexual partners, early ages at first intercourse, or histories of venereal diseases, as well as among those whose male sexual partners reported multiple partners (15). These observations led investigators to search for transmissible etiologic agents, including gonorrhea. syphilis, and herpes simplex type 11. However, neither experimental nor epidemiologic data were sufficient to confirm any of these as the main causal agent for this disease. Concurrently, it became clear from a variety of studies that other factors (e.g., high parity, oral contraceptives, and smoking) might also play a role in the etiology of invasive cancer. However, since most of these factors correlate to some extent with sexual behavior. it was not clear in the earlier studies if their role in the etiology of cervical cancer was independent or just a reflection of their correlation with the underlying infectious agent. Over the last 15 years, a vast amount of evidence has ac- cumulated, suggesting that certain types of human papilloma- virus, in particular, types 16. 18, 31. and 33, are the main causal agents of cervical cancer and its precursors. Despite important indications from animal models and experimental data. the ini— tial epidemiologic studies were flawed by the low sensitivity and specificity of the early DNA hybridization assays developed to test for HPV. The latest polymerase chain reaction (PCR) methods have been very useful in clarifying the role of the viruses and will permit the study of the independent role of other factors, after adjusting properly for the presence of HPV. The more recent case—control studies, which have employed PCR techniques for detection of HPV, have demonstrated rela— Table l. SEER] stage distribution of invasive cervical cancer by age group and race, 1986 through 1991 cal intraepithelial neoplasia (CIN) and invasive cancer (Table 3). In addition, two cohort studies (25,26) indicate that the presence of HPV precedes disease. In the most careful studies, HPV is detected in up to 96% of high—grade ClN (27) and in more than 93% of invasive cancers from all geographic areas (28). An early study (29) suggested that HPV-negative cancers were more likely to have mutations of the p53 (also known as TP53) tumor suppressor gene, but this statement was later weakened by contradictory evidence. At present, the existence of HPV—negative cancers is considered uncommon to rare, and among the best known of HPV’s pathogenic mechanisms are the inactivation of p53 and pRB proteins by products of the E6 and E7 regions of HPV 16, respectively (30). A recent meeting of experts organized to evaluate the carcinogenicity of HPV con- cluded that HPV 16 and 18 are already proven to be carcino— genic to humans (31), types 31 and 33 are probably carcino- genic. and other types are possibly carcinogenic. Given the presence of HPV in virtually all cervical cancers and the fact that low-grade lesions (CIN l/koilocytotic atypia) are now considered one of the manifestations of infection. it ap- pears that the initial event in most instances of cervical neo- plasia, including invasive cancer, is infection with one of these viruses (32). Although there is some evidence for nonsexual routes of transmission (33), the initial infection appears to be common venereal transmission occurring shortly after the initiation of sexual activity. In most populations studied, young women ap- pear to have the highest prevalence of HPV DNA in the cervix. For example, in a group of cytologically normal women from Portland, OR, prevalence among women ages 16—24 years was 32%. declining steadily with increasing age, to a prevalence of Table 2. Percent of invasive cervical cancer patients surviving 5 years. by stage and race (SEER' data). 1986 through 1991 Stage Whites <50 y Whites 250 y Blacks <5()y Blacks 250 y Stage Whites <50 y Whites 250 y Blacks <50 y Blacks 250 y Locali/ed 68 37 50 29 Locali/ed 93.3 88.1 89.4 84.6 Regional 23 42 36 44 Regional 55.4 49.1 40.4 43.0 Distant 4 12 7 18 Distant 9,2 8.3 5.9 10.4 Unstaged 5 9 8 9 Unstaged 77.5 52.9 70.8 49.7 4 Journal ofthc National Cancer institute Monographs No. 21, 1996 Table 3. Recent selected polymerase chain reaction—based case—control studies of CIN and cancer* In\'estigator(s) (ref. No. )/country Cases Bosch ct al. (/6) Spain CIN 3 Colombia CIN 3 Schiffman et a1. (/7). United States CIN 2-3 Becker et al. (/8). United States CIN 2-3 Olsen et al. (/9). Norway CIN 3-3 Liaw c1 al. (20). Taiwan CIN 23 or CilllCCl‘ Mufio/ et al. (2/) Spain Cancer Colombia Cancer Eluf-Neto et al. (22). Brazil Cancer Cancer HPV types OR (95% CIH‘ All tested 569 (248—1306) HPV 16 295.5(443-19466) All tested 15.5 (24.2-29.4) HPV 16 27.1 (10.6-69.5) All tested 42(15.3-124.3) HPV 16/18 180(49—630) A11 tested 20.8(108-402) HPV 16 9.96.4483) A11 tested 7124 (276-1913)) HPV 16 181464116161) All tested High-risk Mediumwisk Louurisk 1223 (385-3889) 1279.9(185588298) 98.8 (23.84105) 48 (0.5478) Alltcsted 46.2(185—1 15.1) HPV 16 45.8(15.8-132.‘)) All tested 15.6(69-347) HPV 16 17.3 (74-404) All tested 37.1 (19.6-70.4) HPV 16 74.9(325-173) 16/33 32.9(7.7-|41.l) Peng et al. (23). China *Modified from (24). CIN = cervical intraepithelial neoplasia; HPV = human papillomavirus: OR = odds ratio: CI = confidence interval. ”(Adjusted ORs and 95% CIs. 4% among those older than 45 years (34). The reduction in the prevalence of detection of HPV DNA as age increases. which is unlikely to be explained by a cohort effect. indicates that the majority of women are able to eliminate the infection by im— munologic mechanisms. This appears to be the case even for women who develop cytologic evidence of CIN. particularly in the case of low-grade lesions. which are associated with very high regression rates (35). In a recent prospective study. 342 patients with abnormal Pap smears were followed every 3—4 months with cytology. colpos- copy. and repeated testing for HPV. After a mean follow-up time of 16.5 months. 19 women developed progressive CIN 3. All of these women had I-IPV that was detectable at the begin- ning of follow—up. while no progression was seen among women in which HPV DNA was not detected (25). Thus. HPV infection appears as a necessary but not sufficient cause in most cases ofcervical cancer. Epidemiologic research now focuses on determining factors responsible for persistence of infection and progression to high-grade disease (36). Persistence of an infection with an oncogenic type of HPV appears as an important determinant for progression from nor— mal or CIN 1 to CIN 2 or CIN 3. The question remains as to the factors that make tnost women clear their infections. even with oncogenic types. while others develop persistent infections and progress to more severe disease. One of the studies (37) address- ing the issue of determinants of persistence found older age and oncogenic viral types as the main factors. Some women with certain immunologic characteristics may be less efficient at Clearing infections with these viruses or other agents requiring cellular immune response. For example. the presence of certain human leukocyte antigen haplotypes has been associated with increased risk of CIN (38). In addition. it is now clear that Journal ofthe National Cancer Institute Monographs No. 21. 1996 patients with acquired immunodeficiency syndrome are more likely to develop cervical disease. as are patients receiving im- munosuppressive treatments (15). Hormonal influences are another group of potential cofactors. as indicated by strong associations with parity and some indica— tions of increased risk among users of hormonal contraceptives. These two have surfaced as independent factors in recent case— control studies (22,39). after adjusting for detection of viral DNA. Smoking has been extensively studied and its role as a cofactor appears biologically plausible. given evidence from animal studies of interaction of papillomaviruses with chemical agents for carcinogenesis and the fact that tobacco products are detectable in cervical mucus of smokers. However. the correla- tion of smoking and sexual behavior has made it particularly difficult to disentangle its independent effect. In several studies (19.21.22) that have been able to adjust for the effect of HPV. the association with smoking disappeared after adjustment. Chronic cervical inflammation associated with bacterial vaginosis or chlamydial infection may also be a cofactor with other sexually transmitted diseases (40). Finally. the role of nutritional factors requires reanalysis in the context of studies where PCR techniques are used for HPV detection (4/). The clarification of the role of certain HPV types in the etiology of cervical cancer has opened the door to the search for prophylac- tic and therapeutic vaccines. In the meantime. efforts should continue to make adequate screening accessible to underserved populations in both developed and developing countries. References (l) Pisani P. Parkin DM. Ferlay J. Estimates of the worldwide mortality from eighteen major cancers in 1985. Implications for prevention and projec- tions of future burden. Int J Cancer 1993;55:891-903. Ru (3 (4 (5 (6 V (7) (9 (/0) (/3) (/4) (l5) (/6) (l7) (/8) (I9) (20) Herrero R. Brinton LA. Hartge P. Reeves WC. Brenes MM. Urcuyo R. et a1. Determinants of the geographic variation of invasive cervical cancer in Costa Rica. Bull Pan Am Health Organ 1993:27: I545. Beral V. Hermon C. MunOI N. Devesa SS. Cervical cancer. Cancer Surv 1994;19-20365—85. Hakama M. Louhivuori K. A screening programme for cervical cancer that worked. Cancer Surv I988;7:4()3-16. Ponten J. Adami HO. Bergstroin R. Dillner J. Friberg LG. Gustafsson L. et a1. Strategies for global control of cervical cancer [see comment citation in Medline]. Int J Cancer 1995:6021-26. Kjaer SK. Brinton LA. Adenocarcinomas of the uterine cervix: the epidemiology ofan increasing problem. Epidemiol Rev 1993;15:486-98. Brinton LA. Reeves WC. Brenes MM. Herrero R. de Britton RC. Gaitan E. et a1. ()ral contraceptive use and risk of invasive cervical cancer. Int J Epidemiol 1990;19:4-1 1. Ursin G. Peters RK. Henderson BE. d‘Ablaing G 3d. Monroe KR. Pike MC. Oral contraceptive use and adenocarcinoma of cervix [we comment citation in Medlinel. Lancet 1994;344:1390-4. Kosary CL. Ries LA. Miller BA. Hankey BF. Harras A. Edwards BK. editors. SEER cancer statistics review. 1973-1992: tables and graphs. Na— tional Cancer Institute. NIH Pub] No. 96—2789. Bethesda. MD. 1995. Wingo PA. Tong T. Bolden S. Cancer statistics. 1995 [published erratum appears in CA CancerJ Clin 1995 Mar—Apr145:127—8]. CA Cancer J Clin 1995:4528-30. Schairer C. Brinton LA. Devesa SS. Ziegler RG. Fraumeni JF Jr. Racial differences in the risk of invasive squamous-cell cervical cancer. Cancer Causes Control 1991.2:283—90. Miller BA. Kolonel LN. Bernstein L. Young JL Jr. Swanson GM. West D. et al.. editors. Racial/ethnic patterns of cancer in the United States 1988- 1992. Bethesda (MD): National Institutes of Health. National Cancer In- stitute; 1996: DHHS Publ No. (NIH) 96—4104. Zheng T. Holford TR. Ma 2. Chen Y. Liu W. Ward BA. et al. The continu- ing increase in adenocarcinoma of the uterine cervix: a binh cohort phenomenon. IntJ Epidemiol 1996;25:252-8. Plat]. CE. Benda JA. Female genital tract cancer. Cancer 1995;75:270—94. Palefski JM. Holly EA. Molecular virology and epidemiology of human papillomavirus and cervical cancer. Cancer Epidemiol Biomarkers Prev 1995:41415-28. Bosch FX. Munol, N. de Sanjosc S Navarro C. Moreo P. Ascunce N. et al. Hutnan papillomavirus and cervical intraepithelial neoplasia grade III/car— cinoma in situ: a case-control study in Spain and Colombia. Cancer Epidemiol Biomarkers Prev 1993:415-22. Schifftnan MH. Batter HM. Hoover RN. Glass AG. Cadell DM. Rush BB. et al. Epidemiologic evidence showing that human papillomavirus infec- tion causes most cervical intraepithelial neoplasia [we comment citations in Medline]. J Natl Cancer Inst 19931852958474. Becker TM. Wheeler CM. McGough NS. Partnenter CA. Jordan SW. Stid» ley CA. et a1. Sexually transmitted diseases and other risk factors for cervi- cal dysplasia among southwestern Hispanic and non-Hispanic white women. JAMA 1994;271:1181—8. Olsen AO. Gjoen K. Sauer T. ()stavik I. Naess O. Kjerulf K. et al. Human papillomavirus and cervical intraepithelial neoplasia grade II-III: a popula- tion-based case~control study. Int J Cancer l995:6:312—5. Liaw KL. Hsing AW. Chen CJ. Schiffman MH. Zhang TY. Hsieh CY. et al. Human papillomavirus and cervical neoplasia: a case-control study in Taiwan. Int J Cancer 1995;62:565-71. Muno/ N. Bosch FX. de Sanjose S. Tafur L. IzarLuga/a I. Gili M. et al. The causal link between human papillomavirus and invasive cervical can- cer: a population—based case-control study in Colombia and Spain. Int J Cancer 1992;52:743~9. Eluf-Neto J. Booth M. Munoz N. Bosch FX. Meijer CJ. Walboomers JM. Human papillomavirus and invasive cervical cancer in Bra/.il. Br J Cancer 1994:6921 14-9. Peng HO. Liu SL. Mann V. Rohan T. Rawls W. Human papillomavirus types 16 and 33. herpes simplex virus type 2 and other risk factors for cer- vical cancer in Sichuan Province. China. Int J Cancer 1991147271 1-6. Munoz N. Bosch FX. Current views on the epidemiology of HPV and cer- vical cancer. Papillomavirus Rep. In press. (29) (3(1) (3/) (33) (34) (35 (36 (37) (38) (39) (40) (Ht Remmink AJ. Walboomers JM. Helmerhorst TJ. Voorhorst FJ. Rozendaal L. Risse EK. et al. The presence of persistent high-risk HPV genotypes in dysplastic cervical lesions is associated with progressive disease: natural history up to 36 months. Int J Cancer 1995;61:306-11. Liu T. Soong SJ. Alvarez RD. Butterworth CE Jr. A longitudinal analysis of human papillomavirus Io infection. nutritional status. and cervical dysplasia progression. Cancer Epidemiol Biomarkers Prev 1995:4:373—8(). Matsukura M. Sugase M. Identification of genital human papillomaviruses in cervical biopsy specimens: segregation of specific virus types in specific clinicopathological lesions. Int J Cancer 1995;61:13-22. Bosch FX. Manos MM. Munol. N. Sherman M. Jansen AM. Peto J. et al. Prevalence of human papillomavirus in cervical cancer: a worldwide per- spective. International Biological Study on Cervical Cancer (IBSCC) Study Group [we comment citation in Medline]. J Natl Cancer Inst 1995: 87:796-802. Kessis TD. Slebos RJ. Han SM. Shah K. Bosch XF. Munoz. N. et al. p53 gene mutations and MDMZ amplification are uncommon in primary car- cinomas of the uterine cervix. AmJ Pathol 1993114311398405. mr Hausen H. Rosl F. Pathogenesis of cancer of the cervix. Cold Spring Harbor Symposia on Quantitative Biology 1994. Vol LIX. New York: Cold Spring Harbor Laboratory Press. 623-8. IARC. Monographs on the evaluation ofcarcinogenic risks to humans. vol 64: Human papillomavirus. Lyon: International Agency for Research on Cancer. 1995. Schiffman MH. Bauer HM. Hoover RN. Glass AG. Cadell DM. Rush BB. et al. Epidemiologic evidence showing that human papillomavirus infec- tion causes most cervical neoplasia [.m' comment citations in Medlinel. J Natl Cancer Inst 1993;85:958-64. Cason J. Kaye J. Best JM. Non‘sexual acquisition of human genital papil- lomaviruses. Papillomavirus Rep 1995:121-7. Bauer HM. Hildesheim A. Schiffman MH. Glass AG. Rush BB. Scott DR. et al. Determinants of genital human papillomavirus infection in low—risk women in Portland. Oregon. Sex Transm Dis 1993;20:274-8. Wright TC. Kurnian RJ. A critical review of the morphologic classification systems of preinvasive lesions of the cervix: the scientific basis for chang- ing the paradigm. Papillomavirus Rep 1994:.5: 175—82. Schiffman MH. New epidemiology of human papillomavirus infection and cervical neoplasia leditorial] [we comment citation in Medline]. J Natl Cancerlnst 1995;87:1345-7. Hildesheim A. Schiffman MH. Gravitt PE. Glass AG. Greer CE. Zhang T. et al. Persistence of type—specific human papillomavirus infection among cytologically normal women [we comment citation in Medline]. J Infect Dis 19941169223540. Apple RJ. Becker TM. Wheeler CM. Erlich HA. Comparison of human leukocyte antigen DR-DQ disease associations found with cervical dys- plasia and invasive cervical carcinoma. J Natl Cancer Inst 1995;87:427—361. Bosch FX. Muno/ N. de Sanjose S. Izarzuga7a1.Gili M. Viladiu P. et al. Risk factors for cervical cancer in Colombia and Spain. Int J Cancer 1992: 52:750-8. de Sanjose S. Munoz N. Bosch FX. Reimann K. Pedersen NS. Orfila J. et al. Sexually transmitted agents and cervical neoplasia in Colombia and Spain. Int J Cancer 1994;56:358—63. Potischtnan N. Nutritional epidemiology of cervical neoplasia. J Nutr 1993;123:424-9 Notes lEt/itnr'x note: SEER is a set of geographically defined. population-based central tumor registries in the United States. operated by local nonprofit or— ganizations under contract to the National Cancer Institute (NCI). Each registry annually submits its cases to the NC] on a computer tape. These computer tapes are then edited by the NCI and made available for analysis. I thank Drs. Allan Hildesheitn and Nubia Munoz for their valuable assis- tance. Journal of the National Cancer Institute Monographs No. 2 I. 1996 Screening Patterns for Cervical Cancer: How Best to Reach the Unscreened Population Carol L. Brown * Although the mortality rate for cervical cancer in the United States has declined steadily since the introduction of the Pap smear for screening in 1945, recent statistics show a rising incidence, with the number of new cases expected in 1996 representing a record high since the mid-1980s. Part of the rising incidence may be because of increasing numbers of women in the United States who did not receive screening or having inadequate screening with the Pap smear. This paper will examine the recent patterns of cervical cancer screening in the United States, with particular attention to defining which populations are not being screened. Barriers to screening in these populations will be defined and grouped into four categories: lack of knowledge, economic, cultural, and belief system; and logistical. Successful approaches that have been used to overcome these barriers in screening programs targeted at the “hard to reach” population will be described. [Monogr Natl Cancer Inst 1996;21:7-11] Although the mortality rate for cervical cancer in the United States has declined steadily since the introduction of the Pap smear for screening in 1945. recent statistics show a rising in- cidence. with the 15 700 new cases expected in 1996 repre- senting a record high since the mid l980s (I). Although this increased incidence may be partially accounted for by the human papillomavirus epidemic. part of the increasing in- cidence may be because of increasing numbers of women who did not receive screening Pap smears. Women from ethnic minorities. Native Americans. African-Americans. and His- panics have higher rates of cervical cancer than whites. and. along with elderly women. excess mortality (l—5). The percent— age of women diagnosed with advanced stage disease is also higher for minorities and the elderly (l.2.4.6). at least suggest- ing that underutilization of screening Pap smears in these populations may contribute to the observed higher incidence and mortality rates. Cervical cancer might be described as an ideal disease for which to implement mass population—based screening. A pro- longed asymptomatic phase permits early detection of prein- vasive disease that is potentially 100% curable. making invasive cervical cancer theoretically a completely preventable disease. The screening test (Pap smear) is simple and risk free and has acceptable levels of sensitivity and specificity. Although the ef- fectiveness of population—based Pap smear screening has never been demonstrated in a randomized. controlled trial. extensive evidence from historical case—control studies and the experience of large screening programs in northern Europe and Canada have shown that Pap smears can decrease the incidence of and Journal ol'the National Cancer Institute Monographs No. 21. 1996 mortality from cervical cancer by as much as 60%-9()%. 1n the United States. it is estimated that abolishing Pap smear screen— ing would increase the incidence of invasive cervical cancer by twofold to threefold (7). Although it is universally accepted that Pap smear screening is highly effective at decreasing the incidence of invasive cancer in a screened population. no universal consensus has been reached for how frequently and for what age range Pap smear screening should be performed. Current published recommenda— tions from a variety of agencies reflect the debate. The Canadian Task Force and the US. Preventive Services Task Force recom- mend screening every 1—3 years beginning at age 18 years or onset of sexual activity and suspending screening in older women if two consecutive normal Pap smears are seen (8. 9). In 1988. the American College of Obstetricians and Gynecologists (ACOG) along with several other professional medical societies issued a consensus recommendation for screening that em- phasized annual screening for younger women. leaving any change in frequency up to “the discretion of the physician." and setting no upper age limit (10). Recently. the ACOG issued a technical bulletin reiterating this recommendation and em— phasizing the bias of gynecologists to perform annual screening by listing a set of risk factors that should trigger “more fre— quent.” i.e.. annual. Pap smears in a woman who has any of these factors (II). A study (7) advocating a 3-year screening interval for Pap smears is based on data from countries with centralized nation- al cytology laboratories with modern tracking systems and uniform quality control. Concerns about accepting less frequent than annual Pap smear screening in the United States are based on the observed 20% false—negative rate for Pap smears in this country as well as our lack of centralized laboratories and the mobility and poor compliance of women returning for follow-up smears (11.12). Because of the lack of a nationwide centralized screening program for cervical cancer in the United States. data reflecting the actual rates of screening must be gathered from a variety of sources; no one source accurately reflects all women at risk. Most available data are based on patient self-reporting of Pap smears. which usually overestimates compliance ([2). The Na- tional Health Interview Survey (NHIS). a large national inter- view study conducted annually by government agencies. contains data regarding the self—reported use of Pap smears. In Table 1. data compared from the 1987 and 1992 NHlS surveys *(‘nrrmpmnlz'm'v Ia: Carol L, Brown. M.l)., Gynecology Service. Department of Surgery. Memorial Sloaanettering Cancer Center. 1275 York Ave.. New York. NY 10021. Table 1. Prevalence of Pap smear screening in the United States: National Health Interview Survey. 1987 and 1992 (ll/5) ‘/( ever having Pap smear ‘4 having Pap smear in last year 1992 Age. y 1987 1992 1987 18-44 90 91 47 50 45-64 92 95 30 41 265 79 X3 17 22 Total 89 91 3X 43 reveal two important trends: the use of Pap smear screening in the United States has increased over time but decreases in direct proportion to increasing age (13-15). The proportion of US. women having ever been screened with Pap smears approaches 90%. which is in keeping with the goals set out in the Healthy People 2000 Program (16). However. the proportion having recent screens. and thus the number of women returning for fol- low-up screens. is much lower. reflecting the problem with com- pliance. What summary data from surveys such as the NHIS do not clearly reflect is the identity of the subpopulations of women in the United States who are not receiving any or regular screening Pap smears. These women account for the majority of new cer- vical cancer cases in this country. with 50% of cancers occur— ring in women who have never been screened and 60% in women with no smear in the past 5 years (12.17). Women with invasive cervical cancer and no history of a recent Pap smear tend to have more advanced stage disease at time of presentation compared with women who have had recent smears (12.17.18). Who are the women in the United States who are not receiv- ing any or adequate screening Pap smears? These “unscreened” populations consist of older women. uninsured and impov- erished women. members of ethnic minorities. particularly Hispanic and older African-American women. and women residing in rural areas. As shown in Table 2. the percentage of women from all of these groups who have had a recent Pap smear falls far below the Healthy People 2000 goals for Pap smear screening (16.19). Since these NHlS data rely on self— reporting. it is likely that the true impact of age. ethnicity, and socioeconomic status on Pap smear screening is even more sub— stantial. Table 2. Effect ofethnicity. age. and socioeconomic status on Pap smear screening rates: National Health Interview Survey ([6. l9) ‘/( having Pap test in last 3 years Group 1987 1992 2000* White 65 65 85 African-American 67 72 80 Hispanic 57 63 80 Rural 62 60 80 Poor:L 64 65 80 Age >70 y 44 46 70 *Goal for year 2000 outlined in Healthy People 2000 Program. 'l'Household income less than 315 10 000 per year, As reflected in the NHIS data. older women have consistently lower rates of screening with Pap smears. Twenty—five percent of cervical cancer cases and 41% of deaths occur in women older than age 65 years (20). In a study where women older than 65 years were offered Pap screening in a general medical clinic as an outreach intervention. more than 1500 women were screened. Twenty-five percent had never had a Pap smear before. and 75% had had no Pap in the last 5 years (4). In several studies of Pap smear screening patterns in the United States. increasing age is directly correlated with the likelihood of never having had a Pap smear (17.18.21). Although they receive Pap smears less frequently than younger women. older women have the same or a greater number of recent contacts with health care providers (22.23). This suggests that the elderly are not being offered screening at the time of their general medi— cal contact. and for the elderly. most of these contacts are with internists and family physicians. not gynecologists (24—26). It has been said that socioeconomic status is the most power- ful determinant of health status (2). and the effect of this vari- able on Pap smear screening rates supports this statement. In the 1985 NHIS study. screening rates were 10%-|3% lower among poor women than nonpoor women for all ages and ethnic sub- groups (27). ln an analysis of the 1987 NHIS data. poor women were found to be twice as likely to have never heard of a Pap smear and to not have had recent screening (5). Screening rates among rural women reflect a similar pattern. with only 57% of women in rural Texas and 55% in Appalachia having had a Pap smear within 3 years of the survey (28). In an interesting study. Katy. and Hofer (29) compared screening rates in women from Ontario. Canada. with the US. NHIS data from 1990 and found that lower socioeconomic status still had a profound negative ef— fect on screening rates. even in Ontario where universal access to care is at least financially available. That the effect of poverty on Pap smear screening is not due solely to the lack of access to care is illustrated by two additional studies in the United States that found that in women covered by Medicaid. 40% had had no Pap test in the last 3 years (28) and 90% of women with no Pap test in the last 4 years were covered by Medicaid (30). The effect of ethnicity on Pap smear screening rates is com— plex because ethnicity is often a marker for circumstances that also reflect low socioeconomic status. In the United States. 20% of African-Americans and 30% of Hispanics are uninsured and the lack of insurance has been shown to be a negative predictor for Pap smear screening in most studies (27). In an analysis of Pap smear use among Caribbean—born blacks. 25% of Haitian— born women were compliant with follow—up smears compared with 86% of women born in the English—speaking Caribbean countries. and the authors attributed this discrepancy to differen- ces in socioeconomic status and language barriers rather than “ethnicity" (3/). Interestingly. data reveal that young black women have made substantial gains in Pap smear screening rates. and in many studies. their rates of screening approach or exceed those of young white women (5.19.22.27.3233). How- ever. elderly black women remain much less likely to be screened: in one study (34). more than 40% of African— American women over 65 years of age had never had a Pap smear. Reasons for the disparate rates of Pap smear screening between elderly African—American women and elderly white Journal of the National Cancer lustitute Monographs No. 21. 1996 women may relate to differences in knowledge and perceptions about cancer prevention due to cultural attitudes (35). Members of other ethnic minorities in the United States have excessively high incidence and mortality rates for cervical can— cer. yet data reflect that they have among the lowest use of screening Pap smears. From analysis of the 1987 NHIS. Harlan et al. (5) found that 20% of Hispanic women had never even heard of a Pap smear. In a multivariate analysis of these data that used several contributing factors. having Spanish as the primary language was a negative predictor for compliance with screening. In another study (36). only 46% of the Mexican— American women had had a Pap test in the last 2 years. and the most important predictor of having had a recent Pap test was the number of close friends in the woman‘s social network. Native American women are also highly “unscreened”; 40% in a defined area of California had had a Pap test within the last year and only 22% in the last 3 years (37). Data from studies such as those described above allow iden- tification of groups of American women who are not adequately screened with Pap smears. What are some of the barriers to screening encountered by these women? Barriers to Pap smear screening fall into at least four categories: lack of knowledge on the part of the patient or health care provider. economic barriers. barriers inherent in the group‘s culture or belief system. and logistical barriers. In the analysis by Harlan et al. (5) of NHIS data from 1987. the most common reason given by women of all ages. eth— nicities. and income levels for not utilizing screening was not knowing the purpose of the Pap smear—not understanding it is intended to prevent cervical cancer. Other studies (26,28) reflect the importance of this barrier. particularly among women of low socioeconomic status and women of ethnic minorities. The second most common reason for lack of screening given in this survey of more than 12000 women was the physician never having told the woman she needed a Pap smear (5). This is par— ticularly important because many of the women who are not being screened are receiving care from internists or family prac- titioners who may not be aware of the high risk and lack of prior screening among their patients. particularly the elderly and poor women (2124.25.28.30). The cost of Pap smear screening is an obvious barrier for the poor and uninsured. Many of these women. particularly those who live in rural areas. may not have access to screening programs (28). For those living in poverty. the trials of day-to- day living. such as securing food and shelter. may make seeking preventive health services such as cancer screening a low priority (2.28.38). Less tangible factors involving the relation- ship between the patient and her health care provider can also be barriers to screening among poor and uninsured women. In a recent study (39). women were asked what they perceived as barriers to health care. Women receiving public assistance com- plained that physicians and other health care workers treated them with less respect and caring because they were on welfare. The attitudes and beliefs of a particular population about can- cer are an important barrier to obtaining Pap smears or any type of cancer screening that has been identified in several studies. Fatalistic attitudes about cancer seem to dominate the beliefs of many poor women and women of ethnic minority groups Journal of the National Cancer Institute Monographs No. 2 l. 1996 (5.40.41). Women who believe cancer is not curable avoid screening or any interaction that might result in being told they have cancer. Misconceptions about effects of treatment. such as the belief that surgery will cause cancer to spread by exposing it to air. were also found to be commonly expressed by women in lower socioeconomic groups and ethnic minorities (40.41). Lan— guage is another important barrier identified in many studies; Spanish-speaking women avoid contact with the English-only clinics and staff and thus avoid screening (5,42). Specific cul- tural barriers among Native American and Hispanic women may include a strong sense of the private nature of one’s state of health or disease (36.42.43). Inherent mistrust of the health care system based on previous negative experiences. such as the Tus- kegee experiment. may also prevent women of ethnic minorities from seeking cancer screening services (35.39). Logistical factors are frequently cited barriers to Pap smear screening. Lack of transportation. lack of child care. and being unable to leave duties of running the household or work. are all expressed as reasons for not being screened by the women inter- viewed in many studies. More subtle are the effects of the dif- ficulty negotiating the complex medical system for the poor and elderly. including long waiting times and multiple appointments. which may also be significant deterrents to screening (30.39— 42). Several published reports describe methods of overcoming the barriers to Pap smear screening in “hard to reach" popula- tions. Successful methods of overcoming the multitude of bar- riers to Pap smear screening facing elderly women and women of low socioeconomic status have generally followed one of three approaches. One method called “inreach” is based on the principle of offering screening at the time of any health care en- counter. Using nurse practitioners who offered same—day educa- tion and screening to elderly women attending an inner-city general medicine clinic. one group was able to markedly in- crease the percentage of women over age 65 years having Pap smears from 18% to 74% (44). In the more traditional “out- reach” intervention. computerized letters. phone calls. and other reminders have been successful in increasing the numbers of women obtaining screening Pap smears. in two studies by sevenfold (45.46). Another highly successful approach recruits community members as lay health workers who then help plan and execute Pap smear screening education (33.46-51). Focus groups have been successful tools for determining the needs of a particular population and attitudes toward cancer before plan— ning any intervention strategy (48). Successful interventions to increase Pap smear screening among the “unscreened” also must be targeted at cultural or eth- nic barriers unique to the population at risk. Examples are using the importance of the church in the urban black community to design an outreach program that trains church leaders to become lay health educators (33). The Talking Circle Project uses a cul- turally acceptable mode of communication and incorporates ap- propriate stories. myths. and legends to increase screening among Native American women (43). A common feature of most successful interventions is involving community members from the initial phases of the project and empowering women to become more active agents in determining their health status (33.37.39.41.43.46-51). To overcome economic barriers to Pap smear screening, many screening programs offer free services. transportation. and babysitting services (44-51). However. the problems of poverty and lack of access to care in the United States require broader. more far-reaching solutions that are beyond the scope of any single targeted screening program. The increasing spread of managed care into our medical care system may have a positive effect on Pap smear screening rates. as two recent studies (52,53) suggest. In an analysis of the 1992 NHIS data. women participating in health maintenance organizations (HMOs) were significantly more likely to have adequate Pap smear screening than those with private insurance or the uninsured (52). A review of more than 10000 patient records from a single HMO in California found that the number of prepaid patients per primary care physician was the only factor positively correlated with performance of Pap smears. The author concluded that large managed care organizations may have tnore effective methods to encourage Pap smear screening by primary care pro- viders (53). Despite the widespread use of Pap smears in the United States. many women remain unscreened or underscreened for cervical cancer: these are the elderly. women of low socio— economic status. and members of ethnic minorities. particularly older African—American. Hispanic. and Native American women. The barriers to these groups obtaining screening Pap smears include lack of knowledge about screening on the part of the patients and the physicians delivering their general medical care. poverty and lack of access to care. culturally based fatalis— tic attitudes toward cancer. and language barriers. What can be done to improve screening rates for cervical can- cer in these populations? Organized educational efforts targeted to the health care providers of elderly and poor women to in- crease awareness of their high risk for cervical cancer and high rate of being unscreened or underscreened should be developed. Community—based screening programs designed to address the specific needs of each underscreened group should be solicited and funded by governmental and private agencies. Programs tar- geted toward Hispanic women are desperately needed since they represent the fastest growing segment of our population and have among the lowest rates of cervical cancer screening. but the highest incidence rates. Research studies of other potentially underscreened populations. such as southeast Asian-Americans (54), other immigrants. and lesbians (55). who are known to have lower rates of interaction with the health care providers than other women should be performed. By addressing these is- sues. the low screening rates for cervical cancer in certain sub— groups of the U.S. population can be improved. thereby resulting in lowering of incidence of and mortality rates among these women. References (l) Parker SL. Tong T. Bolden S. Wingo PA. Cancer statistics. 1996 lulu com- ment citation in Medlinel. CA Cancer] Clin 19961351527. (2) Freeman HP. Cancer in the socioeconomically disadvantaged lsw com» ment citation in Medlinel. CA Cancer] Clin 1989;39:266—811. (3) Mahoney MC. Michalek AM. Cummings KM. Nasca PC. Emrich LT. Cancer mortality in a northeastern Native American population. Cancer 1989;64:187-90. (4) (5 (6) (7 ((3') (9) (I(} (II) (/2 (/3) (/4) (/51 (/6) (I71 (/8) (/9) (2(1) (2/) Mandelblatt J. Gopaul I. Wistreich M. Gynecological care of elderly women. Another look at Papanicolaou smear testing. ]AMA 19861256: 3677 |. Harlan LC. Bernstein AB. Kcssler LG. Cervical cancer screening: who is not screened aitd why? Atn] Public Health 1991;81:885-90. Chen F. Trapido E]. Davis K. Differences in stage at presentation of breast and gynecologic cancers among whites. blacks. and Hispanics. Cancer 1994;73:2838-42. Eddy DM. Screening for cervical cancer lu'e comment citations in Med- linel. Ann Intern Med 1990;113:214-26. Miller AB. Anderson G. Brisson J. Laidlaw ]. Le Pitrc N. Malcolmson P. et al. Can Med Assoc] 199111-15: 130125. US. Preventive Services Task Force: Guide to Clinical Preventive Ser» vices. Baltimore: Williams & Wilkins. 1989:26-31. American College of Obstetricians and Gynecologists. Cervical cytology: evaluation and management of abnormalities. ACOG technical bulletin number 183. Washington (DC): American College of Obstetricians and Gynecologists. 1993. American College of Obstetricians and Gynecologists. Recommendations of frequency of Pap test screening. ACOG committee opinion number 152. Washington (DC): American College of Obstetricians and Gynecologists. 1995. Boyce .IG. Fruchter RG. Romanzi L. Sillman FH. Maiman M. The fallacy of the screening interval for cervical smears. Obstet Gynecol 1990;76:627- 32. National Center for Health Statistics. Public use file documentation. Na- tional Health Interview Survey of topics related to cancer control. 1992. Hyattsville (MD): National Center for Health Statistics. 1993. 1987 National Health Interview Survey Cancer Control. Public use record. Hyattsville (MD): National Center for Health Statistics. 1989. Benson V. Marano MA. Current estimates from the National Health Inter— view Survey. Vital Health Stat 1994;1(1: 1-269. National Center for Health Statistics. Healthy People 2000 Review. 1993. Hyattsville (MD): Public Health Service. 1994. ]anerich DT. Hadjiinichael (). Schwart/ PE. Lowell DM. Mcigs ]W. Merino M]. et al. The screening histories of women with invasive cervical cancer. Connecticut [sec comment citation in Mcdline]. Am ] Public Health 1995;85:791-4. Nasca PC. Ellish N. Caputo TA. Saboda K. Met/.ger B. An epideiniologic study of Pap screening histories in women with invasive carcinoma of the uterine cervix. N Y State] Med 1991;91:152-6. Anderson LM. May DS. Has the use of cervical. breast. and colorectal can- cer screening increased in the United States“? Am ] Public Health 1995;85:840-2. Mortality. In: Vital Statistics of the United States. 1976. Vol 2. part A. 1976. Mamon ]A. Shediac MC. Crosby CB. Sanders B. Matanoski GM. Celentano DD. [liner-city women at risk for cervical cancer: behavioral and utiliza- tion factors related to inadequate screening. Prev Med 1990;19:363-76. Hayward RA. Shapiro MF. Freeman HE. Corey CR. Who gets screened for cervical and breast cancer? Results from a new national survey. Arch 1n- tern Med 198811-1821 177-81. Norman SA. Talbott EO. Kuller LH. Stolley PD. Dunn MS. Baffone K. et al. The relationship of Papanicolaou testing and contacts with the medical care system to stage at diagnosis of cervical cancer. Arch Intern Med 1991;151:58-64. Celentano DD. Shapiro S. Weisman CS. Cancer preventive screening be» havior among elderly women. Prev Med 198211 1:454-63. Montano DE. Phillips WR. Cancer screening by primary care physicians: a comparison of rates obtained from physician self-report patient survey. and chart audit. Am] Public Health 1995;85:795-800. Schapira DV. I’amies R]. Kumar NB. Herold AH. Van Durtne DJ. Woodward L]. el al. Cancer screening Knowledge. recommendations. and practices ofphysicians. Cancer 1993;71:839—43. Makuc DM. Freid VM. Kleinman ]C. National trends in the use of preven- tive health care by women. Am] Public Health 1989;79:21-6. Breast and cervical cancer screening among underserved women. Baseline survey results from six states. The National Cancer Institute Cancer Screening Consortium for Underserved Women. Arch Fam Med 199514: 617-24. Kat/ S]. Hofer TP. Socioeconomic disparities in preventive care persist despite universal coverage. Breast and cervical cancer screening in ()ntario and the United States. JAMA 1994;272:5304. Hiininelstein DU. Woolhandler S Care denied: US residents who are un» able to obtain needed medical services. Am] Public Health 1995;85:341-4. Habenstreit B. A comparative study of health care behavior among three black ethnic groups. Diss Abstr Int 1988149: 1501. Ackerman SP. Brackbill RM. Bewerse BA. Cheal NE. Sanderson LM. Cancer screening behaviors among U.S. women: breast cancer. 1987— | 989. Journal of the National Cancer Institute Monographs No. 21. 1996 (33 (34) (35) (36) (37) (38) (39) (40) (43) and cervical cancer. 19884989. MMWR CDC Surveill Summ 1992:4117- 25. Davis DT. Bustamante A. Brown CP. Wolde-Tsadik G. Savage EW. Cheng X. et al. The urban church and cancer control: a source of social in— fluence in minority communities. Public Health Rep I994:l()9:5()()—o. Calle FF Flanders WD. Thun M]. Martin LM. Demographic predictors of mammography and Pap smear screening in US women, Am J Public Health |993z83253—60. Jepson C. Kessler LG. Portnoy B. Gibbs T. Black-white differences in can- cer prevention knowledge and behavior. Am J Public Health l99l;8| :50I- 4. Suarel. L. Lloyd L. Weiss N. Rainbolt T. Pulley L. Effect of social net- works on cancer»screening behavior of older Mexican-American women. J Natl Cancer Inst 1994;86:775-9. American Indian Health Care Association Urban Native American Women's Cancer Prevention Project. St. Paul Minnesota, 1989. Weinrich S. Coker AL. Weinrich M. Eleazer GP. Greene FL. Predictors of Pap smear screening in socioeconomically disadvantaged elderly women. J Am Geriatr Soc 1995;43:267-70. Leppert PC. Partner SF. Thompson A, Learning from the community about barriers to health care. Obstet Gynecol [9962871404. Gregg J. Curry RH. Explanatory models for cancer among African- American women at two Atlanta neighborhood health centers: the implica- tions for a cancer screening program. Soc Sci Med |994;39:519—26. Loehrer PJ Sr. Greger HA. Weinberger M. Musick B. Miller M. Nichols C. et al. Knowledge and beliefs about cancer in a socioeconomically disad— vantaged population. Cancer I991 :68:2665-7l, Marcus AC. Crane LA. Kaplan C. Barriers to health care among low in» come Latino and black women with abnormal Pap smears (meeting abstract). North American Conference on Cancer in Hispanics. September 12—I4. l99l. Tucson (AZ): Arizona Cancer Center. the University of Arizona College of Medicine A14. [991. Hodge FS. Fredericks L. Rodriguez B. American Indian Women's Talking Circle: a cervical cancer screening and prevention project. Cancer 1996: 78:]592-7. Journal ofthe National Cancer Institute Monographs No. 2|. I996 (4-1) (46) (47) (4s (49) (5!) (5/) 'u. Is) (53 l (54) (55) Mandelblatt J. Traxler M. Lakin P. Thomas L. Chauhan P. Matseoane S. et al. A nurse practitioner intervention to increase breast and cervical cancer screening for poor. elderly black women. The Harlem Study Team. J Gen Intern Med l993;8: 173-8. Lantl. PM. Stencil D. Lippert MT. Beversdorf S. Jaros L. Remington PL. Breast and cervical cancer screening in a low-income managed care sample: the efficacy of physician letters and phone calls. Am J Public Health 1995;85:834-6. Ansell D. Lacey L. Whitman S. Chen E, Phillips C. A nurse~delivered in- tervention to reduce barriers to breast and cervical cancer screening in Chicago inner city clinics. Public Health Rep l994;l()9:l()4-| I. Brownstein JN. Cheal N. Ackermann SP. Bassford TL. Campos-Outcolt D. Breast and cervical cancer screening in minority populations: a model for using lay health educators. J Cancer Educ [99:17:321-6. Dignan M. Michielutte R. Sharp P. Bahnson J. Young L. Beal P. The role of focus groups in health education for cervical cancer among minority women. J Community Health 1990; [4:369-75. Rudolph A. Kahan V. Bordeu M. Cervical cancer prevention project for inner city black and Latina women. Public Health Rep 1993108: l56—60. Curry RH. Moen JB. Morris SJ. Scheivelhud L. Community—directed can— cer screening program [we comment citation in Medlinel. Cancer Pract 1994:22256-63. Banner RO. DeCambra H. Enos R. Gotay C. Hammond 0W. Hedlung N. et al. A breast and cervical cancer project in a native Hawaiian community: Wai'anae cancer research project. Prev Med I995z242447-53. Makuc DM. Freid VM. Parsons PE, Health insurance and cancer screening among women. Advance Data No 254: Vital Health Stat l995:l5:l-7. Harris JM Jr. Papanicolaou smear recommendations. patient complaints. and patient satisfaction in managed-care medical organizations. Med Care I995z332272-9. Carey P. Gjerdingen DK. Follow-up of abnormal Papanicolaou smears among women ofdifferent races. J Fain Pract 1993;37:583-7. Bybee D. Michigan Lesbian Health Survey: a report to the Michigan or- ganization for human rights and the Michigan Department of Public Health. March 1990. The Bethesda System, the Pathology of Preinvasive Lesions, and Screening Technology Michael Henry* The Bethesda System (TBS) of Nomenclature for Cervical Smears The effective communication of results between pathologist and clinician is a critical process that has a direct impact on patient care. It is the pathologist’s responsibility to provide a diagnostic interpretation in terminology that clearly conveys to the patient’s physician the significance of the cellular findings. In the late 1980s. despite the well-recognized deficiencies of the original Pap classification in communicating relevant diagnostic information. many laboratories routinely included a numeric class designation in the laboratory report (1). While many of these laboratories also used descriptive diagnostic terminology in conjunction with the Pap classification, often the class num— ber took precedence and dictated the clinical management of the patient. In addition. a proliferation of poorly defined descriptive terms. such as benign atypia. keratinizing atypia. and koilocytic atypia, were being used loosely with widely varying connota- tions. The resulting ambiguity made it difficult to compare results from different laboratories and pathologists and. at times. jeopardized patient care. In December 1988. the National Cancer Institute (NCI) spon- sored a workshop to address this diagnostic chaos (2) and to develop a uniform descriptive diagnostic terminology for cervi- cal—vaginal cytopathology. At the start of the meeting. the par— ticipants unanimously concluded that use of the Pap classification was no longer acceptable medical practice. The workshop then focused on the goal of developing a uniform ter- minology that would provide clinically relevant diagnostic in— formation. What emerged was not simply a listing of terms, but rather a format for reporting cervical—vaginal cytologic diag- noses that became known as TBS (3). A second NCI meeting was convened in April 199! to evaluate the impact. advantages. and disadvantages of TBS in actual practice. A criteria committee was formed to address the lack of uniform cytologic criteria needed to implement TBS. This committee developed the criteria for both the specific diag- nostic terms used in TBS and those used in determining specimen adequacy. These findings were published as a TBS reference atlas with accompanying photomicrographs (4). The advantages of TBS can be summarized as follows: 1) TBS provides uniform diagnostic terminology to facilitate un- ambiguous communication between the laboratory and the clinician: 2) TBS incorporates specimen—adequacy evaluation as an integral part of the report. and 3) TBS eliminates Pap class numbers. Journal of the National Cancer Institute Monographs No. 2 l. 1996 TBS has three basic elements: 1) statement of specimen ade— quacy. 2) general categorization. and 3) descriptive diagnoses. Specimen adequacy is evaluated both by the physical nature of the specimen and by the clinical information supplied by the submitting clinician. Samples are declared either totally satisfac- tory for evaluation. satisfactory but limited due to some factor, or unsatisfactory for evaluation. Both limited and unsatisfactory specimens must have an explanation of the limiting factor(s). Unsatisfactory samples may be due to nonreparably broken slides, slides or requests lacking the patients” identification, uninterpretable samples due to obscuring or distorting factors. or samples that lack enough squamous cells. Unsatisfactory samples do not receive a morphologic diagnosis. Limited samples may be due to lack of clinical information on the re— quest. lack of an endocervical component on the slide(s), or samples that, while interpretable. have some limiting obscuring or distorting artifacts such as air drying or inflammation. Satis- factory but limited specimens do receive a descriptive diagnosis. The general categorization consists of: 1) within normal limits. 2) benign cellular changes: see descriptive diagnosis. and 3) epithelial cell abnormality; see descriptive diagnosis. This is provided as a clerical device to aid clinicians in prioritizing results. The benign cellular changes category includes both in— fectious and reactive entities and does not indicate any degree of epithelial abnormality. Any cellular abnormalities are reported under epithelial cell abnormality and must be accompanied by a descriptive diagnosis. TBS breaks epithelial abnormalities down into squamous and glandular origins. Squamous abnormalities are divided into four categories: atypical squamous cell of undetermined significance (ASCUS). low-grade squamous intraepithelial lesion (LSIL). high-grade squamous intraepithelial lesion (HSIL), and squamous cell carcinoma. Atypia has long been an overused term meaning anything from benign reactive changes to preinvasive cellular changes. In TBS, atypia is not to be used for specimens interpreted to be benign; instead. the diagnosis of ASCUS is restricted to those cases in which cellular changes exceed those of benign. reactive processes. but which fall short of a definitive diagnosis of squamous intraepithelial lesion (SIL). Laboratories should strive to minimize use of this term; it is not a license for equivocation. *Conm/mmlence to: Michael Henry. M.D.. Department of Laboratory Medicine. Anatomic Pathology/Cytopathology. National Naval Medical Center, 890] Wisconsin A\'e.. Bldg. 9. Rm. 3497. Bethesda. MD 20889-5000. 13 This diagnostic category encompasses a variety of diagnostic dilemmas including 1) cases of probable marked atypical reac- tive/reparative changes that cytologically go beyond the range of confidence that the changes are unequivocally benign. 2) cases in which there are features suggestive of SIL. but criteria for definitive diagnosis are lacking. and 3) atypical cellular changes associated with atrophy. When possible. the diagnosis of ASCUS should be further qualified as to whether a reactive or SIL process is favored. Such comments may help guide the clinician in patient management. SIL is the only new term introduced in TBS; this was not done casually. but only after considering the strengths and weaknesses of previous terminology. In TBS. LSIL and HSIL encompass the spectrum of precursors to squamous cell car- cinoma of the cervix. LSIL includes changes of human papil— lomavirus (HPV) and mild dysplasia/CIN I. HSIL incorporates moderate dysplasia/CIN II. severe dysplasia/CIN III. and car— cinoma in situ/CIN III. The inclusion of HPV within SIL re— quires that one use strict criteria for the diagnosis to avoid overdiagnosis and unnecessary treatment of women for non— specific morphologic changes. Overdiagnosis of HPV. in part due to (werinterpretation of any cytoplasmic halo as koilocytosis and use of nonclassic cytologic signs of condyloina. is a signif- icant problem (5). Significant glandular lesions in TBS include atypical glan- dular cells of undetermined clinical significance (AGUS) and adenocarcinoma. Slides diagnosed with AGUS contain glan— dular cells demonstrating changes beyond those encountered in benign reactive processes yet which are insufficient for a diag— nosis of invasive adenocarcinoma. This diagnosis should be fur- ther qualified. if possible. to indicate whether the cells are thought to be of endocervical or endometrial origin. This category includes a broad morphologic spectrum. rang- ing from atypical-appearing, reactive processes all the way to adenocarcinoma in situ (AIS). Therefore. lesions falling into the AGUS category should be further subclassified according to whether a reactive or a premalignant/malignant process (AIS) is favored. In some cases. it may be difficult to differentiate SIL with gland involvement from AIS. Additionally. squamous and glandular intraepithelial lesions may coexist. Criteria for separating atypical endometrial cells of undeter- mined significance into probably reactive versus probably pre— malignant/malignant categories are not well-defined; therefore. this diagnosis is not subdivided further. Most invasive endocervical adenocarcinomas are preceded by precursor lesions that have been designated AIS. As noted above. lesions thought to reflect AIS are reported as atypical en- docervical cells. probably premalignant/malignant. However. there is considerable morphologic overlap between this category and invasive adenocarcinoma. Criteria indicating invasion— tumor diathesis and macronucleoIi—may be absent in the majority of well-differentiated. early adenocarcinomas. There— fore. the differentiation of atypical endocervical cells of undeter- mined significance, probably AIS from invasive carcinoma. often requires histologic evaluation. 14 Pathology of Preinvasive Disease The rationale for the two—tiered SIL terminology is based on current knowledge of the pathogenesis of cervical cancer includ- ing I) inherent limitations of light—microscopic morphologic diagnosis to distinguish lesions and predict progression. 2) cur- rent concepts of the molecular virology of HPV. and 3) the clinical behavior and management of cervical lesions (6). Numerous studies have demonstrated the lack of reproduci- bility in cervical cytology in subdividing squamous intra— epithelial lesions. Perhaps the most difficult distinction by light microscopy is that between HPV and slight dysplasia or cervical intraepithelial neoplasia I (CIN I) (7). This lack of repro— ducibility also extends to cervical biopsies. Recently. it has been suggested that a two—tier system be used for histopathology as well as cytology. In the past. it was thought that all cervical preinvasive disease progressed in a slow fashion from mild dysplasia (CIN I) through moderate dysplasia (CIN II) to severe dysplasia (CIN III) and finally to invasive carcinoma. Recent advances in molecular biology have revealed that most. if not all. prein- vasive disease is directly related to HPV infection. Following an initial infection. the virus may either undergo replication or be— come latent. The cellular features of a productive infection cor— relate with those of a low-grade intraepithelial lesion. In many cases. these lesions either resolve or become latent. In some cases. further transformation of the cells can lead to a high— grade lesion that then may either progress. regress. or possibly resolve (Fig. l). The subtype of the virus appears to have a great effect on the progression of the lesion. Low—. intermediate—. and high—risk types have been identified (Table 1). HPV molecular analyses have revealed a heterogeneous distribution of low- and high-risk HPV types in both HPV only and CIN I lesions (8.9). By contrast. HSILs show a more homogeneous pattern of ex— pression of high-risk HPV types. notably HPV 16. Longitudinal studies of HPV only and slight dysplasia/CIN I lesions have shown virtually identical rates of progression to high-grade lesions (10). Therefore. if HPV alone and slight dysplasia/CIN I cannot be reliably distinguished morphological— ly. and if these lesions demonstrate similar patterns of behavior. then it is reasonable that they be managed similarly. Current management of high-grade lesions is based principally on the / Initial Infection \ Latent Infection <——> Productive Infection (LSIL) Transformation (HSIL) 7/ Resolution of Disease Invasive Carcinoma Fig. 1. Possible outcomes of human papillomavirus infection. HSIL = high— grade squamous intraepithelial lesion: LSIL = low—grade squamous in- traepithelial lesion. Journal ofthe National Cancer Institute Monographs No. 21. I996 Table l. Hutnan papillomavirus (HPV) trophisms" Location Disease HPV types Mucosa Laryngeal carcinoma 30 Oral papilloma 32.57 Bowenoid papulosis 55 Cervix Condyloma 6.11.4254 lntraepithelial lesion Low risk 6.1 1.42.43.44 Intermediate risk 31.33.35.51525859 16.18.4556 High risk *Modified from (8). size. location. and extent of the abnormality rather than the pathologic grade ofCIN II or CIN 111. Screening Technology Cervical smears were traditionally obtained with a modified Ayers wooden spatula and cotton swab. Recently. new collec— tion devices have been developed. Some of these devices enable the clinician to obtain a better endocervical sample (endocervi— cal brushes). and others enable the collection of both an ectocer— vical and endocervical sample with one device (cervical brooms). Using an endocervical brush together with a spatula increases the detection of high-grade squamous lesions and yields increased numbers of smears with an endocervical com- ponent (1]). The use of these devices has been widely accepted by clinicians. Currently. cervical samples are smeared on one or two glass slides and then fixed with either alcohol or a spray fixative. Op- timal visualization requires well—fixed cells that are evenly spread out on the slide. Poor fixation. thick smears. and obscur— ing inflammation or blood commonly make evaluation difficult. Recent innovations in specimen processing may eliminate some of these problems. Methods using liquid—based collection and thin—layer processing are currently under development for cervi— cal samples. Several studies (l2,]3) have indicated that not only are the liquid—based preparations less likely to be unsatisfactory. but they also significantly increase the cellular yield. The downside of these preparations is the increased cost involved and an apparent slight change in the cellular criteria for squam— ous intraepithelial lesions. The concept of computer—assisted or automated screening for cervical smears has been a dream of cytopathologists for years. However. the complexity of the computer programs and the im— aging capability required for accurate diagnosis has made automated evaluation of Pap smears impossible until recently. In the past few years. advances in computer technology and image analysis have led to the development of several automated screening devices (ASD) (14-18). In September and November 1995. the Food and Drug Ad- ministration (FDA) approved two of these devices for computer- assisted selection of cases for manual rescreening. One ASD (AutoPap 300 QC System. Neopath. 1nc.. Redmond. WA) was approved for quality-control rescreening. In brief. the AutoPap 300 reviews negative cases and selects a population with an in— creased risk of abnormality. The individual laboratory can then Journal of the National Cancer Institute Monographs No. 21. 1996 decide what percentage of these cases to manually rescreen. This procedure replaces the random 10% rescreen performed by all laboratories. A study (19) indicates that. at a 109% rescreen rate. more than 50% of missed abnormal cases will be detected. The other ASD (PAPNET. Neuromedical Systems. Inc.. Suf— fern. NY) was approved as an adjunct test following manual screening. Cases to be rescreened by PAPNET are sent to a PAPNET laboratory where the slides are screened and the most abnormal 128 images on each case are placed on a data-storage tape. The slides and tape are returned to the original laboratory where the tape is revrewed by a cytotechnologist. Abnormal cases detected on the tape are then confirmed or rescreened manually. Both of these machines have shown that they are capable of detecting abnormalities. and thus it appears to many that a breakthrough in automated screening is at hand. However. to consider that ASDs are currently a safe and cost- effective method for routine cervical—smear screening is not yet true. First and foremost. no ASD is yet approved by the FDA for primary screening. Clinical trials are currently being performed and the initial data seem to indicate that ADSs. as currently con- figured. may show some use in routine screening. However. most of the testing of ASDs has been retrospective and per— formed on small populations of known false-negative cases or manually screened negative cervical smears (20.21). Additional data obtained in large—scale. prospective. controlled studies will be necessary before these devices can be used routinely in pri— mary screening. The cost-effectiveness of these machines in a rescreening mode must also be questioned. A recent evaluation of various methods for cervical—smear rescreening suggests that the cost of finding additional abnormal cases (LSIL and up) would range from a low of $348 per case for a manual rapid rescreen to a high of $4486 per case for PAPNET (22) (Table 2). In addition. it was estimated that 100% manual rescreening by a cytotech- nologist would yield tnore new abnormal cases (233 for manual rescreen. 224 for PAPNET. and 222 for AutoPap 300) than ASD rescreen and would cost less per abnormal case found ($1049 for manual rescreen. $4486 for PAPNET. and $2197 for AutoPap 300). These estimates are very conservative and the true cost of detecting new abnormal cases by ASD would probably be much higher. In conclusion. the technology for automated Pap smear screening is currently in its infancy. Advances are being made rapidly but the use of these devices remains to be determined. Table 2. Cost of rescreening strategies (50 000 cases/year) Abnormal Cost per Cost per additional Method cases found* abnormal case abnormal case No rescreen 938 $267 Not applicable 10(/( rescreen 961 $286 $1049 Directed rescreen 1000 $31 1 $981 100% rescreen 1 171 $423 $1049 Rapid rescreen 1055 $276 $348 AutoPap 300111 10% 1060 $490 $2197 PAPNET 1062 $760 $4486 *Assumes a prevalence of low—grade squamous intraepithelial lesion and greater of 2.5% and a false—negative rate of 25%. Modified from ref. ( I7). References (I) (2) (3) (4) (5 (6) Maguire NC. Current use of the Papanicolaou class system in gynecologic cytology. Diagn Cytopathol l988;4:169-76. Koss LG. The new Bethesda System for reporting results of smears of the uterine cervix. J Natl Cancer Inst 1990;82:988-91. The 1988 Bethesda System for reporting cervical/vaginal cytologic diag— noses. National Cancer Institute Workshop. JAMA 1989;262:9314. Kurman RJ, Solomon D. The Bethesda system for reporting cervical/ vaginal cytologic diagnosis: definitions. criteria. and explanatory notes for terminology and specimen adequacy. New York: Springer-Verlag. 1994. CeCchini S. Confortini M. Bonardi L. Cipparrone G. Galante L. lossa A. et a1. Nonclassic signs of cervical condyloma. A case—control study. Acta Cytol 1990;34:781—4. Kurman RJ. Malkasian GD Jr. Sedlis A. Solomon D. From Papanicolaou to Bethesda: the rationale for a new cervical cytologic classification. Obstet Gynecol 1991;77:779-82. Sherman ME. Schiffman MH. Kurman RJ. Erozan YS. Wacholder S. The Bethesda System: interobserver reproducibility of cytopathologic diag» noses. The Bethesda System Second Conference. April 1991. Meisels A. Morin C. Cytopathology of the uterine cervix. Chicago: ASCP Press. 1991. Stoler MH. Correlations of HPV gene expression with The Bethesda Clas— sification of low grade squamous intraepithelial lesions. The Bethesda Sys- tem Second Conference. April 1991. Kiviat N. Koutsky L. Critchlow C. Brockway J. et a1. Prospective evalua- tion of clinical utility of reporting changes of condyloma separately from (/7) (I8) (/9) those of low-grade dysplasia. The Bethesda System Second Conference. April 1991. Boon ME. de Graff Guilloud JC. Rietveld WJ. Analysis of five sampling methods for the preparation of cervical smears. Acta Cytol 1989;33:843-8. McGoogan E. Reith A. Would monolayers provide more representative samples and improved preparations for cervical screening? Acta Cytol 1996;4(1: 107-19. Sprenger E. Schwarzmann P. Kirkpatrick M. Fox W. Heinzerling RH. Geyer JW. et al. The false negative rate in cervical cytology. Comparison of monolayers to conventional smears. Acta Cytol 1996;40:81-9. Grohs DH. Gombrich PP. Domanik RA. Meeting the challenges in cervical cancer screening: the AcCell series 2000 automated slide handling and data management system. Acta Cytol 1996;411:2630. Zahniser D]. Sullivan PJ. CYTYC Corporation. Acta Cytol |99624():37—44. ) Patten SF Jr. Lee 13. Nelson AC. Neopath. Inc. NeoPath AutoPap 300 Automatic Pap Screener System. Acta Cytol 1996;40:45-SZ. Mango LI. Neuromedical Systems. Inc. Acta Cytol 1996;40:53-9. Knesel EA. Roche Image Analysis Systems. Inc. Acta Cytol 1996;40:611—6. Colgan T]. Patten SF Jr. Lee 15. A clinical trial of the AutoPap 3110 QC System for quality control of cervicovaginal cytology in the clinical laboratory. Acta Cytol 1995:3921 191-8. Ashfaq R. Liang Y. Saboorian MH. Evaluation of PAPNET system for rescreening of negative cervical smears. Diagn Cytopathol 1995; 13131-6. Slagel DD. Zaleski S. Cohen MB. Efficacy of automated cervical cytology screening. Diagn Cytopathol 1995;13:26-30. Hutchinson ML. Assessing the costs and benefits of alternative rescreening strategies [editorial]. Acta Cytol 1996;40:4—8. Journal of the National Cancer Institute Monographs No. 21. 1996 Cervical Human Papillomavirus Infection and Intraepithelial Neoplasia: a Review Michele Fallen Mitchell, Guillermo Tortolero-Lima, Thomas Wright, Asis Sarkar, Rebecca Richards-Kortum, Waun K. Hang, David Schottenfeld * Cervical human papillomavirus (HPV) infections and in- traepithelial neoplasias are precursors to cervical cancer, the second most common cancer in women worldwide. HPV satisfies the epidemiologic criteria for causality; the role of other cofactors is under study. Natural history studies show that most low-grade lesions (productive HPV infections) regress or persist, whereas high-grade lesions (those with in- tegrated HPV DNA) progress. Immunobiologic studies demonstrate that infection peaks in the early 205, leading to a 10- to 20-year period of persistent infection, before finally progressing to a preinvasive or invasive lesion. Papanicolaou (Pap) screening has lowered the morbidity and mortality from cervical cancer in every country in which screening programs have been introduced. The diagnostic strategy for an abnormal Pap smear includes colposcopy; the role of HPV DNA testing in screening or diagnosis remains unclear. Patients are treated with cervical ablation, cone biopsy, or chemopreventive agents. Efforts to strengthen screening and prevention, as well as new directions for research, are needed. [Monogr Natl Cancer Inst 1996;21:17-25] Definition and Classification Cervical neoplasia is a term used to describe both premalig- nant (cervical intraepithelial neoplasia [CIN]) and invasive (in- cluding tnicroinvasive) lesions of the uterine cervix. The term CIN was proposed by Richart (l) to define the spectrum of in- traepithelial changes that precede invasive cervical cancer (ICC). These changes were classified into three grades: CIN-I. consisting of undifferentiated cells extending one third of the distance from the basement. membrane to the surface epithelium (mild dysplasia); CIN-2. characterized by extension to two thirds of this distance (moderate dysplasia); and CIN-3. charac— terized by extension through more than two thirds of this dis— tance (severe dysplasia and carcinoma in situ [CIS]). In contrast. the term ICC is used to describe those lesions in which the malignant cells penetrate the underlying basement membrane and infiltrate the stroma (2). The pathologic diagnosis of in- traepithelial neoplasia is based on seven morphologic criteria: increased nuclear size. altered nuclear shape. increased nuclear stain uptake. nuclear pleomorphism. increased mitosis. abnor— mal mitosis. and disordered or absent maturation (2). As the pathobiology of these precursors increasingly focused on the role of human papillomavirus (HPV). its hallmark. koilocytotic Journal ofthe National Cancer Institute Monographs No. 21. I996 atypia, became a recognized part of the spectrum. preceding CIN-l (3) (Fig. I). With the purpose of increasing the agreement among ob- servers and taking into account the role that HPV may play in the pathologic continuum. a new classification system. the Bethesda System. was introduced in 1988 for reporting cervical smears (4). In this system. noninvasive squamous lesions are classified as atypical squamous cells of undetermined signif— icance (ASCUS). including lesions with cells presenting abnor- mal nuclear characteristics but without changes suggestive of koilocytotic atypia or CIN-l; low—grade squamous intra- epithelial lesions (LGSIL). including koilocytotic lesions and CIN—l; and high—grade squamous intraepithelial lesions (HGSIL). including CIN-Z. CIN-3. and CIS. This cytologic clas- sification is now being used to describe histopathologic lesions. The histopathologic correlates are atypias and inflammation for ASCUS. LGSIL for LGSIL. and HGSIL for HGSIL. The clinicopathologic role of atypias and inflammation in the spec- trum ofdisease is unknown. Incidence and Prevalence Cervical cancer is the second most common cancer in women worldwide. exceeded only by breast cancer. In 1985, ap- proximately 437 300 women were diagnosed with cervical can- cer. accounting for 12% of all cancers diagnosed in women worldwide (5). Incidence of and mortality rates for cervical can- cer are higher in many developing countries. where ap— proximately 8()% of all cases occur (5,6). In the United States, cervical cancer is the third most common neoplasm of the female genital tract. In 1995, 15 900 cases of ICC and 65 000 cases of CIS were expected to be diagnosed among US. women. and approximately 4900 US. women were expected to die of this neoplasm (7). The incidence of and mortality from *Af'fi'liuliunx of authors: M, F. Mitchell. 0. Tortolero-Luna (Department of Gynecologic Oncology). W. K. Hong (Department of Medical Oncology). The University of Texas M. D. Anderson Cancer Center. Houston: T. Wright. Department of Obstetrics/Cynecology and Pathology. Columbia University. New York, NY; A. Sarkar. Department of Veterinary Science. The University of Texas M. D. Anderson Cancer Center. Bastrop: R. Richards—Kortum. Depart- ment of Computer and Electrical Engineering. The University of Texas. Austin: I). Schottcnfeld. Department of Epidemiology. University of Michigan School of Public Health. Ann Arbor. Coi'rm/mm/em'c In: Michele Follen Mitchell. M.D.. M.S.. Department of Gynecologic Oncology. Box ()7. The University of Texas M. D. Anderson Can- cer Center. [515 Holcombe Blvd. Houston, TX 77030. See “Notes" section following "References." Squamous Intraepithelial Lesion (SIL) Low Grade High Grade Condyloma Cervical |ntraepithelia| Neoplasia (CIN) Grade 1 Grade 2 Grade 3 Moderate Very Mild-Mild Dysplasra as: Micromvasrve Carcinoma Fig. I. l’athologic continutmi of cenical intraepithelial neoplasia (CIN). (‘er\i- cal squamous carcinoma precursors: schematic representation of cervical cancer precursors and the different terminologies that have been used to refer to tltem. The risk of developing microinvasion from different states of squamous in- traepithelial lesions is arbitrarily represented and is not necessarily proportional to that illustrated in this scheme. Reproduced with permission from Springer- Verlag (32). cervical cancer are about two times higher among Hispanic and African-American women than among white women. Except for CIS. the precursor lesions of ICC. CIN is not a mandatory reporting condition, and the incidence and prevalence of CIN are therefore unknown. CIN is far more com— moit than ICC. The highest prevalence rates of CIN are ob— served in women attending sexually transmitted disease clinics. whereas the lowest rates are observed in adolescents and women attending family-planning clinics and obstetrics and gynecolog clinics (8). In the United States. prevalence estimates of CIN range front 1.05% in women attending family—planning or obstetrics and gynecology clinics to 13.7% in those attending a sexually transmitted disease clinic (8). It is estimated that. in 1992. approximately 250 000 to 1 million women in the United States were diagnosed with CIN (9). whereas Kurman et a1. (10) estimated that approximately 2.5 million women in the United States will have low—grade cervical abnormalities each year. Data from the National Breast and Cervical Cancer Early Detec- tion Program suggested an estimated 10.5% prevalence of cytologic abnormalities (11). Approximately 5.4% of the women with these abnormalities had a diagnosis of ASCUS. and 5.1% had a diagnosis of squamous intraepithelial lesions (4% LGSIL and 1.1% HGSIL). The prevalence of cervical abnor- malities. both ASCUS and squamous intraepithelial lesions. was higher among younger women (<30 years old) and decreased with increasing age. The prevalence of cervical abnormalities varied slightly by racial/ethnic group (11). Risk Factors Epidemiologic studies have long suggested a sexually trans- mitted etiology for cervical neoplasia. The risk of cervical 18 neoplasia is higher among women with multiple sexual partners. women who had their first sexual intercourse at an early age. and women whose sexual partners were more promiscuous (12). In addition. low socioeconomic status. reproductive history. smoking habits. oral contraceptive use. lack of diaphragm con- traceptive use. dietary factors. characteristics of sexual partners. immunosuppression. and decreased frequency of Papanicolaou (Pap) smears have been implicated as risk factors for cervical neoplasia (13). Early studies have focused on the etiologic role of several sexually transmitted agents. including herpes simplex virus type 2. Cltlunrvdiu truc/mmalix. T1‘1('/I()1)I()11(l.\' vaginal/s. cytomegalovirus. Neirseriu gmmrrlmcae. and Trepmmna pul— litlum (8). Since the middle 1970s. substantial evidence has ac— cumulated supporting the role of some types of HPV in the etiology of cervical neoplasia (14.15). Current epidemiologic data support a strong and central role of HPV in the etiology of cervical neoplasia in women worldwide. Furthermore. HPV in— fection appears to explain many of the established risk factors for cervical neoplasia. including sexual behavior and cigarette smoking. The association is particularly strong with specific high-risk HPV types. with increasing viral load. and with coin— fection with a different HPV type (16.17). This association has been independent of the HPV assay method used and the epidemiologic study design (15-21). Of the more than 70 types of HPV described today. more than 35 are associated with anogenital disease. and 2() or more are as— sociated with cancer (22). The most common HPV types detected in cervical lesions are those classified as high—risk HPV types (including types 16. 18. 31. 45. and 56). intermediate-risk types (including types 33. 35. 39. 51. and 52). and low-risk types (including types 6. 11. 42. 43. and 44). HPV DNA has been detected in more than 79% of specimens from women with definite cervical disease (23). The prevalence of HPV increases with the sensitivity of the detection technique: higher prevalence estimates are observed with the use of polymerase chain reac- tion (PCR) and with the use of repeat testing (up to fivefold higher prevalence estimates are observed) (24). In a recent study by the International Biological Study on Cervical Cancer Study Group (22). HPV DNA was detected in 93% (range. 75%— 1()()%) of cancer specimens collected from women in 22 countries and analyzed by the PCR technique. Similarly. HPV DNA has been detected by PCR in up to 94% of women with preinvasive lesions (CIN) (17.18,2().21.25) and in up to 46% of women with cytologically normal tissue (24). HPV type 16 is by far the most prevalent HPV type in both cytologically normal and abnormal women worldwide: it is present in approximately 50% of HGSIL and cancer (22-24). The association between HPV infection and cervical neoplasia satisfies all the criteria for causality in epidemiologic research: strength. consistency. and specificity of the associa- tion; a dose—response and temporal relationship; and biologic plausibility (15). However. while HPV is the most important risk factor for cervical neoplasia. the presence of other cofactors seems to be necessary for the development of disease (23.26). The difference between the high prevalence of HPV infection (1 1%-80%) in young healthy women and the low incidence of cervical neoplasia. as well as the low rate of progression of un— treated CIN lesions. supports the hypothesis that HPV may be a Journal ofthe National Cancer Institute Monographs No. 21. I996 necessary but insufficient factor for cervical neoplasia. Cofac— tors that may be important for the development of disease may be humoral and cell—mediated immunity to HPV. age at expo— sure to HPV. type of exposure, presence of other sexually trans— mitted diseases. and hormonal status. The introduction of more accurate and effective methods for detection of HPV infection has improved the opportunity for conducting larger epidemiologic studies to reassess the independent or joint effect of previously es— tablished risk factors with HPV. In addition. the role of HPV per- sistence in the progression of cervical neoplasia and the determinants of HPV persistence need further evaluation (/3). Natural History The concept that ICC develops from intraepithelial precursor lesions was proposed in the early l900s by clinicians who recognized that intraepithelial lesions that are histologically similar to ICC were frequently identified adjacent to invasive cancers. If left untreated. these lesions, which were referred to as CIS. were found to have a high capacity to progress to ICC (27.28). With the introduction of cytologic screening programs for cervical cancer. it became obvious that other types of in— traepithelial lesions of the cervix exist that are histologically less severe than CIS but that. if left untreated. also can progress to ICC; these lesions were referred to as dysplasia (3). In the 1960s. on the basis of the results of follow-up studies. electron microscopy. clonality studies. and tissue culture studies. Richart proposed that precursor lesions formed a biologic con- tinuum. which he termed CIN (1,29). The CIN continuum concept was based on the premise that precursor lesions share a similar etiology and biology and that, if left untreated. all have the capacity (albeit to varying extents) to progress to ICC. The natural history of cervical cancer precursors has been the subject of two reviews, which are summarized in Table l (29.30). Other sources of information regarding natural history are clinical studies of patients who returned for evaluation after having been lost to follow—up, retrospective studies of recur— rence after primary treatment, and population—based studies of screening programs yielding prevalence/incidence ratios and mathematical models [reviewed in (31)]. All these sources of infor- Tahle 1. Natural history of cervical intraepithelial neoplasia mation support the concepts that less than 2% of CIN lesions (all grades excluding C IS) progress to invasive cancer. that 16%-36% of CIN—3/C IS progress to invasive cancer. that treatment of C IS lowers the rate of progression to less than 2%. that the transition from un- treated CIN to C IS to ICC occurs over a 10- to 20—year period. that probably 25% or more of untreated C IS progresses to ICC. and that the transition time is longer at younger ages and shorter at older ages. Factors that may influence progression include the following: 1) viral type. viral load. and length of infection: 2) host immune func— tion. both humoral and cell mediated; and 3) environmental fac— tors. such as age. sexual behavior. diet. contraceptive use, and. possibly. smoking. Biology During the last decade. with the recognition that cervical can- cer precursor lesions are uniformly associated with HPV. it has become clear that the basic tenet underlying the CIN—continuum terminology is incorrect (3] ). The intraepithelial lesions referred to as CIN do not represent a single entity. but instead they ap— pear to consist of lesions of two types. Those lesions that are cytologically and histologically high grade are monoclonal pro- liferations of squamous epithelial cells that are usually aneu- ploid and almost always associated with “cancer—associated" HPV types such as types l6. 18. 31. 33. 35. 45. 51. and 58 (3/- 33). Prospective follow-up studies have demonstrated that these HGSIL are true “neoplasia” and tend to either persist or progress if left untreated. Those lesions with the highest "viral load" may be the most likely to persist or progress (34). Histologically and cytologically low grade lesions are charac- terized by marked HPV cytopathic effects. appear to represent productive HPV infections. and usually contain large numbers of virions. In contrast to HGSIL. LGSIL are quite hetero- geneous, They can be either monoclonal or polyclonal. are usually diploid or polyploid. and can be associated with any of the more than 25 types of HPV that infect the anogenital tract. LGSIL have a heterogeneous biologic behavior; although more than half spontaneously regress in the absence of therapy. ap- proximately 16% have the capacity to progress to high-grade lesions or ICC if they are left untreated (31.33.35). (CIN) in cohorts of untreated patients from two reviews“ Behavior of lesion ‘/r lesions progressed ‘/( lesions '/( lesions % lesions ‘Vp lesions to higher progressed progressed Study regressed persisted grade CIN to C13 to ICC Studies clustered by study design (29) All grades CIN followed by Pap smear only 34 41 25 10 | All grades CIN followed by Pap smear and biopsy 45 3] 23 14 I4 CIS followed by biopsy 7 — 36 7 36 Studies clustered by grade of CIN (31) CIN] 57 32 7 l l 7 CIN-2 43 35 7 22 7 CIN-3 32 56 — l2 7 Overall all grades ofClN 7 7 7 — 1.7 *CIS = carcinoma in situ: ICC = invasive cervical cancer; Pap : Papanicolaou: 7 = not available. Journal of the National Cancer Institute Monographs No. 2 l. |996 l9 The high-risk oncogenic HPVs produce two oncoproteins from the early—region open reading frames (ORF) E6 and E7. These oncoproteins interact with endogenous cell cycle regu— latory proteins; E6 interacts with p53. and E7 interacts with pRB. p107, and cyclin A (32.36). The interaction of virally derived proteins with cell cycle regulatory proteins leads to deregulation of cell cycle progression and appears to be a criti— cal step in cervical squamous carcinogenesis. Integration of the HPV DNA into the host genome frequently disrupts the E2 ORF. which controls transcription: this disruption results in the further overexpression of E6 and E7 (32.36). The E5 protein has weak transforming activity. whereas E4 is involved in matura- tion and viral replication. The capsid proteins are produced by late region ORFs‘. Ll is the major capsid protein, and L2 is the minor capsid protein. While HPV plays a critical role in the process of cervical car- cinogenesis. interactions with other oncogenes. such as ras and myc. may be important for the final transition to cancer (Fig. 2). Several groups are studying cervical carcinogenesis by looking at the relationship of these biomarkers to cervical lesions as they progress to cervical cancers. Markers under study include quan- titative histopathologic markers. proliferation markers. regula- tion markers. differentiation markers. and general genomic instability markers. When these markers can be shown to corre- late highly with the development of cancer. they are termed “surrogate end-point biomarkers ofcancer" (37) (Table 2). Immunobiology Prevalence studies of HPV in cytomorphologieally normal smears demonstrate age dependence. with peak infections oc- curring in women aged 20-24 years (38). Clinically, if the Pap smear is normal and no lesion is present colposcopically. these infections are referred to as "latent" or “subclinical.” Once a lesion is present cytologically and colposcopically. the HPV in- fection is assumed to be productive. and the HPV DNA may be episomal or integrated. These infections are believed to persist for 10-20 years. after which some regress. some persist. and some progress. Many groups have investigated the humoral response to these infections (39.40). Substantial evidence suggests that both the spontaneous re- gression and persistence of cutaneous HPV—associated warts are related to immunologic responses in the individual (4]). In general. in the majority of viral infections. evidence of im- munity is based on demonstrating the presence of virus—specific antibodies. However. the serologic analysis of HPV—associated mucocutaneous infections (clinical. subclinical. and latent) in patients has been hampered by the lack of a suitable antigen as a target and also the inability to propagate the virus. The initial at— tempts to detect antibodies to HPV used HPV virions prepared from pooled skin—wart biopsy specimens as the antigen (41.42). The advent of recombinant DNA technology allowed several groups to use bacterially expressed HPV viral proteins such as Ll. L2. E6. E7. and E4 as antigens (43-47). Other approaches for detecting HPV—specific antibodies included use of virus-like particles (48) and synthetic peptides corresponding to various HPV proteins as antigens (49). Although some studies demonstrated a statistically significant increase in E7 antibodies in patients with HPV—associated cervi— cal lesions as compared with normal individuals. other studies could not establish such correlations (39). Several investigators believed that the differences in findings were methodologic. Mutagenic co-Iactors Genomic instability induced by increased viral gene expression Upregulation and amplification Regression by cell-mediated immune functions (inhibited in acquired immunodeficiencies e.g. (AIDS) of viral DNA Fig. 2. Biologic role of human papillomavirus. Reproduced \ with permission from Aca- demic Press. Inc. (2). V'ral DNA Modification of Integration an d/or Mutation in - ' dditional host cell Infectlon _.) . _) host cell genes . . . _> a ' Persnstence regulating viral V'ra' ratifiagsnomic genes affecting gene expression mo ' Ica '°"s differentiation and angiogenesis p53? RB? Others? --------- )> . ) ) > _) .__) _) subclinical low grade squamous high grade squamous invasive cancer enhanced intraepithelial neoplasia intraepithelial neoplasia invasive growth, metastases ( ( <( --------- 20 Journal of the National Cancer Institute Monographs No. 2 I. I996 Table 2. Classes ofquantitative biomarkers ofcarcinogenesis in cervical epithelium* Histopathologic and cytologic markers Nuclei (abnormal size. shape. texture. and pleomorphism) Nuclcoli (abnormal number, size. shape. position. and pleomorphism) Proliferation markers PCNA Ki-67 Thymidine labeling index Mitotic frequency (MPM-2) Regulation markers Oncogenes (ras, myc) Tumor suppressors (p53. pRB) Polyamines Altered growth factors and receptors (EGFR. RAR) Differentiation markers Fibers (cytokeratins, involucrin. cornifin, actin microfilaments. and microtubules) Adhesion molecules (cell—cell: gap junctions and desmosomes; cell—substrate: integiins. cadherins. laminins. fibronectin. proteoglycans. and collagen) Glycoconjugatcs (mucins. blood group substances. and glycolipids) General genomic instability DNA abnormalities (point mutations, gene amplification. and LOH) Chromosome aberrations (micronuclei. double minutes. deletions. insertions. translocations. inversions. and isochromosomes) Ancuploidy (flow cytometry and image analysis) *PCNA = proliferating cellular nuclear antigen; MPM = mitotic phosphoprotein monoclonal antibody: EGFR = epidermal growth factor receptor; RAR = retinoic acid receptor: LOH : loss of heterozygosity, since some of the techniques used recognized linear epitopes (fusion protein immunoblotting and synthetic peptides) and others used recognized conformational epitopes (immuno- fluorescence. immunoprecipitation. and enzyme—linked im— munosorbent assays with whole proteins or viral particles). Since then. many studies have demonstrated statistically sig- nificant differences between control subjects and patients with cervical cancer or CIN. with relative risks ranging from 1.3 to 15.] (39,40). It will be important to investigate the immunobiol— ogy and natural history of lesions as they progress from initial infection through persistent infection to the formation of a lesion. The rejection of animal tumors expressing virally encoded tumor antigen is primarily mediated by cell-mediated immune responses (50). It has been reported that individuals with im— paired cell-mediated immune responses have an increased prevalence of HPV-induced lesions. including condyloma and CIN (51.52), indicating that T cells participate in the control of HPV—associated precancers in humans. Also. decreased produc- tion of the cytokine interleukin 2 diminished proliferative responses of patients‘ peripheral blood mononuclear cells against nonspecific mitogens and antigens (51.53). and down- regulation (i.e.. decreased expression) of cell—surface major his— tocompatibility complex (MHC) expression (54) was widely observed in CIN and ICC patients. In designing efficient immunoprevention and immunotherapy of HPV infection. it is important to understand the role of T-cell responses directed against various HPV proteins. Viral proteins present on the infected cell surface in the form of peptides. in the context of MHC molecules, activate autologous T cells (both helper T cells and cytotoxic T cells) expressing appropriate T- cell receptors. The E6 and E7 oncoproteins found in cervical biopsy specimens are therefore considered as attractive can- didates for directing immune intervention strategies. Chen et al. (55.56) have shown that a mouse melanoma cell line expressing Journal of the National Cancer Institute Monographs No. 2|. I996 either HPV-16/l8 E6 or E7 was specifically rejected by mice immunized with MHC-l syngeneic fibroblasts expressing E6 or E7. respectively. With the use of both an in vitro and in vivo ap— proach (57—6l). several peptides from the E6 and E7 on- coproteins have been identified that are capable of inducing HPV—specific T-cell responses in both animal models and humans. Results of humoral immunity studies suggest that initial cer— vical infections might be prevented by a humoral vaccine aimed at the capsid proteins L] and L2. possibly administered before the onset of sexual activity in population-based vaccination programs. Once persistent infection has occurred or a lesion is present. efforts might be better aimed at cell-mediated immunity using HPV type-specific peptides from E6 and E7 with the in- tention of causing regression of disease. These vaccines would be therapeutic and would be used in a well—defined group of women with disease. Screening and Diagnosis of Cervical Neoplasia The Pap smear is the primary tool for the screening of cervi- cal neoplasias. Although its efficacy has never been evaluated in a randomized trial, the decline in the number of advanced lesions and in the mortality rate for patients with cervical cancer observed in the last 40 years, as well as the increase in the early detection of invasive lesions and CIN. has been attributed main- ly to the introduction of this screening test. Approximately 50 million Pap smears are performed annually in the United States (10). and it is estimated that more than 90% of women 18 years old or older have had at least one Pap smear in their lifetime and that more than 67% have had one within the previous 3 years (1]). Identifying the barriers to screening utilization and strategies to overcome these barriers are important areas for fur- ther research. The Pap smear meets the requirements for an effective screening test and is the most cost—effective cancer—screening method (62). While high specificity (up to 99.8%) and high sen- sitivity (up to 85%) have been reported for the Pap smear. a tecent meta— analysis suggests an overall sensitivity of 58% fo1 screening and 66% for follow-up and a mean specificity of 69% for screening and 66% for follow—up (63.64). The Pap smear has an estimated false-negative rate of 5%-40‘7( (62.65.66). In recent years. the accuracy of cervical cytologic screening has been questioned (67.68). Inadequate sampling of the endocervix (67) and inadequate examination of smears are the most important factors contributing to the false- negative 1ates( (69) )Sampling error may be responsible for up to 60% of the false—negative smears. and examination error may account for up to 40% of the false-negatives (63). Once an abnormal Pap smear is detected. the standard of care is referral for colposcopy. Most physicians refer ASCUS and LGSIL for colposcopy. although the clinical yield is low and most lesions regress. A National Cancer Institute-funded study (the ASCUS/LGSIL Trial) has been initiated to address the This trial will include 7000 women nationally and will evaluate colposcopy. cervicography. and HPV DNA testing as a triage strategy for lesions more like— ly to progress to higher grades. management of low—grade lesions. At the first colposcopic visit. the patient should have a com— plete medical history taken as well as a physical examination. a pelvic examination. a repeat Pap smear. and a pan—colposcopic examination (including vulva. perineum. vagina. and cervix). The colposcopic examination includes the application of 3%—6% acetic acid followed by a careful 3— to 6-minute examination of the cervix identifying the endocervical canal. squamo—columnar junction. transformation zone. and lesions (Fig. 3). Typically. lesions turn aceto-white and may or may not exhibit angio— genesis. for which a descriptive classification has been de— veloped. Biopsy specimens are taken from abnormal areas. and an endocervical curettage is performed. Additional abnormal areas in the vagina. vulva. and perianal area are inspected. and biopsy specimens are taken from these areas. Several approaches. including targeted rescreening. colpo» scopic screening. cervicography (in which a picture of the cer— vix after acetic acid placetnent is reviewed by an expert). HPV DNA testing (differentiating low—risk and high—risk viral types). speculoscopy (viewing the cervix under chemiluminescent il— lumination). digital-imaging colposcopy. fluorescent spectros- copy. and computer-assisted automated image analysis of Pap smear readings. have all been considered to improve the perfor- mance of screening cytology and/or to better diagnose high— grade lesions at the time of the initial visit (70-78). While several studies have been performed and demonstrate that the addition of any of these techniques improves sensitivity. few studies have used receiver operating curves to assess improve— tnents. and fewer yet have assessed cost-effectiveness. In those studies that have used receiver operating curves. the curve is shown to improve: however. in studies of cost. the cost of addi- tional testing is as expensive as colposcopy (79.80). 22 Columnar epithelium Atypical vessels Mosaicism 51:: Squamous Early mosaicism e ith ' p ellum Coarse punctation Fine punctation Transformation zone Thick white lesion Squamo-columnar junction B C f l . 6-8 mm l Includes Includes transformation transformation zone and zone and lesion lesion Fig. 3. A) Cervix as seen at colposcopy. B) (‘el'\ ix rcmmed with ablation tech nique. C) Cervix removed with com/anon tcclunques. Treatment High—grade intraepithelial lesions. once identified. are treated with ablative methods. cone biopsy. or chemopreventive agents. Low—grade intraepithelial lesions are treated similarly or are simply followed every 6 months with repeat Pap smears and colposcopy. Biopsy—proven atypias. the histopathologic corre« late of ASCUS. are followed without treatment. The ASCUS/LGSIL trial will also address the natural history of these lesions. hopefully giving us guidance as to which lesions require treatment or close follow—up. Typically. patients are tri— aged. as diagrammed in Fig. 4. to ablation (Fig. 3. B) or cone biopsy (Fig. 3. C). The most popular techniques of ablation in the United States are cryotherapy. laser ablation. and large—loop electrosurgical excision of the transformation zone (LLETZ). While the LLETZ is excisional in nature. it removes as much tissue as is destroyed with cryotherapy and laser therapy and is therefore an equivalent therapy. All three techniques destroy a portion of the cervix for which the superior and deep margin is estimated to be 6—8 mm deep around the endocervical canal. and the inferior and ectocervical margin includes the transformation zone of the ectocervix. I11 nonrandomized series. the failure rates of all three therapies are 10% over a 2-year period (8/). Ran— domized studies of laser therapy and cryotherapy show similar Journal ofthc National Cancer Institute Monographs No. 2 l. 1996 ASCUS favoring dysplasia LGSIL, HGSIL l Review Pap Repeat Pap Co|poscopy Endocervical curettage Cervical biopsy if lesion SlL on biopsy Satisfactory exam (Entire lesion seen) (Squamo—columnar junction seen) Negative endocervical curettage No suspicion of invasion on colposcopy. Pap smear, or biopsy No adenocarcinoma-in-situ or adenocarcinoma Compliant patient l Cryotherapy, laser ablation, LLETZ (ectocervical specimen) & option of conservative follow-up for LGSIL No SIL on biopsy, Pap HGSIL Unsatisfactory exam (Entire lesion not seen) (Squamo—columnar junction not seen) Positive endocervical curettage Suspicion of invasion on colposcopy, Pap smear, or biopsy Suspicion of adenocarcinoma- in-situ or adenocarcinoma Non-compliant patient l Conization with scalpel, cautery, laser, or loop electrosurgtcal excision (ectocervical & endocervical specimens) or option of conservative follow-up for LGSIL Fig. 4. Triage scheme for patients with abnormal Papanicolaou (Pap) smears. ASCUS z atypical squamous cells of undetermined significance; LGSIL : low~ grade squamous intraepithelial lesions: HGSIL = high—grade squamous in- traepithelial lesions; SIL = squamous intraepithelial lesions; and LLETZ = large-loop electrosurgical excision of the transformation lone. long—term cure rates and complications [reviewed in (8/)[. A recent randotnized trial of the three modalities with 27 months of follow—up showed failure rates of ll%-12% for all three treatments and no significant differences in complications, either short- or long—term (Mitchell MF: unpublished observa- tions). Cone biopsies remove the endocervical canal: specimens include a superior and deep margin 2—3 cm in height around the canal, and the inferior and ectocervical margin includes the transformation zone of the ectocervix. Because the endocervical canal is included. these treatments are diagnostic as well as therapeutic. Cone biopsies in the United States are performed with the use of the scalpel (“cold-knife" cone). electrocautery (“hot—knife" cone). laser. and loop electrosurgical excision. A loop electrosurgical excision cone removes the cervix in two to three specimens. while the other techniques remove the specimen in one piece. (The additional two specimens are por— tions of the canal removed with an endocervical loop.) The ad— vantage of a loop electrosurgical excision cone is that it can be performed on the patient under local anesthesia in the clinic in less than 5 minutes. whereas the other techniques are typically performed on the patient under general anesthesia in the operat- Journal of the National Cancer Institute Monographs No. 2 l, 1996 ing room. One randomized study (82) showed many uninter- pretable specimens using loop electrosurgical excision. but other studies show comparable success rates and specimens using loop electrosurgical excision as for other types of cones [(83); Mitchell MF: unpublished observations]. Chemoprevention refers to the use of micronutrients or phar— maceuticals to prevent or delay the development of cancer in healthy populations. Chemopreventive agents have the advan— tage of affecting the patient systemically and may be particulah ly advantageous in the patient with multifocal vulvar. vaginal, and cervical disease or in the patient who is immunodepressed. Several chemoprevention trials have been carried out in the cer— vix with both micronutrients and pharmaceuticals. None of the micronutrient studies have demonstrated statistically significant regression of CW lesions in randomized trials [reviewed in (37)[. Topical Irans-retinoic acid caused statistically significant regression of CIN—2 but not CIN—3 in the trial conducted by Meyskens et al. (84). A SOC/r regression rate of ClN—3 was demonstrated in a phase I, dose—finding study of oc—difluoro— methylornithine, a polyamine synthesis inhibitor (Mitchell MF: unpublished observations). The response rate will be tested shortly in the context of a phase II study. The Future How can we strengthen efforts for prevention and early detec- tion of cervical cancer? Improvements in early detection could be made by increasing the use of the Pap smear (particularly among women of low socioeconomic status. Hispanic wotnen, and women older than 50 years). decreasing the false-negative rate of the Pap smear. automating the screening process to lower the cost, and better defining which patients need to be referred for expensive colposcopy. Preventing HPV infections with a humoral-based vaccine should receive wide interest. Treating lesions with a cell—mediated—based vaccine should be equally pursued. Once the role of other cofactors (i.e., smoking. oral contraceptive use, and diet) is assessed, specific behavioral interventions will be of interest. In general, the use of barrier contraception, delaying the onset of sexual activity, and limiting the number of partners should be encouraged as preventive be— haviors. Prompt evaluation and treatment of cervical cancer pre— cursors should be encouraged. and technologies that automate or shorten the process should be vigorously developed. The most promising areas of research will include those that focus on un- derstanding the natural history and immunobiology of HPV in— fections as they progress to intraepithelial lesions, vaccine trials and their effect on the natural history of lesions. and under- standing the process of cervical carcinogenesis and how it can be interrupted with Chemopreventive agents. References (I) Richart RM. Cervical intraepithelial neoplasia. Pathol Annu I973:8:3()l— 28. (2) Regan JW, Fu YS. The uterine cervix. In: Silvcrberg SC, editor. Principles and practice of surgical pathology. Vol 2. New York: Wiley. 1983:1633— 89. (3) xur Hausen H. Human papilloinaviruses in the pathogenesis of anogenital cancer. Virology [9911184943. (4 (5 (6 (7 (8 (9 (IO (l1) (l2) (l3) (l4) (l5) (16) (I7) (l8) (I9) (20) (2/) (22) (25) (26) (27) (28) (29) The 1988 Bethesda System for reporting cervical/vaginal cytological diag- noses. National Cancer Institute Workshop. JAMA 1989;262:9314. Pisani P. Parkin DM. Ferlay J. Estimates of the worldwide mortality from eighteen major cancers in 1985. Implications for prevention and projec» tions of future burden. IntJ Cancer 1993;55:891-903. Coleman MP. Esteve J. Darniecki P. Arslan A. Renard H. Trends in cancer incidence and mortality. IARC Sci Publ 1993;121:1»8()6. American Cancer Society. Cancer facts and figures. Atlanta (GA): American Cancer Society. l996:7. Paavonen J. Koutsky LA. Kiviat N. Cervical neoplasia and other STD-re- lated genital and anal neoplasias. In: Holmes KK. Mardh PA. Sparling PF. Wiesner PJ. editors. Sexually transmitted diseases. 2nd ed. New York: Mc- Graw—Hill. 1990:561-92. Routine cancer screening. ACOG Committee Opinion: Committee on Gynecologic Practice number 128~October 1993. Int J Gynaecol Obstet 1993;43:344-8. Kurman RJ. Henson DE. Herbst AL. Noller KL. Schiffrnan MH. Interim guidelines for management of abnormal cervical cytology. The 1992 Na» tional Cancer Institute Workshop. JAMA 1994;271:1866—9. CDC surveillance summary. Trends in cancer screening~United States. 1987 and 1992. Morbidity Mortality Weekly Report 1995;45:57-61. Munoz N. Bosch FX. Epidemiology of cervical cancer. lARC Sci Publ 1989:9429—39. Brinton LA. Epidemiology of cervical cancer—overview. lARC Sci Publ 1992;119:3—23. Koutsky L. Role of epidemiology in defining events that influence transmis» sion and natural history of anogenital papillomavirus infections leditoriall [we comment citation in Medline].J Natl Cancer Inst 1991;83:978-9. Munoz N. Bosch FX. HPV and cervical‘neoplasia: review of case-control and cohort studies. lARC Sci Publ 1992;] 191251431. Bosch FX. Munoz N. de Sanjose S. Navarro C. Moreo P. Ascunce N. et al. Human papillomavirus and cervical intraepithelial neoplasia grade III/car- cinoma in situ: a case—control study in Spain and Colombia. Cancer Epidemiol Biomarkers Prev 1993;2z415-22. Morrison EA. Ho GY. Vermund SH. Goldberg GL. Kadish AS. Kelley KF. et al. Human papillomavirus infection and other risk factors for cervi- cal neoplasia: a case—control study [see comment citation in Medline]. IntJ Cancer 1991;49:6-13. Becker TM. Wheeler CM. McGough NS. Stidley CA. Parmenter CA. Dorin MH. et al. Contraceptive and reproductive risks for cervical dysplasia in southwestern Hispanic and non-Hispanic white women. Int J Epidemiol 1994;23:913-22. Koutsky LA. Holmes KK. Critchlow CW. Stevens CE. Paavonen J. Beckmann AM. et al. A cohort study of the risk of cervical intraepithelial neoplasia grade 2 or 3 in relation to papillomavirus infection. N Engl J Med 1992;327:1272-8. Munoz N. Bosch FX. (16 Sanjose S. Vergara A. del Moral A. Munol. MT. et al. Risk factors for cervical intraepithelial neoplasia grade III/carcinoma in situ in Spain and Colombia. Cancer Epidemiol Biomarkers Prev 1993; 2:42.131. Olsen A0. Gjoen K. Sauer T, Orstavik I. Naess O. Kierulf K. et al. Human papillomavirus and cervical intraepithelial neoplasia grade “-111: a popula- tion-based case-control study. Int] Cancer 1995;61:312-5. Bosch FX. Manos MM. Munoz N. Sherman M. Jansen AM. Peto J. et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International biological study on cervical cancer (IBSCC) Study Group [see comment citation in Medline]. J Natl Cancer Inst 1995; 87:796—802. Lorincz AT. Reid R. Jenson AB. Greenberg MD. Lancaster W. Kurman RJ. Human papillomavirus infection of the cervix: relative risk associa— tions of 15 common anogenital types. Obstet Gynecol 1992;79:328-37. Schiffman MH. Recent progress in defining the epidemiology of human papillomavirus infection and cervical neoplasia. J Natl Cancer Inst 1992: 84:394—8. Schiffman MH. Bauer HM. Hoover RN. Glass AG. Cadell DM. Rush BB. et al. Epidemiologic evidence showing that human papillomavirus infec- tion causes most cervical intraepithelial neoplasia [sec comment citations in Medline]. J Natl Cancer Inst 1993;85:958»64. Schneider A. Koutsky LA. Natural history and epidemiological features of genital HPV infection. IARC Sci Publ 199211 19:25-52. Koldstad P. Follow—up study of 232 patients with stage lal and 411 patients with stage la2 squamous cell carcinoma of the cervix (mieroin- vasive carcinoma). Gynecol Oncol 1989;33:265-72. Koss LG. Stewan FW. Foote FW. Jordan MJ. Bader GM. Day E. Some histological aspects of behavior of epidermoid carcinoma in situ and re- lated lesions of the uterine cervix. Cancer 1963:16:1 160-21 1. Mitchell MF. Hittelman WN. Hong WK. Lotan R. Schotterifeld I). The natural history of cervical intraepithelial neoplasia: an argument for inter- (34) (35) (36) (37) (38) (39) (40) (4/) (42) (43) (44) (45) (46) (47) (48) (49) (50) (5/) (52) (53) (54) mediate endpoint biomarkers. Cancer Biomarkers Prev 1994;}:619-26, Ostor AG. Natural history of cervical intraepithelial neoplasia: a critical review. ltit J Gynecol Pathol 1993;12:186-92. Mitchell MF. Schottenfeld D. The natural history of CIN and management of the abnormal Papanicolaou smear. In: Rubin SC. Hoskins WJ. editors. Cervical cancer and preinvasivc neoplasia. New York: Lippincott-Raven. 199610343. Wright TC. Kurman RJ. Ferenczy AF. Precancerous lesions of the cervix. In: Kurman RJ. editor. Blaustein‘s pathology of the female genital tract. 4th ed. New York: Springer-Verlag. 1994:229-77. Wright TC. Kurman RJ. A critical review ofthe morphologic classification systems of preinvasive lesions of the cervix: the scientific basis of the paradigm. Papillomavirus Rep 19945: I 75—8 I. Lungu 0. Sun XW. Felix J. Richart RM. Silverstein S. Wright TC Jr. Relationship of human papillomavirus type to grade of cervical in— traepithelial neoplasia. JAMA 1992:2672493—6. Ho GY. Burk RD. Klein S. Kadish AS. Chang CJ. Palan P. et al. Persistent genital human papillomavirus infection as a risk factor for persistent cervi- cal dysplasia [we comment citation in Medline]. J Natl Cancer Inst 1995;87:1365-71. Park TW. Fujiwara H. Wright TC. Molecular biology of cervical cancer and its precursors. Cancer 1995:76( 10 Suppl): 1902-13. Mitchell MF. Hittelrnan WK. Lotan R. Nishioka K. Tortolero—Luna G. Riehards-Kortuni R. et al. Chemoprevention trials and surrogate end point biomarkers in the cervix. Cancer |995z76t IO Suppl):l956-77. Melkert PW. Hopman E. van den Brule AJ. Risse EK. van Diest PJ. Bleker 0P. et al. Prevalence of HPV in cytomorphological1y normal cervical smears. as determined by the polymerase chain reaction. is age~dependent. Int J Cancer 1993;56:919-234. Dillner J. Antibody responses to defined HPV epitopes in cervical neoplasia: review. Papillomavirus Rep 1994:5235—41. Dillner J. Immunobiology of papillomavirtls. Prospects for vaccination. CancerJ 1992;5z I 81-7. von Krogh G. Warts: immunologic factors of prognostic significance. Int J Dermatol 1979;18:195-204. Viac J. Staquet MJ. Miguet M. Chabonon M. Thivolet J. Specific im— munity to human papilloma virus (HPV) in patients with genital warts. Br JVener Dis 1978;54:172-5. Li CC. Shah KV. Seth A. Gilden RV. Identification of the human papil— lomavirus type 6b LI open reading frame protein in condylomas and cor- responding antibodies in human sera. J Virol 1987;61:2684-90. Galloway DA. Jenison SA. Characterization of the humoral immune response to genital papillomaviruses. Mol Biol Med 199();7:59-72. Mandelson MT. Jenison SA. Sherman KJ. Valentine JM. McKnight B. Daling JR. et al. The association of human papillomavirus antibodies with cervical cancer risk. Cancer Epidemiol Biomarkers Prev 1992;12281-6. Dillner J. Mapping of linear epitopes of human papillomavirus type 16: the E1. E2. E4. E5. E6 and E7 open reading frames. Int J Cancer 1990;46:703- 1 I. Lehtinen M. Leminen A. Kuoppala T. Tiikkainen M, Lehtinen T. Lehtovir- ta P. et a1. Pre- and posttreatment serum antibody responses to HPV 16 E2 and HSV 2 ICPX proteins in women with cervical carcinoma. J Med Virol 1992;37:180—6. Heim K. Christensen ND. Hoeptl R. Wartusch B. Pinzger G. Zeimet A. et al. Serum IgG. IgM. and IgA reactivity to human papillomavirus types I I and 6 virus-like particles in different gynecologic patient groups. J Infect Dis 1995;172:395-402. Hamsikova E. Novac J. Hofmannova V. Munoz N. Bosch FX. de Sanjose S. et al. Presence of antibodies to seven human papillomavirus type 16— derived peptides in cervical cancer patients and healthy controls. J Infect Dis 1994:1702 1424—3 1. Hellstrom KE. Hellstrom I. Cellular immunity against tumor antigens. Adv Cancer Res 1969;12:167-223. Levy S. Kopers/tych S. Musatti CC. Souen JS. Salvatore CA. Mendes NF. Cellular immunity in squamous cell carcinoma of the uterine cervix. Am J Obstet Gynecol 1978;130:1604. Ishiguro T. Sugitachi I. Katoh K. T and B lymphocytes in patients with squamous cell carcinoma of the uterine cervix. Gynecol Oncol 1980:9280- 5. Carson LF. Twiggs LB. Fukushitna M. Ostrow RS. Faras AJ. Okagaki T. Human genital papilloma infections: an evaluation of immunogenic corn- petence in the genital neoplasia-papilloma syndrome. Am J Obstet Gynecol 1986;155:7844). Cromme FV. Snijders PJ. van den Brule AJ. Kenemans P. Meijer CJ. Wal- boomers JM. MHC class 1 expression in HPV 16 positive cervical car- cinomas is post-transcriptionally controlled and independent from c~myc overexpression. Oncogene 199318196945. Epidemiol Journal ofthe National Cancer Institute Monographs No. 21. 1996 (55) (56) (57) (58) (5‘)) (60) (6/) (68) (69) (70 (7]) Chen LP. Thomas EK. Hu SL. Hellstrotn I. Hellstrom KE. Human papil- lomavirus type 16 nucleoprotein E7 is a tumor rejection antigen. Proc Natl Acad Sci U S A 1991;88:110-4. Chen L. Mizuno MT. Singhal MC. Hu SL. Galloway DA. Hellstrom I. et al. Induction of cytotoxic T lymphocytes specific for a syngeneic tttmor expressing the E6 oncoprotein of htttnan papillomavirus type 16. J Im— munol1992;148:2617-21. Statiss HJ. Davies H. Sadovnikova E. Chain B. Horowitz N. Sinclair C. In- duction of cytotoxic T lymphocytes with peptides in vitro: identification of candidate T-cell epitopes in human papillotna virus. Proc Natl Acad Sci U S A 1992;89:7871-5. Feltkamp MC. Smits HL. Vierboom MP. Minnaar RP. de Jongh BM. Drijthout JW. et al. Vaccination with cytotoxic T lymphocyte epitope-con— taining peptide protects against a tumor induced by human papillomavirus type Io-transformed cells. EurJ lmmunol 1993;23:2242-9. Ressing ME. Sette A. Brandt RM. Ruppert J. Wentworth PA. Hartman M. et al. Human CTL epitopes encoded by human papillomavirus type 16 E6 and E7 identified through in vivo and in vitro itnmunogenicity studies of HLA-A*02()I-binding peptides. J Itnmttnol 1995154593443. Ressing ME. van Driel WJ. Cells E. Sette A. Brandt MP. Hartman M. et a1. Occasional memory cytotoxic T—cell responses of patients with human papillomavirus type 16-positive cervical lesions against a human leuko- cyte-A *0201—restricted E7-encoded epitope. Cancer Res 1996: 56:582—8. Sarkar AK. Tortolero—Luna G. Nehete PN. Arlinghaus RB. Mitchell MF. Sastry KJ. Studies on in vivo induction of cytotoxic T lymphocyte re— sponses by synthetic peptides from E6 and E7 oncoproteins of human papillomavirus type 16. Viral Immunity 1995:8:165~74. Davison JM. Marty JJ. Detecting premalignant cervical lesions. Contribu- tions of screening colposcopy to cytology. J Reprod Med 1994;39:338-92. Wilkinson EJ. Pap smears and screening for cervical neoplasia. Clin Obstet Gynecol 1990;33:817-25. Fahey MT. Irwig L. Macaskill P. Meta’analysis of Pap test accuracy. Am J Epidemiol 1995;141:6809. Tawa K. Forsythe A. Cove JK. Salt] A. Peters HW. Watring W0. A com- parison of the Papanicolaou smear and the cervigram: sensitivity. specificity. and cost analysis. Obstet Gynecol 1988171 :229—35. Yobs AR. Swanson RA. Lamotte LC Jr. Laboratory reliability ol‘ the Papanicolaou smear. Obstet Gynecol 1985;65:235-44, Koss LG. The Papanicolaou test for cervical cancer detection. A triumph and a tragedy [we comment citation in Medline]. JAMA 19891261 :737— 43. Richart RM. Vaillant HW. Influence of cell collection techniques upon cytological diagnosis. Cancer 1965118: 1474—8. Koss LG. Cervical (Pap) smear. New directions. Cancer 1993:7114 Suppl): 1406—12. Ferris DG. Cervicographyian adjunct to [review]. Am Fam Physician 1994;50:363—70. Meijer CJ. van den Brule AJ. Snijders PJ. Helmerhorst T. Kenemans P. Walhoomers JM. Detection of human papillomavirus in cervical scrapes by the polymerase chain reaction in relation to cytology: possible implica- tions for cervical cancer screening. IARC Sci Publ 19911 192271—81. Papanicolaott screening Journal olithe National Cancer Institute Monographs No. 21. I996 (72) (73) (74) (75) (76) (77) (78) (79) (80) (8/) (82) (83) (84) Mango LJ. Computer-assisted cervical cancer screening using neural net- works. Cancer Lett 1994:77:155»62. Haraul. G. Chiu DK. MacAuIay C. Palcic B. Probabilistic inference in computer—aided screening for cervical cancer: an event covering approach to information extraction and decision rule formulation. Anal Cell Pathol 1994;6z37-50. Stall A. Cervicography: a new method for cervical cancer detection. Am J Obstet Gynecol 1981;139:8535. Cox JT. Schit'fman MH. Winzelberg AJ. Patterson JM. An evaluation of human papillomavirus testing as part of referral to colposcopy clinics. Obstet Gynecol 1992;80:389-95. Beral V. Day NE. Screening for cervical cancer: is there a place for incor— porating tests for human papillomavirus? IARC Sci Publ 199211 19:26:19. Ramanujam N. Mitchell MF. Mahadevan A. Thomsen S. Silva E. Richards-Kortum R. Fluorescence spectroscopy: a diagnostic tool for cer- vical intraepithelial neoplasia (CIN). Gynecol Oncol 1994;52:31-8. Ramanujam N. Mitchell MF. Mahadevan A. Warren S. Thotnsen S. Silva E. et al. In vivo diagnosis of cervical intraepithelial neoplasia using 337« nm-excited laser-induced fluorescence. Proc Natl Acad Sci U S A l994:91:1()193»7. Reid R. Greenberg MD. Lorincz A. Jenson AB. Laverty CR. Husain M. et a1. Should cervical cytologic testing be augmented by cervicography or human papillomavirus deoxyribonucleic acid detection? Am J Obstet Gynecol 1991:164: 1461-79: discussion 1469-71. Wright TC. Sun XW. Koulos J. Comparison of management algorithms for the evaluation of women with low-grade cytologic abnormalities. Obstet Gynecol 1995;85:202—10. Mitchell MF. Preinvasive diseases ol~ the female lower genital tract. In: Gershenson DM. DeCherney AH. Curry SL. editors. Operative gynecol— ogy. Philadelphia: Saunders. 1993:231-56. Mathevet P. Dargent D. Roy MR. Beau G. A randomized prospective study comparing three techniques of conization: cold knife, laser. and LEEP. Gynecol Oncol 1994154: I75~9. Alvarez RD. Helm CW. Edwards RP. Naumann RW. Partridge EE. Shingleton HM. et al. Prospective randomized trial of LLETZ versus laser ablation in patients with cervical intraepithelial neoplasia. Gynecol Oncol 1994;52:175-9. Meyskens Fl. Jr. Stirwit E. Moon TE. Childers JM. Davis JR. Dorr RT. et al. Enhancement ol~ regression of cervical intraepithelial neoplasia II (moderate dysplasia) with topically applied alI-rruIt.\‘—retinoic acid: a ran- domized trial [see comment citation in Medline]. J Natl Cancer Inst 1994: 86:539-43. Notes Supported by Public Health Service contracts N()ICN25433A and N()|CN25433B from the National Cancer Institute. National Institutes of Health. Department of Health and Human Services. We thank David Gershenson for helpful continents. Sunita Patterson for edit« ing. and Pat Williams for preparation of the manuscript. Histopathologic Prognostic Factors in Early Stage Cervical Carcinoma Jo Ann Bendu * The pathologist makes extensive assessments in cases of early stage carcinoma of the cervix. The following param- eters are known to affect prognosis: capillary—lymphatic space invasion, depth of invasion, size of gross tumor. his- tologic type, grade (in adenocarcinoma). involvement by tumor of pelvic or para-aortic lymph nodes, and lymphoid response and immune status. These findings, along with clinical stage. best predict tumor behavior. Other observa- tions are under investigation to evaluate their usefulness. [Monogr Natl Cancer Inst 1996;21:27-34] Extensive observations are made by the pathologist when evaluating cervical carcinomas. particularly in early stage dis— ease. Each observation adds some prognostic information that will assist the oncologist in planning therapy and counseling the patient. Pathologic parameters that have been demonstrated to have value in predicting behavior include stage. capillaryilym— phatic space invasion. depth of invasion. size of gross tumor. histologic type. grade (for adenocarcinoma only). pelvic and para—aortic lymph node status. and type of lymphoid response in draining lymph nodes. The observations can be made with routine handling of tissue specimens in the laboratory. Newer techniques. such as immunohistochemistry for cellular proteins. flow cytometry for ploidy analysis. DNA analysis for oncogenes and viruses. and image analysis. are being explored as predictors of behavior. As yet, none of these techniques has added suffi- cient predictive information to become standard assessment parameters for low stage cervical disease. Some, however. may prove useful as predictors of tumor behavior and may expand our knowledge of disease biology. Clinical Stage Staging provides a standard for extent of disease by which worldwide discussion of survival statistics for different therapies and populations can be compared. The International Federation of Gynecology and Obstetrics (FlGO) suggests that staging of gynecologic cancers not be the sole guide for therapy (I). How- ever. staging does provide information about disease behavior and its response to given therapeutic modalities. Therefore. it is important to consider when advising patients about treatment and outcome. The histopathologic subdivisions of early stage cervical can- cer have been developed over a long period. with many quan- tification methods offered as possibilities for separating tumors with different behavior. Since 1960. FIGO has changed the definition of early stage disease at least five or six times. The Journal of the National Cancer Institute Monographs No. 21. 1996 staging scheme in Table 1 is the most recent effort from the FIGO Congress in Montreal convened in September 1994 to more specifically define subsets of early stage cervical cancer (I). The major changes in the new staging scheme are the follow— ing: 1) a more specific definition of stage 1A1 and 2) the separa— tion of stage [B into two categories based on the size of the clinical lesion. Stage 1A1 in the prior scheme was defined only as minimal microscopic invasion. It is now quantified as less than 3.0 mm in depth and less than 7.0 mm in width. Many studies have shown that “larger" or “bulky" lesions have a worse prognosis. but the new division of stage [B into tumors 4.0 cm or smaller than or larger than 4.0 cm in diameter provides a common measurement for investigators to compare results. The more specific definitions of stages 1A1 and 1A2 may help to decrease the use of the term “microinvasion” in reference to cervical squamous cell carcinoma. “Microinvasion” and "super— ficial" invasion have been used in an attempt to identify a car— cinotna of small enough proportions to require less than radical therapy. However. there have been multiple. different defini— tions in the literature with resulting confusion. Because of potential treatment differences. the pathologist should provide the actual measurements of tumors as well as the other relevant pathologic information so that the oncologist can make the most informed recommendation for therapy. In a large cohort (16 861 cases) analyzed from the Surveil— lance. Epidemiology. and End Results (SEER)l registry data ap— plying FlGO staging (1985) (2). the 5-year survival for all patients with stage [A carcinoma of the cervix was 97%. that for patients with stage 18 was 78.9%. and that for patients with stage IIA was 54.9%. With stage [A assigned a relative risk of 1.00 for death from disease. the relative risk for stage [B was 3.46 and that for stage IIA was 7.36. There are little data for the specific new categories of the 1994 staging scheme; however. on the basis of historical data using 3—mm and 5-mm depth of invasion. one would expect some difference in behavior be— tween stages 1A1 and 1A2. Tumors under 1 mm in depth would be expected to have a risk of positive lymph nodes in only 0.4% of cases and no recurrence risk (3). The same data show that patients with tumors 1.1-3 mm in depth have a 1.3% risk of positive lymph nodes and 0.5% recurrence risk. whereas those *Ajf'iliminn afumlmr: Departments of Pathology and Obstetrics and Gynecol— ogy. University of Iowa. Iowa City. C(M'I‘eSpUIll/(‘IN'l’ to: Jo Ann Benda, M.D.. Department of Pathology. Univer- sity of Iowa Hospitals and Clinics. 200 Hawkins Drive—5244A RCP. Iowa City. IA 52242-1009. Sl'l‘ “Note" section following “References." 27 Table l. Staging of early invasive carcinoma of uterine cervix“ FIGOT stage Description TNM stage: I Cervical carcinoma confined to the cervix (extension to the corpus should be disregarded) TI IA Preclinical invasive carcinoma. diagnosed by microscopy only TIa IAI Measured invaxirm afxtroma no greater than 3.0 mm in depth and no wider than 7. 0 mm TlaI (I985 definition—minimal microscopic stromal invasion) 1A2 Tumor with invasive component greater than 3 mm and less than 5 mm in depth. taken from the base of the epithelium, and 7 mm or Tlal less in horizontal spread IB All clinically (grossly) visible tutnors (even with superficial invasion) or tumors larger than 1A2 confined to cervix TIb IBI Clinical lesions no greater than 4.0 cm in size, confined to the cervix IB2 Clinical lesions" greater than 4.0 cm in size. ('nnflnml to the cervix ll Cervical carcinoma that extends beyond the uterus but not to pelvic wall or to the lower third of the vagina T2 IIA Tumor extends in upper two thirds of vagina. but without parametrial invasion T2a *Changes since I985 in italics. 'I'Intemational Federation ofGynecology and Obstetrics. iNX = regional lymph nodes cannot be assessed; N0 = no regional lymph node metastases; NI = regional lymph node metastases ( I). with tumors 3—5 mm in depth have a 6.8% risk of positive lymph nodes and a recurrence risk of 2.3%. Only patients in the latter category had an increased risk of death of 1.5% as a result of tumor (3). Nearly all investigators use parameters other than stage alone (such as capillary—lymphatic space invasion, conflu- ent growth pattern. or histologic type) in deciding therapy. Although stage overrides other predictors of tumor behavior. many other parameters provide valuable information in thera— peutic decision—making within each stage. These other param— eters are particularly important (and controversial) in early stage disease. Capillary—Lymphatic Space Invasion In stage IA cervical carcinoma, the presence of capillary— lymphatic space invasion has been the subject of controversy for some time. Correlations have been documented between depth of invasion and the finding of capillary—lymphatic space involve- ment. In lesions under 3 mm in depth. the presence of capillary— Iymphatic space involvement is reported in 4%-20% of cases; in lesions between 3.1 and 5 mm in depth. it is reported in 8%- 50% (4-8). The definition of capillary—lymphatic space invasion is tumor in round or oval spaces lined by flat cells resembling endo- thelium (4). Other investigators (9) added the criterion that the tumor extends into these spaces from contiguous stroma. lin- munohistochemical studies for endothelium have been at— tempted, but they were found to be helpful in only occasional cases ([0). Some of the variation in the amount of capillary— Iymphatic space involvement reported may have to do with the stringency of the observer in requiring that endothelium be recognized before identifying tumor in an apparent space as rep- resenting capillary—Iymphatic space involvement. One report (9) that suggested that capillary—lymphatic space involvement is not an independent variable and is no more predictive than depth used the added caveat noted above. There have also been correlations made between the presence of capillary—lymphatic space invasion and positive lymph nodes (4-8). Because of the correlation between metastatic disease in lymph nodes and the presence of capillary—lymphatic space in- vasion. the current recommendation by the Society of Gyne- cologic Oncologists is that patients with capillary—lymphatic 28 space invasion. regardless of the depth of invasion, undergo radical therapy (surgical or radiation) (3). One group of inves- tigators (6) suggested that it is important to further evaluate the biologic behavior of a cancer with a demonstrated ability to in— vade lymphatic spaces by examining the pelvic lymph nodes. In a study by the Gynecologic Oncology Group (11) of 732 patients with stage I cervical squamous cell carcinomas who had more than 3 mm of invasion, capillary—lymphatic space invasion correlated with the presence of metastatic disease in pelvic lymph nodes in both univariate and multivariate analyses (P<.()00I). When no capillary—lymphatic space invasion was present, 8.2% of the patients had positive lymph nodes. com- pared with 25.4% of the patients with positive lymph nodes when capillary—lymphatic space invasion was identified. Dis- ease-free interval at 3 years was also affected by the presence of capillary—lymphatic space invasion, with those demonstrating capillary—lymphatic space invasion having a disease-free inter— val at 3 years of 77% compared with those not demonstrating it having a disease—free interval at 3 years of 88.9% (12). At 5 years. the disease-free interval was 83% in patients not showing capillary—lymphatic space invasion versus 70% in those show- ing capillary—lymphatic space invasion (P = .05) (13). Other in— vestigators ([4) have suggested that the significance of positive capillary—lymphatic space invasion disappears when patients are also stratified by the presence of lymph node metastases. Capillary—lymphatic space involvement in adenocarcinomas has not been as extensively studied. but it is considered to be an indicator of poorer prognosis (3.15). Depth of Tumor Invasion The common conclusion of all studies measuring depth of tumor invasion is that less involvement of the cervical stroma is associated with fewer pelvic lymph node metastases. better dis- ease-free interval or survival. and better local control of disease. Specifically defining stage lAl as tumors under 3 mm in depth of invasion and stage 1A2 as tumors 3—5 mm in depth provides an opportunity to compare data with more consistent parameters in this very early stage of disease. In stage IB squamous cell carcinoma of the cervix, depth of tumor invasion beyond 5 mm is important for predicting the presence of pelvic lymph node metastases and disease—free in— Journal of the National Cancer Institute Monographs No. 2 I. 1996 terval (Table 2) (11,12). The depth of invasion is significant whether measured in specific millimeters or by thirds of the cer- vical stroma. Depth of invasion, clinical tumor size, and capil- lary—lymphatic space invasion constitute a group of variables that can be used in combination to identify a group of inter- mediate-risk tumors (11,12). Studies evaluating the significance of depth of tumor invasion in adenocarcinoma have been less numerous and have not measured as many variables. No large, multi-institutional studies have been done. One institutional study indicated that adenocarcinomas with a volume of less than 500 mm3 resulted in no lymph node metastases (16). In studies that have attempted to evaluate small adenocar- cinomas (stage IA), the frequency of lymph node metastases has been higher than that of comparable cases of squamous cell car- cinoma. A measurable amount of adenocarcinoma that predicts a sufficiently better prognosis so that less than radical therapy can be offered is not yet clearly defined (3). Most patients with any amount of invasive adenocarcinoma are treated with radical therapy. The evaluation of the lymph nodes gives added infor- mation concerning the biology of the tumor (17). One study of patients treated with radical hysterectomy (17) has shown that adenocarcinomas with less than 50% thickness of cervical stroma have a 3% recurrence rate compared with those with 50% or more thickness, which have a 21% recurrence rate. However, some specific histologic subtypes of adenocar- cinoma, such as villoglandular, can have deep stroma] invasion and still have a good prognosis (3). Extension beyond the cervix increases risk of metastasis and recurrence, whether discovered at the time of histologic evalua- tion of a surgical specimen, detected only at the time of attempted radical surgery, or evident on initial clinical examination (iden- tified as stage II) (18). Size of Gross Tumor Studies of stage IB squamous cell carcinoma of the cervix by the Gynecologic Oncology Group (11,12) demonstrated that the size of the gross tumor is related to the presence of pelvic lymph node metastases and the 3-year disease-free interval. In patients with tumors 3 cm or less in diameter, pelvic lymph node metas- tases occurred in 15% of patients with a 3-year disease-free in- terval of 85.5%. In patients with tumors greater than 3 cm in diameter, pelvic lymph node metastases occurred in 23% and the 3-year disease—free interval was 68.4%. Other investigators Table 2. Relationship of depth of tumor invasion to pelvic lymph node metas— tases and disease-free interval in stage IB cervical squamous cell carcinoma Depth of tumor invasion Disease—free survival at 3 y, %* Positive pelvic lymph nodes. %+ 6—10 mm 86 15 11-15 mm 75 22 16-20mm 72 39 >21 mm 60 23 Inner third 94 5 Middle third 85 13 Outer third 74 26 *Reference (12). lReference (11). (14) have found that clinical tumors larger than 4 cm are associated with a greater risk for lymph node metastases. Local control by radiation therapy is also affected by size, with 100% of 0- to 3-cm tumors having local control compared with 85% of 4- to 5-cm tumors and 73% of tumors 6 cm or larger (19). A similar study (20) showed local control in 99% of cases with tumors smaller than 5 cm with a corresponding 88% survival, compared with 97% local control for exophytic tumors 5 cm or larger but under 8 cm (25<8 cm) with a 76% survival and 91% local control in endophytic tumors 25<8 cm with a 66% survival. The addition of subsets of stage IB cervical carcinoma to reflect tumors that are less than or greater than 4 cm in diameter recognizes the need to separate lesions by clinical size for pur- poses of comparing outcome. Many studies have looked at size of the primary lesion, and all found “bigger” to be worse, though the definition of bigger ranged from greater than 3 cm to greater than 6 cm, or a simple description of “bulky" or “barrel- shaped” cervical disease. Additional studies using the new stag- ing classification will help clarify these findings. Histologic Type The World Health Organization classification of histologic types of cervical carcinoma has recently been revised (21). Three basic subsets are recognized: squamous cell carcinoma, adenocarcinoma, and other epithelial types (Table 3). Mesen- chymal tumors are also found in the cervix; however, as a group, they represent a small number of cases. Patients with squamous cell carcinoma have a better 5-year survival than Table 3. Histologic types of cervical carcinoma“ Adenocarcinoma A. Mucinous Squamous cell carcinoma A. Keratinizing B. Nonkeratinizing 1. Cervical C. Verrucous 2. Intestinal D. Papillary transitional 3. Signet ring E. Lympho—epithelioma like C. Clear cell D. Minimal deviation E, Villoglandular F. Serous G. Mesonephric B. Endometrioid Other epithelial carcinomas A. Adenosquamous 1. With glandular structures 2. With mucin production only B. Glassy cell C. Adenoid cystic D. Adenoid basal E. Carcinoid F. Small cell G. Undifferentiated *World Health Organization classification. Journal of the National Cancer Institute Monographs No. 21, 1996 29 I Squamous E1 Adenocarcinoma I Adenosquamous Relative Survival 5-Year Stage IB FIGO Stage Stage "A Fig. l. Five~year relative survival for cervical carci~ noma patients by International Federation of Gynecol- ogy and Obstetrics stage. Surveillance. Epidemiology. and End Results Program. |‘)73— l987. Reproduced with permission from (2). patients with adenocarcinomas, as reported in most studies and in the population-based cohort of the SEER Program (Fig. 1) (22—25). In low stage disease. a poorer prognosis has been reported for patients treated with either radical hysterectomy or radiation therapy (25.26). In one study comparing stage IB adenocarcinomas with stage IB squamous cell carcinomas treated with radiation therapy in tumors larger than 4 cm. the squamous cell carcinoma patients had a 5—year survival of 73%. whereas the adenocarcinoma patients‘ survival was only 59% (25). Two reports (24,25) also suggested a greater number of distant recurrences in patients with adenocarcinoma without ap- parent differences in risk factors other than histologic type. Many recent reports in the literature have discussed an in- creasing “incidence” of adenocarcinoma of the cervix. How— ever. the SEER data covering three 5-year time periods from 1973 to 1987 showed no change in age-adjusted or age—specific incidence rate (Figs. 2 and 3) (22). What is evident from the SEER data is that the absolute numbers of squamous cell car- cinomas are decreasing and the absolute numbers of adenocar— cinomas are increasing (Fig. 4) (22). If only raw numbers are compared. the proportion of adenocarcinomas is increasing. but the incidence is not. Subtypes of adenocarcinoma and squamous cell carcinoma exist. Some of them have survival statistics at variance with those of the group as a whole. One squamous subtype. ver— rucous carcinoma. has a very good prognosis. Three subtypes of adenocarcinoma (i.e.. viIIogIanduIar. adenoid basal. and mini- maI deviation carcinoma) have a good prognosis when matched for stage and other parameters of extent to those of other cervi- cal carcinomas. Serous adenocarcinoma appears to have a worse prognosis than other cervical carcinomas have (3.27.28). Some ongoing discussion has centered on the behavior and prognosis of adenosquamous carcinoma. Several investigators (24.29-31) have found a poorer prognosis for adenosquamous carcinoma compared with the prognosis for squamous cell carcinoma. Within the other epithelial types, small-cell carcinoma has a very poor prognosis. regardless of whether or not it can be iden- tified as having neuroendocrine features. Glassy—cell carcinomas and adenoid cystic carcinomas also have a reportedly poorer prognosis (3.27). 1 4 q, .1 9 7 3 - 7 7 ‘5 1 2 A1 9 7 a - a 2 0 O1 9 8 3 - 8 7 g 1 o o ._, 3 T o 8 E 3 Fig. 2. Age-adjusted incidence rate for cervical carcinoma in E 6 three time periods. Surveillance. Epidemiology. and End Results 7;, Program. l973-I9X7. :i 5* 4 ‘F a) m < 2 T 0 . All Squamous Adeno- carcinoma cancor carcinoma 30 Journal of the National Cancer Institute Monographs No. 2 I. I996 Rates 0 u— U 0 Q. (I) I 0 U) ‘1 a) v cn v 07 v ca v 07 .— N N m to <1- v L0 m u; o in 0 Ln 0 Ln 0 Ln .- N N m to v v in an Age Fig. 3. Age-specific incidence rates for adenocarcinoma (AC) and squamous chI carcinoma (SQ). Surveillance. Epidemiol- ogy, and End Results Program. |973-|987. 60-64 65-69 70—74 75-79 80-84 85+ Better studies of specific histologic subsets and their behavior patterns may help identify patients who would benefit from ad— ditional or adjuvant therapy (e.g.. patients with small-cell car— cinoma) or those who do not require aggressive therapy (e.g.. patients with villoglandular carcinoma). Grade Squamous cell carcinoma of the cervix has most commonly been graded with a modification of the Broder system (3). Well- differentiated tumors (grade 1) consisted of mature—appearing squamous cells with minimal mitotic activity and pleomorph— ism. Abundant keratinization was often present. Moderately dif— ferentiated tumors (grade 2) showed less abundant cytoplasm. greater pleomorphism and mitotic activity. and indistinct cell borders as opposed to identifiable intercellular bridges. Poorly differentiated tumors (grade 3) contained primitive cells with markedly increased nuclear/cytoplasmic ratios and high mitotic activity (3). However, this system has not been uniformly pre— dictive of behavior. In a study from the Mayo Clinic. where Broder‘s grading system has been employed for several decades. Goellner (32) analyzed 196 patients with invasive squamous cell carcinoma of the cervix and found no significant difference in prognosis among histologic grades of tumors within any stage. He also noted several other studies with similar findings. He described some of the problems with the grading system. includ- ing interobserver variability, assignment usually by predominant type but modified by a small amount of what is considered a poorer type. and the problems in grading tumors in which only a small biopsy specimen is present compared with a large resec- tion. However. a most interesting finding not often noted from this study was that no cases were grade I. 7.6% were grade 2. 74.5% were grade 3. and 18.4% were grade 4. demonstrating the lack of separation of groups ofcases using this system (32). Zaino et al. (/3) undertook a study with the Gynecologic On— cology Group to compare multiple. different grading systems and found none of them to be useful in predicting behavior. They also found poor interobserver reproducibility. They sug- gested that tumor grade, unless use of parameters defined in a given system is strictly followed. can be influenced by other fac- 700 600 (n 3500 N 0 ._400 O 5300 E :200 Z 100 o a: 4 cm) as a surrogate for volume in the 1995 stratifica— tion of stage 18 into [Bl and 182. primary tumor volume is not considered in the FIGO system. Evaluation of lymph node status is also largely irrelevant in the F100 staging classification. The American College of Surgeons Commission on Cancer con- ducted a Patient Care Evaluation survey for patients with cervi— cal cancer in the United States that quantitated the use of various pretreatment diagnostic assessments used during the study years 1984 and 1990 (5). On the basis of a total of 9338 patients with known clinical stage (approximately 30% of the estimated total national cases for the 2 study years). this study revealed a rapid- ly widening gulf between the assessments actually obtained in clinical practice and those allowed under FIGO staging rules (Table 1). Qualitatively similar observations have been reported by the American College of Radiology (6) in patients treated primarily with radiation therapy. Between 1984 and 1990. the use of cystoscopy. proctosigmoidoscopy. lymphangiography. excretory urography (lVP). bone scan. and barium enema ex- aminations each declined by tnorc than 25%. while the use of Table I. Pretreatment diagnostic assessments among 9338 patients with invasive cervical cancer Test 1984. ‘4 1990. (/1 Change. ‘4 Examination utider anesthesia 41.8 43.4 +3.8 Cystoscopy 40.8 30.1 725.7 Proctosigmoidoscop) 29.4 21.7 7727.2 Chest radiography 78.8 75.0 —4.8 Excretot‘y urograph) (NP) 63.7 31.5 50.5 Barium enema 37.5 16.7 455.5 Isotope bone scan 12.9 8.1 737.2 Lymphangiogram 6.0 2.3 461.7 Fine-needle aspiration biopsy 2.8 2.6 —7.1 Computed tomography 38.2 55.1 +443 Magnetic resotiance imaging 0.9 5.6 +6222 *Af/i‘liutiun alum/tor: Department of Obstetrics and Gynecology. School of Medicine. University of California at Davis. CA. (‘orrm/mmlt'm't’ to present (Ir/drum: Anthony 11. Russell. M.D.. Radiological Associates of Sacramento Medical Group. Radiation Oncology Center. Mercy General Hospital. 4001 .1 St. Sacramento. (‘A 95819. 35 Table 2. Use of specific imaging assessments correlated with initial modality of treatment* 1984 1990 Initial hysterectomy? Initial radiotherapyi: Initial hysterectomy'l Initial radiotherapyIE: Test (n: 1914) (n=3406) (n=2447) (n=2686) CXR 75.1% 81.5% 71.8% 77.2% NP 54.0% 66.8% 32.1% 30.9% BAE 23.2% 43.4% 12.2% 21.0% Bone scan 59% 15.5% 3.8% l 1.3% CT 170% 47.2% 30.8% 72.6% MR1 0.9% 0.9% 3.4% 7.4% LAG 2.9% 8.0% 1.7% 3.9% FNA 1.6% 3.0% 3.1% 1.8% *CXR = chest radiography; IVP = intravenous pyelogram excretory urography; BAE = barium enema; CT = computed axial tomography ot~ abdomen or pelvis: MR1 = magnetic resonance imaging; LAG = lymphangiography: FNA = fine-needle aspiration biopsy. Tlncludes patients of known clinical stage treated with initial hysterectomy or extended hysterectomy and patients treated with hysterectomy and postoperative radiotherapy. :Includes patients of known clinical stage treated with radiotherapy without hysterectomy and patients treated with preoperative radiotherapy followed by hysterectomy. computed tomography (CT) and magnetic resonance imaging (MRI) increased dramatically. With the exception of excretory urography (IVP) in 1990, each diagnostic imaging method in both study years was substantially more frequently obtained in patients treated with initial radiotherapy than in patients treated with initial hysterectomy (Table 2). Patients with more ad- vanced stage (FIGO IIB-IVB). who are usually treated with radiotherapy, were more likely to have abnormalities attributed to cancer on their imaging studies than patients of lower clinical stage (FIGO IA-IIA). who are often treated with initial surgery (Table 3). F100 staging guidelines for permitted pretreatment staging assessments did not change between 1984 and 1990. Reconciliation of the discrepancies between sanctioned imaging studies and actual clinical practice will be a complex. but neces- sary, task if the FIGO staging system is to remain relevant to the nuances of medical practice in the developed countries. Sensible extrapolation of research and treatment results from technologic societies to less-developed countries will ultimately require adoption of more sophisticated pretreatment assessment and patient stratification than is currently feasible with FIGO stag— ing. On the basis of volumetric analysis of radical hysterectomy and lymphadenectomy specimens from patients with stages 1B, 11A, and 11B cervical cancer. tumor volume has been shown to be a more accurate predictor of clinical behavior and lymph node metastasis than clinical stage (7). Contemporary imaging technology may provide the means to measure true tumor volume in vivo on the basis of good correlation oftumor volume Table 3. Yield analysis,* 1984 and 1990 combined: percentage of imaging assessments abnormal for cancer correlated with patient clinical stage'r Stage: [A IB-IIA IIB»IVA IVB CXR .76/(1 .770 2.1% 30.1% (n = 7121) 6/893 23/3328 53/2521 114/379 IVP 3.2% 4.4% 22.2% 47.5% (n = 4402) 1 1/340 98/2242 359/1616 97/204 BAE .7% 2.8% 9.1% 24.6% (n = 2480) 1/148 32/1133 97/1061 34/138 Bone scan 33% 2.2% 4.7% 46.3% (n = 946) 2/61 6/278 23/486 56/121 CT 31.8% 43.2% 74.5% 88.5% (n=4215) 68/214 718/1662 1477/1983 315/356 MR1 20.0% 52.1% 76.6% 87.5% (n=263) 3/15 61/117 82/107 21/24 LAG 7.7% 21.2% 52.0% 93.8% (n=393) l/13 29/137 118/227 15/16 FNA 11.1% 15.7% 45.6% 88.9% (n=218) 1/9 Il/70 47/103 32/36 >“Includes only patients known to have undergone the test and for whom the test result was known. Denominator = number ot‘ patients having the test. Nutnerator = number of those patients in whom the test revealed evidence of cancer. +CXR = chest radiography: IVP = intravenous pyelogram excretory urography; BAE = barium enema; CT = computerized axial tomography of abdomen or pelvis; MRI = magnetic resonance imaging; LAG = lymphangiography: FNA = fine—needle aspiration biopsy. Modified from (5). iExcludes I I 1 1 patients in 1984 and 1272 patients in 199010r whom clinical stage was not known or not recorded. 36 Journal of the National Cancer Institute Monographs No. 2 I. 1996 estimates with hysterectomy specitnens (8.9). The ability to tneasure tumor volume could improve the quality of prospective clinical trials by providing a reliable and reproducible index for patient stratification prior to random assignment. The sagittal MRI images in Figs. 1-3 reveal the potential limitations of eval— uating tumor volume by estimating tumor diameter on vaginal and bimanual physical examination. Metastasis to lymph nodes has long been recognized to be of prognostic importance in patients treated surgically for cervical cancer (/0). and the anatomic level of metastasis also appears to convey prognostic information for patients treated either with surgery or with radiation therapy (3,11). Until the advent of the' lymphangiogram. the only reliable means of detecting nodal in— volvement was surgical evaluation. For patients with medically inoperable or surgically unresectable disease. diagnostic imag— ing offers the opportunity to identify nodal metastases and their level. Fine—needle aspiration biopsy under fluoroscopic or CT guidance offers the opportunity to confirm suspected lymph node metastases. Despite conveying an ominous prognosis. per— haps as many as 25(7r-3()(7r of the patients with histologically verified para—aortic metastases may be salvaged through ex— tended-field radiation (12). which may be done with reasonable safety if the appropriate technique is used (/3). Detection of para-aortic lymph nodes by lymphangiography can be ac— complished in 79% of the patients in whom such metastases can be detected by surgical sampling (/4). CT and MRI will detect lymph node metastases only when some degree of nodal enlarge— ment is present and are thus less sensitive for the detection of minimal volume micrometastases. Enlarged. hyperplastic lymph nodes may be misinterpreted as neoplastic. Lymphography is more sensitive for the detection of small—volume lymph node metastases but is also vulnerable to a false—positive interpreta— tion from intranodal glandular inclusions. Since these methods Fig. I. Midplane sagittal magnetic resonance imaging (MRI) of a patient with International Federation of Gynecology and Obstetrics stage IIB cancer of the cervix. ()n the basis of physical examination. the tumor was judged to be 7 cm in largest diameter. did not extend to the vaginal fornices. but was judged to in- filtrate thickened parametria. The MRI reveals the cancer to be exophytic and flat. with minimal extension up the endocervical canal. The normal /.ona| anatomy of the titerine corpus is intact. The MRI did not show parametrial exi tension. Image specifications: 1.5 Tesla T2 weighted image. TR 2000. TH 70. Matrix 256 x 128 2.0 NEX. 34-cm field of view. 57mm image thickness. Journal ofthc National Cancer Institute Monographs No. 21. 1996 Fig. 2. Midplane sagittal magnetic resonance imaging of a patient with Interna- tional Federation of Gynecology and Obstetrics stage IIB cancer of the cervix. ()n the basis of physical examination. the tumor was judged to be 7 cm in largest diameter. The cancer is bulky and substantially invades cervical stroma. The normal Ional anatotny of the uterine corpus is intact. Image specifications: 1.5 Tesla Fast Spin Echo Sequence; TR 4000'. TE 16: Matrix 256 X 128; 20 NEX; 38-cm field of view: 4—mm image thickness. may differ in their sensitivity and specificity. any modification of the FIGO staging system that might integrate the results of diagnostic imaging evaluation of regional nodes would require imaging modalities employed for patients with disease staged node nega— prudent specification and standardization of the live and should additionally require histologic or cytologic con- firmation for patients staged lymph node positive. Fig. .1. Midplane sagittal magnetic resonance imaging (MRI) of a patient with International Federation of Gynecology and Obstetrics stage IIB cancer of the cerv ix. ()n the basis of physical examination. the tumor was judged to be 4 cm in largest diameter. extended to the vaginal fornices. and minimally involved medial parametria. The cancer obliterates the endocervical canal and replaces the entire cervical stroma. It extends into the uterine cavity and by transmural in— \asion of the myometrium of the lower uterine segment extends to the vesicouterine pouch. Medial parametrial thickening appreciated on physical ex- amination was not demonstrated by MRI. but oftlaxis transverse images were not obtained. Image specifications: 1.5 Tesla Fast Spin Echo Sequence. TR 4000. TE 102. Matrix 256 X 192. 20 Mix. 34»cnt field of view. 5-mm image thickness. 37 Imaging Evaluation in Patients With Potentially Operable Disease 1n the United States. patients with FIGO stages 18 and 11A of limited volume who have no medical contraindication to surgery and who have access to a surgeon skilled in the performance of radical hysterectomy and lymphadenectomy will usually be treated with initial surgery because of perceived lesser chronic morbidities than those associated with radical radiation therapy. For such patients. there is no generally accepted role for presur— gical diagnostic imaging. and the intensity of preoperative diag- nostic imaging in these patients declined between 1984 and 1990 (Table 2). The value of preoperative imaging will be contingent on sur— gical policy: If surgical policy is to abandon radical hysterec— tomy when either macroscopic nodal metastases or minimal medial parametrial extension of tumor is detected intraopera— tively. preoperative diagnostic imaging may be of great value. Patients destined to be treated with radiation therapy could potentially be spared the expense and morbidity of a laparotomy if these manifestations of disease spread could be reliably deter— mined preoperatively. lf surgical policy is to continue with radi— cal hysterectomy. even when macroscopic nodal metastases are present or medial parametrial extension is discerned. then results of preoperative imaging become moot. Because patients with limited volume stage 113 cancers are treated surgically in other countries with favorable results (15). it is not universally ac- cepted that full-thickness stromal penetration with minimal parametrial infiltration constitutes an absolute contraindication to hysterectomy. Similarly. optimal treatment for patients with stage IB/llA cervical cancer and macroscopic nodal involve— ment at the time of exploration for planned radical hysterectomy is an issue in dispute (11,16). and there are no persuasive data to support the contention that such patients fare better when treated with radiation therapy. Advocates of abandoning hysterectomy and radiation therapy do so only after removal of all gross adenopathy (16). a maneuver both diagnostic and potentially therapeutic (17), and a possible reason why patients so treated may have a better prognosis than those closed and irradiated (18). In February 1996. 300 of the attendees at the annual meet- ing of the Society of Gynecologic Oncologists in New Orleans. LA. were asked to participate in a patient management survey with instantaneous tabulation of responses by use of an interac- tive audience response system. One patient was a 50—year-old woman with a 3-cm diameter stage 181 cervical cancer explored for a radical hysterectomy. lntraoperatively. a 2—cm pelvic lymph node was detected and confirmed to contain metastatic tumor. The members were asked how they would proceed. Sixty—seven percent indicated that they would complete a hysterectomy. and 87% of those said they would do a radical hysterectomy and therapeutic lymph node dissection. It seems logical to conclude that these same surgeons are unlikely to ob- tain preoperative imaging to evaluate retroperitoneal nodes. CT and MR1 will not detect lymph node metastases unless nodes are enlarged. Among 732 patients with FIGO stage IB squamous cancer of the cervix subjected to laparotomy in a Gynecologic Oncology Group study (19). there were 148 (20.2%) patients with either gross or microscopic metastases to pelvic or para- 38 aortic lymph nodes. Of 119 (16%) patients with pelvic lymph node metastases without para-aortic node metastases. only 19 (16%) had enlarged lymph nodes seen at the time of surgery. Thus. CT or MRI could be expected to detect retroperitoneal nodal involvement in approximately 3‘/(-7‘/( of the total popula- tion of stage 18 patients. suggesting that routine preoperative screening for node metastases by CT or MRI would be ineffi— cient and costly. The yield of preoperative screening in patients with stage [8 disease may be higher in patients with tumors 2 cm or larger in diameter (20). The results of lymphangiography in assessing patients with operable cervical cancer have been somewhat disappointing (21—23). and in 1990 this examination was performed in only 1.7% of 2447 patients treated with initial hysterectomy in the American College of Surgeons survey (Table 2). Assessment of the parametria by MR1 (24-26) or endoluminal ultrasound (27,28) may be more accurate than by physical ex— amination, which itself is inaccurate (29) when compared with surgical evaluation. However. the maximal usefulness of these imaging modalities may depend on imager experience and the use of special imaging sequences and techniques not in univer- sal use. For patients with FIGO stage 182. barrel shaped. or bulky stage ”A cancers. treatment may consist of combined—modality therapy with initial radiation (30) or neoadjuvant chemotherapy followed by surgery (31). or by radiation therapy alone (30). In patients for whom surgery will not be the initial intervention. pretreatment diagnostic imaging may be of value in the deter— mination of appropriate target volumes for radiotherapy and in the design of radiation treatment ports in a fashion analogous to the use of imaging for patients with stages llB-lVA disease. Since the risk of lymph node metastasis increases with increas— ing volume of primary tumor (7). lymphangiography might be expected to have a higher yield in these patients than in patients with stage 181 tumor. Accurate determination of primary tumor volume by MRI or endocavitary ultrasound could result in more refined size stratification for patients entering prospective clini- cal trials of neoadjuvant therapy. Imaging Assessment in Patients Who Are Treated Initially by Radiotherapy For patients who are to receive attempted curative treatment with radiation-based therapy. critical initial parameters to deter- mine are the target volumes for gross disease (both primary dis— ease and its extensions as well as gross adenopathy) and potential microscopic or occult disease. Discrepancies between FIGO clinical stage and pathologic stage determined by opera— tive or imaging assessment with fine—needle aspiration biopsy are of academic interest but of lesser practical importance than assurance that all the cancer lies within the anticipated treatment volume. Pretreatment surgical assessment with histopathologic definition of initial disease extent remains the standard against which other evaluations must be measured (14). although this assessment must itself be recognized as a sampling with some inherent risk of error with failure to detect occult disease that may subsequently become manifest (1/). 1n the American Col- lege of Surgeons Patient Care Evaluation Study of cervical can- Journal ofthe National Cancer Institute Monographs No. 21. 1996 cer (5), approximately one third of the patients with invasive cervical cancer of known clinical stage were 60 years of age or older. The rising incidence of cotnorbidities with age might render surgical evaluation of disease extent unwise for many of these patients. In the American College of Surgeons study. pretreatment diagnostic surgery was performed in 10.8% of the 3073 patients treated with radiation therapy without hysterec— tomy in 1984. In I990. this had risen to 15.2% of 2427 patients. On the basis of surgical assessment of para—aortic lymph nodes in a study performed by the Gynecologic Oncology Group (I4). 24% of the patients with stages llB-lV cervical can- cer will have node metastases cephalad to the pelvic lymph nodes treated with standard pelvic radiation ports. Only 34% of the patients with para—aortic lymph node metastases were iden— tified by CT scanners using criteria used in the mid-1980s. Lymphangiography identified 79% of these patients. with a falsevnegative rate of only 8%. The cost of surgical assessment. either by laparoscopy or laparotomy. will be severalfold that of minimally invasive diagnostic imaging. and surgery may delay the initiation of treatment with radiotherapy. Substantially more than half of the patients with disease contaminating the para- aortic lymph nodes will die from disease (12). It is controversial whether the identification of a comparatively small number of patients with para—aortic adenopathy not identifiable by imaging can justify the morbidity and expense of routine use of surgical assessment of these nodes. Routine extension of the radiation treatment volume to en— compass the para—aortic lymph nodes will result in a modest im- provement in survival (32) in patients with bulky stage IB and 11A cervical carcinomas. but the routine use of extended fields will risk additional acute and chronic morbidity for the majority of patients with negative extrapelvic lymph nodes. It would seem more sensible to reserve surgical evaluation of the nodes for patients in whom imaging is otherwise unable to define the treatment volume. In Sacramento, California. in the absence of major medical contraindications. it has been a long-standing policy to ad- minister extended-field radiation therapy (pelvic and para-aortic node volumes) to any patient with a lymph node metastasis (pel— vic or para—aortic) identified by imaging and confirmed by fine— needle under fluoroscopic or CT guidance. Surgical assessment of lymph nodes is offered only to those patients in whom the imaging assessment is negative or cytologically unconfirmed. Patients with a negative lymph node assessment have only the pelvis irradiated. Employing a cascade of diagnostic assess— ments in a sequential. decision—tree fashion (Fig. 4) derives the information necessary to make treatment decisions without sub— jecting each patient to the expense and morbidity of each diag— nostic assessment (33). Assessments progress from the less to the more invasive. The majority of patients will not have opera— tive evaluation of their retroperitoneal lymph nodes. either be— cause the imaging studies render such evaluation superfluous or because medical contraindications or patient refusal supervene. Cancer of the cervix is an eminently radiocurable disease. but unforgiving. Patients not cured by their initial course of therapy are rarely salvaged. Vital to successful treatment is inclusion of ABDOMINO-PELVIC MRIJ (+) j CT GUIDED FNA ADENEXA (+) (") CT 0R FLUOROSCOPIC FNA (+) (+) EVALUATION ? LYMPHANGIOGRAPHY (+) (') (—) SURGICAL OR "INOPERABLE" <— LAPARASCOPIC —>"OPERABLE" (+) Fig. 4. Diagnostic cascade for definition of lymph node target volume for initial external—beam radiotherapy: ln- ternational Federation of Gynecology and Obstetrics stages lBZ-IVA. (+) = an imaging result suspicious for retroperitoneal lymph node metastasis or cytologic con— firmation of metastasis, (7) = a normal examination or absence of malignant cells on biopsy. Adenexa (+) = ab— normal imaging findings in the fallopian tubes or ovaries. either potentially neoplastic or inflammatory: FNA = fine—needle aspiration biopsy [from (33)]. v INVOLVED FIELDS: PELVIC PELVIS + SUSPICIOUS NODES UNCONFIRMED NODES ALONE EXTENDED VOLUME: PELVIC+PARA-AORTIC LYMPH NODES Journal of the National Cancer Institute Monographs No. 2|. I996 the entire cancer and its regional extensions in the irradiated volume. Bipedal lymphangiography will prove useful in the sizing and contouring of radiation fields (34). In the absence of a lymphangiogram. boney landmarks (33) and palpation of the femoral arteries (35) will give reasonably accurate target vol— umes for pelvic and para—aortic nodes. Body imaging (MRI or CT) may be very helpful in the design of radiation teletherapy ports. Two studies (36,37) have shown that conventional lateral fields. as described and illustrated in reference texts. may be in- adequate in their coverage of locally extensive cancers in ap— proximately 25% of the patients. and that this may adversely affect local control and survival. MRI would appear to be the preferred initial imaging modality in patients with extensive cancers because it is approximately isoeffective with CT in detection of nodal metastases. has superior ability to measure tumor size and evaluate local extensions of disease (9). and can routinely image in transverse. coronal, and sagittal planes. The ability of MRI to delineate the internal architecture of the uterus and illustrate extension of cancer to the endometrial cavity and myometrium exceeds that of CT. Although the patients whose sagittal MRI images are seen in Figs. I and 3 both had stage [IB cancer. based on physical examination. it should be apparent to even the novice brachytherapist that somewhat different brachytherapy dose distributions would be preferred in each patient. Either body imaging method has the potential to help in— dividualize and optimize treatment by remote afterloading brachytherapy so that dose distributions more accurately encom- pass identifiable disease and spare portions of adjacent normal structures. Standard radiation dose distributions. made on the basis of idealized anatomy and developed in an era prior to sophisticated body imaging. will eventually be replaced by tailored dose distributions that more closely conform to actual tumor distribution. However. routine use of either CT or MRI for dose optimization with intracavitary or interstitial brachy— therapy appliances in situ awaits the commercial availability of applicator hardware compatible with these imaging methods. Diagnostic imaging can be useful during the course of thera— py as a tool to define reduced treatment volumes for boost therapy to groups of lymph nodes or residual gross primary can- cer after shrinkage has been accomplished by initial therapy to a larger volume. Efforts to improve locoregional control and sur- vival by shortening overall treatment time through the use of tactics. such as the administration of a concurrent boost dose to gross disease or a second daily treatment fraction to a reduced volume. can only be sensibly implemented if the reduced volume can be specified and targeted with precision. This neces- sarily will involve diagnostic imaging in some fashion. Treatment Failure and Salvage Therapy Despite appropriate surgical or radiotherapeutic treatment. some patients will relapse locally. regionally. or distantly. Sal— vage therapies frequently result in more morbidity than primary therapies and yield fewer cures. It is both kind and economical to use diagnostic imaging to perform a thorough search for dis- seminated disease prior to launching a hazardous and expensive salvage effort. In irradiated patients, MRI may be helpful in dis- tinguishing recurrent tumor from radiation fibrosis and may be 40 valuable in the preoperative assessment of potential candidates for salvage exenterative surgery. Future Directions Thoughtful appraisal of patterns of failure will lead to refine- ments in existing therapies and new strategies to address the deficiencies of established treatments that are neither uniformly effective nor uniformly safe. There is no established role yet for routine diagnostic imaging in the surveillance follow—up of pa— tients treated for cervical cancer. However. integration of diag- nostic imaging evaluations into surveillance follow—up of patients entered in clinical trials might serve to identify first sites of failure with greater precision and timeliness and add substantial insight into the interpretation of results. In an era of increasingly complex. multimodality therapy. historic patterns of failure in patients undertreated (by contemporary standards) may not be the best guide for development of innovative new therapies or combinations of established modalities. Patterns of failure. as revealed by pelvic examination and plain chest radiography. will not be as informative as patterns of failure defined by these traditional means and supplemented by CT or MRI. The true rate of failure in pelvic or para-aortic nodes will likely be underestimated if one relies on the development of pain and lymphedema to suspect the diagnosis. Once the presence of disseminated or otherwise incurable disease has been established. the patient may be a candidate for systemic cytotoxic chemotherapy or other novel systemic therapies. Evaluation of therapeutic efficacy usually entails objective measurement of tumor response. Diagnostic imaging may pro- vide the means to make such measurements for patients whose disease is not accessible to physical examination. Final patterns of failure and cause of death have been tradi- tionally determined by postmortem examination. For all but coroner‘s cases. the autopsy has virtually vanished from the American medical landscape. Antemortem body imaging has become the autopsy of the 90s. Finally. patterns of failure can also shed light along the pathway to improved future therapies. From this brief survey. it is apparent that diagnostic imaging plays integral and growing roles in the initial assessment and clinical management of patients with cervical cancer. The roles of old imaging assessments will likely decline further. to be in- creasingly supplanted by more sophisticated. sensitive. and ac— curate imaging methods. ln the past. a deficiency in the introduction of new imaging methods has been the lack of a sys— tematic. dispassionate approach to evaluate their usefulness in various clinical settings and. when appropriate. a program to en— sure their integration into existing diagnostic and staging algo- rithms. Body imaging (CT and MRI) has been adopted into the clinical management of patients with cervical cancer without a consensus concerning when and how such studies should be employed. Entrepreneurial claims of usefulness by equipment manufacturers and partisan advocacy ofdiagnostic imagers have conflicted with the skepticism. bordering on nihilism. of parties interested in constraining burgeoning medical costs or promot— ing competitive patient assessments. The increased use of these modalities between 1984 and 1990 (Table l) suggests that many clinicians found these studies useful. but the methodology of the Journal of the National Cancer Institute Monographs No. 21. I996 American College of Surgeons study does not allow us to know why these imaging studies were obtained or how their results in- fluenced patient management. In the future, medicine would be better served by early. comprehensive. prospective evaluation of new imaging modalities to avoid some of the needless adver- sarial controversies that have attended the development of modern body imaging. The work of the Gynecologic Oncology Group (/4) is testimony to the feasibility of performing such evaluations in a multi-institutional context that may more ac- curately reflect the potential value and realistic limitations of body imaging rather than a multiplicity of sometimes conflicting assessments from single institutions. References (9) (/0) (II) (/2) (/3) (I4) (/5) (I6) Parkin DM. Pisani P. Ferlay J. Estimates of the worldwide incidence of eighteen major cancers in I985. lnt J Cancer 1993;54:594-606. Piver MS. Chung WS. Prognostic significance of cervical lesion size and pelvic node metastases in cervical carcinoma. Obstet Gynecol |‘)75;45: 507-10. Stehman FB. Bundy BN. DiSaia PJ. Keys HM. Larson JE. Fowler WC. Carcinoma of the cervix treated with radiation therapy I. A multi-variate analysis of prognostic variables in the Gynecologic Oncology Group. Can— cer [991:6722776-85. Homesley HD. Raben M. Blake DD. Ferree CR. Bullock MS. Linton EB. et al. Relationship of lesion size to survival in patients with stage IB squamous cell carcinoma of the cervix uteri treated by radiation therapy. Surg Gynecol Obstet 19801150252931. Russell AH. Shingleton HM. Jones WB. Fremgen A. Winchester DP. Clive R, et al. Diagnostic assessments in patients with invasive cancer of the cervix: a national pattern of care study of the American College. Gynecol Oncol. In press. Montana GS. Hanlon AL. Brickner TJ. ()wen JB. Hanks GE. Ling CC. et al. Carcinoma of the cervix: patterns of care studies: review of I978. 1983. and 1988-1989 surveys [we comment citation in Medline]. Int J Radiat Oncol Biol Phys 1995;32:1481-6. Burghardt E. Pickel H. Local spread and lymph node involvement in cervi» cal cancer. Obstet Gynecol [978:52: I 38—45. Hawnaur JM. Johnson RJ. Buckley CH. Tindall V. lsherwood l. Staging. volume estimation and assessment of nodal status in carcinoma of the ccr- vix: comparison of magnetic resonance imaging with surgical findings. Clin Radiol l994:49:443—52. Subak LL. Hricak H. Powell CB. Azizi L. Stern JL. Cervical carcinoma: computed tomography and magnetic resonance imaging for preoperative staging. Obstet Gynecol 1995;86:43-50. Meigs JV. Radical hysterectomy with bilateral pelvic lymph node dissec— tions: a report of 100 patients operated on five or more years ago. Am J Obstet Gynecol [95 | :622854—66. Kinney WK. Hodge DO. Egorshin EV. Ballard DJ. Podratz KC. Surgical treatment of patients with stages [B and HA carcinoma of the cervix and palpably positive pelvic lymph nodes. Gynecol Oncol l995z57: [45-9. Rottnan M. Axil H. Eifel PJ. Irradiation of pelvic and para-aortic nodes in carcinoma ofthe cervix. Semin Radiat Oncol 1994141234). Russell AH. Jones DC. Russell KJ. Gerdes AJ. Figge DC. Greer BE. et al. High dose para-aortic lymph node irradiation for gynecologic cancer: tech» nique. toxicity. and results. Int J Radiat Oncol Biol Phys 1987;13:267-71. Heller PB. Maletano JH. Bundy BN. Bamhill DR. Okagaki T. Clinical- pathologic study of stage 11B. 111. and WA carcinoma of the cervix: ex- tended diagnostic evaluation for paraaortic node metastasis#a Gynecologic Oncology Group study. Gynecol Oncol 1990;38:425-30. Burghardt E. Baltzer J. Tulusan AH. Haas J. Results of surgical treatment of 1028 cervical cancers studied with volumetry. Cancer 1992170134855. Potter ME. Alvarel. RD. Shingleton HM. Soong SJ. Hatch KD. Early in- vasive cervical cancer with pelvic lymph node involvement: to complete or not to complete radical hysterectomy"? Gynecol Oncol 1990;37:78—8 l. Journal of the National Cancer Institute Monographs No. 2 l. 1996 (I7) ([8) (I9) (20) Iv \. (34) (25) (26) (27 (28) (29) (30) (3/) (33) (34) (35) (36) (37) Potish RA. Downey GO. Adcock LL. Prem KA. Twiggs LB. The role of surgical debulking in cancer of the uterine cervix. Int J Radiat Oncol Biol Phys 1989;17:979-544. Olson N. Koh W. Greer B. Cain J. Tamitni H. Figge D. Macroscopic lymph node metastases in stage IB carcinoma of the cervix explored for radical hysterectomy [abstract]. Gynecol Oncol 1992;46:267. Delgado G. Bundy BN. Fowler WC Jr. Stehman FB. Sevin B. Creasman WT. et al. A prospective surgical pathological study of stage I squamous carcinoma of the cervix: a Gynecologic Oncology Group study, Gynecol Oncol 1989:352314—20. Hricak H. Powell CB. Yu KK. Washington E. Subak LL. Stern JL. et al. Invasive cervical carcinoma: role of MR imaging in pretreatment work- up%ost minimization and diagnostic efficacy analysis. Radiology 1996; 198:403-9. Beyersdort’f D. Bahnsen J. Frischhier HJ. Nodal involvement in cancer of the uterine cervix: value of lymphography and MRI. EurJ Gynaecol Oncol l995;|6:27-l-7. Boie H. Jakohsen A. Petersen J. Sell A. Diagnostic value of lymphography in cervical cancer stage lb. EurJ Gynaecol Oncol l989; l():393-5. Swan E. Bouma J. Schuur K. The clinical value of lymphography in cervi— cal cancer. FlGO—stage lb-Ila. EurJ Gynaecol Oncol l989:l():85—9(). Sironi S. Belloni C. Taccagni GL. DelMaschio A. Carcinoma of the cer- vix: value of MR imaging in detecting parametrial involvement. AJR Am J Roentgenol1991:l56:753—6. Lien HH. Blomlie V. lversen T. Trope C. Sundfor K. Abeler VM. Clinical stage I carcinoma of the cervix. Value of MR imaging in determining in— vasion into the parametrium. Acta Radiol 1993;34:130—2. Togashi K. Nishimura K. Sagoh T. Minami S. Noma S Fujisawa I. et al. Carcinoma of the cervix: staging with MR imaging. Radiology |‘)89:l7l: 245-5I. Houvenaeghel G. [)elpero JR. Rosello R. Resbeut M. Viens P. Jacquemier J. et al. Results of a prospective study with comparison of clinical. en- dosonographic. computed tomography. magnetic resonance imaging and pathologic staging of advanced gynecologic carcinoma and recurrence. Surg Gynecol Obstet 1993;177:2314). Mamsen A. Ledertoug S. Horlyck A. Knudsen HJ. Rasmussen KL. Nyland MH. et al. The possible role of ultrasound in early cervical cancer. Gynecol ()ncol 1995;56: I 87-90. Van Nagell JR Jr. Roddick JW Jr. Lowin DM. The staging of cervical can- cer: inevitable discrepancies between clinical staging and pathologic find- ings. Am J Obstet Gynecol l97|;l [0:973—8. Perez CA. Grigsby PW. Camel HM. Galakatos AE. Mutch D. Lockett MA. Irradiation alone or combined with surgery in stage IB. 11A. and 118 car— cinonta of the uterine cervix: update of a nonrandomiled comparison [see comment citation in Medlinel. Int J Radiat Oncol Biol Phys 1995;31:703- 16. Sardi J. Sananes C. Giaroli A. Bayo J. Rueda NG. Vighi S. et al. Results of a prospective randomized trial with neoadjuvant chemotherapy in stage lB. bulky. squamous carcinoma of the cervix [see comment citations in Med- linel. Gynecol Oncol 1993;49:156~65. Rotman M. Pajak TF. Choi K. Clery M. Marcial V. Grigsby PW. et al. Prophylactic extended-field irradiation of para»aortic lymph nodes in stages [[8 and bulky IB and [IA cervical carcinomas. Teneyear treatment results of RTOG 79-20 [sec comment citation in Medlinel. JAMA 1995; 2742387493. Russell AH. Contemporary radiation treatment planning for patients with cancer of the uterine cervix, Semin Oncol 1994;21:30—41. Pendlebury SC. Cahill S Crandon AJ. Bull CA. Role of bipedal lymph- angiogram in radiation treatment planning for cervix cancer. Int J Radiat Oncol Biol Phys l9‘)3:27:959-62. Greer BE. Koh WJ. Figge DC. Russell AH. Cain J. Tamimi HK. Gynecologic radiotherapy fields defined by intraoperative measurements Gynecol Oncol 1990382434. Russell AH. Walter JP. Anderson MW. Zukowski CL. Sagittal magnetic resonance imaging in the design of lateral radiation treatment portals for patients wth locally advanced squamous cancer of the cervix. Int J Radiat Oncol Biol Phys l992;23:449—55. Kim RY. McGinnis LS. Spencer SA. Meredith RF. Jennelle RL. Salter M. Conventional four-field pelvic radiotherapy technique without computed toinography-treatment planning in cancer of the cervix: potential geographic miss and its impact on pelvic control. Int J Radiat Oncol Biol Phys [9953110942. 41 What Is the Appropriate Management of Early Stage Cervical Cancer (International Federation of Gynecology and Obstetrics Stages I and HA), Surgical Assessment of Lymph Nodes, and Role of Therapeutic Resection of Lymph Nodes Involved With Cancer? David H. Moore, Frederick B. Stehman* It is well recognized that lesion size, tumor volume, depth of stromal invasion, and lymphatic space permeation are all important predictors of lymph node involvement in early stage cervical cancer. Pelvic lymph node involvement is the most important (negative) predictor of survival for these patients with early stage cervical cancer. The number of in- volved nodes and the size of involved and unresected nodes may also be prognostically significant. It is uncertain whether lesion size, tumor volume, depth of stromal in- vasion, or lymphatic space permeation are independent negative predictors of survival when correcting for lymph node positivity. Lymphadenectomy has traditionally been considered a diagnostic procedure. There is accumulating evidence to suggest that lymphadenectomy may have therapeutic benefit for patients with cervical cancer meta- static to lymph nodes. This hypothesis awaits further evidence. [Monogr Natl Cancer Inst 1996;21:43-6] Cervical cancer may spread by vascular embolization or by transperitoneal dissemination. The major route of metastatic spread. however. is direct tumor extension and lymphatic per- meation. None of these processes occurs in isolation. Larger tumors with deeper stromal infiltration are more likely to be associated with vascular—lymphatic space involvement and pel— vic lymph node metastases. Surgical treatment of cervical cancer originated 100 years ago. Working with cadaveric specimens. Ries (I) was the first to propose that routine lymphadenectomy be performed as part of the abdominal surgery with curative intent. Wertheim (2). who acquired extensive clinical experience with and popularized the radical hysterectomy procedure that bears his name, performed lymph node biopsies only on a selective basis. During the first half of the 20th century. radical hysterectomy lay dormant. replaced by radiation therapy. The procedure experienced a rebirth following the report of Meigs (3) to the prestigious American Gynecological Society. Meigs presented 5-year fol— low—up data on 100 consecutive patients operated on by himself without perioperative mortality. He routinely performed pelvic lymphadenectomy in combination with radical abdominal hysterectomy. influencing others to do the same. Journal ofthe National Cancer Institute Monographs No. 21. 1996 Our understanding of the metastatic spread patterns in early stage disease may be attributed to these and subsequent surgical pioneers. There are multiple descriptive studies and time series using this intervention. Lymphadenectomy has become incor— porated in the primary surgical therapy without a rigorous con- trolled trial. however. Predictors of Lymph Node Metastases in Early Stage Disease Approximately 15% of patients with International Federation of Gynecology and Obstetrics (FIGO) stage IB squamous cell cancer of the cervix have pelvic lymph node metastases (4). There is one subset of patients with invasive squamous cell car— cinoma of the cervix for whom the risk of pelvic lymph node metastasis is minuscule. The incidence of lymph node metas- tases is sufficiently rare for cancers invading 3 mm or less ofthe cervical stroma and without lymphatic space involvement to eliminate the need for routine lymph node dissection. Lymph node metastases are infrequent with depth of invasion up to 5 mm (5-8). Burghardt and Pickel (9) performed cross-sectional area measurements of radical surgery specimens and found an increase in lymph node metastases with increasing tumor area: less than 100 mm2 (0%), 100—599 mm2 (19%). and 600 mm2 or more (46%). Using a three-dimensional approach. Baltzer et a1. ([0) correlated lymph node metastases with increasing tumor volume and reported no lymph node metastases in patients with tumors 500 mm3 or less in size. Although volumetric measure— ments have prognostic importance. they are labor intensive and remain unpopular in the United States. Interestingly, Burghardt et al. reanalyzed their series according to new FIGO staging criteria. Among 344 women with stage IAl disease. only one (0.29%) developed recurrent cancer 12 years after cervical con- ization. Among women with stage 1A2 disease. cancer recurred in five (4.9%) of 101 (ll). *Afliliulimi of uulhmzr: Department of Obstetrics and Gynecology. Indiana University School of Medicine. Indianapolis. CtII‘I‘é’A/NHYU'UIICE’ m: Frederick B. Stehman. M.D., Department of Obstetrics and Gynecology. Indiana University School of Medicine. 550 N. University Blvd. UII-2440. Indianapolis. IN 46202—5274. 43 In stage IB cervical cancer. the propensity for lymph node dissemination is a function of tumor size or volume. depth of stromal invasion. and the permeation of lymphatic spaces. It would appear that clinical lesion size and depth of stromal in- vasion are adequate surrogates for volumetric tumor measure— ments. Several investigators (4.]2-I4) have reported a positive correlation between increasing lesion size and pelvic lymph node metastases. Moreover. the prospective study (/5) con— ducted by the Gynecologic Oncology Group (GOG) demon— strated a strong correlation between depth of stromal invasion. the risk of pelvic lymph node metastases. and disease-free sur- vival. In a re—evaluation of this GOG data. Zaino et al. (/6) con— firmed the reproducibility and prognostic significance of both stromal and lymphatic invasion and stated that these should be reported routinely. Others (17,18). besides the GOG. have described the impor— tance of lymphatic space invasion as a predictor of lymph node metastases. For squamous cell cervical carcinoma. histologic grading has no prognostic significance (/6). Our ability to identify patients with lymph node metastases prior to surgery is suboptimal. The most commonly used imag— ing study. computed tomography (CT). has an unacceptably high cost and limited benefit (/9). Heller et al. (20). reporting for the GOG. evaluated CT prospectively in patients with local— ly advanced disease (and a higher rate of node metastases). They found a sensitivity of 34.4% and false—negative rates of 17.8% for para-aortic nodes and 25% for pelvic nodes. Imaging studies are little or no better than clinical surrogates. such as tumor volume. for predicting nodal metastasis. Lymph Node Metastases: Survival Several investigators (4.13.20 have further defined that. for early stage cervical cancer. the presence of pelvic lymph node metastases is the most important poor prognostic factor for sur— vival. For patients who undergo complete surgery and who have negative pelvic lymph nodes. the 5—year survival rate is 88%- 93% compared with 40%—63% for patients with positive pelvic lymph nodes. The clinical significance attached to increasing tumor size. stromal invasion. and lymphatic space involvement partially reflects the predictive value ofthese parameters for pel— vic lymph node metastases. Using multivariate analysis. Alvarez et al. (13) found that both age and tumor size contributed addi- tional risk of adverse outcome for patients with positive lymph nodes. Delgado et al. (15), reporting the GOG data. found that size and lymph—vascular involvement were significant for pa— tients with negative lymph nodes. Histology (squamous versus nonsquamous) and tumor size were significant in multivariate analysis by Kamura et al. (2]). Many studies lack sufficient power to determine whether these parameters have independent prognostic significance for survival. In addition to the presence of pelvic lymph node metastases. the number of involved lymph nodes and extension of tumor to the para-aortic lymph nodes have been reported to have prog- nostic significance. Even though the GOG ([5) was unable to find a correlation between the number of involved lymph nodes and disease-free survival in a subgroup of resected patients. others have observed a worse outcome for patients with two or 44 more involved lymph nodes. Burghardt et al. (22) reported a 5- year survival rate of 88.7% for patients with no pelvic lymph node metastases. 64.3% for one. 54.1% for two. and 33.8% for three to eight lymph node metastases. Alvarez et al. (13) reported that the probability of relapse was doubled when three or more lymph nodes were involved compared with a single lymph node. Extracapsular extension or more than a single lymph node was found to be significant by Tinga et al. (23). This difference persisted for the subsets of patients with only pelvic nodes involved and for only surgically occult nodes in— volved. These findings are corroborated by Inoue and Morita (24) and Kamura et al. (2/). The latter used the number of nodal groups involved rather than individual nodes. Approximately 50% of the patients with positive pelvic lymph nodes will prove to have metastases to the aortic lymph nodes. Only 25% of these patients will ultimately survive (25.26). Survival approaches zero if there is distant spread. There is no standard approach when metastatic disease is en— countered at the time of radical hysterectomy. In the GOG report. l()0 (l5.5%) of 645 patients with stage IB disease had microscopically positive lymph nodes (4). An additional [9 patients had grossly enlarged lymph nodes. Thus. the number of patients in whom an intraoperative decision is necessary is small. The rarity of the event may explain why there is little evidence available in the literature to help make this decision. The available studies often reflect individual institutional bias. There is. however. clear evidence in the literature that excellent central tumor control can be obtained with radiation therapy alone without hysterectomy (27,28). Therapeutic Lymph Node Resection Lymphadenectomy has been traditionally viewed as a diag— nostic procedure. Discrepancies between clinical staging and surgical—pathologic findings occur in about one third of the patients. prompting several investigators (29-31) to recommend pretreatment surgical staging for patients at risk for aortic lymph node involvement. Theoretic advantages to surgical staging in- clude the following: I) to identify extrapelvic disease sites. 2) to remove inflamed adnexa or myomata that may complicate radiotherapy. and 3) to downstage patients and allow for more appropriate treatment (29). The morbidity associated with the combination of surgery and irradiation can be ameliorated by the use of an extraperitoneal technique (30.32). Pretreatment surgical staging is unnecessary for most early stage cervical can— cers but is a potential consideration for stage [82 (bulky) tumors in which the incidence of positive aortic lymph nodes may be significant (3/). Whether surgical resection of involved lymph nodes has therapeutic benefit is uncertain. There are no randomized trials to aid the gynecologic oncologist when he or she encounters bulky lymph nodes intraoperatively. Does the therapeutic ad— vantage of debulking these nodes outweigh potential morbidity? Those data that are available are subject to the same selection bias encountered in most studies of ovarian cancer debulking. Although unconfirmed by a randomized controlled trial. the survival of patients with early stage cervical cancer treated with surgery or radiation therapy appears to be equivalent. These sur— Journal ofthc National Cancer Institute Monographs No. 2!. I996 vival data could be imbalanced, however. Reports of the use of radiation therapy tend to use “intent to treat” methodology; that is, that all patients who were assigned to therapy are accounted for. Many surgical studies do not account for patients who were explored but not resected. In addition, surgical trials may in— clude a disproportionate representation of favorable prognosis patients (4). These data do imply that radiation therapy alone is no less effective in treating nodal disease. Clear evidence that radiation therapy is able to eradicate metastatic disease in lymph nodes was provided by Lagasse et al. (33). This group randomly assigned 1 18 patients to node dis- section prior to or following radiation therapy. There was a 50% reduction in the number of patients with positive nodes after radiation therapy (22.4% versus 11.7%). Radiation therapy would be expected to be effective treatment of micrometastases but less effective in the treatment of bulky nodal disease. They concluded that micrometastases but not bulky positive nodes could be sterilized with radiation therapy. Several studies (34-38) have reported enhanced survival with lymph node resection; however. patients found to have lymph node metastases generally received radiation therapy. Survival differences may have reflected the underdiagnosis of lymph node metastases rather than any therapeutic advantage to the debulking procedure. Inoue et al. (34) noted significantly poorer survival among patients whose nodal disease was unresectable. In their experience, 16 (94%) of 17 patients with unresected bulky pelvic lymph nodes treated with radiation therapy developed recurrent cancer. Kjorstad et al. (35) at the Norwegian Radium Hospital eval— uated the value of lymphangiography to direct lymphadenec- tomy. When there were four or more lymph nodes remaining on a postoperative radiograph, there were more pelvic recurrences than if three or fewer lymph nodes remained. They were unable to prove an advantage in a complete dissection, however. A lower rate of recurrence was observed by Tulzer and Kupka (36) among a cohort of patients who had obligatory lymphadenec- tomy compared with a historical comparison group who had selective lymphadenectomy. Among patients with advanced cervical cancers treated with radiation therapy, Vigliotti et al. (37) reported 5—year survival rates of 50% with microscopic disease. 33% with residual dis— ease less than 2 cm, and 0% with residual disease of 2 cm or more in the para—aortic lymph nodes. In 1976, Keys and Park (38) reported a 5-year survival rate of 75% for patients with stage IB cervical cancer who underwent resection of gross nodal disease. There may be a therapeutic advantage to surgical debulking of pelvic nodal metastases. In a series reported by Potish et al. (39), one third of 159 patients with advanced cervical cancers had grossly positive pelvic lymph nodes that were debulked. More than one half of these patients survived without evidence of disease recurrence versus no survivors among the patients with unresectable lymph nodes. Furthermore. the disease—free survival in patients with microscopic nodal metastases was the same as that for patients with grossly positive pelvic nodes that were completely resected. In preparation for this Consensus Development Conference. those investigators provided us with an update of their series. The series now includes 267 patients; Journal of the National Cancer Institute Monographs No. 2 l. 1996 136 with negative nodes, 40 with micrometastases. 78 with bulky nodes resected. and 13 with unresectable nodes (groups A—D, respectively). The 5—year survival rates (Kaplan—Meier) for groups B and C are similar, 47% and 50%, respectively. Both are superior to the zero survival rate among group D (Cosin JA, Twiggs LB. Potish RA: personal communication). Hacker et al. (40) resected bulky positive lymph nodes detected at the time of radical hysterectomy (23 patients) or by CT (11 patients). One patient did not complete planned post- operative radiotherapy. Of the remaining 33 patients. 28 received pelvic and para—aortic radiation and 23 also received cisplatin chemotherapy. For patients undergoing radical hysterectomy. the overall survival was 80% with positive pelvic nodes and 48% with positive aortic nodes. Survival was com— parable to that for patients with microscopic lymph node in- volvement (40). All of these results are intriguing. Verification of the thera— peutic effectiveness of lymphadenectomy or nodal debulking awaits the collection of further evidence. Summary 1) Lesion size. tumor volume, depth of stromal invasion, and lymphatic space permeation are important predictors of lymph node involvement in early stage cervical cancer. 2) Pelvic lymph node involvement is the most important negative predic- tor of survival in early stage cervical cancer. The number of in- volved nodes and the size of involved and unresected nodes may also be prognostically significant. 3) It is uncertain whether lesion size, tumor volume, depth of stromal invasion. and lym- phatic space permeation are independent negative predictors of survival when correcting for lymph node positivity. 4) Lym- phadenectomy has traditionally been considered a diagnostic procedure. 5) There is accumulating evidence to suggest that lymphadenectomy may have therapeutic benefit for patients with cervical cancer metastatic to lymph nodes. This hypothesis awaits further evidence. References (I) Ries E. Eine neue operations methode des uteruscarinoms. A Geburtshilfe Gynaekol 1895;32:266-74. (2) Wertheim E The extended abdominal operation for carcinoma uteri. (Translated by 11. Grad). Am J Obstet Dis Women and Children 1912: 66:1697232. (3) Meigs JV. Radical hysterectomy with bilateral pelvic lymph node dissec— tions: a report of 100 patients operated on five or more years ago. Am J Obstet Gynecol 1951;62:854-661. Delgado G. Bundy BN. Fowler WC Jr, Stehman FB, Sevin B. Creasman WT. et al. A prospective surgical pathological study of stage I squamous carcinoma of the cervix: a Gynecologic Oncology Group study. Gynecol Oncol 1989;35:314-20. (5) Averette HE. Nelson JH Jr. Ng AB. Hoskins WJ. Boyce JG. Ford JH Jr. Diagnosis and management of microinvasive (stage 1A) carcinoma of the uterine cervix. Cancer 1976;38:414-25. (6) van Nagell JR Jr. Greenwell N. Powell DF, Donaldson ES, Hanson MB, Gay EC. Microinvasive carcinoma of the cervix. Am J Obstet Gynecol 1983;145:981—91. (7) Creasman WT. Fetter BF. Clarke-Pearson DL. Kaufmann L. Parker RT. Management of stage IA carcinoma of the cervix. Am J Obstet Gynecol 1985;153:164-72. (8) Sedlis A. Sall S. Tsukada Y, Park R, Mangan C. Shingleton H. et al. Microinvasive carcinoma of the uterine cervix: a clinical-pathologic study. Ath Obstet Gynecol 1979;133:64—74. (4 45 (9) (l0) (1/) (/2 (/3) (I4) (/5 (/6) (l7) (/8) (/9) (2m (2]) 46 Burghardt E. Pickel H. Local spread and lymph node involvement in cervi— caI cancer. Obstet Gynecol 19785213845. Baltzer J. Lohe KJ. Kopcke W. Zandet' J, Histologic criteria for the prog- nosis in patients with operated squamous cell carcinoma of the cervix. Gynecol ()ncol [9242;13:184-941. Burghardt E. Girardi F. Lahousen M. Pickel H. Tainussino K. Microin- vasive carcinotna of the uterine cervix (International Federation of Gynecology and Obstetrics stage IA). Cancer l99l;o721037-45. Piver MS. Chung WS. Prognostic significance of cervical lesion siIe and pelvic lymph node metastases in cervical carcinoma. Obstet Gynecol I975;46:507-l0. Alvarez RD. Soong SJ. Kinney WK. Reid GC. Schray MF. Podratz KC. et al. Identification of prognostic factors and risk groups in patients found to have nodal metastasis at the time of radical hysterectomy for early-stage squamous carcinoma ofthe cervix. Gynecol Oncol 1989351305. Fuller AF Jr. Elliott N. Kosloff C. Lewis JL Jr. Lymph node metastases from carcinoma of the cervix. stages IB and IIA: implications for prog- nosis and treatment. Gynecol ()ncol 1982;13:165—74. Delgado G. Bundy B. Zaino R. Sevin BU. Creasman WT. Major F. Prospective surgical»pathological study of disease-free interval in patients with stage IB squamous cell carcinoma of the cervix: a Gynecologic On- cology Group study. Gynecol ()ncol |99();38:352—7. Zaino RJ. Ward S. Delgado G. Bundy B. Gore H. Felter G. et al. His- topathologic predictors of the behavior of surgically treated stage IB squamous cell carcinoma of the cervix. A Gynecologic Oncology Group study. Cancer [992:6921750—8. van Nagell JR Jr. Donaldson ES. Wood EG. Parker JC Jr. The significance of vascular invasion and Iymphocytic infiltration in invasive cervical can— cer. Cancer I97X:4I 2228-34. Barber BR. Sommers SC. Rotterdam H. Kwon T, Vascular invasion as a prognostic factor in stage IB cancer of the cervix. Obstet Gynecol I978: 52:343-8. Moore DH. Dotters DJ. Fowler WC Jr. Computed tomography: does it really improve the treatment of cervical carcinoma? Am J Obstet Gynecol |992;I67:768»71. Heller PB. Maletano JH. Bandy BN. Barnhill DR. Okagaki T. Clinical- pathologic study of stage IIB. III. and IVA carcinoma of the cervix: ex- tended diagnostic evaluation for paraaortic node metastasis—a Gynecologic Oncology Group study, Gynecol Oncol 1990;324:4530. Kamura T. Tsukamoto N. Tsuruehi N. Saito T. Matsuyama T. Akazawa K. et al. Multivariate analysis of the histopathologic prognostic factors of cer- vical cancer in patients undergoing radical hysterectomy. Cancer I992;69:18l-6. Burghardt E. Pickel H. Haas J. Lahousen M. Prognostic factors and opera— tive treatment of stages IB to IIB cervical cancer. Am J Obstet Gynecol l987:|56:988-96. Tinga DJ. Timmer PR. Bouma J. Aalders JG. Prognostic significance of single versus multiple lymph node metastases in cervical carcinoma stage IB. Gynecol Oncol 1990:39: I 75-80. lnouc T. Morita K. The prognostic significance of number of positive nodes in cervical carcinoma stages IB. IIA. and IIB. Cancer I990zoS: I923- 7. (29) (30) (3/) (32) (33) (34) (35) (16) (37) (38) (39) (40) Berman ML. Keys II. Creasman W. DiSaia P. Bondy B. Blessing J. Sur» vival and patterns of recurrence in cervical cancer metastatic to periaortic lymph nodes (a Gynecologic Oncology Group study). Gynecol Oncol |9X4:|9:X—lo. Stehman FB. Bundy BN. DiSaia PJ. Keys HM. Larson JE. Fowler WC, Carcinoma of the cervix treated with radiation therapy. |. A multi—variate analysis of prognostic variables in the Gynecologic Oncology Group. Can» cer l99| :6722776-85. Thorns WW Jr. Eifel PJ. Smith TL. Morris M. Delclos L. Wharton JT. et al. Bulky endocervical carcinoma: 1123—year experience |.\‘¢'(' comment cita- tion in Medlinel. IntJ Radiat Oncol Biol Phys 1992;23:49l—9. Perez CA. Grigsby PW. Camel HM. Galakatos AE. Mutch D. Lockett MA. Irradiation alone or combined with surgery in stage IB. IIA. and 118 car- cinoma of the uterine cervix: update of a nonrandomi/ed comparison [we comment citation in Medline]. lntJ Radiat Oncol Biol Phys |995z312703— l6. Averette HE. Donato DM. Lovecchio JL. Seven BU. Surgical staging of gynecologic malignancies. Cancer I987;60:20|0-20. Ballon SC. Berman ML. Lagasse LD. Petrilli ES. Castaldo TW. Survival after experimental pelvic and paraaortic Iymphadenectomy and radiation therapy in cervical carcinoma. Obstet Gynecol |981z57190«5. Gallup DG. Jordan GH, Talledo OE. Extraperitoneal lymph node dissec- tions with use ofa midline incision in patients with female genital cancer, AmJ Obstet Gynecol I986: l55z559-64. Weiser EB. Bundy BN. Hoskins WJ. Heller PB. Whittington RR. DiSaia PJ. et al. Extraperitoneal versus transperitoneal selective paraaortic lym- phadenectomy in the pretreatment surgical staging of advanced cervical carcinoma (a Gynecologic Oncology Group study). Gynecol Oncol 1989; 33:283-9. Lagasse LD. Smith ML. Moore JG. Morton DG. Jacobs M. Johnson GH. et al. The effect of radiation therapy on pelvic lymph node involvement in stage I carcinoma ofthe cervix. Am J Obstet Gynecol l974:l 19328—34. Inoue T. Chihara T. Morita K. The prognostic significance of the size of the largest nodes in metastatic carcinoma from the uterine cervix. Gynecol ()ncol l984;l9: 187-93. Kjorstad KE. Kolbenstvedt A. Stricken T. The value of complete lyin- phadenectomy in radical treatment of cancer of the cervix. stage IB, Can— cer 1984;54:2215-9. Tulzer H. Kupka S. The effectiveness of obligatory Iymphadenectomy in treating carcinoma of the cervix. Int J Gynaecol Obstet 1978—79116:l97— 203. Vigliotti AP. Wen BC. Hussey DH. Doornbos JF, Staples JJ. Jani SK. et al. Extended field irradiation for carcinoma of the uterine cervix with positive periaortic nodes. lntJ Radiat Oncol Biol Phys l992:23:50I—9. Keys H. Park RC. Treatment and survival of patients with cancer of the cervix and nodal metastases. Int J Radiat ()ncol Biol Phys [976:] I:l()9l-7. Potish RA. Downey GO. Adcock LL. Prem KA. Twiggs LB. The role of surgical debulking in cancer of the uterine cervix. Int J Radiat Oncol Biol Phys I989:l7:979»84. Hacker NF. Wain GV. Nicklin JL. Resection of bulky positive lymph nodes in patients with cervical carcinoma. Int J Gynecol Cancer I99515: 25041. Journal ofthe National Cancer Institute Monographs No. 2|. 1996 Management of Stage IA Cervical Carcinoma Mitchell M()rris* Approximately 10%-15% of women with stage I cervical cancer have microinvasive lesions (stage IA). In studying these patients, we aim to identify the cancers that have little or no chance of harboring metastatic disease from the pri- mary site. These patients may be treated by more conserva- tive means, thereby lowering morbidity and cost and preserving the women’s fertility. The diagnosis of stage I cervical cancer is surrounded by controversies concerning the diagnostic criteria, the low level of agreement among pathologists interpreting the same material, and the limited evidence in the literature for definitions of risk factors. The available information suggests that women with squamous cell lesions that invade at a 3-mm depth or less and who have no evidence of lymph—vascular invasion may be successfully treated with cervical conization. Women with a depth of in- vasion of more than 3 mm or with lymph—vascular invasion should be treated with methods that address the potential for disease in the lymph nodes and paravaginal tissues. [Monogr Natl Cancer Inst 1996;21:47-52] Most women with invasive carcinoma of the uterine cervix can be distinguished from those with preinvasive lesions by visual inspection alone. followed by a confirmatory biopsy. Among these patients. however, are women who have occult in- vasive lesions that cannot be diagnosed visually. A subset of these includes women who have (MIC) of the cervix. In 1947. Mestwerdt (I) first described microcarcinoma of the cervix. Since then. the histologic and staging criteria for MIC have remained a matter of controversy. which has resulted in an inconsistent treatment philosophy for patients with MIC. Invasive carcinoma of the cervix primarily spreads by direct extension and lymphatic dissemination. This places the pelvic lymph nodes, upper vaginal and paravaginal tissues. and the suspensory ligaments of the uterus at risk for harboring meta— static disease. Most patients with invasive carcinoma require microinvasive carcinoma therapy that addresses not only the tumor site but also the ad- jacent tissues and lymph nodes. This is accomplished by either radical hysterectomy and pelvic Iymphadenectomy or radiation therapy. These treatments have a high cost in terms of potential for morbidity and mortality, health care dollars. and work time lost. Most important. perhaps. is the loss of potential fertility for younger women. Arriving at a consensus regarding MIC is a high priority, therefore. because it may permit us to use more conservative therapy for women who have this diagnosis. If we could identify a subset of women who have little or no risk of lymph node metastasis or spread beyond the cervix to ad- jacent tissues. a more conservative approach. such as simple Journal ofthe National Cancer Institute Monographs No. 2 1. I996 hysterectomy or cone biopsy. could be employed. How best to define this subject is the topic of this review. It is difficult to estimate the incidence of MIC among the total number of women who have invasive cervical cancer. At the M. D. Anderson Cancer Center. among women who presented with clinical International Federation of Gynecology and Obstetrics (FIGO) stage I cervical cancer between I959 and 1985. 13% were considered to have a superficially invasive tumor (2). The literature on the incidence of MIC is confusing because of differing diagnostic methods used. Some series in- cluded patients with a superficially invasive lesion that was grossly visible and diagnosed by a punch biopsy of the cervix. This group had. by definition. stage IB cervical cancer. MIC is often diagnosed incidentally after cone biopsy for what was believed to be preinvasivc disease. In a review of 756 cone biopsies. Matseoane et al. (3) found a 3.6% incidence of microinvasion in that population. The reported incidence of MIC may be increasing as a result of the more widespread use of cervical cytology screening programs. Problems we encounter when reviewing MIC are controver- sies concerning the diagnostic criteria. the low level of agree— ment among pathologists interpreting the same material. and the limited evidence in the literature for the definition of risk fac- tors. Staging The staging of cervical cancer has changed little in the past 30 years. except for the definition of MIC, which changes every few years (further confusing the interpretation of the literature). In I961. the FIGO suggested the division of stage I cervical can- cer into two subgroups: stage 18 for clinically evident cancer. and stage IA for denoting early stromal invasion, In 1963. Frick et al. (4) described the criteria for early stromal invasion in more detail and proposed a revision of the staging system into two subgroups of IA. Stage IAI would classify invasion up to 3 mm from the basement membrane. and stage IA2 would include in— vasion of 3—5 mm from the basement membrane. The FIGO cancer committee further defined stage IA in I975 when they decided that only patients with histologic evidence of early stromal invasion should be included. This definition re— quired diagnosis to be made on microscopic evaluation of tissue from a biopsy. cone biopsy. trachelectomy. or hysterectomy specimen. In 1985. the FIGO Congress again modified the staging sys- tem. trying to take into account not only the depth of invasion but also the lateral spread of cancer. which could not exceed 7 *('orn'xpomlmu"U m: Mitchell Morris. M,I).. The University of Texas M. I). Anderson Cancer Center. l5|5 Ilolcombe Bl\d.. Box 43. Houston. TX 77030. 47 mm. Stage lAl consisted of lesions that displayed early stromal invasion. Stage 1A2 was used to describe lesions that had a measurable depth of invasion of up to 5 mm. The new FIGO definition provoked much controversy in the United States. where physicians believed that patients with more than 3 mm of invasion or lymph—vascular space involvement should not be treated conservatively. The addition of the 7—mm lateral spread parameter was also criticized. since different pathology labora- tories may produce differing diagnoses. In 1995, partly because of the controversy that surrounded this version of the staging system. the FIGO once again modified the definition of stage IA cervical cancer as follows (5): stage IA (invasive cancer identified only microscopically; all grossly observed lesions. even those with superficial in— vasion. are stage IB cancers; invasion is limited to measured stromal invasion with a maximum depth of 5 mm and no wider than 7 mm): stage IAl (measured invasion of stroma no greater than 3 mm in depth and no wider than 7 mm); and stage 1A2 (measured invasion of stroma greater than 3 mm and no greater than 5 mm in depth and no wider than 7 mm). This new FIGO classification now approximates more closely a functional definition of microinvasion proposed by the Society of Gynecologic Oncologists (SGO) in 1974 (6). They defined MIC as a lesion in which the neoplastic epithelium invades the stroma in one or more places to a depth of 3 mm or less below the base of the epithelium and in which lymphatic or blood ves- sel involvement is not demonstrated. This is the definition that most commonly guides therapy today. Although the new FIGO definition recognizes that more than 3 mm of invasion seems to confer greater metastatic potential. it has been criticized for not addressing the issue of lymphatic invasion. It should be noted that the SGO functional definition of microinvasion is not a staging system. Gynecologic oncologists employ the FIGO sys- tem for staging, but many will use 3 mm as their cutoff point for conservative therapy. Pathology MIC represents but one stage in a spectrum of disease. In- traepithelial preneoplastic lesions may progress to the point at which carcinoma in situ advances to the next step: abnormal cells are found below the basement membrane of the epithelium, representing MIC. This is the defining point of invasive cancer. which has the potential for malignant behavior and metastatic spread. The study of MIC is primarily concerned with identify— ing the pathologic variables that correlate with little or no risk of metastatic disease. The primary defining characteristic of MIC is the depth of in— vasion. It is this parameter that has the greatest correlation with the potential for metastatic spread. For that reason. MIC is re- lated exclusively to squamous cell lesions. Microinvasive adenocarcinoma is difficult to define because no one knows where measurements should be taken to define the greatest depth of invasion. Indeed, Brown et al. (7) found that recurrent invasive cancer following conization with apparently negative margins was most commonly seen in adenocarcinoma in situ. thus illustrating the difficulty of distinguishing the in situ lesion from early invasion with glandular lesions and the difficulty of 48 interpreting cone margins of these lesions. The remainder of this review will be concerned with squamous cell lesions only. Diagnostic Reproducibility Reproducibility of MIC diagnosis has been a major issue among pathologists. For example. a clinicopathologic study car— ried out by the Gynecologic Oncology Group was compromised when roughly half of the 265 patients with MIC entered in the study were ruled ineligible on pathologic review (8). The in- eligible patients were believed either not to have invasion in the first place or to have invasion in excess of 3 mm. The choice of the exact point at which to take measurements of depth and lateral spread is part science and part art. The somewhat subjec— tive nature of pathologic interpretation makes it difficult to cor— relate pathologic features with clinical outcomes. Microscopic Features The diagnosis of MIC is based on the appearance of one or more microscopic tongues of tissue penetrating the basement membrane or by clusters of tumor cells seen below the basement membrane. These lesions are nearly always associated with sur- rounding intraepithelial lesions. It is not uncommon to find that the invading cells are differentiated better than the surrounding intraepithelial lesion. Keratin pearls are occasionally seen. The area of MIC can usually be distinguished from carcinoma in situ by its irregular border when compared with the smooth edge of the basement membrane. Depth of Invasion Depth of invasion of squamous cell carcinoma should be measured from the basement membrane to the point of deepest tumor infiltration. When the invasion comes from an endocervi— cal gland. the measurement should still be made from the base— ment membrane at the point of invasion (Fig. 1). Use of a calibrated micrometer is recommended to aid in the measure— ment. One can readily see that the site of sectioning the biopsy for the slide and the pathologist‘s choice of the measuring point may lead to significant variation. It is important, therefore, to follow a careful protocol when preparing the cervical cone specimen for examination. The general agreement is that the greater the depth of inva— sion, the greater the potential for metastatic spread. By examin— ing the data from several studies of pelvic lymphadenectomy performed in the treatment of MIC. we can evaluate the risk of nodal metastases on the basis of depth of invasion at the primary site. As seen in Table 1, increasing depth correlates with in— creasing risk of lymph node metastasis. Risk increases sig— nificantly when the depth of invasion goes beyond 3 mm. Although sporadic anecdotal reports have appeared in the litera- ture of lymph node metastases in patients with less than 3 mm of invasion, these are quite rare. Lateral Spread Burghardt et al. (9) advanced the concept that, compared with a more detailed analysis of tumor volume. a single measurement of depth ()f invasion does not provide enough prognostic infor— mation. With an extensive step—sectioning technique, Burghardt Journal of the National Cancer Institute Monographs No, 21. I996 Fig. l. A) Early invasion of the cervical stroma can be seen beneath an epithelial layer. consistent with carcinoma in situ. B) More extensive invasion is present with no identifiable epithelial layer. et al. were able to calculate not only the depth of invasion but indeed the entire volume of the tumor and to correlate this with the risk of recurrent disease. Few investigators have been either able or willing to adopt this technique, and so it remains incom— pletely evaluated. Although the FIGO staging system adopts the somewhat arbitrary limit of 7 mm of lateral spread as the upper limit for stage IA. most pathologists find it difficult to accurate- ly estimate lateral spread because MIC lesions do not grow as a sphere but often have tongues of abnormal cellularity that ex— tend for variable distances in different directions. Evaluating the role of multifocal lesions has been similarly difficult. Maiman et al. (/0) also criticized the technique of estimating total tumor volume. since many processing techniques are available: the ex— pected shrinkage from fixation with formalin is 1()%-2()%. Maiman et al. suggested, instead, that counting the number of quadrants involved in microinvasive disease is a more practical and reproducible approach. This should be further evaluated. Lymph—Vascular Invasion When tumor forms within endothelium-lined spaces, lymph— vascular invasion (LVI) is considered to be present. The re- Journal of the National Cancer Institute Monographs No. 21, 1996 Table l. Depth of invasion and risk for lymph node metastasis Incidence of nodal metastasis, ‘70 lnvestigatorts) No. of (Ref. No.) patients 3.()-5.0 mm Averette et al. (/9) I98 0 — A Maiman et al. (/0) 95 0 1.5 13.3 Sevin et al. (2/) HO 0 0 56 van Nagell et al. (20) 84 — 0 9.3 Simon et al. (/2) 69 — 0 3.9 Hasumi et al. (25) I35 — 0.9 11.9 H2 ~— 0 0 Sedlis etal, (8) ported incidence of LVI varies greatly in the literature, the variance probably depending somewhat on how hard the pathologist looks for it. From the Norwegian Radium Hospital, Kolstad (11) reported an I 1% overall incidence of LVI in women with FIGO stage IA cervical carcinoma. Burghardt et al. (9), in contrast, reported a 39% incidence of LVI in Austrian women with stage IAI cervical carcinoma. In the Austrian series, an increasing incidence of LVI was not observed as depth of invasion increased, but most other investigators have figures more in line with Kolstad’s in noting that the incidence of LVI increased with the depth of invasion ([0,1]). Information is limited on the risk of lymph node metastasis in relation to LVI. In a study by Maiman et al. (10), of 95 patients treated for stage IA cervical carcinoma with radical hysterectomy and pelvic lymphadenectomy. those who had no LVI had a 3% incidence of lymph node metastasis. For patients who had LVI, the risk in- creased to 25%. The total number of patients. however, was small. Pattern of Invasion Several investigators have examined patterns of invasion and attempted to further define the risk for recurrent disease in MIC by characterizing the patterns with descriptive terms such as “confluent," "stranding,” “bulging," and “solid.” For most of these patterns. there is little agreement among pathologists and much interobservational difference. If the description of the pat- tern of invasion is limited to pushing versus infiltrating, then some consistent evaluations can be made. Yet these observa- tions are unlikely to ever provide clinically useful information (2.12). Processing of the Conization Specimen That MIC can be diagnosed only on a cone or hysterectomy specimen is well established. A punch biopsy is inadequate, since additional and possibly more deeply invasive disease may be present in the area immediately adjacent to the biopsy site. Similarly, patients whose cone biopsy showed a positive margin have not been completely assessed, and the possibility that in- vasive disease is present in the residual uterus is significant. Selected patients may be assessed with a second conization. but this may be technically difficult and the margins troublesome to evaluate. Most patients with positive margins should be con— sidered as having a stage IB lesion, and they should be treated with the appropriate radical therapy. A confounding feature in 49 the literature is that many of the larger series permitted patients whose cone biopsies showed positive margins to be included in the series of patients with microinvasion (ll, 13). In general. a conization specimen should be opened at 12 o'clock and serially sectioned into a minimum of 12 slices. The endocervical and exocervical margins should be inked to aid in the interpretation. Whether additional routine sectioning is beneficial remains to be determined. Diagnosis As one would expect. women with MIC do not have sig— nificant signs or symptoms related to the lesion. and most cases are detected on routine cervical screening. Lohe et al. (/4) found that nearly half of their patients had either vaginal discharge. bleeding. or pain. Similarly. 32% of the women in Kolstad‘s study (1/) complained of irregular bleeding with or without dis- charge. It is likely. however. that these symptoms. although un- related to MIC. brought the patient to the physician‘s office for examination. Since these patients are at high risk of contracting sexually transmitted diseases. appropriate screening should be done when MIC has been diagnosed. The Papanicolaou (Pap) smear is the diagnostic test that leads most patients to undergo further evaluation. Cervical cytology has low sensitivity for detecting MIC. with many smears mis— read instead as severe intraepithelial disease. In a study of 127 Australian women with preclinical cervical cancer. Rome et al. (/5) found that early stromal invasion was accurately diagnosed on Pap smear in only 35%. When patients had a stage 1A2 lesion, the diagnosis was made accurately on Pap smear in 38% of the cases. Most "undercalls" were read as severe dysplasia or carcinoma in situ. however. leading in turn to further evaluation and finally the correct diagnosis. How often MIC is associated with a completely normal Pap smear is unknown. but the num- ber is probably very low. The major cytologic features of MIC are coarse. irregular chromatin and a background of cellular debris. Whenever a patient’s Pap smear shows the possibility of invasive cancer but no clinically evident disease. colposcopy is indicated. Likewise. in cases that are “undercalled” on cytologic examination (but there is cytologic evidence of an intraepithelial lesion). colpos— copy is performed routinely. Colposcopic findings of MIC in- clude bizarre vessels with an abnormal branching pattern. Unfortunately. colposcopy often fails to detect the presence of actual invasion and in fact misses many cases. If a colposcopi- cally directed biopsy is interpreted as MIC. the patient still re— quires a cone biopsy to evaluate the remainder of the cervix for widespread or more deeply invasive lesions. Most patients who are eventually found to have invasive car— cinoma will have an unsatisfactory colposcopy for one reason or another. and they will undergo conixation (/5-I8). Rigid ad- herence to a colposcopy protocol is recommended. therefore. for patients whose examinations have been unsatisfactory or whose cytologic findings have been worse than the biopsy results. and such patients should be managed with conixation. Cone biopsy. using the cold-knife technique. is currently the standard method for making the definitive diagnosis of MIC. Other techniques for cone biopsy. such as laser excision or loop 50 electrosurgical excision procedure (LEEP). must be carefully evaluated for the potential of thermal artifact at the margins and difficulty in orienting the specimen to calculate lateral spread. This is especially true when operators performing the LEEP remove the specimen in more than one piece. If MIC is sus- pected before conilation, use of the cold—knife technique should be strongly considered unless the operator has great confidence in his or her LEEP skills. The operator must also assist the pathologist in defining the anatomic orientation of the resultant specimen. Management The literature regarding MIC is difficult to interpret. since many of the reported studies employed a variety of definitions of microinvasive disease. and nearly all were retrospective. The information can be evaluated in two ways. The first is to look at patients who underwent radical therapy with either a radical or modified radical hysterectomy with pelvic lymphadenectomy. and to look at the incidence of pelvic nodal metastasis in addi- tion to the risk of recurrent disease. Second. we can assess the risks of conservative therapy by examining a series of patients who were treated with either simple extrafascial hysterectomy. vaginal hysterectomy. or cervical conization for microinvasive lesions. Results of Radical Therapy In a multi—institutional study. Averette et a1. (19) examined 162 patients with MIC of the cervix. All patients had invasion of less than 1 mm and no lymph—vascular space involvement. All patients were treated with radical hysterectomy and pelvic lymphadenectomy. In none was a nodal metastasis noted, and none had recurrent disease. van Nagell et al. (20) treated 84 patients with stage IA disease with radical hysterectomy and pelvic lymphadenectomy. Among 52 patients who had lesions 3 mm or less in depth. no patient had lymph node metastasis. and none had recurrent disease. In contrast. three of 32 patients with stromal invasion of 3.1-5.0 mm experienced lymph node metas— tasis; all three had invasive recurrences and two died. In a series of 1 17 patients treated at the Downstate Medical Center, New York (10). women with invasion of 1 mm or more were treated with either radical or modified radical hysterectomy with pelvic node dissection. Although the Downstate study noted a pelvic node metastasis with 2 mm of invasion in one patient. no in— vasive recurrences occurred in any patients in that series. Among 125 women with stage IA squamous cell carcinoma treated by Simon et a1. (/2). 69 (55%) underwent both radical hysterectomy and pelvic lymphadenectomy. Only one of the 69 patients had a metastatic tumor. and there were no recurrences. These results lead us to conclude that radical hysterectomy with pelvic node dissection is an efficacious therapy for cervical MIC with invasion of up to 5 mm in depth. On the basis of low incidence of pelvic node metastasis in women whose lesions in- vaded 3 mm or less. we may also question the need for such a radical intervention. Unquestionably. radical surgery adds sig— nificantly to the morbidity and cost of treating these lesions and results in permanent loss of fertility. Journal ofthe National Cancer Institute Monographs No. 21. I996 But radical surgery is clearly an effective treatment. Of 110 patients treated with radical hysterectomy by Sevin et al. (2]). 54 had a lesion that fulfilled the SGO definition of microin- vasive cancer. and none recurred. Forty—two of the patients had lesions that fulfilled the FIGO definition of MIC. and. likewise. none of these recurred. Twenty—seven patients could be categorized in either group. and they also remained free of dis— ease. The disease recurred in four patients in the Sevin et al. series; all had depth invasion of more than 3 mm and a lateral spread of between 13 and 22 mm. In the Gynecologic Oncology Group study (8). 74 patients treated with radical hysterectomy had lesions that fulfilled the SGO definition of rnicroinvasive disease. No patients in this series had positive nodes. Two patients died of recurrent disease. but both had extensive LVI. Results of Conservative Therapy In the past few decades. some investigators have felt confi— dent performing a simple extrafascial hysterectomy for MIC. In a series reported by Popkin et al. (22). of l25 women without LVl and with lesions up to 5 mm in diameter. 106 were treated with abdominal hysterectomy. eight with trachelectorny. eight with cone biopsy. and three with radical hysterectomy. None of the patients had recurrent disease. In a study by van Nagell et al. (20). 93 patients were treated with either vaginal or abdominal hysterectomy. Only one. a patient with 2.6 mm of invasion. had recurrent disease. Greer et al. (/3) described 50 patients with stage IA MIC who first underwent conization and then hysterectomy. Surprisingly. 33 of 50 patients in this series had positive cone margins. Of the l7 patients in whom the negative cone margins were negative. four were found to have residual disease in the radical hysterec— tomy specimen. The interpretation was made difficult by the fact that none of the operations were done at the authors~ institu— tion. and slides were not reviewed: only pathology reports were available. In 138 patients treated with hysterectomy. Kolstad (1/) found no patients with recurrent disease when he used the 860 definition of microinvasion. Since extrafascial hysterectomy has been accepted as a safe and effective definitive treatment for microinvasion as defined by the 500. can conization alone be considered adequate treat- ment? By accepting hysterectomy as complete therapy. we evi— dently believe that the parametria. pelvic lymph nodes. and upper vagina have little or no chance of harboring metastatic disease. The same argument is made for the residual cervix and uterine corpus when conization is used as definitive therapy. To accept the argument. we must be reasonably sure that conixation with a negative margin ensures that no disease is present in the residual uterine tissue. Unfortunately. again. the literature offers little information on which to base a decision. Several reports describe conization as definitive therapy for early cervical cancer. The 500 definition is not often used. and conclusions are difficult to draw. Simon et al. (/2) included eight patients treated with conization alone in their total series of |25 patients. None of the eight developed recurrent disease. but the depth of invasion in this group is not described. Kolstad (1/) studied 48 patients who underwent conization as definitive therapy for stage IA cervical cancer. Thirtyvthree of these had only early stromal invasion. and IS had FIGO stage 1A2 disease. Journal of the National Cancer Institute Monographs No. 2 l. 1996 Four of the 48 patients developed recurrent disease. but it is not clear whether all of these were invasive recurrences or what the group‘s initial pathologic features were. Burghardt et al. (9) described 18 patients treated for stage 1A2 lesions with coniza— tion alone. Three patients had recurrent disease. and one patient died of it. However. none of these patients fulfilled the SGO criteria for MIC. In a study at The University of Texas M. D. Anderson Cancer Center. 14 women who had stage IA cervical cancer and fulfilled the $00 criteria were treated with coniza- tion as the sole therapy (23). The follow-up period in that study was brief. but to date no patient has developed recurrent in— vasive disease. Andersen et al. (24) described 31 Danish patients with MIC who fulfilled the $60 definition and were successfully treated with laser conization as definitive therapy. During a median fol- low-up period of 36 months. no patient developed recurrent dis- ease. The mean depth of invasion. however. was only 0.4 mm. Treatment Recommendations All patients with suspected MIC should undergo conization for definitive diagnosis. Women with clinically evident invasive lesions have. by definition. stage [8 disease and are not can- didates for conservative therapy. The cone specimen should be systematically sectioned and thoroughly evaluated. In areas where microinvasion is found. additional sections should be taken to determine the extent of lateral spread and to search carefully for LVI. If the margins of the conization are positive for invasive disease. the patient should be considered to have stage IB cervical MIC and be treated with the appropriate radi— cal therapy. If the margins are positive for preinvasive disease. a second coni/.ation should be considered to rule out the presence of multifocal invasion. unless other definitive therapy is planned. If the margins are negative. the patient may be a can— didate for conservative therapy. Patients With 3-mm or Less Depth of Invasion but Without LVI Patients with depth of invasion of 3 mm or less may be treated conservatively with abdominal hysterectomy. vaginal hysterectomy. or conization. Consideration of conization re- quires careful patient selection. and patients who wish to main- tain fertility should be fully counseled about this treatment approach. The surgeon should also provide close surveillance for persistent disease and expect such patients to comply with a follow—up regimen. especially because cone margins may be misinterpreted in some cases. Hysterectomy may be the best treatment for lesions with extensive lateral spread or multiple in— vasive foci. Unfortunately. in the “real world." cone margins may not be evaluable with the degree of certainty we desire. In these cases. it may be best to err on the side of more complete therapy while taking the wishes of the patient into careful com sideration. For patients who are managed with conization alone. Pap smears should be performed at frequent intervals. The impor- tance of endocervical curettage and colposcopy in the follow—up regimen is not known; their use in a surveillance strategy should minimize the advancement of persistent disease before its detection. 5] Patients With LVI or 3.1- to S-mm Depth of Invasion These patients have a small but significant risk for lymph node metastasis. The tissues at risk include the parametria as well as the node—bearing areas. For that reason. it is recom- mended that such patients be treated with either radical or modified radical hysterectomy with pelvic lymph node dissec- tion. For the patients who are not surgical candidates, pelvic radiation therapy is also an effective treatment. Until further prospective clinicopathologic studies are performed. it is dif- ficult to justify pushing the envelope of conservative therapy further than recommended above. Because of the changing definition of stage IA MIC and the uncommon nature of this lesion. it will remain a subject of con- troversy until prospective clinicopathologic studies can be con— ducted. Unresolved issues in this disease include the risk of extensive lateral spread and the presence of multifocal invasive lesions. The method of evaluating LVI and the degree to which LVI affects the prognosis also require additional study. Some is- sues for future research include development of reproducible methods of assessing tumor volume. assessment of the patient‘s risk of developing a second primary cancer of the cervix, and evaluation of the reproductive potential of women who have had conservative therapy for stage IA cervical cancer. References (I) Mestwerdt G. Die Fruhdiagnose des Kollumkarzinoms. Zentarlbl Gynakol 1947;69:198-202. Copeland L. Silva E. Gershenson D. Morris M. Young D. Wharton JT. Su- perficially invasive squamous cell carcinoma of the cervix. Gynecol Oncol 1992;45:307—12. Matseoane S. Williams SB. Navan‘o C. Hedriana H. Mushayandebvu T. Diagnostic value of conization of the uterine cervix in the management of cervical neoplasia: a review of 756 consecutive patients. Gynecol Oncol 1992;47:287—91. (4) Frick H. Janovski N. Gusberg S. Taylor H. Early invasive cancer of the cervix. AmJ Obstet Gynecol l963;85:927-39. Shepherd JH. Staging announcement. FIGO staging of gynecologic can- cers; cervical and vulvar. lntJ Gynecol Cancer 1995:5319. (2 (3 Ln (6) Seski J. Abel] MR. Morley GW. Microinvasive squamous carcinoma of the cervix: definition. histologic analysis. late results of treatment. Obstet Gynecol 1977;50:4l0—4. 52 (7) (8) (9) (I0) (1/) (/2) (I3) ([4) (l5) (/6) (/7) (IX) (/9) (20) (2/) (22 (23) (24) (25) Brown JV. Peters WA. Corwin DJ. Invasive carcinoma after cone biopsy for cervical intraepithelial neoplasia. Gynecol ()ncol l99l :40225-8. Sedlis A. Sall S. Tsukada Y. Park R. Mangan C. Shingleton H. et al. Microinvasive carcinoma of the uterine cervix: a clinical»pathologic study. Am J Obstet Gynecol |979zl33:64-74. Burghardt E. Girardi F. Lahousen M. Pickel H. Tamussino K. Microin- vasive carcinoma of the uterine cervix (International Federation of Gynecology and Obstetrics Stage IA). Cancer 1991;67:1037—45. Maiman M. Fruchter RG. DiMaio TM. Boyce JG. Superficially invasive squamous cell carcinoma of the cervix. Obstet Gynecol 1988;72:399—403. Kolstad P. Follow-up study of 232 patients with stage Ial and 41 1 patients with stage [212 squamous cell carcinoma of the cervix (microinvasive car- cinoma). Gynecol Oncol 1989;33:265—72. Simon NL. Gore H. Shingleton HM. Soong SJ. Orr JW Jr. Hatch KD. Study of superficially invasive carcinoma of the cervix. Obstet Gynecol 1986;68: l9-24. Greer BE. Figge DC. Tamimi HK. Cain JM. Lee RB. Stage 1A2 squamous carcinoma of the cervix: difficult diagnosis and therapeutic dilemma. Am J Obstet Gynecol l990; I62: [406-9. Lohe KJ. Burghardt E. Hillcmanns HG. Kaufmann C. Ober KG. Zander J. Early squamous cell carcinoma of the uterine cervix. II. Clinical results of a cooperative study in the management of 419 patients with early stromal invasion and microcarcinoma. Gynecol Oncol 1978:6231-50. Rome RM. Chanen W. Ostor AG. Preclinical cancer of the cervix: diag- nostic pitfalls. Gynecol Oncol 198512230242. Benedet J. Anderson GH. Boyes DA. Colposcopic accuracy in the diag— nosis of microinvasive and occult invasive carcinoma of the cervix. Obstet Gynecol 1985;65:557-62. Paraskevaidis E. Kitchener HC. Miller ID. Mann E. Jandial L. Fisher PM. A population-based study of microinvasive disease of the cervix—a colpo- scopic and cytologic analysis [we comment citation in Medline]. Gynecol Oncol 1992;45:9-12. Madej J. Madej JG Jr. Colposcopy as a method of management strategy in CIN and microinvasive cancer. EurJ Gynaecol Oncol l99();2:l 17-22. Averette HE. Nelson JH Jr. Ng AB, Hoskins WJ. Boyce JG. Ford JH Jr. Diagnosis and management of microinvasive (stage IA) carcinoma of the uterine cervix. Cancer l97oz38z4l4-25. van Nagell JR Jr. Greenwell N. Powell DF. Donaldson ES. Hanson MB. Gay EC. Microinvasive carcinoma of the cervix. Am J Obstet Gynecol 1983;145:981-91. Sevin B. Nadji M. Averette H. Hilsenbeck S. Smith D. Lampe B. Microin— vasive carcinoma of the cervix. Cancer 1993;70:2121—8. Popkin DR. Pilorge R. Latour JP. The treatment of microinvasive squam- ous cell carcinoma of the uterine cervix. Gynecol Oncol 1979;25:84-6. Moms M. Mitchell MF. Silva EG. Copeland LJ. Gershenson DM. Cervical conization as definitive therapy for early invasive squamous carcinoma of the cervix. Gynecol Oncol 1993;51:193-6. Andersen E. Husth M. Joerjenson A. Nielsen K. Laser conilation from microinvasive carcinoma of the cervix: short-term results. Int J Gynecol Cancer 1993;32183-5. Hasumi K. Sakamoto A. Sugano H. Microinvasive carcinoma of the uterine cervix. Cancer 1980;45:928—31. Journal of the National Cancer Institute Monographs No. 2|. I996 Primary Surgery for Stage IB-IIA Cervical Cancer, Including Short—term and Long—term Morbidity and Treatment in Pregnancy Javier F. Magrirza:k Radical hysterectomy and pelvic lymphadenectomy are indi- cated for the treatment of cervical carcinoma that is local- ized clinically to the cervix and upper vagina. Intraoperative complications have been reported in 1.l%-7.4% of patients. Long-term complications include bladder dysfunction (2% at 3 years). urinary fistula (vesical. 0.8%; ureteral, 1.2%), stress urinary incontinence (29%), ureteral stricture (1%), rectal dysfunction (80%), severe constipation (5.3%), lym- phocysts (20% by ultrasonography; 2% clinically), and lymph- edema (10%). The operative mortality is 0.7%. The 5-year survival rate for patients with stage IB disease is 85.7% and for stage [IA is 69.6%. The recurrence rate is 27.2%. Recur- rences are distributed equally between the pelvis and ex- trapelvic sites. Radical hysterectomy is the treatment of choice for pregnant patients with early cervical cancer. It af- fords termination or delivery of the pregnancy at the same time that the treatment is provided. For patients with stage 1 disease treated with radical hysterectomy, the survival rate is 92.1%. [Monogr Natl Cancer Inst 1996;21:53-9] Surgery has been used more frequently in the treatment of early cervical carcinoma in the last few years. An American College of Surgeons Patterns of Care Study involving 5817 patients with cervical cancer treated in 1990 revealed that some type of hysterectomy was part of the treatment in 48.2% of the patients (1). In the state of Connecticut. an overall decline of 19% in the use of radiotherapy was noted among 1286 patients receiving treatment for cervical carcinoma from 1983 to 1990. When only local-stage tumors were considered. the decline for the same period was 23% (2). The availability of more gyne— cologic oncologists and an earlier stage of diagnosis may ac- count at least partly for this trend. Radical Hysterectomy and Pelvic Lymphadenectomy Technique It is not correct to equate "Wertheim hysterectomy" with the current “radical hysterectomy." The original technique de— scribed by Wertheim (3) corresponds to the present modified radical hysterectomy technique (4). The original technique did not include pelvic lymphadenectomy. As described previously (5). the technique of radical hysterec- tomy and pelvic lymphadenectomy consists of 1) central resec— tion. which includes the removal of the cervix and immediate Journal of the National Cancer Institute Monographs No. 21, 1996 surrounding tissues. and 2) resection of the regional lymph nodes. The importance of the surgery relies on the adequate removal of the cervix and paracervical and paravaginal tissues: lymphadenectomy has therapeutic value for only about half of the patients with positive nodes (6. 7). Indications Radical hysterectomy and pelvic lymphadenectomy are indi- cated for the treatment of cervical carcinoma that is clinically localized to the cervix and upper vagina. In combination with radical vaginectomy. the surgery is applicable to cervical car— cinomas that extend to the lower vagina, that exclude the lower third. In this case. most gynecologic oncologists would prefer radiation therapy to avoid the additional operation needed for vaginal reconstruction. Surgical treatment has several advantages over radiation therapy for the treatment of stage IB—IIA cervical carcinoma. Our approach is to consider radiation therapy for only those patients for whom surgery is contraindicated. Radical surgery for stage IB—IIA cervical carcinoma is as likely to produce a cure as is surgery in combination with irradiation or chemotherapy (or both), at least until more effective agents are introduced (8). 1) Surgery delineates the pathologic extent of the tumor. This has significance with regard to prognosis and therapy. A Gynecologic Oncology Group (GOG) study that compared clinical versus surgical staging demonstrated a 24% surgical difference for patients with stage I cervical carcinoma and a 50% difference for those with stage [1 disease (9). In another GOG study, 12% of 940 patients with stage I cervical carcinoma did not have radical hysterectomy because of gross extrauterine disease or positive aortic nodes (10). 2) The remaining vagina is normal. both histologically and functional- ly. 3) In young patients, ovarian preservation is possible. The risk of ovarian metastases is low: 0.5% or less for patients with squamous lesions (11.12) and l.3%-4.0% for those with adenocarcinomas (Ii/4). 4) Radical hysterectomy selectively removes the malignant tissues, thus preserving the integrity of the remaining healthy tissues. The surgery is performed in 11/2 to 41/: hours ([5). with a blood loss of 1000-1900 mL (16—18). The average hospital stay is 7.4 days (17); in patients without complications. recovery is complete within 6 weeks. 5) Most of *('m‘respmldom'e m: Javier F. Magrina. M.D.. Department of Gynecology. Mayo Clinic Scottsdale. 13400 E. Shea Blvd. Scottsdale. AZ 85259. 53 the complications due to the surgery are apparent within the first 3—4 postoperative weeks. They usually are easy to manage, and good results are obtained (I). 6) Pelvic recurrences after radical hysterectomy can be treated with irradiation. 7) In comparison with irradiation, long—term morbidity with surgical treatment is less frequent. After irradiation. actinic sequelae may appear years after the therapy. Among these sequelae is an increased in— cidence of second cancers, including uterine sarcoma and car— cinomas of the bladder, vagina. and overlying integutnent (19). 8) The surgery has a salvage role for patients with incomplete irradiation of the pelvis, persistent cancer after completion of a therapeutic dose of irradiation. or a small cervical recurrence after irradiation. 9) Radical hysterectomy is of value in pregnant patients with cervical cancer. For patients electing to have im- mediate therapy early in pregnancy, the surgery affords removal of the cancer and, simultaneously, of the pregnancy. For patients wanting a mature infant. a planned cesarean radical hysterec- tomy meets the double objective of providing a viable infant and immediate cancer therapy. 10) Radical hysterectomy affords a potential cure for patients with contraindications for irradiation. such as pelvic inflammatory disease or previous bowel obstruc- tion. Contraindications In some cases. surgical treatment may not be optimal for patients with early cervical carcinoma. The contraindications to surgical therapy are related to the patient‘s medical condition and wishes and to the surgical findings. 1) Patients with severe medical conditions are not candidates for surgical treatment when a nonsurgical therapy is available. 2) ln morbidly obese patients, the adequacy of the resection is compromised and the risk of wound infection is markedly increased. 3) Some patients clearly state a preference for a nonsurgical method of therapy. 4) In the presence of peritoneal or distant metastases. surgical treatment is of little value and less radical forms of palliative therapy should be used. 5) The finding of unresectable pelvic or aortic nodes should constitute grounds for discontinuing the radical hysterectomy. 6) Although gross parametrial extension is not an absolute contraindication, it preferably is treated with ir- radiation. Removal of the pelvic nodes is curative in about 50% —60% of the patients with positive pelvic nodes (6. 7). Because of the small number of patients with early cervical cancer and positive pelvic nodes, this curative value affects only eight of |0() patients with early cervical carcinoma. If one accepts an in- cidence of positive pelvic nodes in 14.5% of the patients with stage IB cervical carcinoma (20), lymphadenectomy was not therapeutically beneficial for the other 85%. Among the 15% of the patients with positive pelvic nodes, the 5—year cure rate has been reported at 50%—60%. even with adjuvant pelvic irradia— tion (6. 7). Eight of 100 patients remain for whom the removal of positive pelvic nodes is therapeutically curative. Although the presence of grossly positive but resectable aor— tic nodes causes many gynecologic oncologists to abandon the planned radical hysterectomy, it should not constitute a con— traindication to radical hysterectomy, because a significant sur- vival has been noted in these patients. [n 25 patients in whom radical hysterectomy was discontinued because of grossly posi— 54 live aortic nodes and treatment was completed with irradiation. the 5—year survival rate was 26% (2/). ln nine of these patients, the only site of recurrence was in the pelvis. Complications l. lntraoperative complications. A. Pelvic lymphadenectomy. lntraoperative complications related to pelvic lymphadenectomy are usually caused by injury to neural or vascular structures. injury to the main vessels is un- common and immediately recognizable: it can be associated with major blood loss. Unless the bleeding is in an inaccessible area, it is easily controlled. Neurologic injuries associated with surgical treatment of gynecologic cancers are uncommon and reported in 1% of patients (22). Some nerve injuries are related to the use of a self retractor, transverse incision, or the position of the patient. Nerve injuries associated with pelvic lymphadenectomy usually involve the lateral pelvic nerves. The genitofemoral nerve is easily injured because of its location and size. Transection of the obturator nerve has serious consequences. producing marked weakness of adduction and stability of the lower extremity when standing or walking. Reanastomosis of the obturator nerve is possible and should be attempted when it is transected. The sciatic nerve can be incorporated in the sutures during an at- tempt to control pelvic bleeding lateral to the iliac vessels. B. Radical hysterectomy. Complications arising during the performance of a radical hysterectomy are more common, oc— curring in 1.1% -7.4%, of patients (23—25). They usually involve the bladder. ureter(s), rectum, or pelvic wall vessels. Ureteral injuries have been reported in ().6%w—l.7% of radical hysterectomies. Bladder or rectal injuries are not a major prob— lem when recognized and appropriately repaired; they have been reported in ().4%—4,3% and in l.4%7-2.l%,. respectively, of patients undergoing radical hysterectomy. Bleeding from the pelvic veins has been reported in 3.5% of the patients undergo— ing radical hysterectomy (23). II. Postoperative complications. A. Short-term complications. The most common short-term complications are urinary tract infection (l6.l% of patients), wound infection (6. | %). and deep—vein thrombosis (5%) (Table l), B. Long-term complications. 1. Radical hysterectomy complications.—Bladder. The most common long—term complications of radical hysterectomy Table l. Short—term complications of radical hysterectomy and pelvic lymphadenectomy in 423 patients:*‘ Complication % of patients Urinary tract infection lo.l Wound infection 6.1 Deepwcin thrombosis 5.0 llcus 2.4 Pulmonary embolism 2| Pelvic abscess l .9 Pulmonary infection 1.7 Cardiac. central nertous system 1.7 iiiModified from Webb and Symmonds (20). By permission of the American College of Obstetricians and Gynecologists. Journal ol’the National Cancer Institute Monographs No. 2l. I996 are vesical. with stress ttrinary incontinence and bladder dys- function being the tnost common ones. Major vesical complica- tions can be separated into bladder dysfunction, vesicovaginal fistula. and stress urinary incontinence. a) Bladder dysfunction.~Bladder dysfunction is common. occurring in 60%—7()% of the patients immediately after the surgery (24,26). It is short-lived. and resolution is the rule for most patients. Bladder dysfunction after radical hysterectomy can be categorized as primarily sensory. motor. or mixed. It results from the changes secondary to surgical trauma and par— tial disruption of the sympathetic and parasympathetic nerve supply to the bladder (27-29). 1) Sensory dysfunction—About 65% of patients experience some degree of loss of bladder sensation and. occasionally. dis— appearance of the sensation of urinary urgency. even with a full bladder. This sensory loss is usually temporary (days to months) and full recovery typically occurs in 6-24 months. The incidence of bladder dysfunction decreases to 16% at 6 months (Method MW. Sevin BU. Rodriguez M. Lee W. Penalver M: abstract. 27th meeting of the Society of Gynecologic Oncologists. New Orleans. LA. February 11—14. 1996) and to 11% at 1 year. At the end of 3 years. only 2% of the patients have markedly decreased bladder sensation (l5). 2) Motor dysfunction.~Motor dysfunction is a more serious. but less common. sequela of radical hysterectomy. oc- curring in 1.7% of patients (30). Increased myogenic detrusor activity and decreased bladder compliance are noted in the first days to weeks after radical hysterectomy. Subsequently. a decrease in detrusor contractility. as evidenced by the maximal flow rate and residual urine. and inability to have normal void- ings are noted (28). A return to normal bladder accommodation and capacity usually occurs 6—8 weeks postoperatively (3/). 3) Mixed dysfunction.~lt is uncommon to experience corn— plete bladder sensation with inability to void. However. dif— ferent combinations of sensory and motor dysfunction are not uncommon in patients with bladder dysfunction. b) Stress urinary incontinence.—The reported incidence of stress urinary incontinence after radical hysterectomy in retrospective series frotn the literature ranges from 0% to 75% (mean. 35%) (28). In three prospective studies that included pre- operative and postoperative urodynamic evaluation of a total of 45 patients. the incidence was 28.8% (range. 22.7%-41.6%) (32- 34). This is higher than the incidence of bladder dysfunction at the end of 1 year. The development of stress urinary incontinence is related primarily to urethral pressure changes and not to a disturbance of the urethrovesical angle. at least in patients undergoing the customary resection of the upper vagina associated with radical hysterectomy. However. the urethrovesical angle appears to be disrupted as the length of t'esected vagina increases. The in- cidence of urethral instability was higher when the upper two thirds of the vagina was excised as part of a radical hysterec- tomy (35). c) Vesicovaginal fistula.—Vesicovaginal fistulas are uncom— tnon after radical hysterectomy. The incidence has decreased remarkably through the years. most likely because of a com- bination of factors rather than a specific one. Among these fac— tors are refinement of surgical technique. improvement of the Journal of the National Cancer Institute Monographs No.21. 1996 surgical skills of gynecologists because of the availability of training programs. the use of sharp dissection for the separation of the bladder from the cervix and vagina. and the prophylactic use of antibiotics. Among 3084 patients reported in the litera— ture. the incidence of vesicovaginal fistula after radical hysterec- totny was 0.8% (range. 0%-3.3% (Table 2). 2. Radical hysterectomy complications—ureter a) Ureterovaginal fistula.~Ureterovaginal fistulas are more common than vesicovaginal fistulas. Among 3084 patients reported in the literature. the incidence of ureterovaginal fistula was 1.2% (range. ()%-3.3%) (Table 2). b) Stricturesr—Ureteral strictures are uncommon after radi— cal hysterectomy. occurring in about 1% of patients. Early stric- tures (<3 months postoperatively) are related to ureteral fibrosis and result from ischemia caused by dissection and devas- cularization of the distal ureter. Late strictures (>3 months pos- toperatively) are related to tumor recurrence. unless proved otherwise (38). 3. Radical hysterectomy complications—rectum. The most common rectal complication after radical hysterectomy is rectal dysfunction. The manometric findings are similar to those of bladder dysfunction (39). Clinically significant constipation has been noted in 5.3% of patients (24). Rectovaginal fistulas are not a complication of radical hysterectomy. at least in nonir— radiated patients. 4. Pelvic lymphadenectomy complications. Postoperative complications directly attributable to lymphadenectomy are the formation of lymphocysts and the development of lower ex— tremity lymphedema. a) Pelvic lymphocysts.—The incidence of pelvic lym— phocysts prospectively detected with ultrasonography is about 16.3%—22% and with clinical examination. 6% (40.41). How- ever. clinically significant lymphocysts are uncommon. In a prospective series of 100 patients with a 16.3% incidence of lymphocysts detected ultrasonographically. only one (1%) patient had symptoms attributable to the lymphocysts (41). In 1570 patients collected from retrospective series. the in- cidence of pelvic lymphocysts was 2% (range. 1%—4.3%) (1520.24.36.42). b) Lymphedema.——Lymphedema of the lower extremities is a sequela of pelvic lymphadenectomy. lts incidence is asso- ciated with the extent of the lymphadenectomy. With extensive lymphadenectomy. the incidence reportedly is as high as 23%, Table 2. Stage lB~l|A cervical cancer: urinary fistulas associated with radical hysterectomy Fistula (No. of patients. %) 1m estigator(s)/Ref No. No. of patients Ureteral Vesical Symmonds ( I5) 72 0 (0) 0(0) Webb and Symmonds (2(1) 423 6 ( 1.4) 3 (0.7) Shingleton and (iusberg (36) 444 6 ( 1.4) 1 (0.2) Lee et al. (25) 954 11 (1.2) 10 ( 1.2) Kenteret a1. (24) 213 7 (3.3) 7 (3.3) Averette et al. (37) 978 9 (0.9) 5 (0.5) Total 3084 39 (1.2) 26 (0.8) 55 whereas with more standard lymphadenectomy. it is about 89’— 10% (15.20.241.43). However. severe lymphedema is rare. Operative Mortality A review of the literature of the period from 1956 through 1993. including more than 4000 patients. revealed a 0.7% in— cidence of operative mortality (range. 0%—1.4% ) (Table 3). Results Survival For 3396 patients reported in the literature. the 5-year sur— vival rate for patients with stage IB cervical cancer was 85.7% (range. 80%—90.5%) (Table 4). For 452 patients with stage ”A cervical carcinoma reported in the literature. the 5-year survival rate was 69.6% (range. 65.3%—79%) (Table 5). Recurrences A survey of the literature encompassing 2338 patients with stage IB-IIA cervical cancer revealed a recurrence rate of 27.2% (range. |1.9%v—55.1%)(I6.2(),24.37.42.47). Analysis of the site of recurrence can be an indicator of the appropriateness of the operation. Among 823 patients with stage lB—IIA cervical cancer collected from two studies. recurrences were detected in the pelvis in 67 (8.1%) patients: they were ex- trapelvic in 62 other patients (7.5%) (Table 6). Of the 67 patients with recurrence in the pelvis. it was located centrally in 24 (2.9%) and in the pelvic sidewall in 43 (5.2%). Isolated central recurrences in the absence of extrapelvic disease were noted in only 1.6% of patients (48). The low rate of central and pelvic wall recurrence indicates the adequacy of central tissue resection and node dissection with a properly performed radical hysterectomy and pelvic lymphadenectomy. A Less Radical Operation for Selected Patients With Stage IBl Disease The standard central resection of radical hysterectomy has been applied systematically for years to cervical lesions ranging from 3 mm of invasion to bulky lesions of 6 cm or more in diameter. Similar to the less radical treatment for breast and vul— var cancer. recent studies have shown the safety of a less radical central resection for selected patients with stage IB1 cervical cancer. In the Mayo Clinic experience with 83 patients with cervical cancer less than 2 cm (mean depth of invasion. 0.6 cm: range 0.4-1.8 cm) and without lymphovascular involvement. the 5— year disease-free survival was 97.6%: median follow-up was 9.8 years (49). Modified radical hysterectomy was the treatment in four patients. none of whom had recurrence. A similar experience has been reported from other medical centers. In three other studies (42.46.50) involving 373 patients with cervical lesions with a depth of invasion up to 1 cm. the 5- year survival rate was 94.6% (range. 93.7%—97%). These data support the need for a prospective evaluation of a less radical 56 Table 3. Stage lB~llA cervical cancer: mortality associated with radical hysterectomy lm‘cstigutor(s)/ No. of Operative mortality Ref. No. patients (No. of patients. %) Synmonds ( I5) 101 0 (0) Webb and S_\ Inmonds (20) 610 2 (0.3) lander et al. (44) I092 10 ( 1.0) Shingleton and (itisberg (.36) 444 1 (0.2) Lee ct al. (25) 954 4 (0.4) Averette ct a]. (37) 973 13 (1.4) Total 4174 30 (0.7) Table 4. Stage 18 cervical cancer: survival after radical hysterectomy ln\'esligator(s)/ No. of 5-_v survival Ref. No. patients (No. of patients. %) Webb and Symmonds (20) 205 184 (90) Zander ct al. (44) 747 631 (84.5) Delgado c1 al. (45) 645 516 (80) Fuller el al. (42) 295 253 (86.0) Lee ct al. (25) 237 204 (86.1) Alvare/ clam-16) 401 340 (85.0) Averette et al. (37) 866 784 (90.5) Total 3396 2912 (85.7) Table 5. Stage 11A ccn ical cancer: survival after radical hysterectomy lnvesligator(s)/ No. of 5-y survival Ref. No. patients (No. of patients. ‘74) Webb and Symmonds (2U) 72 45 (62.5) Fuller et al. (42) I36 98 (72) Lee et al. (25) 106 76 (71,7) van Nagell el al. (47) 43 34 (79) Axerette et al. (37) 95 62 (65.7) Total 452 315 (69.7) Table 6. Stage lB-llA cervical cancer: site of recurrence after radical hysterectomy Pelvis (No. of patients. %) , r " ' '—‘ rem Distant 1nvestigator(s)/ No, of (No. of Ref. No. patients Central Sidewall patients. %) Webb and Symmonds (48) 610 17 (2.7) 37 (6.1) 46 (7.5) Kentereta|.(24) 213 7(3.2) 6(28) 16(7.5) Total 823 24 (2.9) 43 (5.2) 62(7.5) treatment for cervical cancer with a depth of invasion less than 2 cm. Treatment in Pregnancy Stage for stage. the outcome of the treatment of cervical can— cer in pregnant patients is similar to that of nonpregnant ones (151—55). The determining prognostic factor is the stage of the disease at the time of diagnosis. Also. survival is not affected by Journal ofthe National Cancer Institute Monographs No. 21. 1996 the gestational age at which treatment is instituted (51,52,55). The frequency of pregnancy and coexisting cervical cancer is 0.024%—0.046% (54,56). Timing of Treatment Traditionally, the timing of therapy for patients with early lesions is based on the gestational age at the time of diagnosis. If the gestational age is 20 weeks or fewer, immediate therapy is recommended without concern for the pregnancy. If the diag- nosis is established after 20 weeks of gestation, a waiting period till fetal maturity is allowed. The basis for this recommendation is that a long delay in therapy could result in local or systemic dissemination of the cervical cancer. worsening the prognosis for the mother. Indeed, from a review of 60 pregnant patients with cervical carcinoma reported in the literature and for whom therapy was delayed from 1 to 40 weeks, only two (3.7%) had disease progression (Table 7). In a prospective follow-up study of eight pregnant patients with stage lBl cervical carcinoma (<25 cm) and a mean therapeutic delay of 26.7 weeks (range, 3-40 weeks), no clinical progression occurred in any of the patients (62). It appears that in patients with a small (<2 cm) cervical car— cinoma and desirous of a viable fetus, prospective follow—up with clinical and magnetic resonance imaging at intervals of 2—4 weeks during pregnancy is feasible, without having a pro— nounced detrimental effect on the prognosis for the mother. The safety of delaying treatment for patients with bulky stage I lesions or a more advanced stage of disease has not been estab— lished. Type of Delivery The type of delivery recommended is cesarean section. be— cause of fear of dissemination of cervical cancer and worsening of the prognosis for the mother and because of the associated risks of infection. bleeding. and obstructed labor. Anecdotal reports exist of recurrent cervical carcinoma at the episiotomy site (52) and at the site of the cesarean incision on the ab- dominal wall (54). Numerous reports have shown no adverse prognostic effects from vaginal delivery, and the cure rates for mothers after vaginal delivery were the same for mothers after cesarean sec— tion, for all stages of cervical carcinoma (I). For stage 1 disease, the survival rate for 1 12 patients after vaginal delivery was 70%. similar to the rate of 75% for 40 patients after abdominal Table 7. Stage 1 cervical cancer and pregnancy: result of delayed therapy No. of lnvestigator(s)/Ref. No patients Delay, wk Maternal survival, % Prem et al. (57) 5 11—17 100(3-5 y) Dudan et al. (56) 23 8-24 2 (progression) Lee et al. (58) 8 1-12 No change in stage Greer et al. (59) 5 6-17 100 Monk and Moritz (60) 3 l()— 16 1()() (40 mo) Duggan et al. (6]) 8 7-30 100(23 mo) Sorosky et al. (62) 8 340 No progression Total 60 1-40 2 (3.7%, progression) Journal of the National Cancer Institute Monographs No. 21, 1996 delivery (Table 8). For patients with stage 11 disease, a survival advantage was noted for those who had had vaginal delivery (45%) as compared with those who had had cesarean section (23%) (Table 8). However, no firm conclusions can be drawn because these were retrospective studies and known prognostic factors such as lesion size were not considered. It is likely that a bias toward vaginal delivery could exist for patients with small lesions and could influence the results. It is possible that the reported risk of significant hemorrhage and obstructed labor (67) is associated with friable or bulky lesions. In such cases. cesarean delivery should be considered. Type of Treatment Although the survival rate after radiation therapy is similar to that after surgical treatment for patients with stage 1 cervical carcinoma (1), surgical treatment offers concurrent delivery and therapy as well as preservation of ovarian and vaginal function. It is the treatment of choice when there are no known contrain— dications. In early pregnancy, radical hysterectomy removes the pregnancy and the disease, whereas in late pregnancy, it imme- diately follows cesarean delivery. The complications of radical hysterectomy in pregnant pa— tients are not greater than in nonpregnant ones (54,60). As ex— pected. the amount of blood loss with radical hysterectomy in a pregnant patient is greater when associated with cesarean sec— tion (60). but when performed in the first trimester or postpar- tum, it is not different from that in a nonpregnant patient (53). Surgical Treatment Results In a review of the literature, the survival rate for 141 pregnant patients with stage I cervical cancer treated by radical hysterec— tomy and pelvic lymphadenectomy was 92.1% (range, 77.7%~ 100%) (Table 9). References (I) Shingleton HM. ()rr JW Jr: Cancer of the cervix. Philadelphia, Lippincott Company. 1995,11. pp 123-154; b, pp 332—235. (3) Polednak AP. Declining use of radiotherapy for invasive cervical cancer in Connecticut: 1983-1990. Gynecol ()ncol l995:58:226-9. Table 8. Stage I and ll cervical cancer and pregnancy survival according to type of delivery" Survival, % Vaginal delivery Abdominal delivery lnvestigator(s)/Ref. No. Stage I Stage II Stage 1 Stage ll Waldrop and Palmer (63) 75 0 70 44 Bosch and Marcial (64) 50 () 45 38 Prem ct al. (57) 67 (l 63 0 ()‘Lcary et al. (65) 50 100 67 0 (‘reasman et al. (5]) 8‘) 50 87 64 Lee et al. (58) 92 40 91 43 Nisker and Shuhat (66) 64 N/A 65 N/A Total 75 23 70 45 *N/A 2 not available. Modified from Shingleton and Orr ( I). By permission of Lippincott. Table 9. Stage 1 cervical cancer and pregnancy: survival after radical hysterectomy Investigator(s)/ Ref. No. No. of Survival Length of patients (No. of patients. %) follow-up. mo Dudan et al. (56) 1() 3311 et al. (53) Lee et al. (58) Hopkins and Morley (52) 21 Monk and Montz (60) 21 Duggan et al. (6/) Sivanesaratnam et a1. (54) 18 Total 236 23-260 12-2120 24—2120 40 l~ | 24 24~ I56 9 (9()) 28 (96.5) 19 (95) 18 (85.7) 2()(95) 22(1()()) l4(77.7) 130 (92. l) 29 2t) 22 141 (3 (4 (5 (6 (7 (8 (9 (10) (11) (12) (13) (l4) (I5) (16) (/7) (18) (/9) (20) (2]) (22) 58 Wertheim E. The extended abdominal operation for carcinoma uteri (based on 500 operative cases). Am J Obstet 1912166: 169—232. Magrina JF. Goodrich MA. Weaver AL. Podratl. KC. Modified radical hysterectomy: morbidity and mortality. Gynecol Oncol 1995;59:277-82. Podratz KC. Symmonds RE. Radical hysterectomy and pelvic lymph- adenectomy. ln: Surwit EA. Alberts DS. editors. Cervix cancer. Boston: Martinus Nijhoff. 1987:67-88. Morrow CP. ls pelvic radiation beneficial in the postoperative management of stage lb squamous cell carcinoma of the cervix with pelvic node metas» tasis treated by radical hysterectomy and pelvic lymphadenectomy? A report from the Presidential Panel at the 1979 Annual Meeting of the Society ofGynecologic Oncologists. Gynecol Oncol 1980110: 105—10. Kinney WK. Alvarez RD. Reid GC. Schray MF. Soong SJ. Morley GW. et al. Value of adjuvant whole-pelvis irradiation after Wertheim hysterec» tomy for early-stage squamous carcinoma of the cervix with pelvic nodal metastasis: a matched-control study. Gynecol ()ncol 1989;34:258-612. Shingleton HM. Surgical treatment of cancer of the cervix. Eur J Gynaecol Oncol 1992;13:45-52. Lagasse LD. Creasman WT. Shingleton HM. Ford JH. Blessing JA. Results and complications of operative staging in cervical cancer: exper- ience of the Gynecologic Oncology Group. Gynecol Oncol 198():9:9()-8. Delgado G. Bundy BN. Fowler WC Jr. Stehman FB. Sevin B. Creasman WT. et al. A prospective surgical pathological study of stage 1 squamous carcinoma of the cervix: a Gynecologic Oncology Group study. Gynecol Oncol 1989;35:314-20. Baltzer J. Lohe KJ. Kopcke W. Zander J. Formation of metastases in the ovaries in operated squamous cell carcinoma of the cervix uteri [German]. Geburtshilfe Frauenheilkd 1981;41:672-3. Toki N. Tsukamoto N. Kaku T. Toh N. Saito T. Kamura T. et al. Micro- scopic ovarian metastasis of the uterine cervical cancer. Gynecol Oncol 1991;41:46-51. Brown JV. Fu YS. Berek JS. Ovarian metastases are rare in stage 1 adenocarcinoma ofthe cervix. Obstet Gynecol 1990;76:623-6. Kjorstad KE. Kolbenstvedt A. Stricken T. The value of complete lyin- phadenectomy in radical treatment of cancer of the cervix. Stage IB. Can- cer1984z54z2215—9. Symmonds RE. Morbidity and complications of radical hysterectomy with pelvic lymph node dissection. AmJ Obstet Gynecol 1966;94:663—714. Hoskins WJ. Ford JH Jr. Lutl. MH. Averette HE. Radical hysterectomy and pelvic lymphadenectomy for the management of early invasive cancer of the cervix. Gynecol Oncol 1976:4:278—9(). Kennedy AW. Peterson G. Tuason LJ. Belinson JL. Webster KD. Radical hysterectomy for cervical cancer: the effect of shorter length of stay on outcome. Cleve Clin J Med 1995;62:193-7. Orr JW Jr. Shingleton HM. Hatch KD. Mann WJ Jr. Austin JM Jr. Soong SJ. Correlation of perioperative morbidity and conilation to radical hysterectomy interval. Obstet Gynecol 1982;159:7268 1. Hoffman M. Roberts WS. Cavanagh D. Second pelvic malignancies fol- lowing radiation therapy for cervical cancer. Obstet Gynecol Surv 1985: 40:61 1-7. Webb MJ. Symmonds RE. Wertheim hysterectomy: a reappraisal. Obstet Gynecol 1979;54:140-5. Olson N. Koh M. Greer B. Cain J. Tamimi H. Figge D. Macroscopic lymph node metastasis in stage IB carcinoma of the cervix explored for radical hysterectomy. Gynecol Oncol 1992;46:2-4. Hoffman MS. Roberts WS. Cavanagh D. Neuropathies associated with radical pelvic surgery for gynecologic cancer. Gynecol Oncol 1988131: 462—6. (23) E, «k (26) (27 (28) (30) (3/) (32) (33) (1’4) (.15) (.16) (37) (.18) (.19) (4(1) (4/) (42) (43) (44) (45) (46) Creasman WT. Weed JC Jr. Complications of radical hysterectomy. ln: Schaefer G. Grabcr EA. editors. Complications in obstetric and gyne— cologic surgery: prevention. diagnosis. and treatment. Hagerstown (MD): Harper & Row. 198138998. Kentcr (JG. Ansink AC. Hcintx AP. Aartsen liJ. Delemarrc Jl“. Hart AA. Carcinoma of the uterine cervix stage 1 and 11A: results of surgical treat ment: complications. recurrence and survival. Eur J Surg ()ncol 1989115: 55470. Lee YN. Wang KL. Lin MH. Liu CH. Wang KG. Lan CC. et al. Radical hysterectomy with pelvic lymph node dissection for treatment of cervical cancer: a clinical review of 954 cases. Gynecol ()ncol 1989;32:135-412. Ralph G. Winter R. Michelitsch L. Tamussino K. Radicality of parametrial resection and dysfunction of the lower urinary tract after radical hysterec- tomy. EurJ Gynaecol ()ncol 1991:12227-30. Sasaki H. Yoshida T. Noda K. Yachiku S. Minami K. Kaneko S. Urethral pressurc profiles following radical hysterectomy. Obstet Gynecol 1982: 59:1(11-4. Vervest HA. Barents JW. Haspcls AA. Debruyne FM. Radical hysterec» tomy and the function of the lower urinary tract. Urodynamic quantifica— tion of changes in storage and evacuation function. Acta Obstet Gynecol Scand 1989;68:331-40. Westby M. Asmusscn M. Anatomical and functional changes in the lower urinary tract after radical hysterectomy with lymph node dissection as studied by dynamic urethrocystography and simultaneous urethrocys- tometry. Gynecol Oncol 1985;21:261-761. Kadar N, Saliba N. Nelson J11. The frequency. causes and prevention of severe urinary dysfunction after radical hysterectomy. Br J Obstet Gynaecol l983:9():858~63. Scski JC. Diokno AC. Bladder dysfunction after radical abdominal hysterectomy. Am J Obstet Gynecol 1972;128:643-51. Brown JB. Copas PR. McDonald TW. Urodynamic studies before and after radical hysterectomy. J Tenn Med Assoc 1994;87:332-4. Forney JP. The effect of radical hysterectomy on bladder physiology. Am J Obstet Gynecol 198();138:374»82. Scotti RJ. Bergman A. Bhatia NN. Ostergard DR. Urodynamic changes in urethrovesical function after radical hysterectomy. Obstet Gynecol 1986;68:111-20. Loran OB. Pushkar DU. Urethral instability after radical hysterectomy. J Urol (Paris) 1992:982210—2. Shingleton HM. Gusberg SB. Radical hysterectomy. ln: Gusberg SB. Shingleton HM. Deppe G. editors. Female genital cancer. New York: Churchill—Livingstone.1988:535-53. Averette HE. Nguyen HN. Donato DM. Penalver MA. Sevin BU. Estape R. et al. Radical hysterectomy for invasive cervical cancer. A 25-year prospective experience with the Miami technique. Cancer 1993;7ltSupp1 4):1422-37. Hatch KD. Parham G. Shingleton HM. ()rr JW Jr. Austin JM Jr. Ureteral strictures and fistulae following radical hysterectomy. Gynecol Oncol 1984;19:17-23. Barnes W. Waggoner S. Delgado G. Maher K. Potkul R. Barter J. et al. Manometric characterilation of rectal dysfunction following radical hysterectomy [.wc comment citation in Medlinel. Gynecol ()ncol 1991242: 1 16-9. Conte M. Panici PB. Guariglia L. Scambia G. Greggi S. Mancuso S. Pelvic lymphocele following radical para-aortic and pelvic lymphadenectomy for cervical carcinoma: incidence rate and percutaneous management. Obstet Gynecol 1990;76:268-71. Lopes AD. Hall JR. Monaghan JM. Drainage following radical hysterec- tomy and pelvic lymphadenectomy: dogma or need? Obstet Gynecol 1995: 86:960-3. Fuller AF Jr. Elliott N. Kosloff C. Hoskins WJ. Lewis JL Jr. Determinants of increased risk for recurrence in patients undergoing radical hysterec- tomy for stage IB and 11A carcinoma of the cervix. Gynecol Oncol 1989:33:34»9. Martimheau PW. Kjorstad KE. Kolstad P. Stage 1B carcinoma of the cer- vix. the Norwegian Radium Hospital. 1968—1970: results of treatment and major complications. 1. Lymphedcma. Am J Obstet Gynecol 19782131: 389—94. Zander J. Bah/er J. Lohe KJ. Ober KG. Kaufmann C. Carcinoma of the cervix: an attempt to individualilc treatment. Results of a 20-year coopera- tive study. Am J Obstet Gynecol 1981;139:7524). Delgado G. Bundy B. Zaino R. Sevin BU. Creasman WT. Major F. Prospective surgical»pathologica| study of diseased‘rec interval in patients with stage IB squamous cell carcinoma of the cervix: a Gynecologic On— cology Group study. Gynecol Oncol 1990;38:352-7. Alvarez RD. Potter ME. Soong SJ. Gay FL. Hatch KD. Partridge EE. et al. Rationale for using pathologic tumor dimensions and nodal status to sub— Journal ofthe National Cancer institute Monographs No. 21. 1996 (47 (48) (4‘) (56) classify surgically treated stage [13 cervical cancer patients. Gynecol Oncol I99] :43zltJ8—l2. van Nagell JR Jr. Rayburn W. Donaldson ES. Hanson M. Gay EC. Yoneda J. et al. Therapeutic implications of patterns of recurrence in cancer of the uterine cervix. Cancer l‘)79;44:2354—6|. Webb MJ. Symmonds RE. Site of recurrence of cervical cancer after radi— cal hysterectomy. Am J Obstet Gynecol 1980;138:2“3-7. Kinney WK. Hodge DO. Egorshin EV. Ballard DJ. Podrat/ KC. ldentit'ica» tion of a low-risk subset of patients with stage IB invasive squamous can- cer of the cervix possibly suited to less radical surgical treatment [see comment citation in Medline]. Gynecol ()ncol |995z57:3-6. lnoue T. Prognostic significance of the depth of invasion relating to nodal metastases. parametrial extension. and cell types, A study of 628 cases with stage 18. HA. and 118 cervical carcinoma. Cancer 1984;54:3035-42. Creasman WT. Rutledge FN. Fletcher GH. Carcinoma of the cervix as- sociated with pregnancy. Obstet Gynecol 1970;36:495-501. Hopkins MP. Morley CW. The prognosis and management of cervical can» cer associated with pregnancy. Obstet Gynecol 1992;80:9-13. Sal] S. Rini S. Pineda A. Surgical management of invasive carcinoma of the cervix in pregnancy. Am J Obstet Gynecol 1974:] 181—5. Sivanesaratnam V. Jayalakshmi P. Loo C. Surgical management of early invasive cancer of the cervix associated with pregnancy. Gynecol ()ncol 1993;48:68-75. Symmonds RE. Carcinoma of the cervix associated with pregnancy. In: Lewis GC. Wentz LB. Jaffe RM. editors. Symposium on new concepts in gynecological oncology (a Hahnemann symposium). Philadelphia: F. A. Davis. 1966:18l-92. Dudan RC, Yon YL Jr. Ford JH Jr. Averette HE. Carcinoma of the cervix and pregnancy. Gynecol Oncol I973; l 1283—9. Journal of the National Cancer Institute Monographs No. 2]. I996 (57) (58) (59 (60 ,3 W (64) (65 (66) (67) Prcm KA. Makowski EL. McKelvey JL. Carcinoma of the cervix as— sociated with pregnancy. Am J Obstet Gynecol |‘)66:95:99-l()8. Lee RB. Neglia W. Park RC. Cervical carcinoma in pregnancy. Obstet Gynecol IOXI :582584-9. Greer BE. Easterling TR. McLennan DA. Benedetti TJ. Cain JM. Figge DC. et al, Fetal and maternal considerations in the management of stage I— B cervical cancer during pregnancy. Gynecol ()ncol l‘)89:34:61-5. Monk BJ. Mont/ FJ. Invasive cervical cancer complicating intrauterine pregnancy: treatment with radical hysterectomy. Obstet Gynecol 1992;80: 199-203. Duggan B. Muderspach Ll. Roman LD. Curtin JP. d'Ablaing G 3d. Mor— row CP. Cervical cancer in pregnancy: reporting on planned delay in therapy. Obstet Gynecol l993:82:598-6()2. Sorosky JI. Squatrilo R. Ndubisi BU. Anderson B. Podczaski ES. Mayr N. et al. Stage I squamous cell cervical carcinoma in pregnancy: planned delay in therapy awaiting fetal maturity. Gynecol ()ncol 1995;59:207-l0. Waldrop GM. Palmer JP. Carcinoma of the cervix associated with preg- nancy. Am J Obstet Gynecol l963;86:202-12. Bosch A. Marcia] VA. Carcinoma of the uterine cervix associated with pregnancy. Am J Roentgenol Radium Ther Ntlcl Med 1966;96:92-9. O'Leary JA. Munnell EW. Moore JG. The changing prognosis of cervical carcinoma dtiring pregnancy. Obstet Gynecol l966:28:46()—8. Nisker JA. Shuhut M. Stage ”3 cervical cancer and pregnancy: report 0149 cases. Am J Obstet Gynecol l983:l45:2()3—6, Hacker NF. Berek JS. Lagasse LD. Charles EH, Savage EW. Moore JG. Carcinoma of the cervix associated with pregnancy. Obstet Gynecol l982159z735-46. Primary Radiotherapy for Stage IE or IIA Cervical Cancer Perry W. Grigsby* Patients with carcinoma of the cervix (stages lB-IIA) have a tumor that is confined to the cervix and upper vagina. These patients, however, may have a tumor that ranges in diameter from less than 1 cm to as large as 10 cm. The sur- vival for all patients with stage IB disease is 86%; for those with stage IIA disease, it is 72%. Reports of radiotherapy alone for small stage [B or IIA cervical cancer indicate a cure rate of greater than 90% and a severe complication rate of less than 5%. Bulky cervical cancers of stages IB-IIA treated with high doses of radiation have survival and com- plication rates that are dependent on tumor size, and no sur- vival benefit has been demonstrated with the addition of a hysterectomy. Elective para-aortic irradiation has been demonstrated to be of benefit in this patient population. The quality of life for patients treated with radiotherapy alone depends on many factors. [Monogr Natl Cancer Inst 1996;21:61-4] The best standard treatment for patients with carcinoma of the cervix is radiation therapy alone. without surgery or chemo— therapy. This is supported by data from many institutions and my personal experience of treating hundreds of patients with this condition. No prospective randomized studies have been performed to directly compare radiotherapy with surgery for patients with early stage cervical cancer. Therefore. only retro— spective data are available (1). Information regarding quality of radiation therapy technique. survival outcomes. and quality of life will be presented (2). Radiotherapy Quality The technical aspects of radiotherapy quality and outcomes for carcinoma of the cervix have been extensively evaluated by Patterns of Care Surveys sponsored by the American College of Radiology and the National Institutes of Health. These surveys. conducted during the past 20 years. are based on a random sample of academic and nonacademic radiation oncology prac— tices throughout the United States. Consequences of the publica~ tion of the results of these surveys are that practice patterns have changed and outcomes have improved. Additional data regard- ing diagnostic and therapeutic aspects have been published by others. Diagnostic evaluation of lymph node status is an essential component of patient management. Lymphangiography has been clearly demonstrated by a Gynecologic Oncology Group (GOG) study as being the most sensitive and specific diagnostic imaging study for evaluating lymph node status. This GOG Journal of the National Cancer Institute Monographs No. 21. 1996 study. by Heller et al. (3). demonstrated lymphangiography sen— sitivity and specificity to be 79% and 73%. respectively. These results were significantly better than those obtained by com— puted tomography and ultrasonography. Although it is presumed to be more accurate. outcome benefit has not been demonstrated for surgical lymph node staging. The only survival outcome study evaluating the benefit of surgical lymph node staging is a matched-pair study performed by Potter et al. (4). In their study of 58 patients with surgical lymph node staging and 58 matched patients treated with radiotherapy alone without surgical lymph node staging. the 5—year survivals were no different (P = .48). When surgical lymph node staging is performed by the transperitoneal approach. the postirradiation bowel complication rate was 12% compared with 4% for the retroperitoneal ap- proach. as demonstrated by Weiser et al. (5) in another GOG study. One patient-related aspect that can affect outcome is to trans- fuse patients who are anemic at the time of diagnosis. This was shown in a prospective randomized study by Bush (6). Bush demonstrated that patients with cervical cancer treated with radiotherapy alone had more pelvic failures and poorer survival if their pretreatment hemoglobin level was less than 10 g/mL. Subsequently. in a prospective randomized study. Bush (6) demonstrated that pelvic failure could be decreased and overall survival increased if anemic patients were transfused before treatment. The radiotherapy process begins with the simulation of ir- radiation portals. The 1973 Patterns of Care Survey (7) demon- strated that the postirradiation complication rate was 5% for patients undergoing simulation compared with 14% for patients not undergoing simulation (P = .01). This same study also demonstrated a 14% complication rate if one irradiation field was treated per day compared with 8% if two irradiation fields were treated per day (P = .03) (8). Russell et al. (9) demon- strated that. if “standard" lateral irradiation ports were employed. then up to 24% of tumors may not be included in the treated volume. Kim et al. (10) also evaluated this issue and demonstrated a decreased pelvic control rate if the lateral portal margins are inadequate. The type of treatment equipment has also been demonstrated to be of importance. In a prospective randomized study of patients with cervical cancer. Allt ([1) demonstrated that the survival outcome was better if patients were treated by a high— energy betatron than if they were treated by a Co—60 unit. Also. *CUI‘I‘c’.\'])nll(/(‘Il(‘(‘ to: Perry W. Grighsy. M.D.. M.B.A.. F.A.C.R.. Radiation Oncology Center. Mallinckrodt lnstitute of Radiology. Washington University School of Medicine. 4939 Children‘s Place. Suite 5500. St. Louis. MO 63110. 61 the 1973 Patterns of Care Survey (7) demonstrated that treat— ment administered by a linear accelerator resulted in a lower pelvic recurrence rate (14%) than treatment given by other types of therapy equipment (21% recurrence rate) (P = .02). The total radiation dose has also been shown to be an impor— tant factor in outcome. The Patterns of Care data reported by Montana et al. (12) demonstrated that patients receiving less than 75 Gy to point A had a significantly worse outcome than those receiving 75 Gy or more to point A. If the radiation dose delivered to point A was by whole—pelvis irradiation alone. then the complication rate was 13%,. However. if the radiotherapy consisted of external irradiation and brachytherapy. then the complication rate was only 9% (P = .04) (7). Brachytherapy is. without a doubt. a determinant factor in the outcome of patients with cervical cancer. This is most clearly demonstrated in the longitudinal Patterns of Care Study of 1973. 1978. and 1983. These studies (7.8.1216) identified that some patients did not receive brachytherapy and that their outcomes were poor. Practices changed with the following results for patients with stage 1118 disease. The percentage of patients treated without brachytherapy was 40%. 24%. and 12% (P<.01) for 1973. 1978. and 1983. respectively. The corresponding 5— year survivals were 25%,. 39%. and 47% (P = .02). Another problem identified was the quality of the individual brachytherapy implants. Corn et al. (13) developed quality standards and reviewed the brachytherapy implants in the Pat— terns of Care database. correlated these findings with outcome. and demonstrated that local tumor control and survival were su— perior with implants of good quality than with poor—quality pro— cedures. The duration of overall treatment time has been related to sur— vival in the Patterns of Care studies and in other retrospective databases. One of the first and most elegant demonstrations of this is in a report by Fyles et al. (14). These investigators demonstrated that there was a 1% per day decrease in survival when the overall treatment time was prolonged. Other inves— tigators (1.15.16) have demonstrated similar decreases in sur— vival with prolonged treatment time. Another related factor for these patients is the relationship be- tween facility structure and outcome. Diamond et a1. (17) re— viewed the Patterns of Care database and related outcome to facility structure issues. such as technologists per treatment machine. new patients per physician. and new patients per physicist. Their analysis identified these factors as significant for recurrence. survival. and complications: i.e.. facilities with fewer staff per patient result in poor quality of care for the patient. Survival Outcome The survival of patients with carcinoma of the cervix is re- lated to stage of disease. However. survival is better correlated with size of tumor, irrespective of stage. This is true regardless of whether the patient is treated with radiotherapy or with surgery. Patients with carcinoma in situ are traditionally treated surgi- cally. These patients. however. may be treated with intracavitary irradiation rather than surgery. Excellent results have been 62 achieved. Grigsby and Perez (18) had no treatment failures in a group of 23 patients treated with irradiation alone for carcinoma in situ. Kolstad and Klem (19) had no failures in their group of 55 patients treated with intracavity radium. Patients with stage lA disease are also traditionally treated with surgery. However. this group of patients can be treated in- stead with intracavitary irradiation alone. with excellent results. Grigsby and Perez (18) had no failures at 5 years among 44 patients treated with irradiation alone. Kolstad (20) reported no failures among 36 patients who were treated with irradiation alone for stage 1A disease. Patients with stage IB carcinoma of the cervix are a very heterologous group. based on lesion size. The diameter of these masses can range from less than 1 cm to 10—12 cm. Treatment results for radiotherapy alone for all patients with stage IB dis- ease at the Mallinckrodt Institute of Radiology are 86% progres— sion—free survival at 5 years. These results are comparable to those obtained at other large institutions treating patients with radiotherapy alone. The results for patients with stage IB disease treated with surgery are also 86%. However. the populations of patients treated with radiotherapy and surgery are not the same. Patients treated with surgery tend to be younger. healthier. and less obese than those receiving radiotherapy (21). But the most important difference in these two populations is that those treated with surgery alone generally do not have known positive lymph nodes and have lesions that are usually smaller than 3-4 cm in maximum diameter. The effect of tumor size on survival outcome is clearly demonstrated for patients with stage 18 or 11A disease treated with radiotherapy alone (Table 1). For these stages of disease. patients with lesions less than or equal to 3 cm in diameter have survivals greater than 90%. unsurpassed by surgical results. For lesions larger than 3 cm in diameter. the survival rates fall significantly. and there are no comparable sur- gical results. since these patients usually do not undergo surgery because of their large tumor. The special cases of barrel-shaped cervical cancer are not discussed here. The importance of excellent—quality radiotherapy and ability to cure the disease in the pelvis cannot be overemphasized. This was significantly demonstrated by Fagundes et a1. (22) in a mul- tivariate analysis of 121 1 patients with carcinoma of the cervix who were treated with radiotherapy alone. One of the most im- portant factors in the development of distant metastases was a lack of pelvic tumor control (P<.001). Patients with cervical carcinoma of stage 18 (>4 cm). IIA. or 118 without known positive para—aortic lymph nodes should receive prophylactic para-aortic irradiation. This was demonstrated by Rotman et al. (23). who reported the results of a phase 111 Radiation Therapy Oncology Group study (RTOG Table l. Five—year progression—free survival (radiotherapy alone) Tumor size. cm Stage 18. % Stage ”A. % <1 100 100 l- l .9 93 100 22.9 98 75 3—3.9 83 66 24 76 61 Journal of the National Cancer Institute Monographs No. 21. 1996 79-20) in which these patients were randomly assigned to receive either prophylactic para—aortic irradiation or no para— aortic irradiation. This study demonstrated that. at 10 years. the overall survival was 44% for those not receiving prophylactic para-aortic irradiation (P : .04). Patients with stage 1B or IIA disease with positive para—aortic lymph nodes at the time of diagnosis have a 30% 5—year sur— vival. as demonstrated by Grigsby et al. (24) at the Mallinckrodt Institute of Radiology. Grigsby et al. also demonstrated that. if patients do not receive prophylactic para—aortic irradiation and subsequently have a recurrence in the para-aortic lymph nodes only. then there are no survivors beyond 2 years. Quality of Life The post-treatment quality of life in patients treated for car- cinoma of the cervix is multidimensional. Influencing factors include age. hormonal status. pretreatment physical and psycho- social status. tumor stage. and tumor size. The acute effects of radiotherapy for cervical cancer may in— clude dysuria. diarrhea. and fatigue. These effects usually resolve shortly after therapy is completed. The incidence of severe long-term complications for patients treated with radiotherapy at the Mallinckrodt Institute of Radiology can be related to tumor size. Patients with carcinoma in situ or stage IA cervical cancer treated with intracavitary brachytherapy alone experienced no severe long—term complications. The rate of severe complications among patients with stage IB disease was 1.4% for lesions less than 2 cm in diameter. 6.8% for those 2-5 cm in diameter. and 8.6% for those greater than 5 cm in diameter. For stage IIA. the corresponding incidences were 0%. 13.7%. and 9.6%. respectively. The most frequent and severe complications occur in those patients who receive two forms of radical therapy: radical hysterectomy with lymph node dissection and postoperative pel- vic irradiation. This situation occurs in one of five patients (25.26). The severe complications in the population of patients receiving both radical treatments include bowel obstruction. severe leg edema. and ureteral obstruction. The estimated in- cidence of these complications ranges from 5% to 25%. Soisson et al. (26) reported a 22% incidence of chronic leg edema in their patients who received both forms of radical therapy. Yet. chronic leg edema is a rare complication when patients are treated with radiotherapy alone. without prior surgery. Leg edema may occur in patients undergoing radical hysterectomy. It is often stated that two advantages to performing a radical hysterectomy rather than radiotherapy are ovarian preservation and superior sexual function. But Anderson et al. (27) reported that only 53% of their premenopausal patients undergoing radi— cal hysterectomy and ovarian preservation without radiotherapy maintained normal. uncomplicated ovarian function. Sexual function after irradiation can be poor or nonexistent in patients with advanced stage disease because of destruction of the vagina. cervix. and parametrial tissues by tumor and by the high doses of radiation necessary to control their disease. How- ever. in young patients with small lesions treated with lower doses of radiation. sexual function can be maintained. In a study of patients with stage 1 or II cervical cancer who were randomly Journal of the National Cancer Institute Monographs No. 21. 1996 assigned to treatment with either surgery or irradiation. Vincent et al. (28) demonstrated that there was no difference in post- treatment sexual function in the two groups. Appropriate postir— radiation management for these patients with early stage disease includes psychosexual counseling. hormone replacement thera- py. vaginal hormone creams. and excellent personal hygiene. Conclusion The quality of the radiotherapy that is administered for patients with carcinoma of the cervix is of extreme importance. The American College of Radiology Patterns of Care studies have demonstrated multiple factors in the treatment of these patients with irradiation that will result in an improved outcome. Quality radiotherapy consists of the simulation process. two fields per day. adequate port size. high—energy therapy equip- tnent. total irradiation dose. the use of brachytherapy. and short overall treatment time. The outcome for patients with carcinoma in situ or stage IA disease is 100% progression—free survival at 5 years and no sig- nificant long—term complications. Survival outcome for patients with stage IB or lIA disease is better for those treated with ex- cellent-quality radiotherapy than for those treated with surgery when lesion size is taken into account. All patients with cervical cancer should be treated with radiotherapy alone and not ex- perience the cost. discomfort. and complications of receiving two forms of radical therapy. The quality of life for patients with small cervical lesions is no worse with radiotherapy than with radical hysterectomy. References (I) Perez CA. Grigsby PW. Camel HM. Galaktos AE. Mutch D. Lockett MA. Irradiation alone or combined with surgery in stage 18. 11A. and 113 car- cinoma of uterine cervix: update of a nonrandomized comparison [we comment citation in Mcdline]. lntJ Radiat Oncol Biol Phys 1995;31:703- 16. Grigshy PW. Russell A. Bruncr D. Eifel P. Koh WJ. Spanos W. et al. Late injury of cancer therapy on the female reproductive tract. Int J Radiat Oncol Biol Phys 1995131 : 1281-99. Heller PB. Maletano JH. Bundy BN. Barnhill DR. Okagaki T. Clinical— pathologic study of stage 1113. 111. and IVA carcinoma of the cervix: ex— tended diagnostic evaluation for paraaortic node metastasis—a Gynecologic ()ncology Group study. Gynecol Oncol 199();38:425»30. Potter ME. Spencer S. Soong SJ. Hatch KB. The influence of staging laparotomy for cervical cancer on patterns of recurrence and survival. Int J Gynecol Cancer 1993:3:169-74. Weiser EB. Bundy BN. Hoskins WJ. Heller PB. Whittington RR. DiSaia P]. et al. Extrapcritoneal versus transperitoneal selective paraaortic lym- phadenectomy in the pretreatment surgical staging of advanced cervical carcinoma (a Gynecologic Oncology Group study). Gynecol Oncol 1989;33:283-9. (6) Bush RS. The significance of anemia in clinical radiation therapy. Int .1 Radiat Oncol Biol Phys 1986;12:2047-50. Hanks GE. Herring DF. Kramer S. Patterns of care outcome studies. Results of the national practice in cancer of the cervix. Cancer 1983: 51:959—67. Hanks GE. Diamond .1]. Kramer S. The need for complex technology in radiation oncology. Correlations of facility characteristics and structure with outcome. Cancer 1985:55(9 Suppl )22198-201. Russell AH. Walter JP. Anderson MW. Zukowski CL. Sagittal magnetic resonance imaging in the design of lateral radiation treatment portals for patients with locally advanced squamous cancer of the cervix. lntl Radiat Oncol Biol Phys 1992;23:449-55. Kim RV. McGinnis LS. Spencer SA. Meredith RF. Jennelle RL. Salter MM. Conventional four—field pelvic radiotherapy technique without com- puted tomography-treatment planning in cancer of the cervix: potential lo (3 (4 (5 (7 (8 (9 (10) 63 (ll) (12) (13) (I4) (15) (I6) (/7) (/8) (I9) 64 geographic miss and its impact on pelvic control. Int J Radiat Oncol Biol Phys 1995;31:109-12. Allt WE. Supervoltage radiation treatment in advanced cancer of the uterine cervix. A preliminary report. Can Med Assoc] 1969;100:792—7. Montana GS. Marty. KL. Hanks GE. Patterns and sites of failure in cervix cancer treated in the U.S.A. in I978. Int J Radial Oncol Biol Phys 1991;20:87-93. Corn BW. Hanlon AL. Pajak TF. Owen J. Hanks GE. Technically accurate inlracavitary insertions improve pelvic control and survival among patients with locally advanced cancer of the uterine cervix. Gynecol Oncol 1994;53:294-300. Fyles A. Keane TJ. Barton M. Simm J. The effect of treatment duration in the local control of cervix cancer [we comment citation in Medline]. Radiother Oncol 1992;25:273-9. Girinsky T. Rey A, Roche B. Haie C. Gerbaulet A. Randrianarivello H. et al. Overall treatment time in advanced cervical carcinomas: a critical parameter in treatment outcome. Int J Radiat Oncol Biol Phys 1993; 27:]051-6. Lanciano RM. Pajak TF. Mam. K. Hanks GE. The influence of treatment time on outcome for squamous cell cancer of the uterine cervix treated with radiation: a patterns-of—carc study. Int J Radial Oncol Biol Phys 1993;25:391-7. Diamond JJ. Steint‘eld AD. Hanks GE. The relationship between facility structure and outcome in cancer of the prostate and uterine cervix. Int J Radiat Oncol Biol Phys l99l :21:l()85-7. Grigsby PW. Perez CA. Radiotherapy alone for medically inoperable car- cinoma of the cervix: stage IA and carcinoma in situ. Int J Radiat Oncol Biol Phys 1991;21:375—8. Kolstad P, Klem V. Long—term followup of ll2l cases of carcinoma in situ. Obstet Gynecol [976.482 125—9. (20) (2]) (24) (25) (26) (27) (28) Kolstad P. Follow-up study of 232 patients with stage lal and 41 1 patients with stage la2 squamous cell carcinoma of the cervix (microinvasive car- cinoma). Gynecol Oncol 1989;33:265—72. Piver MS. Marchelli DL. Patton T. Halpern J. Blumenson L. Driscoll DL. Radical hysterectomy and pelvic lymphadenectomy versus radiation therapy for small (S3 cm) stage IB cervical carcinoma. Am J Clin Oncol 1988;11:21-4. Fagundes H. Perez CA. Grigshy PW. Lockett MA. Distant metastases after irradiation alone in carcinoma of the uterine cervix [see comment Citation in Medline]. Int J Radial Oncol Biol Phys 1992;24:197—204. Rolman M. Pajak TF. Choi K. Clery M, Marcial V. Grigsby PW. et al. Prophylactic extended-field irradiation of para-aonic lymph nodes in stages 118 and bulky IB and [IA cervical carcinomas. Ten-year treatment results of RTOG 79-21) [My comment citation in Medline]. JAMA 1995: 274:387—93. Grigsby PW. Vest ML. Perel. CA. Recurrent carcinoma of the cervix ex- clusively in the paraaonic nodes following radiation therapy [see comment citation in Medlinel. lntJ Radiat Oncol Biol Phys 1993;28:451—5. Feeney DD. Moore DH. Look KY. Stehman FB. Sutton GP. The fate of the ovaries after radical hysterectomy and ovarian transposition [we com- ment citation in Medlinel. Gynecol Oncol 1995;56:3-7. Soisson AP. Soper JT. Clarke-Pearson DL. Berchuck A. Montana G. Creasman WT. Adjuvant radiotherapy following radical hysterectomy for patients with stage IB and HA cervical cancer. Gynecol Oncol 1990;37: 390-5. Anderson B. LaPolla J. Turner D. Chapman G. Buller R. Ovarian transposition in cervical cancer. Gynecol Oncol 1993;49:206-14. Vincent CE. Vincent B. Greiss FC. Linton EB. Some marital—sexual con— comitants ofcarcinoma ot' the cervix. South MedJ 1975;68:552-8. Journal of the National Cancer Institute Monographs No. 21. 1996 Stress and Quality of Life Following Cervical Cancer Barbara L. Andersen * Significant progress has been made in understanding the psychologic and behavioral aspects of cervical cancer. Descriptive data indicate acute trauma and disruption with the diagnosis and treatment, yet the majority of women return to precancer life. The notable exception to this is sexual functioning, which remains an area of significant morbidity. Future research will need to test variables that predict which women will be vulnerable to sexual dysfunc- tion. Sexual self-concept (sexual self-schema), which is the extent to which a woman has a positive view of her own sexuality, appears to provide valuable information. The identification of such variables is an important step toward designing interventions for enhancing quality of life for patients with cervical cancer. [Monogr Natl Cancer Inst 1996;21:65-70] The human cost of cervical cancer is staggering. In many countries, it is a leading cause of cancer death for women. Par— ticularly in developing and Third World countries, the rate remains high (e.g., an age-adjusted death rate of 15.9 and 12.5 per IOO 000 in Mexico and Chile, respectively) (I). In the United States, however, there has been a 68% decline in the death rate over the past 30 years (2). While much of the decrease in the United States appears related to advances in early detection, cervical cancer will remain a significant medical problem that will affect the health of women worldwide. Diagnosis An early clinical study (3) suggested that the diagnosis of cancer produces an existential plight, meaning that the news brings shock, disbelief, and emotional turmoil. Today, we know that individuals even become alarmed at the time of medical screening. long before a cancer diagnosis is suggested (4). Fur— thermore. the manner in which the information is disclosed is important. Physicians who communicate hope have patients, in turn. who are more hopeful and who report more favorable over- all emotional adjustment (5). This points to the need for physicians and other health professionals to learn more effective and thorough means of communicating information (e.g., diag— nosis. treatment, and prognosis) to their patients (6,7). Studies have clarified that specific emotions—sadness (de- pression). fear (anxiety), and anger—characterize the period after diagnosis. In terms of clinical problems, it is not surprising that depression-related difficulties are the most common diag- noses. Survey estimates for major depression are in the order of 5%—6”/( (8.9). similar to that of the general population (10). Journal of the National Cancer Institute Monographs No. 21, 1996 However, when major depression and adjustment disorder with depressed mood are considered, prevalence rates are higher (e.g.. 16%—25%. 8,11). Lower rates have been found when diagnostic criteria are used and/or the patients assessed are am- bulatory with good physical functioning (11,12). In general, depression is more common for those patients in active treat- ment rather than those on follow-up. receiving palliative rather than curative treatment. with pain or other disturbing symptoms rather than not, and/or with a history of affective disorder or al- coholism. Treatment Surgery A certain component of the emotional distress occurring at diagnosis is that due to the anticipation of treatment. For patients, the diagnostic period with the selection and anticipa- tion of therapy is a time of multiple occasions of medical stres- sors. There have been few investigations of cancer surgery, but there are numerous descriptive and intervention studies of the reactions of healthy women undergoing surgery for benign con- ditions. The latter studies are consistent in their portrayal of (a) high levels of self—reported preoperative anxiety predictive of lowered postoperative anxiety and (12) postoperative anxiety predictive of recovery (e.g., time out of bed and pain reports). What may distinguish cancer surgery patients are higher overall levels of distress and slower rates of emotional recovery. For ex- ample, Gottesman and Lewis (13) found greater and more last- ing feelings of crisis and helplessness among cancer patients in comparison with benign surgery patients for as long as 2 months following discharge. Considering these data, findings on the interaction patterns of physicians and cancer patients on morning surgical rounds are disturbing. Blanchard et al. (14) found attending physicians on a cancer unit less likely to be supportive and address patients‘ needs than physicians treating general medical patients. The heavier volume of more seriously ill patients common to cancer units might be sources for this unfortunate relationship. Related findings indicate that oncology nurses might find their job sig— nificantly more stressful than other assignments [e.g.. cardiac, intensive care, or operating room nursing (15)] and may limit communication with cancer patients, especially those experienc- ing a recurrence (16). ”z(‘urrmpmulemv to: Barbara L, Andersen, Ph.D., Department of Psychology. The Ohio State University, 1885 Neil Ave. Columbus. OH 432104222, See “Note" section following “References." As noted above. there has been considerable research on the psychologic and behavioral aspects of response to surgery. and many effective interventions have been tested. Components of these interventions include procedural information (i.e., how the surgery is to be performed as well as pre— and post-operative events from the perspective of the patient). sensory information on the actual physical sensations of the surgery or preparatory events. behavioral coping instructions, cognitive coping inter- ventions. relaxation, hypnosis. and emotion-focused interven- tions. In a meta-analysis of this literature. Johnston and Vogele (/7) reported that procedural information and behavioral in— structions show consistent and strong positive effects on post- operative recovery. Effects are significant for a broad band of measures. including ratings of negative affect and pain. amount of pain medication. length of stay. behavioral recovery. and physiologic indices. Radiotherapy Many women receive radiation therapy. and clinical descrip- tions have noted patient’s fears (e.g.. being "burned." sterility). To understand radiation fears. the surgical anxiety studies described above have been a paradigm. Here again. anxiety can often cause more overall distress than physical symptoms (/8) and are predictors of treatment response (/9). lntracavitary treat— ments are particularly difficult for women [we (20) for a discus- sion]. lf interventions to reduce distress (especially anticipatory anxiety) are not conducted. heightened post-treatment anxiety is also found (2/) and might be maintained for as long as 3 months post—therapy. particularly when treatment symptoms linger [e.g., diarrhea, fatigue (22)]. When acute side effects resolve (usually by 12 months post—treatment), there appears to be no higher in— cidence of emotional difficulties for radiotherapy patients than for cancer surgery patients (23). Chemotherapy For some women, the expectation and/or experience of un- pleasant side effects can compromise their quality of life to the point that they may miss treatment appointments and/or be un- willing/unable to continue treatment. regardless if the treatment is curative or palliative (24,25). Noncotnpliance is a behavioral problem that can directly affect the effectiveness of cancer therapy. Noncompliance with treatment has been related to in— creased emotional distress [e.g.. (25.26)]. severity of treatment side effects [e.g., nausea and vomiting (25.27)]. and lower in— come (28). Even when patients are responsible for self-ad- ministration of therapy, such as taking their chemotherapy at home. and they can reduce the number of required, but incon- venient. hospital visits. noncompliance may continue. One Table l. Psychologic interventions to improve patient compliance . Appointment reminders . Clearly written and specific treatment communications given to the patient . Home visits . One‘session intenentions that include a tour ofthe oncology clinic. videotape presentation about the therapy. discussion/question sessions. and/or take-home information JA'JJIJ— 66 report of multidrug therapy with adults with hematologic can- cers indicated that self-reports (versus sera reports) overes- timated compliance by a factor of two (29). Psychologic interventions have focused on a variety of methods to improve patient compliance. as indicated in Table I (29-31). Burish et al. (3/) and others (32,33) suggest that preparatory information can improve coping with treatment. Psychologic Interventions Importantly. psychologic interventions can reduce distress during and immediately following gynecologic cancer treat- tnents. For example, Cain et al. (34) compared individual and group therapy formats. The intervention had eight components. including discussion of the causes of cancer at diagnosis. impact of the treatment(s) on body image and sexuality. relaxation training. emphasis on good dietary and exercise patterns. com- tnunication difficulties with medical staff and friends/family. and setting goals for the future to cope with uncertainty and fears of recurrence. The eight—session program was conducted during individual sessions conducted in the hospital or the women‘s homes or in weekly groups of four to six patients con- ducted at the hospital. Seventy—two women with recently diag- nosed gynecologic cancer or who were currently receiving therapy for gynecologic cancer participated. Post—treatment analyses indicated that all groups improved with time. however. anxiety was significantly lower for the individual—therapy sub— jects. Gains for the intervention subjects were more impressive with the 6—month follow-up data when there were no differences between the intervention formats. but both groups reported less depression and anxiety and better psychosocial adjustment (in— cluding health perspectives. sexual functioning. and use of leisure time) than the no-treatment control group. Thus. the brief intervention appeared to be immediately effective. with gains enhanced during the early recovery months, Recovery and Long-term Survival Emotional Distress Lingering emotional distress from the trauma of diagnosis. treatment. and. more generally, life threat may occur for a small subset. perhaps 5%-IO% of cancer patients. When pronounced, such distress has been likened to post—traumatic stress disorder. [In fact. having residual distress from the diagnosis and treat— ment of a life-threatening illness is now included as one of the circumstances that may precipitate a psychiatric diagnosis, such as Post Treatment Stress Disorder (35). It is unlikely that such extreme distress will occur for the average cervical cancer patient and instead only potentially occur for those who have undergone the tnost difficult of treatment regimens. lengthy toxic chemotherapies or life—altering and/or disfiguring treat— ments (e.g.. pelvic exenteration). Sexuality The psychosocial literature on gynecologic cancer of the last 50 years has focused on sexual outcomes (36). In one of the ear- liest prospective reports (37). 50 women with early stage cervi- cal disease were randomly assigned to receive either radical Journal ofthe National Cancer Institute Monographs No.21. 1996 hysterectomy or radiotherapy treatments. Similar to the results from the retrospective studies. the changes in sexual desire and activity from pretreatment to 6 months post—treatment were comparable for the treatments: estimates of diminished desire were obtained from 24% of the radiation therapy patients and 20% of the surgical patients. Decreased frequency of intercourse was reported by 29% of the radiation therapy patients and 33% of the surgical patients. More recently, Schover et al. (38) reported sexual outcomes for 61 women treated for stage l-lIa cervical cancer. Twenty-six women received radical hysterec- tomy and 35 received radiotherapy: however, 19 of the latter also received surgery. While data were gathered at 6 and 12 months post-therapy, data were available only for 48% and 42% of the subjects, respectively. The rates of sexual complaints were high. For example, at 12 months, 21% noted that vaginal penetration was painful, 45% indicated deep thrusting was pain- ful, and 24% reported postcoital vaginal pain. Our studies have also included a controlled longitudinal study of sexually active and inactive women with stage 1 or 11 cervical or endometrial disease. They were assessed at diagnosis, prior to treatment, and at 4-, 8-. and 12-months post-treatment. The sexual behavior, sexual response cycle, sexual dysfunction, and medical outcomes for the women with cancer were compared with data from two matched comparison groups: women diag- nosed and treated for benign disease (e.g., uterine fibroid treated with simple hysterectomy) and gynecologically healthy women (39). Analyses indicated that the primary sexual behavior dis- rupted by the disease and treatment process for women with malignant or benign disease was the frequency of intercourse. declining from an average of 9.5 occasions per month to six to seven occasions per month during the post—treatment period. Eleven percent of the women with cancer failed to resume regular intercourse. This rate was comparable to that in the com— parison groups; however, one half of these cases in the cancer group were due to apparent disease—related causes (e.g., dyspareunia). It appeared that the women‘s complaints of inter— course pain also contributed to the husbands‘ losing erections during intercourse, which is a common clinical scenario in cases of sexual dysfunction among healthy individuals. The difficulty with sexual excitement for both disease groups was substantial. Following treatment, women with disease reported awareness of fewer signs of sexual excitement and lower arousability for sexual activities with their partner, and the evaluators and the women themselves felt that significant arousal problems were experienced. A likely reason for the arousal deficits was the co- occurrence of significant disruptors (e.g., dyspareunia due in part to radiation effects and/or induced menopause). Also important are the data on the percentage of sexual dys- function diagnoses and their status by 12 months post—treatment. The 16%-20% of cases found to be continuing dysfunction (i.e., sexual dysfunction that continued from 4 to 12 months post- treatment) represent the numbers of women who had chronic problems and who are at greatest risk for permanently impaired sexual functioning. The 10%-15% of cases noted as new late (i.e.. no dysfunction reported at 4 months but difficulties at 12 months) represent the number of cases that could potentially be prevented with therapy during the immediate post—treatment period. The 15%-30% of cases found to be resolved were Journal of the National Cancer Institute Monographs No. 21. I996 heterogeneous. About 50% of the latter cases are women whose dyspareunia improved with continued recovery, whereas the remainder are women who adapted to their sexual problems— they reported the same low responsiveness but they also re- ported less accompanying emotional distress. A companion report (40) provided data on other quality—of— life outcomes. We included the 47 sexually active women described above as well as an additional 18 women who were not sexually active at diagnosis. The data indicated that the emo- tional response to the life-threatening diagnosis and the anticipa— tion of treatment was characterized by depressed. anxious. and confused moods. whereas the response for women with benign disease was anxious only. In both cases, the responses were transitory and resolved post—treatment. There was no evidence of marital dissolution or poorer adjustment; however, 30% of the women treated for disease reported that their sexual partners may have had some difficulty in reaching orgasm (i.e., delayed ejaculation) during the recovery year. There was no evidence of impaired social adjustment. Finally. women retained their employment and occupations; however. their involvement (e.g., hours worked per week) was reduced significantly during recovery. In summary, life areas other than sexuality appeared to rebound or be resilient to major disruption among gyne— cologic cancer survivors, whereas rates of sexual morbidity are in the range of 50% and of moderate severity. Predicting risk. With this basis of descriptive findings, re- search is currently focused on predicting risk for sexual mor— bidity. It is evident that medical factors. such as the extent of disease and magnitude of treatment, appear to play part of the role in determining sexual outcomes for women with gyne- cologic cancer. Models of disease and treatment pathways that can yield differential levels of risk for sexual morbidity have been formulated as summarized in Table 2 [see (41) for a com- plete discussion]. Research suggests five important factors. First. at the onset of cancer, we consider the occurrence of dis- ruptive signs/symptoms. When the disease is detected and produces sexual disruption, this is the first point of psy— chologic/behavioral morbidity. This early change is important. since it increases a woman’s emotional distress and alerts her to the potential for subsequent life changes. Second, the extent of disease and of treatment are major determinants of risk. Third, for those women whose therapy cannot be modified, it is impor- tant to consider the availability of risk—reducing medical inter- ventions. For those with extensive disease/treatment. the availability of such interventions might reduce the level of risk from high to moderate. Examples of rehabilitative medical ef- forts include vaginal reconstruction for pelvic exenteration pa- tients. For patients who have had such interventions, psychologic. behavioral, and sexual outcomes are significantly better, although the interventions are not panaceas (42.43). Table 2. Predictors of risk for sexual morbidity (4/) . Sexual self-schema . Level of precancerous sexual activity (cg. frequency of intercourse) . Extent of disease and treatment . Extent of post-treatment hormonal changes and/or other health problems 411le— 67 Fourth. we consider the occurrence of new health problems (e.g.. hormonal changes and/or continuing stressors from the disease/treatment). Consideration of hormonal changes includes two issues: induced menopause and infertility. Menopausal changes produce significant sexual effects [see (44) for a review]. Also, ovary removal/sterilization ends child-bearing potential—a possible stressor for the young woman with cancer. Thus. multiple pathways can lead to high or moderate sexual morbidity risk. but only those individuals who have limited dis— ease/treatment and who have no new or continuing problems are hypothesized to have the lowest risk. Sexual self-schema as a predictor. In addition to medical contributors. we have focused on identifying sexually relevant psychologic factors in predicting risk for sexual morbidity. Specifically. sexual self-schema (self-concept) is a cognitive view about sexual aspects of oneself; it is derived from past ex- perience. manifested in current experience. and it guides the processing of domain-relevant social information (45). The con- cept includes two positive aspects—an inclination to experience romantic/passionate emotions and a behavioral openness to sex— ual experiences and/or relationships—and a negative aspect— embarrassment and/or conservatism—that appears to be a deterrent to sexual expression. Using a measure developed for this purpose. the predictive power of sexual self—schema was ex- amined in the context of risk of sexual morbidity following gynecologic cancer (46). Consistent with our definition of the construct. we anticipated that women with a negative sexual self-concept. in contrast to women with a positive sexual self- concept. would be at greatest risk for sexual difficulty. Women with a negative sexual schema would be expected to have more difficulties because they are. in general. less romantic/passionate in their emotions. less open to sexual experiences. and more likely to have negative feelings about their sexuality. In a field test of sexual schema in the prediction of gyne- cologic cancer outcomes (46). we assessed 62 women who were currently disease free but who received treatment from 6 months to 5 years previously for stage 1-“ disease. Comparison subjects included 68 women seeking routine gynecologic care. Analysis of the quality-of—life data replicated earlier prospective lon— gitudinal findings (39.40)—specifically. sexuality is the major life area of disruption for the survivor. For example. there were no differences between the groups in the areas of mental health (emotional distress. depression) or social functioning In con- trast. a comparison between the samples in terms of current sexual functioning found significant differences. with the cancer sample reporting lower levels of sexual behavior. sexual respon- siveness. and global evaluations. We tested the use of the schema construct with the cancer sample in the prediction of sexual responsiveness (i.e.. desire. excitement. orgasm. and resolution) and sexual behavior (frequency of intercourse) with regression analyses. Sexual self-schema accounted for a sig— nificant and large portion of the variance (26%) in the prediction of current sexual responsiveness. Marital Adjustment In view of these data on sexual outcomes. there has been the question of the potential impact on marital happiness for couples in whom the woman has been treated for gynecologic cancer. 68 Despite the emotional distress and. for some. accompanying sexual disruption that couples experience. data from the con- trolled longitudinal studies previously discussed indicate that marriages remain intact and satisfactory. and. on occasion. stronger (Baider & Sarell. 1984). Psychologic Interventions A large literature on the use of psychologic interventions to improve the cancer patients quality of life exists (47). The ef- fectiveness of these interventions is robust. since they have reduced distress and enhanced the quality of life of many cancer patients differing as to disease stage as well as disease site. De- spite the challenges of studying these patients, well-controlled investigations have been conducted. Improvements in emotional distress are found at the end of the interventions as well as with continued gains at follow-up. In addition. change in other areas—self—esteem/concept. death perceptions. life satisfaction. and/or locus of control—have been found. Important for quality of life. psychologic interventions have also lowered or stabilized pain reports. The positive outcomes for terminal patients are notable. considering their worsening pain and/or increasing debilitation. While there appear to be unique intervention components for different phases in disease. there are some commonalities. Therapy components have included those noted in Table 3. These components appear to be more important to the outcome than procedural variations. For example. therapy format. such as individual or group. appears to have little impact. There were also null findings for (group) interventions that included no structured content. suggesting that group support alone is insuf— ficient to produce any measurable benefit. How do psychologic interventions achieve these effects? In large measure. the psychologic mechanisms may not differ from those designed for coping with other stressors. That is. confront- ing a traumatic stressor with positive cognitive states. active be— havioral strategies. and. eventually. lowered emotional distress may enhance one‘s sense of control and provide realistic ap— praisals of stresses of the disease or treatment process. Similarly for sexual interventions. information provides realistic expecta— tions for sexuality and specific strategies to manage sexual ac- tivity when it is difficult or impossible. That the interventions produce more than situational improvement and may alter an individual's long—term adjustment processes is suggested by the data indicating that adjustment gains continue (and often in- crease) during the first post-treatment year. Immediate and long- terrn psychologic changes may. in turn. increase the likelihood Table 3. Components ofeffective psychologic therapies that improve quality of life (47) I. An emotionally supportive context to address fears and anxieties about the disease. typically offered in a group meeting format . Provision of information about the disease and treatment and ways to main- tain compliance and working relationships with health care providers . Behavioral coping strategies such as assertive communication . Relaxation training to lower anxiety and/or enhance one‘s sense of control 5. Focused information and interventions for sexual functioning. including in— formation about treatment effects on sexuality. postmenopausal changes. and frank discussion of body changes and distress about body image L) Lb.) Journal ofthe National Cancer Institute Monographs No. 2 1. I996 of changes in behavioral mechanisms. such as increasing the likelihood of adaptive health behaviors (e.g.. complying with medical therapy: improving diet. exercise. etc.). to directly im- prove mental health. adjustment. and. possibly. medical out- comes (48). Recurrence and Death Cancer recurrence is devastating: the magnitude of distress is even greater than that found with the initial diagnosis (49). and studies contrasting cancer patients showing no evidence of dis- ease with those receiving palliative treatment (50) have reported the greatest distress for those with disseminated disease (5]). Difficult decisions (e.g., beginning a regimen that offers little chance for cure and has side effects versus no treatment) are tnade in a context of extreme emotional distress and physical debilitation. The few studies of psychologic interventions for adult patients have indicated that important emotional gains can be achieved during terminal stages (52). At this time of sig- nificant emotional turmoil and physical difficulty. psychologic interventions appear to enhance the quality of life (53.54). Conclusion This paper provides a brief overview of the central findings that have emerged on the psychologic and behavioral aspects of cervical cancer, and the discussion is organized by disease- relevant time points. The time of diagnosis is a period of acute emotional distress. including anxious. depressed. and confused moods. For the majority. these feelings dissipate as treatment is begun and completed. with adjustment stabilizing by 1 year fol— lowing therapy. In contrast. sexual functioning is disrupted during this first year for the majority of women and may remain a source of chronic difficulty for upwards of 40% of women who are successfully treated. A model has been proposed for predicting which women are at risk for sexual morbidity. with the primary predictors including pretreatment level of sexual functioning. extent of treatment. and sexual self-schema. or a woman’s self view of her own sexuality. Finally. psychologic interventions are available to reduce distress and enhance quality of life for cervical cancer patients. References (I) Parker SL. Tong T, Bolden S. Wingo PA. Cancer statistics. 19961314» com» ment citation in Medline]. CA Cancer] Clin 1996;46:5-27. (2) American Cancer Society. Cancer facts and figures—1996. New York. (3) Weisman AD. Worden JW. The existential plight in cancer: significance of the first 100 days. IntJ Psychiatry Med I976:7:I—15. (4) Wardle J. Pope R. The psychological costs of screening for cancer. Psychosom Res 199286260924. (5) Sardell AN. Trierweiler SJ. Disclosing the cancer diagnosis. Procedures that influence patient hopefulness. Cancer 1993;72:3355»65. Sell L. Devlin B. Bourke SJ. Munro NC. Corris PA. Gibson GJ. Com< municating the diagnosis of lung cancer. Respir Med 1993;87:61-3. Woodard LJ. Pamies R]. The disclosure of the diagnosis of cancer. Prim Care 1992;19:657—63. (8) Derogatis LR. Morrow GR. Fetting J. Penman D. Piasetsky S. Schmale AM. et al. The prevalence of psychiatric disorders among cancer patients. JAMA1983;249:751-7. (9) Lansky SB. List MA. Herrmann CA. Ets»Hokin EG. DasGupta TK. Wilhanks GD. et a1. Absense of major depressive disorders in female can- cer patients. J Clin Oncol 198511155360. Journal of the National Cancer Institute Monographs No. 21, I996 (10) (II) (15) (I6) ([7) (18) (22) (23) (24) (25) (26) (27) (28) Locke BZ. Regier DA. Prevalence of selected mental disorders. In: Taube CA. Barrett SA. editors. Mental health in the United States. Rockville (MD): National Institute of Mental Health. 1985:1«6. Massie MJ. llolland JC. Depression and the cancer patient. J Clin Psychiatry 1990;51:12—9. Levin SH. Jones LD. Sack DA. Evaluation and treatment of depression. anxiety. and insomnia in patients with cancer. Oncology I993:7:I 19—25. Gottesman 1). Lewis M. Differences in crisis reactions among cancer and surgery patients] ConsultClin Psychol 1982;50:381-8 Blanchard CG. Ruckdeschel JC. Labrecque MS. Frisch S. Blanchard EB. The impact of a designated cancer unit on house staff behaviors toward patients. Cancer 1987;60:2348—54. Stewart BE. Yarkin KL. Meyerowitz BE. Harvey JH. Jackson LE. Psycho- logical stress associated with outpatient oncology nursing. Cancer Nurs 1982:5:383-7. Wilkinson S. Factors which influence how nurses communicate with can- cer patients. J Adv Nurs 1991;16:677—824. Johnston M. Vogelc C. Benefits of psychological preparation for surgery: a meta-analysis. Anti Behavior Med 1993;15:245-56. Munro AJ. Biruls R. Griffin AV. Thomas H. Vallis KA. Distress as- sociated with radiotherapy for malignant disease: a quantitative analysis based on patients' perceptions. BrJ Cancer 1989;60:370-4. Wallace LM. Priestman SG. Dunn JA. Priestman T]. The quality of life of early breast cancer patients treated by two different radiotherapy regimens. Clin Oncol 1993;5z228-33. Andersen BL. Karlsson JA. Anderson B. Tewfik HH. Anxiety and cancer treatment: response to stressful radiotherapy. Health Psychol 1985353551. Andersen BL. Tewfik HH. Psychological reactions to radiation therapy: reconsideration of the adaptive aspects of anxiety. J Pers Soc Psychol I985;48:|024»32. King KB. Nail LM. Kreamer K. Strohl RA. Johnson JE. Patients' descrip- tions of the experience of receiving radiation therapy. ()ncol Nurs Forum 1985;12:55—61. Hughson AV. Cooper AF. McArdle CS. Smith DC. Psychosocial effects of radiotherapy after mastectomy. Br Med] 1987;294:1515-8. Morrow GR. Hickok JT. Behavioral treatment of chemotherapy-induced nausea and vomiting. Oncology 1993;7:83»9. Richardson JL. Marks G. Levine A. The influence of symptoms of disease and side effects of treatment on compliance with cancer therapy. J Clin Oncol 198816174652. Gilbar O. De»Nour AK. Adjustment to illness and dropout of chemo- therapy. J Psychosotn Res 1989;33:1-5. Lewis C. Linet MS. Abeloff MD. Compliance with cancer therapy by patients and physicians. Am J Med I983;74:673»8. Lebovits AH. Strain JJ. Schleifer SJ. Tanaka JS. Bhardwaj S. Messe MR. Patient noncompliance with self—administered chemotherapy. Cancer 1990'. 65: 17-22. Richardson JL. Marks G. Johnson CA. Graham JW. Chan KK. Selser JN. et al. Path model of multidimensional compliance with cancer therapy. Health Psychol 198726: I 83-207. Anderson RJ. Kirk LM. Methods of improving patient compliance in chronic disease states. Arch Intern Med 1982;142:1673-5. Burish TG. Snyder SL. Jenkins RA. Preparing patients for cancer chemo— therapy: effect of coping preparation and relaxation interventions. J Con- sultClin Psychol 1991;39:518—25. Nail LM. King KB. Symptom distress. Fatigue. Semin Oncol Nurs 1987; 3:257—62. Rainey LC. Effects of preparatory education for radiation oncology patients. Cancer 1985;56:1056—61. Cain EN. Kohorn El. Quinlan DM. Latimer K. Schwartz PE. Psychosocial benefits ofa cancer support group. Cancer 1986;57:183-9. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th edition. Washington (DC): American Psychiatric Association. Andersen BL. van der Does J. Sexual morbidity following gynecologic cancer: an international problem. IntJ Gynecol Cancer 1994;4z225-40. Vincent CE. Vincent B. Greiss FC. Linton EB. Some marital-sexual con- comitants ofcarcinoina of the cervix. South Med J 1975;68:552-8. Schover LR. Fife M. Gershenson DM. Sexual dysfunction and treatment for early stage cervical cancer. Cancer 1989;63:204—12. Andersen BL. Anderson B. deProsse C. Controlled prospective longi- tudinal study of women with cancer: 1. Sexual functioning outcomes. J Consult Clin Psychol 1989;57:683-931. Andersen BL. Anderson B. deProsse C. Controlled prospective longi« tudinal study of women with cancer: ll. Psychological outcomes. J Consult Clin Psychol 1989;57:692-7. Andersen BL. Predicting sexual and psychologic morbidity and improving the quality of life for women with gynecologic cancer. Cancer 1993:71: 1678-90. 69 (42) (43) <44) (45) (46) (47) (48) (49) 70 Andersen BL. Hacker NF. Psychosexual adjustment following pelvic ex- enteration. Obstet Gynecol [983:ol :33 1-8. Berek J. Andersen BL. Sexual rehabilitation: surgical and psychological approaches. In: Hoskins WJ. Perel CA. Young RC. editors. Principles and practice of gynecologic oncology 2nd ed. Philadelphia: Lippincott. In press. Walling M. Andersen BL. Johnson SR. Hormonal replacement therapy for postmenopausal women: a review of sexual outcomes and related gyne- cologic effects. Arch Sex Behav l99l);l9:l l9-37. Andersen BL. Cyranowski JC. Women's sexual self schema. J Pers Soc Psychol 1994;67:1079—l00. Andersen BL. Woods X. Copeland L. Sexual self schema and sexual mor- bidity following gynecologic cancer. J Consult Clin Psychol. In press. Andersen BL. Psychological interventions for cancer patients to enhance the quality of life. J Consult Clin Psychol 1992;60:552-68. Andersen BL. Kiecolt—Glaser J, Glaser R. A biobehavioral model of cancer stress and disease course. Am Psychol 1994;49:389404. Thompson L. Andersen BL. DePetrillo D. The psychological processes of recovery from gynecologic cancer. In: Coppleson M. Morrow P. Tattersall M, editors. Gynecologic oncology. 2nd ed. Edinburgh: Churchill Livingstone. 1992. (50) (5/) (52) t 53 ) (54) Cassiletli BR. Lunk EJ. Strouse TB. Miller DS. Brown LL, Cross PA. A pyschological analysis of cancer patients and their ncxt~of-kin. Cancer 1985;55:72-6. Bloom JR. Psychological response to mastectomy. Cancer 1987;59:189-96. Linn MW. Linn BS. Harris R. Effects of counseling for late stage cancer patients. Cancer 1982149104855. Speigel D. Bloom JR. Yalom I. Group support for patients with metastatic cancer. A randomized outcome study. Arch Gen Psychiatry 19%| :381527- 33. Speigel D. Bloom JR. Group therapy and hypnosis reduce metastatic breast carcinoma pain. Psychosom Med 1983;45:333-9. Note Supported by grants PER-89 and PBR-X9A frotn the American Cancer Society. Journal of the National Cancer Institute Monographs No. 21. I996 New Surgical Approaches to Treatment of Cervical Cancer Kenneth D. Hatch, Alton V. Hallum Ill, Magd'v Nour* Purpose: Our goal was to evaluate laparoscopic pelvic lymph node dissection, para-aortic lymph node sampling, and laparoscopic radical vaginal hysterectomy (Schauta) in the treatment of early stage cervical cancer. Materials and Methods: In a retrospective study of 37 patients treated in the period between October 1993 and February 1996, we evaluated operative time, blood loss, length of hospital stay, lymph node count, and morbidity. Radical abdominal hysterectomy was compared with laparoscopic pelvic lymph node dissection and para-aortic lymph node sampling. 1m- provement over time was analyzed. Results: Mean operative time was 225 minutes, blood loss was 525 mL, and the average hospital stay was 3 days. This information was com- pared with a radical abdominal hysterectomy and pelvic and para-aortic lymph node dissection, where the operative time was 210 minutes, blood loss was 1500 mL, and the hospital stay was 9.7 days. Blood transfusion was required in 11% of patients compared with a range of 35%-95 % reported in the literature for radical abdominal hysterectomy. The mean pelvic lymph node count was 35; the mean para-aortic lymph node count was 11. Two patients had cystotomies repaired at surgery without lengthening hospital stay or subsequent complication. Two patients had ureteral vaginal fistulae treated by a ureteral stent, which was removed 6 weeks later without further operative procedures or urinary damage. When the data were correlated with the length of experience using the analysis of variance test and linear regression, operative time, blood loss, and hospital costs sig- nificantly improved over time. Patient charges averaged $14 868.00 and estimated hospital costs averaged $6449.00. Conclusion: Laparoscopic pelvic lymph node dissection and para-aortic lymph node sampling can be performed with adequate lymph node counts and lower morbidity. Laparo- scopic Schauta allows shorter hospital stay than radical ab- dominal hysterectomy, with significantly less blood loss and markedly fewer blood transfusions. Morbidity is higher early in the surgeon’s experience but decreases over time. [Monogr Natl Cancer Inst 1996;21:71-5] Video laparoscopy has been widely accepted for many opera- tive procedures because it requires a short hospital stay and the patient recovers quickly from the procedure and is able to rapid— ly return to full activity. Its use in cervical cancer was limited 1) by a lack of technique to perform pelvic lymphadenectomy and para—aortic lymph node sampling. 2) by a lack of trained sur— geons to perform the laparoscopic operation, and 3) by the fact that there did not appear to be an advantage to laparoscopic Journal ofthe National Cancer Institute Monographs No. 21. 1996 lymph node dissection. since the standard operation for the primary cervical tumor was radical abdominal hysterectomy. The techniques for performing pelvic lymph node dissection and para-aortic lymph node sampling were developed by Dar- gent and Salvat (I). Querleu et al. (2). and Childers et al. (3,4). Studies comparing laparoscopic lymph node dissection or sam- pling with immediate laparotomy and inspection of the lymph node bed were published by Querleu et al. (2). Childers et al. (3). and Fowler et al. (5). Querleu et al. (2) performed transperitoneal laparoscopic pel- vic lymphadenectomy in 39 patients with cervical cancer. Thir- ty—two patients underwent abdominal radical hysterectomy and evaluation of the completeness of the laparoscopic lymph node dissection. Querleu et al. found that five patients had positive pelvic lymph nodes by laparoscopy but found no further positive lymph nodes by laparotomy. These investigators did not state how many additional lymph nodes were removed. and they did not report para-aortic lymph node sampling. Childers et al. (3) reported pelvic lymphadenectomy and para—aortic lymph node sampling in 16 of 18 patients who were being treated for cervical cancer. Two patients could not have their para—aortic lymphadenectomy completed because of obesity. Eight of the patients had stage [8 disease and under— went laparoscopic staging prior to planned radical hysterectomy. Three of these patients had positive pelvic lymph nodes as as- certained by laparoscopy and were treated with radiation therapy. The remaining five patients had radical hysterectomies immediately following the laparoscopic lymphadenectomy. The average number of lymph nodes removed at laparoscopy was 3]. Two patients had additional lymph nodes found at the time of laparotomy. but no positive lymph nodes were found at laparotomy. Fowler et al. (5) performed laparoscopic lymphadenectomies on 12 patients with cervical cancer. All patients underwent laparotomy after laparoscopy to evaluate completeness of lymph node dissection. An average of 23.3 lymph nodes was removed by laparoscopy. and seven additional lymph nodes were re- moved by laparotomy. Two patients had positive lymph nodes identified by laparoscopy. The radical vaginal hysterectomy (Schauta) was performed in Europe as the treatment of choice for many years because it had '“All'i/iutimi of ail/Inuit: Gynecologic Oncology Division. University of Ari/.ona. Tucson. ('tirrmpmitlt'm'1' Io: Kenneth D. Hatch. M.l).. Department of Obstetrics and Gynecology. (‘ollege of Medicine. University of Arizona. 1501 North Campbell Ave. Tucson. AZ 85724. 71 a lower complication rate than the Wertheim operation. Em— phasis on pelvic lymph node dissection as a staging and prog— nostic finding led to the adoption of the Wertheim—Meigs abdominal operation as the treatment of choice in the United States and in most European centers. The cure rates and com plication rates using Schauta were reported by Massi et al. (6), These investigators reported on 356 patients who underwent the Schauta procedure and on 228 patients who underwent the Wertheim—Meigs procedure during the period from I968 through 1983. Before 1977. most patients had the Schauta operation. In 1977. Massi et al. began performing Wertheim— Meigs for most patients with stage IB ()r “A disease while reserving Schauta for high—risk. older. and obese patients. The 5—year survival slightly favored the Schauta operation (81% ver— sus 75%) but was not statistically significant. The severe corn— plication rate slightly favored the Schauta procedure (4.5% versus 5.3%); there were two deaths in the Wenheim—Meigs group and zero deaths in the Schauta group. Dargent (7) reported on a similar experience in 398 patients subjected to the Wertheim type III or IV procedure versus 210 women subjected to one of the two variants of the Schauta pro— cedure (Stoekel and Amreich). Five-year survival was 83% with the Wertheim procedure versus 84.4% with the Schauta var— iants. and severe complications were “three times higher" in the Wertheim group. including 24% mortality versus 0.6%. Dar— gent proposed that extraperitoneal pelvoscopy could be used to sample the pelvic lymph nodes and to direct the operation to be done by the Wertheim procedure if positive and by the Schauta procedure if negative. Laparoscopic pelvic lymph node dissection and sampling with the Schauta hysterectomy were combined by Dargent (8). Querleu (9), and Hatch et al. (10). Dargent (8) reported 3-year survival of 51 patients as 95.5% in lymph node-negative pa— tients and 80% in lymph node-positive patients with stage [B or IIA disease. Querleu (9) reported on eight patients and found less than 300-mL blood loss. 4.2 hospital clays. and less pain from the elimination of an abdominal or perinea] incision. Hatch et a1. (10) reported on 25 patients treated by laparoscopy and Schauta in 1995. and our report will update that experience for 37 patients. Nezhat et al. (II) and Canis et al. (12) have discussed laparo- scopic radical hysterectomy in case reports. No series are avail— able at this time. Patients and Methods During the period from September [993 through February 1996. 37 patients underwent laparoscopic pelvic lymph node dissection. and 30 underwent para- aortic lymph node sampling followed by a laparoscopically assisted radical vaginal hysterectomy. The uterine arteries and cardinal ligaments were stapled or clipped at the pelvic sidewall. and the ureter was dissected off the peritoneum to the ureteric tunnel. The rest of the procedure proceeded vaginally. Perineal in- cisions were not performed. A mini-laparotomy 4—7 cm in length was performed in six cases to palpate and inspect the lymph node dissection areas. Data were correlated with the length of experience using the analysis-of—variance test and linear regression: a P value less than .05 was considered significant. These 37 patients were not a consecutive series. During the same time frame. seven patients underwent an abdominal radical hysterectomy. Patients were of- fered the laparoscopy followed by Schauta hysterectomy: if they declined this procedure. they were treated with the standard abdominal hysterectomy. Three 72 patients were selected for abdominal radical hysterectomy because they had had prcvious surgical procedures or there was insufficient room in the vagina to allow for the vaginal approach. Results Patient characteristics are shown in Table l. The Quetelet index was calculated by the formula described by Khosla and Lowe ([3); this formula calculates the weight in kilograms divided by the height in centimeters squared. A Quetelet index of 24 is 20% over the ideal body weight, and a Quetelet index of 40 indicates morbid obesity. Lymph node counts were available for 33 patients with pelvic lymph nodes and 26 patients with para—aortic lymph node sampling (mean count of 35.5 and l 1.3. respectively). Four patients had positive pelvic lymph nodes. and none had positive para-aortic lymph nodes. Operative time. hospital stay. estimated blood loss. patient charges. and hospital costs are shown in Table 2 and Figs. 1-4. Operative time. hospital stay. hospital costs. and blood loss decreased with operator experience. The hospital stay range was 2-ll days. The patient who stayed l l days developed an ileus on postoperative day 3 and was diagnosed as having postopera- tive appendicitis. This condition was treated by laparoscopic ap- pendectomy on day 9. when the appendix was adherent to the pelvic lymph node dissection site with marked inflammatory reaction. The patient was discharged 2 days after the appendec- tomy without further complications. Despite the blood loss decrease throughout the 21/2 years of experience. the four trans- fusions were distributed equally throughout the time period. Table 3 lists the morbidity and compares this information with that given in a report from the senior author‘s former in— stitution (University of Alabama at Birmingham) (14). Two in- traoperative urinary bladder injuries and two ureterovaginal fistulas occurred in this study group. Both of these cystotomies and one of the ureterovaginal fistulas occurred in the first nine operations. Cystotomies were repaired vaginally with no se— quelae. Both patients with ureterovaginal fistula were treated as outpatients with ureteral stents left in place for 6 weeks with Table 1. Patient characteristics" Characteristic Mean (range) Patient age. y 42.3 (23-69) Patient weight. kg 68.8 (38-109) Patient height. in 1.63 (1.5—1.7) Quetelet index. kg/m: 26 ( ”13-378) *Of 37 patients. 25 (67%) had Quetelet indices >24 kg/mZ and 33 (89%) had stage IB disease. three had stage lA: disease. and one had stage ”A disease. Table 2. Operative characteristics Parameter Mean Range Operative time. min 226 l48-395 Estimated blood loss. mL 525 200-1400 Hospital stay. days 3 2—11 Estimated hospital costs. US. dollars 6449 3225-14 878 Patient charges. US. dollars [4 868 5980-33 553 Journal ot'thc National Cancer Institute Monographs No. 21. I996 .099 o... 09‘ o P2032 35,000 30000 - - . t 25000 - 20.000 15,000 0 a 10,000 ‘ 1 . ‘ n g l n fiHospnal cogts P<0. 4u ‘ 5,000 ' l v " P " l E E .1 u P<0.004 Fig. 1. Effect of surgeon‘s experience on hospital stay of patient. 1400 . . 7 1200 ~ 1000» ~ - ‘ .. 200 ' ° 0 5 1O 15 20 25 30 P=0 O2 Fig. 2. Effect of surgeon‘s experience on estimated blood loss in patient. 400 - a . l 350 ' . 300 . 250 ‘ 200 > 150 - 100 ' 50 v 0 i l - l l v P=0,04 Fig. 4. Effect of surgeon‘s experience on costs. Discussion The ability of laparoscopic pelvic lymph node dissection to adequately identify all pelvic lymph nodes and to remove a suf- ficient number of lymph nodes for adequate staging has been demonstrated by the studies of Dargent and Salvat (I). Querleu et al. (2). Childers et al. (3,4). and Fowler et al. (5). The American Medical Association Physician Current Procedural Terminology (CPT 95) includes laparoscopic total pelvic lymph- adenectomy and para—aortic lymph node sampling as CPT 56313. As yet. the long-term follow-up is not adequate to make a statement about the survival rates. However, the long-term fol- low—up of large European studies on patients treated with the Schauta operation when the lymph nodes were not dissected suggests that lymph node dissection may not play an important role in survival. The ability of the laparoscopically assisted Schauta to ade- quately treat the primary tumor is likely to be dependent on the surgeon and the patient. The large published series of Massi et al. (6) and Dargent (7) indicate that the Schauta procedure results in survival comparable to that achieved with the Wertheim—Meigs operation. Table 3. Morbidity comparison: Hatch et al. ([0) and Orr et al. ([4) Fig. 3. Effect of surgeon‘s experience on surgery time. complete resolution. Two patients with hypotonic urinary blad- ders required catheter drainage beyond 3 months. One of those patients continues to self-catheterize at 6 months. Journal of the National Cancer Institute Monographs No. 2 l. 1996 Hatch et al. Orr et al. Surgical morbidity (n = 37) (n = 3| l) Operative time. min 225 2l() Estimated blood loss, mL 525 1500 Blood transfusion. No. (‘7r) 4 (l 1) NA* Hospital stay. days 3 9.7 Urinary bladder injury. No. (‘7r) 2 (5) 3(1) Large—bowel injury. No. (‘71) l (3) 0 Ureterovaginal fistula. No. (C/() 2(5) 4 ( l) Small-bowel obstruction. No. (‘7r) 0 8(3) Febrile morbidity. No. (‘70) 9 (24) 105 (34) Postoperative appendicitis. No. (‘70) l (3) 0 Wound infection. No. (7,.) 0 ll (4) Pelvic cellulitis. No. (‘7r) 1 (3) 23(7) Death. No. (”/0 0 l (0.3) Pulmonary embolism, No. (%) 0 4 ( l .3) *NA = not available. 73 The morbidity of the patients reported here was compared with that of 31 1 patients who had had abdominal radical hysterectomy and pelvic lymph node dissection at the senior author‘s former institution (University of Alabama at Birming- ham). The mean blood loss and hospital stay were markedly dif- ferent but. because of low numbers of patients. were not statistically significant. The maximum blood loss was 1.4 L in the present report versus 8.0 L at the University of Alabama at Birmingham. The risk of pulmonary embolus. wound infection. small—bowel obstruction. and febrile morbidity appears to be in— creased. Operative times are similar. The longest operation in this series was 6.6 hours versus 7.6 hours at the University of Alabama at Birmingham. Another comparison was made to a large study in Toronto (I5). where all the data were collected prospectively (Table 4). In that study. the blood loss change was not significant. but transfusion rates were 35% versus 11%. The incidence of uri- nary bladder injury was 3% versus 5%. Neither the study by Orr et al. (14) nor the study by Covens et al. (15) addresses bladder atony. The issue of blood loss and transfusion is very important to patients since the identification of the human immunodeficiency virus. Benjamin et a1. (16) reported their experience with blood transfusions at Memorial Sloan-Kettering Cancer Center. New York. NY. In the period from 1980 to 1991. the transfusion rate was 91% compared with a transfusion rate of 44% in the period 1991 through 1993. These investigators also noted that the patients in the more recent period had a shorter mean postopera- tive length of hospital stay (11 days versus 14 days) and were discharged from the hospital with a significantly lower mean hemoglobin level (9.7 g/dL versus 1 1.4 g/dL). lf laparoscopic pelvic and para-aortic lymph node dissection plus Schauta radical hysterectomy is found to be comparable to the Wertheim—Meigs procedure. the stimulus to perform it lies in the perceived advantage of decreased hospital stay. rapid recovery, and a quick return to normal function. This study did not evaluate the amount of pain or length of recovery. The length of hospital stay serves as an indicator of this parameter. When compared with the studies by Orr et a1. ([4). Covens et al. Table 4. Morbidity comparison: Hatch et a1. (10) and Covens ct a1. (I5) Hatch et a1. Covens et a1. Operative morbidity (n = 37) (n = 405) Operative time. min 225 180 Estimated blood loss. mL 525 910 Blood transfusion. No. (’70) 4(11) 138 (35) Hospital stay, days 3 9" lntraoperative complications. No. ("/1) 3 (8) 22 (5) Urinary bladder injury. No. (%) 2(5) 10(3) Vessel laceration. No. (‘7r) 1 (3) 9 (2) Obturator nerve injury. No. (‘71) O 2 (0.5) Ureter injury. No. (‘71) 0 1 (0.2) Large-bowel injury. No. (‘70) 1 (3) 0 Ureterovaginal fistula. No. (%) 2 (5) l (0.2) Small—bowel obstruction. No. (‘70) 0 1 (0.2) Febrile morbidity. No. (‘70) 9 (24) 84 (21) Postoperative appendicitis. No. (%) 1 (3) 0 Wound infection. No. (%) 0 24 (6) Pelvic cellulitis. No. (‘71) 1(3) 13(3) 74 (I5). and Benjamin et a1. (16). the hospital stay of 3 days com— pares very favorably with the range of9.7—1 1 days. As managed care plans force earlier patient discharge. the length of hospital stay for abdominal operations will also decrease. and com- parisons will need to be tnade. The patient charge and hospital cost analysis dispels the con— cern that laparoscopic surgery is expensive (Fig. 4). The length of the operation. the investment in technical equipment. and the number ofdisposable instruments used are the main expense. As operating time decreases. and the use of disposable instruments and staplers is avoided. the cost decreases. In this study. the costs range from $3225 to $14 878 (mean. $6449). These costs are acceptable. Conclusions Laparoscopic pelvic lymph node dissection and para—aortic lymph node sampling can be performed with adequate lymph node counts and low morbidity. Laparoscopically assisted Schauta hysterectomy allows a shorter hospital stay than the Wertheim procedure. with less loss of blood and lower risk of transfusion. The morbidity from the Schauta operation is higher early in the surgeon‘s experience and decreases over time. The widespread adoption of laparoscopic lymph node dissection. para-aortic lymph node sampling. and Schauta radical hysterec- tomy will depend on acquisition of skills and experience. Train— ing programs that provide consistent experience in laparoscopic skills are needed. References (I) Dargent D. Salvat .1. Envahissencnt ganglionnaire pclvien: place de la pel- viscopie retroperitoneale. Paris: McGraw-Hill. 1989. Querleu D. Leblanc E. Castelain B. Laparoscopic pelvic lymphadenectomy in the staging ofearly carcinoma of the cervix. Am J Obstet Gynecol 1991: 164:579-81. Childers JM. Hatch K. Surwit EA. The role of laparoscopic lymphadenec- tomy in the management of cervical carcinoma. Gynecol ()ncol 1992:47: 38-43. Childers JM, Hatch KD. Tran AN. Surwit EA. Laparoscopic para-aortic lymphadenectomy in gynecologic malignancies. Obstet Gynecol 1993: 82:741—7. (5) Fowler JM. Carter JR. Carlson JW. Maslonkowski R. Byers LJ. Carson LF. et a1. Lymph node yield from laparoscopic lymphadenectomy in cervi- cal cancer: a comparative study. Gynecol Oncol 1993;51:187-92. (6) Massi G. Savino L. Susini T. Schauta-Amreich vaginal hysterectomy and Wertheim-Meigs abdominal hysterectomy in the treatment of cervical can- cer: a retrospective analysis Iscc comment citation in Medline]. Am J Obstet Gynecol 1993:168(3 Pt 1):928-34. (7) Dargent D. A new future for Schauta's operation through a presurgical retroperitoneal pelviscopy. EurJ Gynaecol Oncol 1987;82292—6. (8) Dargent D. Laparoscopic surgery in gynecologic cancer. Curr Opin Obstet Gynecol 1993:52294—300. (9) Querleu D. Laparoscopically assisted radical vaginal hysterectomy. Gynecol Oncol 199315 1 :248-54. ([0) Hatch KD. Hallum AV 3rd. Surwit EA. Childers .1M. The role of laparos- copy in gynecologic oncology. Cancer 1995;76(10Supp1):21 13-6. (II) Nezhat C. Nexhat F. Burrell MO. Benigno B. Welander CE. Laparoscopic radical hysterectomy with paraaortic and pelvic node dissection [letter]. Am J Obstet Gynecol 1994;170:699. (12) Canis M. Mage G. Wattiez A. Puly J. Chapron C. Bruiat M. Vaginally as— sisted laparoscopic radical hysterectomy. J Gynecol Surg 199218: 103-4. (I3) Khosla l. Lowe CR. lndices of obesity derived from body weight and height. BrJ Prev Soc Med 1967;21:122-8. (I4) Orr JW Jr. Shingleton HM. Hatch KD. Mann W] Jr. Austin JM Jr. Soong SJ. Correlation of perioperative morbidity and conization to radical hys- terectomy interval. Obstet Gynecol 1981591726341. (7 (3 (4 Journal of the National Cancer Institute Monographs No. 21. 1996 (I5) Covens A, Rosen B, Gibbons A, Osborne R. Murphy J, DePetrillo A, et al. of transfusion-related human immunodeficiency virus infection. Obstet Differences in the morbidity of radical hysterectomy between gynecologi- Gynecol 1994;84:974-8‘ cal oncologists. Gynecol Oncol 1993;51:39-45. (I6) Benjamin 1. Barakat RR. Cunin JP. Jones WB. Lewis JL Jr. Hoskins WJ. Blood transfusion for radical hysterectomy before and after the discovery Journal of the National Cancer Institute Monographs No. 21. I996 75 Adjuvant Therapy After Primary Surgery for Stage I-IIA Carcinoma of the Cervix Gillian M. Thomas* Radical hysterectomy and bilateral pelvic lymph node dis- section is commonly used as a primary management option for treatment of stage lB/IIA carcinoma of the cervix. Over- all cure rates approach 85%. However, a spectrum of relapse risk exists, depending on the presence or absence of primary tumor and nodal-related prognostic factors. Known factors include number and location of lymph nodes; size of primary, deep invasion in the cervix; capillary lymphatic space involvement; occult parametrial involvement; and positive or close surgical margins. Biologic determinants have yet to be identified. No systematic analysis has ex- amined various combinations of prognostic factors to precisely define associated levels of risk and to predict the sites of relapse. Decreased local control and survival rates in some high-risk subgroups, usually those with nodal positivity, has led to the exploration of adjuvant therapies. Compiled data from retrospective series have defined the overall patterns of failure. Seventy-two percent of those relapsing have a component of pelvic failure, while 42% ex- perience relapse in the pelvis alone. F ifty-eight percent have a component of distant failure but only 28% have distant disease alone. Adjuvant treatment options include pelvic radiotherapy, extended-field radiotherapy, chemoradio- therapy, and chemotherapy. Trials of adjuvant chemo- therapy are too few to evaluate the use of available agents. Pelvic radiotherapy has been shown to reduce the relapse risk when surgical margins are close or positive. It also reduces the risk of pelvic relapse and improves the relapse- free interval but has no apparent impact on overall survival in the groups that have been selected for treatment. The ap- parent lack of benefit may relate to the choice of patients with nodal involvement who, despite high risk of pelvic failure, most likely have a predominant pattern of distant failure. Maximization of the survival benefit of pelvic radiotherapy requires the identification and treatment of the subgroup with a predominant pattern of pelvic failure, such as that examined in Gynecologic Oncology Group protocol 92. These may be patients with primary tumor-related, high- risk factors but negative nodes. Extended-field irradiation for microscopically involved para-aortic nodes provides a cure in 25%-40% of the patients. Further studies of prog- nostic factors and their relationship to sites of failure after surgery are necessary to define the benefits of currently available adjuvant therapies with respect to local control, survival, and quality of life, and also to direct future studies. New, effective systemic agents are required for those at high risk of developing distant disease. [Monogr Natl Cancer Inst 1996;21:77-83] Journal ot'thc National Cancer Institute Monographs No. 2]. 1996 Radical hysterectomy and bilateral pelvic lymph node dissec- tion (RH and PLND) and radical pelvic irradiation are accepted as primary management options for patients with carcinoma of the cervix that is confined either to the cervix alone (stage IE) or involving the upper two thirds of the vagina (stage IlA). Primary surgical therapy is highly successful as the sole management for some patient groups. Overall relapse rates are only approximate— ly |5%. However, it is recognized that the spectrum of patients within stage lB/llA varies from those at low risk of relapse fol— lowing surgery and for whom no adjuvant therapy has been shown to improve either local control or survival to those with increasing risk of relapse. Because it is recognized that local control and survival rates are unsatisfactory in patients at high risk for recurrence when treated with surgery alone. attention has been directed toward offering selected patients some form of adjuvant therapy. either locoregional (usually radiation) or sys- temic (usually chemotherapy) in attempts to improve loco- regional control rates and survival. The previous criterion used to select those for whom adjuvant pelvic irradiation was indi- cated was usually pathologic evidence of pelvic lymph node in— volvement. However. an increasing body of evidence suggests that certain attributes of the primary tumor. independent of nodal status. also may predict for recurrence risk after surgery alone. Since the risk of relapse relates to both nodal— and primary tumor—associated attributes, the possibility that anatomic sites of relapse may vary with specific risk factors must be considered. Furthermore. rational therapeutic strategies for improving out- comes after radical surgery should be tailored to both the level of risk of recurrence and the anticipated pattern of failure. Cur- rent therapies need to be evaluated in the context of knowledge of both risk and predicted patterns of relapse. Prognostic Factors for Relapse The strongest single predictor of outcome remains nodal in- volvement. In a review of nearly 1500 patients by van Bommel et al. (I). the 5—year survival rate for those without nodal invol- vement was 90% compared with only 46% for those with nodal involvement. Within the 18% of the patients who had positive nodes. additional prognostic factors could be identified. While only about 5% of the patients have involvement of common *At'fi/iurimm of'uutlmr: Departments of Radiation Oncology and Obstetrics and Gynecology. University of Toronto. ON. Canada. and Division of Radiation Oncology. Toronto-Sunnybrook Regional Cancer Centre. Toronto. Cun‘m-pmulom't' m.- Gillian M. Thomas. B.Sc.. M.D.. F.R.C.P.C.. Division of Radiation Oncology. Toronto-Sunnybrook Regional Cancer Centre. 2075 Bayvicw Ave. Toronto. ON. M4N 3M5. Canada. 77 iliac or para—aortic lymph nodes. their survival is significantly worse than those with pelvic nodal involvement. Similarly. worse outcomes were reported as the number of involved nodes increased to three or more or the extent of nodal involvement in— creased from microscopic to macroscopic (2-5). Retrospective. single—center analyses and a large. multicenter prospective surgical pathologic database have established in some detail the multiple risk factors associated with failure after RH and PLND (/6). Those factors include increasing size of the primary. deep invasion of the cervical stroma (either abso— lute or relative). capillary/lymphatic space (CLS) involvement. occult parametrial involvement. positive (or close) surgical mar— gins. and nodal involvement. In univariate analyses from multi- ple series. the 5—year survival of patients with tumors smaller than 2-4 cm in diameter was 91% versus 64% in those with tumors larger than 2—4 cm. Survival was 88% in those without CLS involvement compared with only 64% in those with CLS positivity, Another univariate analysis showed a progression- free survival of 98% for those with inner third involvement of the cervical stroma versus 81% for middle and 63% for outer third involvement. In that same analysis (6). the 5-year survival rate for those with tumors with an absolute depth of invasion of 5-7 mm was 86%. with a progressive decline to 54% for those with tumors invading more than 11 mm in depth. In collected series. the survival with parametrial involvement was 74% com— pared with 94% for those without parametrial involvement (Table I) (I). The significance of cell type. that is. adenocar— cinoma or adenosquamous versus squamous cell carcinoma. is unclear. A number of reports suggest a less favorable outcome for the nonsquamous cell variants while others have not (1). Possible molecular and "bio—functional" markers. such as DNA content. apoptosis. oncogenes. and hypoxia. are being inves- tigated for their predictive value in cervical cancers. but none of the data are convincing enough to be clinically applicable. While multiple analyses have revealed the prognostic sig— nificance of the individual factors discussed. many of these fac- tors are interrelated. No analysis has systematically examined Table 1. Outcome according to various prognostic factors in patients with stages lB/llA disease after radical hysterectomy and bilateral pe1\ ic lymph node dissection (I) Factor No. 5-year survival. % Lymph nodes Negative 1233 90 Positive 227 46 Tumor si/c. cm <2—-1 203 91 >2-4 101) 64 Stromal invasion Superficial 331 90 Deep (>10 mm or >5()% ) 91 7t) Capillary/lymphatic space Negative 572 88 Positive 545 64 Grade 1 and 2 590 82 3 224 68 78 the various specific combinations of all known independent prognostic factors to more precisely define associated levels of risk. It is clear. however. that local control and survival rates diminish as the number of risk factors accumulate. This knowledge has led to the use of adjuvant therapies after primary surgery for selected patients with early stage cervical cancer. Sites of Relapse Retrospective analyses (2.4.7-11) have documented the pat— terns of relapse after RH and PLND for stage lB/llA cervical cancer (Fig. 1). Seventy-two percent of those relapsing have a component of pelvic failure. while 42%) have recurrence in the pelvis alone. Although 58% have some component of distant failure. only 28% have disease at distant sites without evidence of pelvic disease. Further knowledge of relapse patterns as- sociated with defined high—risk subgroups would provide rational guidelines for clinical trials of adjuvant therapy specifically tar— geting local. locoregional. and/or systemic disease. Adjuvant Therapy Currently available management options for patients at in- creased risk of relapse after RH and PLND include observation. pelvic radiotherapy. extended-field radiotherapy. chemoradio— therapy. and adjuvant chemotherapy. The optimal choice for postoperative management of the patient with stage [B or 11A carcinoma of the cervix is dependent on the specific risks that the patient carries for failure. the specific pattern of failure ex- pected. depending on the tumor— and nodal—related prognostic factors. the sites of disease that any postoperative therapy would target. and the efficacy of the therapy in reducing risk of relapse in that site. Both Pelvis Distant Pelvis i Distant (A+B): 72% 28% _ n = 323 (“'7‘“) Distant Only (C): 42% Distant i Pelvis (B+C): 58% Pelvis Only (A): . _ } n=161”""9 1” Fig. 1. Sites of relapse after radical hysterectomy and pelvic lymph node dissec— tion in stages lB/llA. Journal oflhe National Cancer Institute Monographs No. 21. 1996 Since only 28% of patients appear to fail with distant metas— tases alone (Fig. 1), even ifa systemic therapy were 100% effec- tive in eradicating metastatic disease, it could potentially only improve survival rates by approximately 25%. If it were highly effective, one could infer that it should also decrease the risk of pelvic failure. Unfortunately, highly effective systemic chemo— therapeutic agents for cervical carcinoma have not been iden- tified, although relatively high short—term response rates are seen. Cisplatin is the most effective single-agent chemothera— peutic drug, with response rates ranging between 20% and 45% (12.13). Unfortunately. the median duration of response is generally very short, less than 5 months for patients with ad- vanced disease. The short duration of response indicates that available agents are ineffective in completely eradicating clonogenic cells. either because of large tumor burden. tumor resistance. and/0r rapid mutation to resistance in residual cells. Other agents tested in phase II trials, such as ifosfamide (l4). methotrexate (l5), vincrinstine (15), and bleomycin ([5). have response rates in the range of 15%. Response rates up to 78% with combinations of these agents in advanced disease have been reported and are significantly higher than with cisplatin alone (16-19). However, no randomized trial in advanced or recurrent disease has demonstrated a survival advantage with combination chemotherapy compared with cisplatin alone. A paucity of studies has examined the role of adjuvant chemo— therapy after hysterectomy for early cervical cancer. A phase II study of postoperative adjuvant cisplatin and radiation therapy showed relapse in four of l 1 patients (20) compared with 18 of 27 unmatched patients selected for adjuvant radiation therapy only. An observational study from Memorial Sloan-Kettering Cancer Center used postoperative adjuvant cisplatin and bleo- mycin before pelvic irradiation in 44 high—risk patients. Risk was defined as the presence of positive nodes (2]). primary size greater than 4 cm (22), CLS positivity (2]). high grade (II), or positive margins or parametria (5). The 5-year relapse—free sur— vival was 71%. but no conclusions can be drawn about how much of the observed survival was attributable to the adjuvant chemotherapy (22). There is only one published randomized study addressing the possible role of adjuvant chemotherapy. The results were compared with adjuvant radiotherapy alone versus that of three courses of adjuvant cisplatin, vinblastine, and bleomycin followed by radiotherapy (23). While the num- ber in this trial was small (71), there was no apparent decrease in distant metastases rate or survival improvement after chemotherapy. Thus, while distant metastases are recognized as a significant problem for patients at high risk for recurrence after surgery, there is a dearth of data pertinent to understanding the use of postoperative adjuvant chemotherapy. It is doubtful, however, that substantial gains would be realized with currently available chemotherapeutic agents. Pelvic irradiation has been the most commonly used adjuvant therapy. Given that more than 70% of the patients who had dis— ease recurrence have some component of uncontrolled pelvic disease (Fig. l) and the fact that pelvic irradiation is highly ef- fective even for control of bulky cervical cancer (2]). it is ra- tional that adjuvant pelvic irradiation has been used in an attempt to improve survival through improving pelvic control. Radiation therapy in doses of 75—85 Gy controls up to 85% of Journal of the National Cancer Institute Monographs No. 21, 1996 cervical cancers measuring up to 6 cm in diameter (24). In the postoperative adjuvant setting. residual tumor burden should be microscopic in extent; radiation dose control data suggest that doses of 45—50 Gy postoperatively are approximately 90% ef— fective in eradicating residual clonogenic cells within the pelvis. Several studies support the assertion that adjuvant pelvic radiotherapy should improve pelvic control: 1) Pelvic radio— therapy appears to effectively reduce the relapse risk when sur- gical margins are close or positive. In the report of Guttmann (25), pelvic relapse was prevented in 21 of 22 patients with positive or close surgical margins. 2) Isolated pelvic relapses after surgery are salvaged with radiation therapy alone or chemoradiation therapy in 15%-45% of the cases (26—29). 3) Disease in lymph nodes would appear to be sensitive to exter— nal-beam pelvic radiotherapy. The frequency of lymph node positivity encountered at surgery is reduced by the use of pre- operative external—beam pelvic radiotherapy (30). Similarly. definitive pelvic radiotherapy for more advanced cervical cancer provides high levels of pelvic control, even in those with known pelvic lymph node metastases. 4) Most reports of postoperative pelvic radiotherapy confirm a reduction in pelvic relapses (4,10,31-34). No randomized trials have examined the use of postoperative pelvic radiotherapy, but Kinney et a1. (34) conducted a study using a matched—pair analysis. One hundred twenty patients with and without adjuvant pelvic irradiation were matched for known prognostic factors, including tumor stage and size and number and location of nodes (Table 2). While overall recurrence num- bers were similar in the two groups. the authors concluded that adjuvant pelvic irradiation reduced pelvic failure from 67% to 27%. The median time to recurrence was 2.1 years in patients receiving radiation therapy compared with 1.4 years if radiation therapy was not given. Despite convincing evidence from this study and retrospec- tive series (4,10,31,33,.5’4) demonstrating a reduction in pelvic recurrence rates when adjuvant pelvic irradiation is employed, there is no evidence that it improves survival rates. Examining possible explanations for the apparent lack of survival benefit may contribute to a better understanding of the circumstances in which adjuvant pelvic irradiation might lead to survival benefits. Typically. adjuvant radiation therapy has been directed to those with positive pelvic lymph nodes. since lymph node positivity is Table 2. Adjuvant radiation therapy in node-positive stage lB/IIA cancer of the cervix after radical hysterectomy and bilateral pelvic lymph node dissection (34)* Sites of recurrence Surgery + Surgery radiation therapy Site No. of patients % No. of patients We Pelvis I4 67 6 27 Distant 6 29 I3 59 Pelvis + distant 1 S . 14 Total 21 22 *Pairs with and without radiation therapy matched by stage. size. number. and location of nodes (n: l 20). 79 the strongest single predictor of risk for recurrence after surgery. Although only 15% of patients with stage IB disease have posi— tive lymph nodes. approximately 45% of these patients are cured compared with 90% of the group with no pelvic nodal in— volvement. It is tempting. therefore. to accept that nodal positivity best defines the high-risk group and address adjuvant therapies to this group. The lack of demonstrated survival benefit from pelvic irradiation, however. may be related to the selection primarily of those with nodal positivity for adjuvant pelvic treatment. Approximately 60% of the patients who ex— perience disease recurrence after RH and PLND have disease beyond the pelvis and could not be expected to have a survival benefit from radiation therapy directed to the pelvis only. If ad— juvant pelvic radiation therapy were 100% effective. only two of five patients in whom postoperative therapy controlled pelvic disease and prevented recurrence would be cured. Thus. if 15% overall of surgically treated patients with stage [8 disease would be expected to relapse, pelvic radiotherapy could prevent. at most, six of the 15 relapses, even if it were 100% effective. No reported study has had the power to detect such a small percent- age of improvement. The potential curative impact of adjuvant pelvic radiotherapy on the overall group of patients with stage IB/lIA disease is small. To maximize the survival benefit of adjuvant pelvic irradia- tion. identification of patients with cancer most likely to recur in the pelvis alone is required. Because nodal involvement is the most significant factor worsening prognosis (6). little attention has been paid to determining whether there are unfavorable tumor-associated factors that add independent information regarding the risk of recurrence. In absolute terms. in fact. an equal number of deaths occur in the node-negative group as in the node-positive group (Fig. 2) (Table 3) (2, 7-9.35,36). It is im— portant to determine which combination of primary tumor-re- lated prognostic characteristics account for most of the relapses in the node—negative group. To rationalize choosing locoregional, systemic adjuvant treatment, or a combination thereof, it would be important to confirm or refute the assurnp— tion of a model (2.37) that node-negative patients relapse more THE PROBLEM: 100 patients, Stages IB/IIA, after Radical Hysterectomy & Node Dissection 1 5 Node Positive 85 Node Negative 45% cured 90% cured V v 8 relapse 8 relapse Fig. 2. Typical frequencies of nodal metastases. cures. and absolute numbers of relapsing patients after radical hysterectomy and pelvic lymphadcnectomy in stages lB/llA cervical cancer. 80 Table 3. Number of recurrences by nodal status Investigators (Ref. No.) N— N+ N7 /N+ Excess Ni Fuller et a1. (2) 56 29 1,93 Hogan et a1. (8) 3-1 18 1.89 Burke et al. (35) 18 13 1.39 N+ = N— Boyce et al, (9) 1 1 9 1.22 Figge and Tamimi (36) It) 12 (1.8} Excess N+. 39 ()2 0.63 Martimbeau et a1, (7)* Total 168 143 1.17 *All patients received central pelvic irradiation (preoperative inlracavitar‘y therapy). N— = node»nega1ive group; N+ : node—positive group. commonly at locoregional sites. whereas the node-positive group exhibits a greater proportion of distant failures. It is pos— sible that a combination of tumor—prognostic factors would define a subgroup in which there is a higher risk of central pel— vic rather than sidewall or distant failure and in which adjuvant pelvic radiation therapy would improve survival. It is this group in which a randomized trial of adjuvant pelvic radiotherapy should be undertaken. While further study is necessary to char- acterize the node—negative. high-risk patients. a recent study by the Gynecologic Oncology Group (GOG) (protocol 92) does ad— dress the issue of the benefit of adjuvant pelvic radiotherapy versus observation in the high—risk. node-negative group. Their definition of high risk is based on tumor size. capillary/lym— phatic space invasion. and depth of invasion. The high-risk group. by their definition, had. in a retrospective review. a greater than 30% risk of relapse within 2 years of surgery. This study (37) is unique in that it is built on the premise that those with nodal positivity. although having a statistically significant- ly higher overall risk of failure, probably have a higher risk of distant failure and are, therefore, least likely to have a survival benefit from pelvic radiation. The group in GOG protocol 92. however. with primary tumor—related. poor—prognostic factors. should have a higher risk of pelvic-alone failure and potentially a greater survival benefit from pelvic irradiation. If patients ac— crued to this study are skewed toward lower risk features. how- ever. the power of the study may be insufficient to detect any survival benefit from pelvic radiotherapy. The end points used to evaluate the benefit of postoperative adjuvant therapy have been mainly the survival rate and the median survival (4). Little attention had been paid to evaluating whether adjuvant therapy improves local control and. if so. whether this has contributed to patients’ quality of life. Future trials evaluating adjuvant therapy should employ these latter measures in addition to survival end points. It is important that locoregional therapies not be inappropriately discarded, even if they do not have an impact on survival. since contributions to quality of life may be substantial. Even in those patients unlikely to have a survival advantage because they are at higher risk of distant disease. the risk of pelvic failure without additional radiation therapy is high (Fig. 1). Almost all reports confirm statistically significant reductions in pelvic failure and a prolonged relapse—free interval Journal of the National Cancer Institute Monographs No. 21, 1996 (2,4,26,34,38) in both node-positive and node—negative patients (Table 1). Complications It is important to select patients for adjuvant pelvic radiation carefully. since its addition to surgery increases complication rates. The risk of recurrence predicted by unfavorable tumor characteristics must be high enough to justify the morbidity of postoperative therapy. Serious complications occur in only 3%— 5% of the patients treated with surgery alone. but this increases to l()%-l3% when adjuvant radiation therapy is employed (38). The degree of risk is radiation dose related. being statistically significantly higher than 55 Gy (2.48.11.25.32). A formal study of the quality of life associated with delivering or withholding pelvic irradiation in those at risk for pelvic recurrence would clarify the associated therapeutic ratios. No formal studies of the addition of concurrent chemothera- peutic agents to radiation therapy in the postoperative adjuvant setting have been performed. Convincing theoretical arguments for the use of concurrent chemotherapy with radiation therapy have launched a series of controlled trials where concurrent chemotherapy and radiation therapy are used as primary therapy in more advanced cervical cancer. If the results of these studies from the GOG. National Cancer Institute of Canada. and the Toronto Group show evidence for better pelvic control with concurrent chemoradiotherapy and complication rates are ac— ceptable. it should also be examined in the postoperative ad- juvant setting. Extended-Field Irradiation Some patients with microscopic involvement of the para—aor— tic nodes metastatic from cervical cancer are cured with ad- juvant extended-field irradiation. The contribution of this treatment to improved cure rates increases in situations where pelvic disease is also likely to be controlled. for example. those with small—volume stage IB/IIA disease. Numerous series (39- 42) have reported approximate 5—year survival rates of 25%— 40% with the use of extended—field irradiation for those with involved para-aortic nodes. While those with proven para-aortic nodal metastases have an increased risk for distant relapse. we can infer from the observed cure rates that a substantial propor— tion of patients have disease confined to the pelvis and para—aor— tic nodes. While these data support extended-field adjuvant irradiation for those with microscopic residual disease in the para—aortic nodes. its role in other groups of patients is undeter- mined. It has become common practice to offer postoperative extended-field irradiation additionally to those whose highest identified level of involvement is in the common iliac nodes or even pelvic nodes where para-aortic node biopsies have been negative. This practice has been justified by the assumption that undetected microscopic disease could be present in the next higher echelon of draining lymph nodes. even if undetected by biopsy. The incremental benefit of this practice is uncertain and must be weighed against associated increased complication rates from extended-field irradiation compared with pelvic irradia- tion. On theoretical grounds. it is expected that approximately Journal of the National Cancer Institute Monographs No. 2 l. 19% one half of the patients with proven pelvic nodes will have para— aortic nodal metastases. If the sensitivity of surgical staging is 90%. 45 of 50 patients with disease would have had it detected at surgery and irradiated. Ofthe remaining five patients with un- detected microscopic disease. probably a maximum of two would be cured with prophylactic treatment of the nodes. In this model. 98 of 100 patients given para-aortic nodal irradiation despite biopsy-proven. negative para-aortic nodes would not benefit frotn this practice. The benefit may be slightly higher for those with high common iliac nodes. where a slightly higher proportion may be expected to have para-aortic nodal disease. Few series report survival after para-aortic nodal irradiation in those with macroscopic para—aortic nodal disease and, indeed. this therapy could not truly be considered adjuvant. The limita- tions to cure in this setting are related to the relative ineffective- ness of a tolerable radiation dose in eradicating macroscopic disease and the high risk for distant metastases that accompanies macroscopic nodal disease in any site. Without controlled data. proponents have been emerging for a practice of initial resection of macroscopic nodal disease followed by adjuvant irradiation (40.43). No data exist to determine whether this will be of addi— tional patient benefit. although occasional long—term survival times have been reported. Summary Overall relapse rates following radical hysterectomy and pel- vic lymph node dissection for stage IB/IIA cervical cancer are approximately 15%. a rate too low to justify the routine use of adjuvant therapy in all patients. Multiple independent prognostic factors have been identified that are associated with an increas— ing risk of relapse following surgery alone. There is a need. however. to better characterize the high-risk group and to iden- tify the specifically associated patterns of failure by constella- tions of risk factors. The exploration of adjuvant therapies. including pelvic irradiation. extended—field irradiation, and radiation therapy in combination with systemic therapy. should consider the patterns of failure. A proposed schema for future studies of adjuvant therapy for patients with stage IB/IIA resected cervical cancer would be to take those at increased risk for distant disease. for example. those with greater than three in- volved pelvic nodes or macroscopic nodal involvement or com- mon iliac involvement, and explore the combined use of pelvic radiation therapy for optimal pelvic control and new systemic agents as they become available. For those patients considered to be at high risk for pelvic failure by constellations of primary tumor factors. it is reasonable to examine the use of adjuvant pelvic radiation therapy alone. Patients with both favorable tumor factors and negative nodes can be observed after surgery. since their risk of relapse is very low. Adjuvant pelvic irradiation clearly improves pelvic control and disease—free interval for those at high risk for pelvic recur— rence. but its impact on survival and quality of life has not been adequately evaluated. Extended-field irradiation may cure patients with proven microscopic para-aortic nodal disease. Sys- temic therapy with currently available agents has no known benefit. but few trials have addressed the possible benefit. 81 While this paper has addressed issues concerning adjuvant therapy following radical surgery for stage 1B/11A carcinoma of the cervix. it has not addressed whether patients at high risk are best served with primary radiation therapy rather than primary surgery with or without selected adjuvant treatment. Given the risk of pelvic relapse in selected high-risk patients after primary surgery compared with primary in‘adiation. making adjuvant postoperative radiation therapy advisable. the use of both treat- ment modalities rather than radiation therapy alone may not be justified (44). References (I) '0 (3 (4 (5 (6 (I0) (l3) 2 «L ([5) (l6) (/7 van Bominel PF. van Lindert AC. Kock HC. Leers WH. Neijt JP. A review of prognostic factors in early-stage carcinoma of the cervix (FIGO 1 B and 11 A) and implications for treatment strategy EurJ Obstet Gynecol Reprod Biol 1987;26:69-84. Fuller AF Jr. Elliott N. Kosloff C. Lewis JL Jr. Lymph node metastases from carcinoma of the cervix. stages 18 and 11A: implications for prog- nosis and treatment. Gynecol Oncol 198211116574. Kneale B. Pelvic lymph node metastases in carcinoma of the cervix. Aust N Z J ()bstet Gynaecol 1967;10:167-72. ls pelvic radiation beneficial in the postoperative management of stage 1b squamous cell carcinoma of the cervix with pelvic node metastasis treated by radical hysterectomy and pelvic lymphadenectomy‘.’ A report from the Presidential Panel at the 1979 Annual Meeting of the Society of Gyne- cologic Oncologists. Gynecol ()ncol 1980110: 105-10. Bttrghardt E. Pickel H. Haas J. Lahousen M. Prognostic factors and opera— tive treatment of stages 18 and 118 cervical cancer. Am J Obstet Gynecol |987'.156:988»96. Zaino RJ. Ward S. Delgado G. Bundy B. Gore H. Fetter G. et al. Histo— pathologic predictors of the behavior of surgically treated stage IB squamous cell carcinoma of the cervix. A Gynecologic Oncology Group study. Cancer 1992169: 1750-8. Martimheau PW, Kjorstad K. lversen T. Stage IB carcinoma of the cervix. the Norwegian Radium Hospital. 11. Results when pelvic nodes are in- volved. Obstet Gynecol 1982;60:215—8. Hogan WM. Littman P. Griner 1. Miller CL. Mikuta JJ. Results of radia~ tion therapy given after radical hysterectomy. Cancer 1982;49:1278-85. Boyce J. Fruchter RG. Nicastri AD. Atnbiavagar PC. Reinis MS. Nelson JH Jr. Prognostic factors in stage 1 carcinoma of the cervix. Gynecol ()ncol 198|:|2(2 Pt I ):154~65. Russell AH. Tong DY. Figge DC. Tamimi HK. Greer BE. Elder SJ. Adv juvant postoperative pelvic radiation for carcinoma of the uterine cervix: pattern of cancer recurrence in patients undergoing elective radiation fol- lowing radical hysterectomy and pelvic lymphadenectomy. Int J Radiat Oncol Biol Phys 19841103114. Jobson VW. Girtanner RE. Averette HE. Therapy and survival of early in- vasive carcinoma of the cervix uteri with metastases to the pelvic nodes. Surg Gynecol Obstet 1980;151:27—9. Bonotni P. Blessing JA. Stehman FB. DiSaia PJ. Walton L. Major FJ. Ratt— domized trial of three cisplatin dose schedules in squamous-cell carcinoma of the cervix: a Gynecologic Oncology Group study. J Clin Oncol 1985;}: 1079-85. Thigpen T. Shingleton H. Homesley H. LaGasse L. Blessing J. “'5. Dichlorodiammineplatinum(11) in the treatment of gynecologic malignan— cies: phase 11 trials by the Gynecologic Oncology Group. Cancer Treat Rep 1979;63:1549-55. Sutton GP. Blessing JA. McGuire WP. Patton T. Look KY. Phase 11 trial of ifosfamide and mesna in patients with advanced or recurrent squamous carcinoma of the cervix who had never received chemotherapy: a Gyne— cologic Oncology Group study. Am J Obstet Gynecol 1993;168:8057. Wasserman TH. Carter SK. The integration of chemotherapy into com- bined modality treatment of solid tumors. V111. Cervical cancer. Cancer Treat Rev |977:4:25—-l6. Buxton EJ. Meanwell CA. Hilton C. Mould JJ. Spooner D. Chetiyawar— dana A. et al. Combination bleomycin. ifosfamide. and cisplatin chemo- therapy in cervical cancer. J Natl Cancer Inst 1989;81:359—61. Friedlander M. Kaye SB. Sullivan A. Atkinson K. Elliott P. Coppleson M. et al. Cervical carcinoma: a drugresponsive tumor—experience with com- bined cisplatin. vinblastine. and bleomycin therapy. Gynecol Oncol 1983: l6:275»81. (l8)Vogl SE. Moukhtar M. Calanog A. Greenwald EH. Kaplan BH. (I9) (24) (3.?) (34) (35) (36) (37) (38) (39) (40) (4/) Chemotherapy for advanced cervical cancer with bleomycin. vincristine. milotnycin C. and (ix—diamminedichloroplatinum(ll) BOMP. Cancer Treat Rep 1980;64:1005-7 Vermorken JB. Colombo N. van der Burg M. De Oliveira C. Kobierska A. Guastalla JP. et a1. Vindesine (E). bleomycin (B). mitomycin C (M) and cisplatin (P) BEMP versus cisplatin in disseminated squamous cell car— cinoma of the uterine cervix (SCUC): an EORTC phase 111 trial. Cairns. Australia: lnt Gynecol Cancer Soc. 1991. Pedttlla F. Centttrioni MG. Foglia G. Ferrari l. ()rsatti M. Vitale V. et al. Treatment of FIGO stage 1b cervical carcinoma with nodal involvement. EurJ Gynaecol Oncol 1994;15:59-64. Perez CA. Grigsby PW. Nene SM. Camel HM. Galakatos A. Koa MS. et al. Effect of tumor size on the prognosis of carcinoma of the uterine cervix treated with irradiation alone. Cancer 1992;69:2796-806 Hakes T. Nori D. Lewis JL. Adjuvant cisplatin/bleomycin (C/B) for high risk stage lB/llA cervix carcinoma patients~a pilot study. Abstract 455. In: chenthal B. editor. Twenty»third Annual Meeting of ASCO. Atlanta (GA): Grune & Stratton. Inc. 1987: 116. Tattersall MH. Ramirez C. Coppleson M. A randomized trial of adjuvant chemotherapy after radical hysterectomy in stage lB-llA cervical cancer patients with pelvic lymph node metastases. Gynecol ()ncol 1992;46:176— 81. Eifel PJ. Morris M. Wharton JT. Oswald MJ. The influence of tumor size and morphology on the outcome of patients with FlGO stage IB squamous cell carcinoma of the uterine cervix. Int J Radiat ()ncol Biol Phys 199-1; 29:9-16. Guttmann R. Significance of post-operative irradiation in carcinoma of the cervix: a ten year survey. Ath Roentgenol 1970: 108: 102-8. Fuller AF Jr. Elliott N. Kosloff C. Hoskins WJ. Lewis JL Jr. Determinants of increased risk for recurrence in patients undergoing radical hysterec- tomy for stage 18 and 11A carcinoma of the cervix. Gynecol Oncol 1989; 33234-9. Mttnnell EW. Bonney WA. Critical points of failure in the therapy of can- cer ofthe cervix. Am J Obstet Gynecol 1961 :8 | 15214. Deutsch M. Parsons JA. Radiotherapy for carcinoma of the cervix recur— rent after surgery. Cancer 197434220515. Thomas GM. Detnbo AJ. Black B. Bean HA. Beale FA. Pringle JR. et al. Concurrent radiation and chemotherapy for carcinoma of the cervix recur- rent after radical surgery. Gynecol ()ncol 1987;27:254—63. Lagasse LD. Smith ML. Moore JG. Morton DG. Jacobs M. Johnson GH. et al. The effect of radiation therapy on pelvic lymph node involvement in stage 1 carcinoma of the cervix. Am J Obstet Gynecol l974:| 192328—34. Alvarez RD. Soong SJ. Kinney WK. Reid GC. Schray MF. Podratz KC. et al. Identification of prognostic factors and risk groups in patients found to have nodal metastasis at the time of radical hysterectomy for early-stage squamous carcinoma ofthe cervix. Gynecol ()ncol 1989;35:130—5. Chting CK. Nahhas WA. Stryker JA. Curry SL. Abt AB. Mortel R. Analysis of factors contributing to treatment failure in stages IB and 11A carcinoma ofthe cervix. Am J Obstet Gynecol 1980;138:5504). Hitnmelman A. Holmberg E. Jansson l. Oden A. Skogsberg K. The effect of postoperative external radiotherapy on cervical carcinoma stage IB and 11A. Gynecol Oncol 1985;22:73»84. Kinney WK. Alvarez RD. Reid GC. Schray MF. Soong SJ. Morley GW. et al. Value of adjuvant whole-pelvis irradiation after Wertheim hysterec— tomy for early-stage squamous carcinoma of the cervix with pelvic nodal metastasis: a matched-control study. Gynecol Oncol 1989;34:25862. Burke TW. Hoskins WJ. Heller PB. Bihro MC. Weiser EB. Park RC. Prog- nostic factors associated with radical hysterectomy failure. Gynecol Oncol 1987;26:153—9. Figge DC. Tamimi HK. Patterns of recurrence of carcinotna following radical hysterectomy. AmJ Obstet Gynecol 1981;140:213-20. Thomas GM. Dembo AJ. Is there a role for adjuvant pelvic radiotherapy after radical hysterectomy in early stage cervical cancer? Int J Gynecol Cancer 1991;]:1—8. Bloss JD. Berman ML. Mukherjee J. Manetta A. Emma D. Ramsanghani NS. et al. Bulky stage IB cervical carcinoma managed by primary radical hysterectomy followed by tailored radiotherapy. Gynecol Oncol 1992: 47121-7. Podczaski E. Stryker JA. Kaminski P. Ndubisi B. Larson J. DeGeest K. et al. Extendedsfield radiation therapy for carcinoma of the cervix. Cancer 1990;66:251-8. Potish R. Adcock L. Jones T Jr. Levitt S. Pretn K. Savage J. et al. The mor- bidity and utility of periaortic radiotherapy in cervical carcinoma. Gynecol Oncol 1983;15:1-9. Nori D. Valentine E. Hilaris BS. The role of paraaortic ttode irradiation in the treatment of cancer of the cervix. lnt J Radiat Oncol Biol Phys 1985;11:1469-73. Journal of the National Cancer Institute Monographs No. 21. 1996 (42) Vigliotti AP. Wen BC. Hussey DH, Doornhos JF, Staples 1]. Jani SK. ct al. Extended field irradiation for carcinoma of the uterine cervix with positive periaonic nodes In J Radiat Oncol Biol Phys l‘)92:23:5()|-‘). (43) Hacker NF. Wain GVt Nicklin JL. Resection of bulky positive lymph nodes in patients with cervical carcinoma. Int J Gynecol Cancer 19955: 2506. Journal of the National Cancer Institute Monographs No. 2 l. 1996 (44) Landoni F. Maneo A. (‘olomho A. Placa F, Zanetta Gt Pellegrino A. et al Radical surgery or radiotherapy for cervical carcinoma stage lB-llAi A randomized study. (Abstract-17L Filth Biennial Meeting of the Internation— al Journal of Gynecological Cancer. Philadelphia: Blackwell Science Inc.. Cambridge I995: I4, Adj uvant Surgery After Radiotherapy Patricia J. Eifel * Investigators disagree about the role of adjuvant hysterec- tomy after irradiation of bulky stage IB cervical carcinomas, although the benefit of combined treatment has never been clearly demonstrated. Studies that have correlated outcome with initial tumor diameter suggest that central recurrences are rare after irradiation of tumors less than 5 cm in diameter, leaving little room for improvement with addition- al local treatment. Early studies suggested that adjuvant hysterectomy may improve pelvic disease control for patients with bulky endocervical tumors, but these results may have reflected the selection of tumors with relatively favorable characteristics for combined treatment. Several studies have suggested that central pelvic disease can be con- trolled in more than 90% of patients with bulky endocervi- cal tumors if they are treated with adequate doses of irradiation. Although published studies are somewhat limited by their retrospective designs, the available data do not support the added cost and morbidity of adjuvant hysterectomy in the routine treatment of patients with bulky early stage cervical carcinomas. [Monogr Natl Cancer Inst 1996;21:85-8] The role of adjuvant hysterectomy in the management of patients with bulky stage IB carcinomas of the cervix has been a subject of controversy for several decades. The theoretical benefit of removing a cervix at risk for harboring residual dis- ease has never been demonstrated in a prospective study. and in~ vestigators have drawn conflicting conclusions from reviews of their retrospective experiences. In 1969. Durrance et al. (I) reported the results of a retrospec- tive review of the M. D. Anderson experience with stages I and II carcinomas of the cervix and concluded that patients with bulky (26 cm) endocervical tumors had a lower risk of pelvic disease recurrence when radiation therapy was followed by an extrafascial hysterectomy. These investigators also concluded that patients with smaller tumors or exophytic lesions did not re— quire combined treatment because they had a very low risk of having a central disease recurrence after treatment with radia— tion alone (1,2). On the basis of these results. the authors recom- mended that adjuvant extrafascial hysterectomy be performed if there was no evidence of parametrial disease after 40-50 Gy of external—beam irradiation. In later publications. the group em- phasized the importance of limiting the surgical procedure to an extrafascial (Type I) hysterectomy. using a careful sharp dissec- tion technique to reduce the risk of serious urinary tract com- plications (3). They also proposed a staging system that placed these bulky endocervical tumors in a special category labeled “stage “BMW" (4). Although this system is still used in a num— Journal ofthe National Cancer Institute Monographs No. 2|. I996 ber of European countries. it was not widely adopted in the United States. In a more recent update of this experience, a number of fac- tors were identified that had influenced the selection of patients for different treatments at M. D. Anderson. potentially causing patients with more aggressive tumors to be treated with radia— tion alone (5). In their review of 37] patients treated for stage lB—llB endocervical tumors 6 cm or larger. Thorns et al. (5) found that prognosis was correlated with tumor volume (<8 ver— sus 28 cm). lymphangiogram results. tumor response after 40—50 Gy. and disease extent after initial external-beam therapy. Be cause patients who were treated with radiotherapy alone more often had positive lymphangiograms. massive tumors. and poor initial responses to irradiation. the authors suggested that these and other biases may have caused the apparent differences in pelvic disease control and survival rates between the two treat— ments. When the authors excluded patients who had massive tumors (28 cm) or evidence of lymph node involvement on lym- phangiogram from the analysis. the pelvic disease control rate for patients treated with irradiation and extrafascial hysterec— tomy was not significantly better than that of patients treated with irradiation alone. However. the authors concluded that the many biases that influenced selection of treatment at M. D. Anderson during this period make it impossible to determine the relative efficacies of the two treatments for this population of patients. A 1991 study from the University of Florida (6) included fewer biases than most retrospective reviews because the authors compared all patients with bulky endocervical tumors larger than 6 cm in diameter who were treated before or after an abrupt change in treatment policy that occurred during the mid 1970s. Patients treated before or after this policy change were compared according to the treatment intent; therefore. 13 patients who had laparotomy but did not undergo hysterectomy because of liver metastasis (two patients) or extensive lymph node involvement (10 patients) and one patient who refused surgery were still included in the combined treatment group, avoiding an important flaw in other retrospective studies. Patients in the two groups were evenly balanced in terms of tumor size. extent, and initial tumor response. However. the authors found no significant differences in the pelvic disease control (75% versus 76%) or cause-specific survival rates (62% versus 55%) between 75 patients treated with radiation alone and 75 patients whose planned treatment included extrafascial hysterectomy. :E:(‘UI’I‘é’XpUIN/t’H(‘(‘ to: Patricia J. Eifel. M.D.. Department of Radiation Oncol» ogy. The University of Texas M. D. Anderson Cancer Center. [515 Holcombe led.. Houston. TX 77030. In a retrospective review of 128 patients with barrel-shaped lesions that were larger than 5 cm in diameter. Perez and Kao (7) concluded that tumor control with radiation alone was com- parable to that with combined treatment if adequate radiation doses were used (Table 1). In another analysis. Eifel et al. (8) reported that 25 patients who were treated with radiation alone for bulky endocervical tumors (at least 6 cm in diameter) but received less than 6000 ing—hrs of brachytherapy after 40 Gy of external—beam irradiation had a higher risk of pelvic disease recurrence than 73 patients who received more than 6000 mg- hrs (33% versus 16%; l’ = .03). Six thousand mg—hrs cor— responded to an average dose to Point A of 50 Gy and a total dose (including external—beam irradiation) of approximately 90 Gy. Patients who received relatively low doses frequently had poor anatomy for intracavitary therapy. usually because of a narrow vagina or a small uterine cavity. The authors suggested that patients who had bulky barrel—shaped tumors limited to the cervix and who had normal tissue anatomy that severely limited the deliverable dose of intracavitary radiation might benefit from adjuvant surgery. although this question has not been studied prospectively. Although most recent studies have suggested that adjuvant hysterectomy does not improve pelvic disease control or sur— vival rates over treatment with irradiation alone. these studies have all been limited by their retrospective design. Many clinicians continue to believe in the benefit of combined treat— ment. questioning the ability of radiation alone to control bulky central disease. The literature has often been confused by differ— ing definitions of “bulky" disease and indications for adjuvant hysterectomy vary widely among clinicians. Until very recently. the International Federation of Gynecol- ogy and Obstetrics (FIGO) categorized all patients with tumors that were more than microinvasive and clinically confined to the cervix in a single category—stage 1B. This classification placed tumors smaller than 1 cm in diameter in the same category as massive 8-10 cm endocervical tumors (9). However. clinicians have long recognized the strong relationship between tumor diameter and regional spread. tumor recurrence. and survival. Since the 1969 report by Durrance et al. (I). the influence of tumor size on the outcome of patients treated with irradiation has been documented by many other investigators ( /()—l7). In these studies. tumor size was often treated as a dichot— omous variable using thresholds ranging between 3 and 6 cm to define a group at increased risk for recurrence. Some authors have included exophytic tumors in their definition of un- Table 1. Relationship between the dose to point A central disease recurrences in patients with barrel-shaped carcinomas of the cerx ix larger than 5 cm in diameter* F1(}() stage Dose to point A. (1y [8 11A 11B <60 2/20'i 2/5 UK 60-80 0/13 2/14 3/18 0/13 favorable stage 1 disease. while others included only barrel— sbaped endocervical tumors. The Gynecologic Oncology Group has usually included all tumors measuring 4 cm or more in clini— cal tumor diameter in its studies of “bulky" stage 1 disease. In 1995. FlGO finally recognized the importance of tumor size in its staging system. dividing stage [B into stages lBl (tumors S4 cm) and 182 (tumors >4 cm) (I8). Although all of these methods divide patients into two groups with differing overall outcomes. they do not fully characterize the relationship be— tween tumor si/.e and outcome. The broad range of definitions of "bulky" cervical cancer has led to considerable confusion in the clinical literature. A review of 1526 patients treated with radiation therapy for FIGO stage IB squamous cell carcinomas of the cervix at M. D. Anderson addressed the relationship between clinical tumor diameter and outcome in more detail (ll). Patients were analyzed according to the clinical estimates of tumor diameter without grouping according to any preconception. All measures of outcome were strongly correlated with tumor diameter (Table 2). Central disease recurrences were rare for any subgroup of patients with tumors smaller than 5 cm in diameter. The overall central disease control rate for patients with F160 [81 tumors (S4 cm) was 99% at 5 years, The pelvic disease control and dis— ease—specific survival rates were 97% and 89%. respectively. for this group of patients. Central recurrences were also uncommon in patients with larger tumors that had a predominantly ex— ophytic morphology. Central recurrence rates were 3% and 9%. respectively. for patients treated with irradiation for exophytic or endocervical tumors measuring 5—7.9 cm in diameter. Central tumor recurrence was more frequent for patients with endocervi- cal tumors 8 cm or larger in diameter. However. in more than half of the patients with these massive stage IB tumors. pelvic wall disease was appreciated after response to initial irradiation facilitated clinical examination of the paracervical region. in— dicating that these patients were frequently understaged at their initial evaluation. Some investigators (l9-22) have suggested that adenocar— cinomas of the cervix may be tnore radioresistant than squamous cell carcinomas. although others (23,24) have failed to find a correlation between histology and outcome. This per- Tahle 2. Outcome of 1494 patients with stage IB squamous cell carcinoma of the cervix according to clinical estimate of ten ical diameter”: Central Pelvic Disease- ttitnor tumor specific No. of control control survival Size. cm patients rate. ‘/1 rate. ‘/r rate. V? Normal 250 100 99 94 Small, post cone"? I54 98 95 XX Enlarged. <4 297 99 98 87 4.0 193 96 97 86 4.5 s} 94 95 85 5.0 175 90 85 71 5.5 62 94 X7 73 606.5 130 90 82 69 7.0-7.5 X4 94 X4 64 28.0 66 75 67 47 >80 0/8 0/5 *From ref. (7). ’1'Number with central disease recurrence/number treated. 86 *From ref. ([0). ’i'Sinall cervix. patient referred after cone biopsy. Journal of the National Cancer Institute Monographs No. 21. 1996 ception has sometimes led investigators to advocate the use of combined treatment for patients with bulky stage IB adenocar- cinomas. Most reports have included too few patients to permit a detailed analysis of tumor control rates in patients with similar-sized tumors. In one of the largest retrospective studies of this subject in patients treated with irradiation. Eifel et a1. (25) compared the outcomes of 229 patients with stage IB adenocarcinomas of the cervix with those of 1538 patients with squamous carcinomas of the satne stage. As with squamous car- cinomas. control of adenocarcinomas was strongly correlated with the initial diameter of the tumors. Central tumor recurrence was somewhat more frequent in patients with larger (24 ctn) adenocarcinomas of the cervix than in those with squamous car— cinomas (12% versus 7%. respectively; P : .04) of similar size. although overall pelvic disease recurrence rates were not sig— nificantly different (17% versus 13%; P = .16). This study sug— gested that the most important factor contributing to the poorer prognosis of patients with adenocarcinomas was their much greater risk of distant disease recurrence Comparison of pelvic disease control rates for patients treated with or without ad- juvant hysterectomy did not demonstrate an advantage for corn- bined treatment after patients with clinically positive pelvic nodes (most of whom were treated with irradiation alone) were excluded. However. the number of patients was insufficient to rule out small differences. The risk of major complications may be greater for patients who have adjuvant hysterectomy. particularly if a radical (type 11 or 111) hysterectomy is performed after high-dose radio- therapy. Rotman et al. (26) found a very high rate of serious uri- nary tract complications (particularly fistula formation) with this combination. as did investigators at M. D. Anderson (3). The M. D. Anderson group emphasized that an extrafascial abdominal hysterectomy performed with careful sharp dissection is the preferred procedure in this setting (3). When hysterectomy was performed after somewhat reduced doses of irradiation. the overall rate of major complications was similar to that following high—dose irradiation. although the rate of fistula formation was approximately doubled (5.27). in another study from the University of Florida. Mendenhall et a1. (6) reported an actuarial rate of major complications of 18% at 6 years for patients treated with combined surgery and radiation compared with 7% for patients treated with radiation alone (P = .027). The only prospective randomized study that has questioned the role of adjuvant hysterectomy was conducted by the Gynecologic Oncology Group. accruing 269 assessable patients with stage 18 tumors at least 4 cm in diameter between 1984 and 1991. The Gynecologic Oncology Group is expected to analyze and report its results soon. However. this study has al- ready been criticized because of the long accrual period. the relatively low dose of radiation used to treat patients with radia- tion alone (80 Gy to Point A). and the inclusion of a large nurn— ber of patients with exophytic tumors or relatively small (4 cm) endocervical tumors. Because central relapses are rare after appropriate radical it'- radiation of exophytic tumors and of endocervical tumors smaller than 5 cm in diameter. the cost and morbidity of com- bined treatment cannot be justified for patients who receive ade- quate irradiation of these more favorable tumors. The role of Journal of the National Cancer Institute Monographs No. 21. 1996 adjuvant hysterectomy in patients with larger endocervical tumors has not yet been established. Studies have failed to demonstrate a major advantage for combined treatment and therefore do not justify its routine use for patients with stage 182 squamous cell carcinomas or adenocarcinomas. However. these studies may not have had the power to detect differences in small subsets of patients who are at particularly high risk for central disease recurrence. If there is any benefit from adjuvant surgery. it is probably limited to those patients with localized bulky FlGO stage 1B2 endocervical tumors who have normal tissue anatomy that prevents high—dose brachytherapy (e.g.. a very narrow vagina) or those with tumors that respond very poorly to initial external-beam irradiation. However. although patients who have a poor initial clinical response to irradiation tend to have a poorer prognosis. several authors (5.17.28) have demonstrated that patients who have gross residual disease in their hysterectomy specimens also have a relatively poor out— come with combined treatment. It has not yet been demonstrated that adjuvant surgery is more effective in achieving a cure of these patients‘ cancers than the additional radiation dose that is given with radical irradiation. Unfortunately. because such patients forrn a very small proportion of those presenting for radical radiation therapy. it may never be possible to determine whether their outcome can be improved with combined treat— ment. References (I) Durrance FY. Fletcher GH. Rutledge FN. Analysis of central recurrent dis— ease in stages 1 and 11 squamous cell carcinomas of the cervix on intact uterus. Am J Roentgenol Radium Ther Nucl Med 1969;106:831-8. (2) Fletcher GH. Cancer of the uterine cervix. Janeway lecture. 1970. Am J Roentgenol Radium Ther Nucl Med 1971:3:225-42. (.1‘) O‘Quinn AG. Fletcher GH. Wharton JT. Guidelines for conservative hysterectomy after irradiation. Gynecol ()ncol 1980;9z68-7‘). (4) Eifel PJ. Morris M. Oswald MJ. Wharton JT. Delclos L. Adenocarcinoma of the uterine cervix. Prognosis and patterns of failure of 367 cases. Cancer 1990;65:2507-14. (5) Thorns WW Jr. Eifel PJ. Smith TL. Morris M. Delclos L. Wharton JT. et a1. Bulky endocervical carcinomas: a 23—year experience [.\'(’l’ comment citation in Medline]. lntJ Radiat ()ncol Biol Phys 1992;23:491-9. (6) Mendenhall WM. McCarty PJ. Morgan LS. Chafe WE. Million RR. Stage 18 or llA-B carcinoma of the intact uterine cervix greater than or equal to 6 cm in diameter: is adjmant extrafascial hysterectomy beneficial? Int J Radiat Oncol Biol Phys 1991;21:899—904. (7) Perez CA. Kao MS. Radiation therapy alone or combined with surgery iii the treatment of barrel-shaped carcinoma of the uterine cervix (stages 18. 11A. 11B). Int J Radiat Oncol Biol Phys 1985;] 121903-9. (8) Eifel PJ. Thorns WW Jr. Smith TL. Morris M. Oswald MJ. The relation- ship between brachytherapy dose and outcome in patients with bulky en- docervical tumors treated with radiation alone. Int J Radiat ()ncol Biol Phys 1994;28:113-8. (9) Eifel PJ. Problems with the clinical staging of carcinoma of the cervix. Semin Radiat Oncol 199-13: 1-8. ([0) Arimoto T. Significance of computed tomography~measured volume in the prognosis ofcervical carcinoma. Cancer 1993;72:2383-8. Eifel PJ. Morris M. Wharton JT. Oswald MJ. The influence of tumor size and morphology on the outcome of patients with FIGO stage IB squamous cell carcinoma of the uterine cervix. Int J Radiat Oncol Biol Phys 1994;29:9-16. Perez CA. Grigsby PW. Nene SM. Camel HM. Galakatos A. Kao MS. et al. Effect of tumor size on the prognosis of carcinoma of the uterine cervix treated with irradiation alone. Cancer l992:69:2796—X()6. Lowrey GC. Mendenhall WM. Million RR. Stage 18 or llA-B carcinoma of the intact uterine cervix treated with irradiation: a multivariate analysis. lntJ Radiat Oncol Biol Phys 1992;24:205-10. ([4) Lanciano RM. Won M. Coia LR. Hanks GE. Pretreatment and treatment factors associated with improved outcome in squamous cell carcinoma of (I3) 87 (15) (/6) (I7) (/8) (I9) (20) (2/) 88 the uterine cervix: a final report of the 1973 and 1978 patterns of care studies. lntJ Radiat Oncol Biol Phys 1991;20:667—76. Kovalic JJ. Perez CA. Grigsby PW. Lockett MA. The effect of volume of disease in patients with carcinotna of the uterine cervix. Int J Radiat Oncol Biol Phys 1991 :21 :9()5—|(). Homesley HD. Raben M. Blake DD. Ferree CR. Bullock MS. Linton EB. et :11. Relationship of lesion size to survival in patients with stage 18 squamous cell carcinoma of the cervix uteri treated by radiation therapy. Surg Gynecol Obstet 1980; 1 5025293 I. Maruyama Y. van Nagell JR. Yoneda J. Donaldson E. Gallion HH. Hig- gins R. et al. Dose—response and failurc pattern for bulky or barrel—shaped stage IB cervical cancer treated by combined photon irradiation and ex- trafascial hysterectomy. Cancer 1989:63270—6. International Federation of Gynecology and Obstetrics. Staging an- nouncement. FlGO staging of gynecologic cancers; cervical and vulva. Int J Gynecol Cancer 1995:5319. Hopkins MP. Morley GW. A comparison of adenocarcinoma and squamous cell carcinoma of the cervix. Obstet Gynecol 1991;77:912-7. Milsom 1, Friberg LG. Primary adenocarcinoma of the uterine cervix. A clinical study. Cancer 1983;52:942-7. Moberg PJ. Einhom N. Silfversward C. Soderberg G. Adenocarcinoma of the uterine cervix. Cancer 1986;57:407-10. «'3 cs (28) Saigo PE. Cain JM. Kim WS. Gaynor JJ. Johnson K. Lewis JL Jr. Prognos— tic factors in adenocarcinoma of the uterine cervix. Cancer 19862571584- 93. Grigsby PW. Perez. CA. Kuske RR. Camel HM. Kao MS. Galakatos AE. et a1. Adenocarcinoma of the uterine cervix: lack of evidence fora poor prog‘ nosis. Radiother Oncol 1988;12:289—96. Rutledge FN. Galakatos AE. Wharton JT. Smith JP. Adenocarcinoma of the uterine cervix. Am J Obstet Gynecol 1975;122:3645. Eifel PJ. Burke TW. Morris M. Smith TL. Adenocarcinoma as an inde- pendent risk factor for disease recurrence in patients with stage 18 cervical carcinoma. Gynecol Oncol 1995;59:38-44. Rotman M. John MJ. Moon SH. Choi KN. Stowe SM. Abitbol A. et al. Limitations of adjunctive surgery in carcinoma of the cervix. Int J Radiat Oncol Biol Phys 19795232732. Eifel PJ. Levenback C. Wharton JT. Oswald MJ. Time course and in» cidence of late complications in patients treated with radiation therapy for FIGO stage IB carcinoma of the uterine cervix [we comment citation in Medline]. 1ntJ Radiat Oncol Biol Phys 1995132: 1 289-300. Hardt N. van Nagell JR. Hanson M. Donaldson E. Yoneda J. Maruyama Y. et a1. Radiation-induced tumor regression as a prognostic factor in patients with invasive cervical cancer. Cancer 1982;49:35-9. Journal of the National Cancer Institute Monographs No. 2 I. 1996 Optimal Management of Locally Advanced Cervical Carcinoma Henry Keys, Susan K. Gibbons * Locally advanced or recurrent cervical cancer is highly respon- sive to treatment and at least moderately curable with effective aggressive treatment. Radiation therapy is the mainstay of treatment for patients with this cancer. The roles for surgery and chemotherapy are as yet unproved, and both modalities are currently under investigation for their potential roles in the management of these conditions. Exenterative surgery clearly has an established utility for central pelvic failures after prior radiation therapy. Postsurgical pelvic recurrences are rarely successfully treated for cure, but considerable palliative effect is possible. The roles of intraoperative irradiation, sensitizing chemotherapy, and radical resection with interstitial irradia- tion are all under investigation at this time. Much has been learned over the past several decades about what parameters are important for successful radiation therapy for cervical can- cers of stages IIB-IVA. While the traditional staging work-up for these patients included excretory urography, barium enema, examination under anesthesia, cystoscopy, and proctos- copy, there is now good evidence that computed tomography scan with intravenous contrast and office examination and biopsy are sufficient, with cystoscopy reserved for those few patients in whom clinical or imaging data suggest a higher risk of involvement. Surgical lymph node staging, especially of para-aortic lymph nodes, may be worthwhile in certain settings (e.g., for entry into research protocols), but it has no demon- strated role in routine clinical practice. Evidence is clear and convincing that effective treatment for these disease stages re- quires the inclusion of intracavitary brachytherapy. The role of interstitial brachytherapy is less clear, although there are some fervent advocates of this procedure. The debate continues about the use of Iow-dose-rate versus high-dose-rate brachy- therapy. Treatment dose, volume, and length of treatment course are all important variables with outcome implications. The central disease requires a total dose of 8000-9000 cGy for maximal control probability, with larger tumors requiring the higher doses. The three-dimensional treatment volume must adequately surround the cancer and its likely routes of spread. Overall treatment time should be kept as short as possible, within the limits of conventional, tolerable fractionation. The potential theoretical advantage of hyperfractionated external- beam irradiation has yet to be verified in this disease but is of interest. It will be tested in an upcoming Gynecologic Oncology Group clinical trial. The negative prognostic significance of hypoxia in cervical cancers in general has been reported recently. While tumor cell hypoxia is almost certainly a prob- lem in this disease, hypoxic cell sensitizers have not yet been found to improve treatment results. In clinical practice, re- oxygenation probably occurs in these tumors. The role of para- Journal of the National Cancer Institute Monographs No. 2 I. I996 aortic lymph node elective irradiation has been of interest for more than 20 years and was the subject of two ran- domized trials with quite different results. The Radiation Therapy Oncology Group trial found significantly improved survival in the treatment group assigned to receive para- aortic irradiation. when compared with the pelvic treatment group. However, a similar study by the European Organiza- tion for Research and Treatment of Cancer found no dif- ference. The results of treatment today are substantially improved from those seen two decades ago. About 75% of patients with stage IIB disease and fully 50% of patients with stage “[8 disease are now cured with conventional ir- radiation alone. Clearly, there is still a need for further im- provement. Of patients with urinary bladder involvement, 10%-20% are long-term survivors, as are 25%-30% of patients with para-aortic lymph node metastases. While these improvements are significant, there is clearly room for further progress. Improved radiation delivery, multi- modality treatment programs, and better biological assays for directing treatment more appropriately are just some of the research directions that hold promise for such future progress. [Monogr Natl Cancer Inst 1996;21:89-92] Locally advanced cancer of the cervix (International Federa- tion of Gynecology and Obstetrics [FIGOI stages IIB-IVA) is best managed with primary radiation therapy. This article will survey the data and opinions that define what is the best current therapy for these stages of disease. We will discuss staging work—up. treatment volume and technique. the role of para-aor- tic lymph node irradiation. dose. time. and the importance of in- tracavitary brachytherapy. Treatment of recurrent pelvic disease after primary treatment depends on whether the primary management was surgical or radiotherapeutic. Although salvage is occasionally possible in either situation. the prognosis is poor and survival is usually relatively short. Staging Work-up The FIGO staging system for cervical cancer is the favored classification worldwide. This system was first promulgated in I950 and subsequently modified several times. most recently in *Al'filiulimm ofumhmzr: II. Keys (Department of Radiation Oncology. Cancer Center). S. K. Gibbons (Department of Radiation Oncology). Albany Medical College. NY. (‘rirrm/uint/envy In: Ilenry Keys. M.D.. Department of Radiation Oncology. Albany Medical College. 47 New Scotland Ave. (A-I37). Albany. NY [2208. Soy “Note" section following "References." 89 1993 (l). The rules for staging have allowed findings on ex- cretory urography and barium enema to be used in stage de- termination. while disallowing findings from computed tomography (CT) scans or lymphangiography. In addition to ex— cretory urography and barium enema examinations. cervical cancer work—up has traditionally included examination under anesthesia with cystoscopy and proctoscopy—sigmoidoscopy. The Gynecologic Oncology Group (GOG) has also insisted that surgically staged para—aortic lymph nodes be pathologically negative before patients could be eligible for entry into GOG treatment protocols for locally advanced cervical cancer (2). Current clinical practice has been evolving toward more ra— tional resource utilization in staging work-up. The yield of posi— tive findings from barium enema and proctoscopy is almost nil: hence. these examinations are used only when symptoms. physi— cal findings. or fecal occult blood makes them indicated. Cys— toscopy is similarly unrewarding in the absence of physical findings and is not routinely employed in the evaluation process. Lytnphangiography is the best imaging technique for lymph node involvement. but it is falling out of use (3). Abdominal— pelvic CT scanning with intravenous contrast provides evalua— tion of the local extent of the primary tumor. the presence of enlarged lymph nodes. and the presence or absence of any hydroureter or hydronephrosis. In our practice. then. patients have a careful office examination. biopsy. routine blood work (including complete blood cell count. kidney function tests. and chemical profile screen). and CT scan with contrast. Compared with the more traditional staging evaluation. this approach has significant time. cost. and potential morbidity savings. Treatment Volume and Treatment Technique Radiation therapy for these extensive cancers consists of two components: external irradiation and intracavitary brachytherapy. The external radiation should be delivered by linear accelerator— produced x rays. preferably at energies of 10 MV or higher unless patients are quite small physically. For smaller women (with separations of <20 cm). 6— or even possibly 4-MV beams may be suitable. Although some authors advocate anterior and posterior. parallel. opposed treatment field arrangement. we find that four— field pelvic treatment techniques almost always allow for more protection of normal tissue by beam shaping and placement. and a recent analysis of GOG protocol-treated patients showed a reduc- tion in grade 3 and 4 complications when a four—field technique was utilized. qu‘allel. opposed treatment technique was associated with a higher complication rate (4)'. The treatment volume must include all known or suspected areas of cancer involvement. lnferiorly. the field margin should be at the inferior part of the obturator foramen or 3 cm below the lowest extension of known disease. whichever is lower. The lateral borders should be positioned at least 1.5 cm outside the pelvic brim. Lateral fields should extend anterior to the external iliac lymph nodes as identified on CT scan. This usually re— quires the most anterior portion of the field to be anterior to the pubic symphysis. lntraoperative mapping by Greer et al. (5) sug— gests that the posterior margin of lateral treatment fields may need to extend more posteriorly into the presacral area than pre- viously appreciated to cover the posterior insertion of the 90 ureterosacral and cardinal ligaments. The 82-3 interspace should be included at a minimum. Typically. those patients with more advanced disease will have treatment up to the middle common iliac lymph node chain area. with a superior border placed at the 1-4—5 interspace. A recent analysis by Greer et al. (6) demonstrates excellent results in the treatment of patients with locally advanced cervi- cal cancer with extended pelvic fields as discussed above. These investigators report a 4—year disease-specific survival of 76% for patients with stage [[8 disease and 53% for patients with stage ”[8 disease. If para—aortic lymph nodes are to be treated. either electively or therapeutically. they should be included in extended pelvic— para—aortic fields. Either the parallel. opposed or the four-field technique may be used. depending on which gives the best dose distribution. We do not recommend matching separate pelvic and para—aortic treatment fields. because the field junctional dosimetry is tricky and could cause overdosing or underdosing structures in that area. Para-aortic Lymph Node Management Elective Treatment The possibility that elective irradiation of the para-aortic lymph nodes might improve the outcome of treatment for cervi— cal cancer patients has been debated for more than 20 years. Es- timates based on patterns of treatment failure suggested that it was unlikely that an improvement in survival or relapse could be detected. with the possible exception of patients with stage [IB disease. where as many as 16% might be aided by the addi— tional treatment (7). Subsequently. two randomized trials were carried out. and results were conflicting. The Radiation Therapy Oncology Group (RTOG) trial found a significant increase in survival among patients randomly assigned to receive para—aortic irradia— tion (8). It is interesting that the difference was most pro— nounced in patients with earlier stage disease and was primarily due to fewer pelvic failures in those patients. The para-aortic and distant failure rates were not affected. Rotman et al. (8) have recommended elective para-aortic therapy for patients with higher risk stage IE or stage [[3 disease who have a fairly good prospect for pelvic disease control. The European Organization for Research and Treatment of Cancer (EORTC) study. on the other hand. found no significant difference in survival or metas— tatic rate between the two treatment arms (9). The GOG cervical cancer treatment protocols dodge the issue by insisting on surgi- cally staged negative lymph node sampling for the para—aortic area as a condition for patient eligibility. Thus. the role of elective treatment of the para—aortic lymph nodes remains uncertain. There may, indeed. be an advantage for some patients with higher risk. early stage disease. but this was shown in only one of the two randomized trials of this ex- tended-field treatment. Treatment for Positive Para-aortic Lymph Nodes Cervical cancer patients with positive para-aortic lymph nodes are at high risk for eventual distant metastases. provided Journal ofthe National Cancer Institute Monographs No. 21. [996 their pelvic disease is effectively treated. However. not all patients with involved para—aortic lymph nodes are doomed. One study (/0) has shown that. with para—aortic irradiation. a significant salvage rate can be accomplished. Long—term dis— ease—free survivals of 25% to greater than 40% have been reported. There is some uncertainty about the significance of the number of positive lymph nodes. whether they are resected or ir— radiated in situ. macroscopic versus microscopic involvement. and the need for scalene lymph node biopsy. irradiation tech— niques have been described. including the use of rotational boost fields. The treatment dose is best kept under 5000 cGy because the risk of complications rises with increasing dose above that level (1/). Importance of Brachytherapy The use of brachytherapy is the most important predictor of treatment success. it is more important than external dose. in— tracavitary dose. or the number of implant applications a patient receives.l Every effort. therefore. should be made to implant cervical cancer patients as part of their treatment program. Evidence is accumulating suggesting that high-dose-rate. re— mote. afterloading brachytherapy may be as effective and as safe as conventional low-dose—rate brachytherapy. provided that individual fraction sizes are not too large. The maximum safe fraction size for high—dose—rate brachytherapy would appear to be around 700 cGy for four fractions. 600 cGy for five fractions. and 500 cGy for six fractions. Brachytherapy Dose Description There are a variety of systems of dose specification that make it possible to prescribe doses and to combine external and in- tracavitary treatment dose. Use of any such system requires both an understanding that the system is merely a shorthand for describing the desired three—dimensional composite dose dis— tribution and experience using that type of dose distribution. Many practitioners use various modifications of the Manchester Point A and Point B System. This allows simple addition of ex- ternal and intracavitary dose contributions to the defined points. Point A is basically a paracervical reference point. whereas Point B is a lateral parametrial dose point near the pelvic lymph nodes. Recommended Doses in patients with locally advanced disease. the bulk of the tumor at presentation will usually preclude placing a tandem and ovoids (or ring) early in the course of treatment. instead. a larger proportion of the central dose is delivered with external- beam irradiation of the pelvis. After a dose of 40-50 Gy. tumor shrinkage is generally adequate to allow implant placement with reasonable dosimetry. If the tumor remains bulky after full ex— ternal-beam treatment. the option of an interstitial central body may be considered. In general. patients with stage [18 disease who are not receiv— ing chemotherapy should be treated to a total central dose of 80- 85 Gy. About 40 Gy is delivered in external—beam irradiation of Journal of the National Cancer institute Monographs No. 2|. I996 the pelvis. and the remainder is delivered with intracavitary im— plants. Depending on the implant contribution to sidewall dose (which is usually quite small). the extent of the initial para- metrial disease. and the presence of suspicious lymph nodes in the pelvis. the sidewall dose can be boosted with further exter— nal—beam irradiation to a total dose of 45-55 Gy. This is ac— complished with antert)posterior—posteroanterior fields. using a custom block that shields midline structures from overdose. The midline block should cover the implant dose distribution out to approximately where the 50% isodose line is located. Patients with stage IIIB disease who are not receiving chemo- therapy require a central dose of 85-90 Gy for optimal local con— trol; about 50 Gy of this dose is delivered with initial whole—pelvis irradiation. A central block can then be utilized to boost the pelvic sidewall dose. as above. Depending on the response to initial treatment. the involved sidewall may be boosted to a total dose of 55-60 Gy. Patients with stage IVA dis- ease are treated similarly. with minor modifications depending on initial disease distribution. Patients who are treated with adjuvant chemotherapy are treated with slightly more conservative doses. We generally use a total central dose of 80 Gy for all of our patients with locally advanced disease who are receiving weekly cisplatin chemo- therapy during radiation therapy. Importance of Overall Time Several investigators have reported the adverse effect of prolongation of overall duration of cervical cancer treatment regimens. Petereit et al. (/2) report their own data and sum— marize reports of Lanciano et al. (/3). Fyles et al. (14). and Girinsky et al. (/5). In the study by Petereit et al.. the median treatment time was 55 days. When all patients were subdivided into those who were treated in fewer than 55 days and those who took 55 or more days to complete treatment. pelvic control was found to be significantly better in the shorter treatment time group (87% versus 72%). and 5-year survival was better in the same group (65% versus 54%; P = .006). Survival decreased 0.6% and pelvic control decreased 0.7% for each additional day of treatment beyond 55 days. Fyles et al. (/4) and Girinsky et al. (/5) report even greater reductions in pelvic control and survival with prolonged time. whereas Lanciano et al. (/3), reporting on patterns of care data. found slightly less negative impact than that reported by Petereit et al. Not all studies have noted this effect. however. The relation— ship between overall treatment time and outcome may be more complicated than simple time—dose compression. In some series. patients with earlier stage disease were more likely to get two intracavitary brachytherapy applications, while patients with stage IIIB disease usually had only one. This could mask or con- found the effect of time. causing patients with stage IIB disease (who have a better prognosis) to have a longer overall treatment time than those with stage lllB disease (in whom the excepted outcome is worse). Petereit et al. (/2) also note that there does not seem to be any association between shorter treatment time and likelihood of developing long—term complications. A recent report by Perez et 9| a1. (16) supports this finding of a lack of increased toxicity with shorter treatment duration. Therefore. the current recommendation is to try to complete all external and intracavitary treatment in as short a time as pos- sible. without accelerating external-beam fraction size beyond 180-200 cGy per day. Treatment interruptions are especially to be avoided. Patients need to be told of the importance of keep— ing the overall treatment time short. They should be given dietary instructions and medications to avoid intestinal toxicity with diarrhea. because this is the most common cause of treat— ment interruptions. Results Expected The outcome of treatment for locally advanced cervical can- cer has improved substantially during the past quarter century. Some 70%-75% of patients with FIGO stage [[8 disease can now expect to be alive and disease free 5 years after treatment (6). The patterns of care study found the national average. 4- year disease—free rate for these patients to be 66% in its 1974 survey of patients treated in the late 1960s (17). The improve- ment for stage IIIB is much more substantial. The same survey found a national average disease—free survival of only 28%. while the results today are about 50% at 5 years. This improve— ment is the result of more frequent use of brachytherapy. possib- ly some effect of shortening overall treatment time. and a trend toward increasing dose for larger cancers. Cervical cancers that recur after primary surgical therapy can occasionally be salvaged with the combination of external and intracavitary or interstitial brachytherapy with or without chemotherapy. Unfortunately. however, most patients with pel— vic relapses derive only palliative benefit from treatment. Patients with cervical or central recurrence after irradiation have a slightly more optimistic outlook if they are proven to be can— didates for total pelvic exenteration. About 20%-3()0/(~ of patients who undergo exenterative surgery will be long—term survivors without further evidence of recurrence ([8). Role of Chemotherapy Considerable experience now exists in the use of combined radiation therapy and chemotherapy for locally advanced cervix cancer. Both concomitant “sensitizing" chemotherapy and neo— adjuvant chemotherapy programs have been studied. Research Directions Current and future clinical research will focus primarily on efforts to increase the rates of clinical clearance of pelvic dis— ease, i.e.. complete response to the initial therapy. There will be continued exploration of the role of combined-modality treat— ment with radiation therapy and chemotherapy. A future GOG trial will evaluate improved standard irradiation and hyperfrac- tionated external irradiation. Recurrent cervical cancer will con— tinue to pose a serious challenge. The role of combined surgery. radiation therapy, and chemotherapy will be studied. Where feasible, intraoperative or interstitial irradiation will be at- tempted. Metastatic cervical cancer remains essentially incur— able at this time. Systemic treatment programs will be tested in phase I and H trials until more effective treatments are iden— tified. References (I) Beahrs (). Henson 1). Hutter B. Kennedy B. editors. Handbook for staging of cancer. 4th ed. Philadelphia: Lippincott. 1993. Stehman FB. Bandy BN. DiSaia PJ. Keys HM. Larson JE. Fowler WC. Carcinoma of the cervix treated with therapy. I. A multi~variate analysis of prognostic variables in the Gynecologic Oncology Group. Cancer 199l :67: 2776-85. Heller P. Malfetano J. Bandy B. Clinical pathologic study of stages MB. III. and IVA carcinoma of the cervix: extended diagnostic evaluation for paraaortic node metastasis ta GOG study). Gynecol Oncol 1990;38:425- 30. Peru CA. Uterine cervix. ln: Perel. CA. Brady LW. editors. Principles and practice of radiation oncology. 2nd ed. Philadelphia: Lippincott. I992: | H3202. Greer B. Koh W. Figgc D. Russell A. Cain J. Tamimi H. Gynecologic radiotherapy fields defined by intraoperative measurements. Gynecol ()ncol 1990;324:434. (()l Greer B. Koh W. Stelzer K. Goff B. Comsia N. Tran A. Expanded pelvic radiotherapy fields for treatment of local-regionally advanced carcinoma of the cervix: outcome and complications. Am J Obstet Gynecol 1996;174: 1 HI v50. Chism S. Park R. Keys H. Prospects for para-aortic irradiation in treatment ofcanccr of the cervix. Cancer 1975135: 1505-9. Rotman M. Choi K. Gu/c C. Marcial V. Hornback N. John M. Prophylac- tic irradiation of the para-aortic lymph node chain in stage 118 and bulky stage IB carcinoma of the cervix. initial treatment results of RTOG 7920. Int J Radiat ()ncol Biol Phys l990:l‘):5l3-Zl. Haie C. Pejovic M. Gerbaulet A. Horiot J. Pout'quier H. Delouche J. et al. ls prophylactic para—aortic irradiation wonhwhile in the treatment of ad» vanced cervical carcinoma? Results of a controlled clinical trial of the HORTC radiotherapy group. Radiother ()ncol |9881l 12101-12. (l0) Berman ML. Keys HM. Creasman WT. DiSaia PJ. Blessing J. Bundy B. San ival and patterns of recurrence in cervical cancer metastatic to periaor- tic lymph nodes. Gynecol ()ncol 1984;19:8-I6. (ll) Cunningham M. Dunton C. Corn B. Noumoff J. Morgan M. King S. et al. Extended-field radiation therapy in early-stage cervical carcinoma: sur— vival and complications. Gynecol ()ncol l99l‘.43:51-4. (I2) Petereit D. Sarkaria J. Chappell R. Fowler J. Hartmann T. Kinsella T. et al. The adverse effect of treatment prolongation in cervical carcinoma. Int J Radiat ()ncol Biol Phys |9‘)5;33: |301~7. (I3) Lanciano R. Pajak T. Marti. K. Hans G. The influence oftreattnent time on outcome for squamous cell cancer of the uterine cervix treated with radia- tion: a patterns-oilcare study. Int J Radiat ()ncol Biol Phys l993:2:39l—8. (I4) Fylcs A. Keane T. Barton M. Sitnln J. The effect of treatment duration in the local control ofcervix cancer. Radiother Oncol 1992;25:273-9. ([5) (iirinsky T. Rey A. Roche B. Haie C. Gerbaulet A. Randrianarivello H. et al. ()verall treatment time in advanced cervical carcinomas. A critical parameter in treatment outcome. Int J Radiat ()ncol Biol Phys l994: 27: I05 I —6. (l6) Perez. C. Grigsby P. Castro-Vita H. Lockett M. Carcinoma of the uterine cervix. ll. Lack of impact of prolongation of overall treatment time on morbidity of radiation therapy. [at] Radiat Oncol Biol Phys 1996;34:3—l I. (I7) Hanks GE. Herring DF. Kramer S. Patterns of care outcome studies. Results of the national practice in cancer of the cervix. Cancer 1983; 5| :959«67. (IX) Lawhead RA Jr. Clark DG. Smith DH. Pierce VK. Lewis JL Jr. Pelvic ex- enteration for recurrent or persistent gynecologic malignancies: a 10-year review of the Memorial Sloan—Kettering Cancer Center experience (I972— l‘)Xl ). Gynecol ()ncol 1989;33:279-82. (7 (3 (4 (5 (7 (X (9 Note lKeys H. Bundy B. Stehman F. DiSaia P. Larson J. Fowler W. Manuscript submitted for publication. Journal ofthe National Cancer Institute Monographs No. 2|. I996 Neoadjuvant Chemotherapy Before Surgery in Cervical Cancer Gary Lee Eddy, Sr. * Eighteen phase II clinical trials have demonstrated the rela- tive ease of administering intravenous neoadjuvant chemo- therapy before radical surgery in patients with cervical cancer. Toxicity has been modest, with the exception of oc- casional severe bleomycin-induced pulmonary toxic effects. All regimens tested have been cisplatin based with no clearly superior combination. Overall objective response rates were 85% for patients with International Federation of Obstetrics and Gynecology stage [B or IIA disease, 88% for those with stage IIB disease, 74% for those with stage III disease, and 47% for those with stage IV disease. There was a cor- responding decrease in clinical complete response rates with increasing stage from 28% for those with stage [B or IIA disease to 7% for those with stage IV disease. Operability was also stage dependent: 94% for patients with stage IB and IIA disease, 74% for patients with stage 113 disease, and 58% for patients with stage III or IV disease. All inoperable patients received primary radiation therapy. Fifty-five per- cent of operable patients received postoperative radiation therapy. Most phase II trials using historical controls demonstrated comparable survival. Only a few trials showed improved survival with the use of neoadjuvant chemo- therapy before radical surgery. To date, there has been only one reported prospective trial of neoadjuvant chemotherapy before surgery. Patients with stage IB disease were randomly assigned to receive radical hysterectomy and pelvic lymph- adenectomy with or without neoadjuvant chemotherapy. A survival advantage was present only in patients with a tumor volume of greater than 60 cm3. Randomized, prospective tri- als of neoadjuvant chemotherapy followed by radiation therapy have shown either no advantage or a statistically significant reduction in survival when neoadjuvant chemotherapy is administered before radiation therapy. In light of this finding and the Operability of only 58% of the patients with stage III or IV disease, it would be prudent to limit neoadjuvant trials to patients with stage I or II disease with an initial tumor volume of greater than 60 cm3. Four such trials either are under way or are approved and await- ing activation. [Monogr Natl Cancer Inst 1996;21:93-9] For nearly 20 years, there has been intense interest in neoad— juvant chemotherapy (1). Initial trials concerned treatment of head and neck cancer. By the early 1980s, efforts were under way to explore the utility of neoadjuvant chemotherapy for can— cer of the cervix (24). The underlying rationale for neoadjuvant chemotherapy in cervical cancer included the following: uncom— promised blood flow, enhanced patient tolerance, improved radiosensitivity secondary to chemotherapy debulking, eradica- Journal ofthe National Cancer Institute Monographs No. 2|, l996 tion of subclinical metastases. and avoidance of radiation therapy— or surgery-induced chemoresistance. Early trials (3-21) served to demonstrate treatment feasibility and to evaluate toxicity and response rate. Although quantitative measurements were not done. there was the impression that neoadjuvant chemotherapy improved the ease and feasibility of radical pelvic surgery in otherwise inoperable cases (3). Since 1984. there have been numerous publications of nonrandom— ized, prospective trials evaluating neoadjuvant chemotherapy before radical surgery (4-21). Only one randomized. prospective comparative trial (22) and one nonrandomized. contempor- aneous comparison (23) of neoadjuvant chemotherapy before surgery have been reported. More information has been gathered. however. on neoad- juvant chemotherapy before radiation therapy. Several ran- domized, prospective trials have failed to show any therapeutic advantage of neoadjuvant chemotherapy in that setting (24-27). In fact, in one trial in patients with stage IIIB disease (28) and in another trial in patients with stage MB or IVA disease (29). the group receiving neoadjuvant chemotherapy had significantly worse survival. These trials have raised serious concerns about chemotherapy-induced tumor radioresistance. accelerated re- growth of tumor during protracted treatment. and interference with prompt delivery of optimal doses of radiotherapy. When neoadjuvant chemotherapy is followed by surgery. there is less concern with radioresistance or delay in radiotherapy. since only a subset of patients will receive postoperative radiation therapy. Even in these cases. most of the tumor will have been surgically removed, reducing the impact of accelerated tumor regrowth. Consequently. it is appropriate to evaluate published clinical trials involving neoadjuvant Chemotherapy before surgery in cervical cancer to see if the initial enthusiasm surrounding this approach is still justified. This review evaluates reported clinical trials of neoadjuvant chemotherapy administered before radical surgery in cervical cancer with respect to the following criteria: appropriateness of research objectives, choice of drugs. route of administration. optimal drug schedule. impact of tumor volume on results, and published proposals for future research. Appropriateness of Research Objectives Nearly all clinical reports have evaluated clinical objective response rates. toxicity of the chemotherapy, and the effect of i“(‘(il'rt‘xpmitlr'nr'0 to: Gary Lee Eddy. Sr.. M.D.. Division of Gynecologic On- cology. Department of Obstetrics and Gynecology, State University of New York—Health Science (‘cntcr at Syracuse. 736 Irving Ave. Rm. 366l—West Tower, Syracuse. NY [32“). 93 chemotherapy on subsequent surgery. Most reports have eval— uated pathologic response and have focused on the central dis— ease. parametrial disease. and regional lymphatics. The minimal duration of follow-up for most of these studies has been less than a year and sometimes only as long as the length of treat— ment ([6). Survival benefit has been assessed for the most part by comparison with historical controls either from the same in— stitution or from the literature. Of the two comparative trials. one employed nonrandomly assigned patients treated concur- rently with surgery alone (23). The other represents the only randomized. prospective trial of surgery with or without neoad— juvant chemotherapy (22). At present. there has been an adequate number of noncom- parative trials to assess the feasibility. response rate. and toxicity of this approach. What is clearly needed are additional ran— domized. prospective trials comparing this modality with cur— rent standard therapy in patients with similar extent of disease. Choice of Drugs Clinical trials involving patients with advanced or recurrent cervical cancer have provided the data required to make rational choices about which drugs to use in combination in the neoad— juvant setting. Table 1 shows response rates derived from reviews of these trials (303/) for each of the agents that have been used in neoadjuvant trials in cervical cancer. Of the drugs with extensive experience. cisplatin is clearly the leading agent. In an excellent review. Vermoken (30) noted that cisplatin com— binations were superior to cisplatin alone (38%—4l (/( versus 25%) in the treatment of advanced or recurrent cervical cancer. Table 2 shows 20 clinical trials in which neoadjuvant chemo— therapy was used before radical surgery in cervical cancer patients. All but one trial utilized cisplatin. Only four trials ex— cluded bleomycin. Only six trials excluded vinca alkaloids. Al— though ifosfamide has been utilized in several neoadjuvant trials involving radiation therapy (1632-34). only one trial employing surgery used this agent ([6). Table l. Single-agent response rates in advanced recurrent cervical cancer for drugs used in combination in ncoadjuvant trials”: No. of patients responding/ Drug total No. treated (‘/r) 23/64 ( 1(3) 238/968 (25) 35/157 (22) lipirubicin (‘isplatin lfosfamide’l‘ Mitomycin (‘ 5/23 (22) Doxorubicin 32/172 ( l9) Vincristine Ill/58 ( l7) Methotrexate 12/7} ( 1()) 30/271 (13) 36/27()(13) I9/I7otll) 2/20 ( It); (‘yclophosphamide Fl ttorou rac i | Blcomycin Vinhlasti ne *Vermokcn. 1993 (30). '1’1‘higpenetal.. 1995 (3/). The majority of studies have demonstrated acceptable toxicity from neoadjuvant chemotherapy. World Health Organization grade [—111 neurotoxicity has been reported with the use of cisplatin and the vinca alkaloids ([0,2].22). Except for the com- bination of carboplatin and ifosfamide (I6). myelosuppression has been acceptable. A few investigators (l.t’.l6.2l) have had concerns about potentially fatal bleomycin toxicity and have ex— cluded this drttg from their combinations. Chambers et al. (35) reported on 20 patients with recurrent or advanced cervical can- cer who had received a combination of bleomycin. vincristine. mitotnycin C. and cisplatin. These 20 patients. all of whom had received prior radiation therapy. received a median bleomycin dose of 90 U (range. 20—220 U). Chambers et al. reported five deaths from pulmonary toxicity (25% ). Among the fatalities. the median bleomycin dose was 130 U and the maximum dose was 220 U. Although most investigators (4-I2.I4.I5.I7.I8.2].22) have employed cisplatin and bleotnycin in combinations without difficulty in the neoadjuvant setting. in one neoadjuvant trial of 30 patients (20). two patients experienced grade 2 and two Table 2. Drugs used in neoadjuvant chemotherapy before radical surgery for cervical cancer"*‘ Authors and year (reference No.) Location Friedlander et alt. 1984 (4) Australia Valle et al.. 1985(5) Brazil Rustin et al.. 1987 (6) UK. Kirsten et al.. 1987 (7) Australia Kim et al.. 1989 (8) Korea Sardi et al.. 1990 (9) Argentina Benedetti Panici et al.. 1991 ([0) Italy Benedetti Panici et al.. 1991 (ll) ltaly Dottino et al.. 1991 (I2) Deppe etal.. 1991 ([3) Chang etal..1992(l4) Fontanelli et al.. 1992 (I5) Kuehnle et al.. 1992 (I6) United States United States Taiwan ltaly Federal Republic of Germany Carlos do Valle et al.. 1992 ( I7) Brazil Zanetta et al.. 1993 (I8) ltaly Shimamoto et al.. 1993 ( I9) Japan Sardi et al.. 1993 (22) Argentina Bloss et al.. 1995 (20) Eddy et al.. 1995 (ZI) United States United States Korea (‘isplatin Bleomycin Vincristine ()ther + + Vinblastine + + Doxoruhicin + + + Methotrexate + + Vinblastine + + Vinhlastine + + + + + Methotrexate + + + + + Mitomycin C + Mitomycin (‘ + + + + + Carhoplatin and ifosfamide + + l)oxorubicin + + + + + Peplomycin + + + + + + + + + + Vinblastine Natnkoong et al.. 1995 (23) *+ : drug used. 94 Journal ofthe National Cancer Institute Monographs No. 21. 1996 patients experienced grade 4 pulmonary toxicity, Significantly. there has been no superiority demonstrated with the use of neoadjuvant drug combinations containing bleomycin. Route of Administration There have been no reports of oral neoadjuvant chemotherapy in the treatment of cervical cancer. One reason for this is that hydroxyurea. an oral agent commonly used as a radiation sen— sitizer. has not produced any objective responses in patients with advanced or recurrent cervical cancer (36). Nearly all reports of neoadjuvant chemotherapy before surgery have in- volved agents administered intravenously or intramuscularly. Although comparable results have been reported in a few trials in which intra—arterial chemotherapy was used. there have been increased route—associated morbidity. increased locoregional toxicity. and decreased ease of access with the intra—arterial route (37,38). Optimal Drug Schedule Table 3 lists cycle characteristics. number of cycles. interval before surgery. and maximal duration of therapy in reported neoadjuvant chemotherapy trials before radical surgery. The median number of drugs employed was three (range. two to four). Chemotherapy was delivered over a median number of 4 days (range. 1-7 days per cycle). Cycles lasted from 7 to 28 days (median. 21 days). The median number of cycles administered was three (range. one to six). In these trials. chemotherapy was delivered over a range of 9-94 days (median duration. 30 days). Subsequent surgery was initiated from 7 days prior to the com- pletion of chemotherapy to 49 days following the completion of chemotherapy. There was a median interval of 19 days from completion of chemotherapy to surgery. Kim et al. (39) reported that there was a significant increase in the number of complete and partial responders with sub— sequent courses of chemotherapy from the first to third course. The median of three courses delivered by these reviewed clini— cal trials seems to acknowledge this finding. It is unknown whether additional courses of chemotherapy could increase the objective or complete response rates. Impact of Tumor Volume on Results Objective and complete response rates reported according to the International Federation of Obstetrics and Gynecology (FIGO) stage for clinical trials that used neoadjuvant chemo— therapy followed by surgery are shown in Table 4. Cumulative objective response rates of 85%. 88%. 74%. and 47% were noted for patients with stage IB and 11A. 118. III. or IVA dis- ease. respectively. Table 5 shows how these objective responses translated to stage—dependent operability. Ninety—seven percent of patients with stage 1B or IIA disease were able to undergo subsequent surgery. This compares with 74% operability for patients with stage 118 disease and 58% operability for patients with stage III or IVA disease. Neoadjuvant chemotherapy is generally less effective in patients with higher disease stage. despite significant differences between investigators. As Table 5 demonstrates. there were significant differences between studies regarding the use of postoperative radiation therapy. Overall. 55% of patients able to undergo surgery received postoperative radiation therapy. As a result. theoretical concerns raised about the adverse effect of neoadjuvant chemotherapy on the efficacy of subsequent radiation therapy apply not only to the high percentage of patients with stage III or IV disease who were still inoperable following neoadjuvant chemotherapy. but also to the majority of patients who were able to undergo radical surgery since they required postopera- Tahle 3. Variation in dose intensity of neoadjuvant chemotherapy before radical surgery (‘ycle characteristics Days of Authors and year (reference No,) chemotherapy Friedlander et al.. 1984(4) Valle et al.. 1985(5) Rustin et al.. 1987 (61 Kirsten et al.. 1987(7) Kim et al.. 1989(8) Sardi et al.. 1990(9) Benedetti Panici et al.. 1991 (/0) Benedetti Panici et al. 1991 (/l) Dottinoet al.. 1991 (/2) Deppe etal.. 1991(l3) Chang etal.. 1992(l4) Fontanelli et al.. 1992(l5) Kuehnle et al.. 1992(I6) Carlos do Valle et al.. 1992(l7) Zanetta et al.. 1993(I8) Shimamoto et al.. 1993(l9) Sardi et al.. 1993 (22) Bloss et al.. 1995(2()) Eddy et al.. 1995 (2/) J‘a'—'4J—\I—Lh’—\IJ——J—\II~)'NIJJ—J—ZI|J—n ,2 2 Interval Maximum Length of No. of before duration of cycle. days cycles surgery. days"‘ therapy. days 21 3 NS 56+ 21 3-5 NS 85+ 14 2 NS 17+ 21 3 +7? 35 21 1-5 NS 82+ 1 1 3 I4 35 21 3 NS 48+ 9 1 14—29 38 21 1 I4 35 21 3 NS 42+ 7 7-3 NS 17+ 9 1 NS 39 28 1-3 NS 56+ 21 3-5 NS 82+ 7 6 <20 54 2| 2 <21 48 l l 3 NS 21+ 1 I 3 I() 33 1 I 3 12-49 70 21 2-5 NS 94+ Namkoong et al., 1995 (23) *NS = not stated, i‘Chemotherapy was completed 7 days after surgery. Journal of the National Cancer Institute Monographs No. 21. 1996 Table 4. Effect ofdisease stage on objective and complete reponse rates for neoadjuvant chemotherapy followed by surgery Authors and year (reference No.) Friedlander et 211.. 1984 (4) Valle et a1., 1985 (5) Rustin et 111.. 1987 (6) Kirsten et 211.. 1987 (7) Kim etal.. 1989 (8) Sardi et 111.. 1990 (9) Benedetti Panici et 211., 1991 ([0) Benedetti Panici et 111.. 1991 (II) Dottinoet a1.. 1991 (I2) Deppe etal.. 1991 (1.?) Chang eta1.. 1992(14) Fontanelli et a1.. 1992 (I5) Kuehnle et 211.. 1992 (I6) Carlos do Valle et 111.. 1992(17) Sardi et 111.. 1993 (22) Bloss et a1.. 1995 (20) Namkoong et a1.. 1995 (23) Eddy (:1 211., 1995 (2]) 18. 11A 5/7 (0/7) 12/16 ((1/16) 18/20 (8/20) 8/9 (3/9) 6/6 (4/6) l0/1()(5/10) 8/10 (3/10) 28/33 (8/33) 13/19(1/19) 68/74 (30/74) 64/77 (22/77) 28/34 (2/34) Total 268/315 (86/315) Clinical objective response rate 85 (27) (clinical complete response rate) Stage* 11B 111 [VA 6/9(1/9) 9/l4(4/14) 7 7 28/350935) — — 1/1 (1/1) 1/4(0/4) 6/13(1/13) 12/17(4/17) 1/1 ((1/1) 13/15(8/15) 7 7 98/107 (28/1 ()7) 32/44 (5/44) -— 31/35(5/35) 23/31 (3/31) 7 11/11(1/11) 6/9(0/9) 7 7/7(3/7) 1()/1()(2/1()) 1/1(()/I) 4/4 (2/4) 1/3 ((1/3) 7 8/8(2/8) — 7 2/210/2) 10/17(2/17) 2/5(1/5) 7 16/21(3/21) 7 18/l9(0/19) 6/7(()/7) 2/4(()/4) I2/15(4/15) — — 216/245 (55/245) 154/209 (32/209) 7/15 (1/15) 88(22) 74(15) 47(7) *The number ofclinical objective responses divided by the number of assessable patients is listed first. The number of clinical complete responses divided by the number ofassessable patients is listed in parentheses. tive radiation therapy. Patients with bulky stage 1B. 11A. or 11B disease would appear to be preferable candidates for neoad- juvant chemotherapy before surgery. since they have the highest rates of operability. Although several authors (59,10) have suggested that sur- vival is improved in patients receiving neoadjuvant chemother- apy before surgery, there have been only two contemporaneous comparative studies (22,23). Sardi et a1. (22) reported in 1993 on 74 patients with stage IB disease who received vincristine. cisplatin. and bleomycin prior to radical hysterectomy and bilateral pelvic lymphadenectomy. When compared with a similar group of patients prospectively. randomly assigned to treatment with surgery alone, superior survival rates were noted only when the initial tumor volume was 60 cm} or greater (87% versus 68%; P = .008). This finding further shrinks the patient population who could potentially benefit from neoadjuvant chemotherapy. Of note. all patients in this trial received post- operative radiation therapy. It is possible that preoperative neoadjuvant chemotherapy would be of greater benefit if only patients at high risk received postoperative radiation therapy. In Table 5. Effect of initial disease stage on operability following neoadjuvant chemotherapy Stage* Postoperative More Authors and year (reference No.) 1B. 11A 11B 111. [VA RTr than R113: Valle et a1., 1985 (5) 0 0 35/35 ()/35 No Kirsten et 211.. 1987 (7) 10/16 |/13 4/18 4/15 No Rustin et a1., 1987 (6) 1/1 0 0/8 0/1 No Kim et a1., 1989 (8) 31/35 19/19 0 8/54 No Sardiet a1.. 1990(9) 0 76/107 18/44 94/94 No Benedetti Panici et 111.. 1991 (10) 8/9 31/35 23/31 <18/62 Yes Benedetti Panici et a1.. 1991 (II) 6/6 10/1 1 5/9 NS Yes Dottino et a1. 1991 (12) 10/10 7/7 11/11 14/26 Yes Deppe et 111.. 1991 ([3) 8/10 1/4 1/3 10/10 No Chang et a1, 1992 ([4) 33/33 0 () 6/33 No Fontanelli et 211.. 1992 (I5) 18/19 7/8 0 4/26 No Carlos do Valle et 211.. 1992 (I7) 0 0 14/21 ()/14 No Sardi et £11., 1993 (22) 74/74 0 0 74/74 No Bloss et al.. 1995 (20) 0 10/19 0/11 10/10 No Namkoong et al.. 1995 (2.1) 77/77 15/15 0 NS No Eddy et al.. 1995(21) 31/31 0 0 8/31 No Total (‘7r) 312/321 (97) 177/238 (74) 111/191 (58) 232/425 (55) No *Each fraction represents the number of patients undergoing radical surgery divided by the number of patients entered in the study. (Each fraction represents the number of patients receiving postoperative radiation therapy (RT) divided by the number of patients undergoing radical surgery. NS = not stated. iAdditional surgery included type IV. radical hysterectomy (RH). rectosigmoid resection. or pelvic exenteration. 96 Journal of the National Cancer Institute Monographs No. 21. 1996 1995, Namkoong et a1. (23) compared 77 patients with stage 1B or “A disease and 15 patients with stage “8 disease who were treated with neoadjuvant chemotherapy prior to radical surgery with a similar group of patients treated with surgery alone at the same institution and during the same time period. Only patients at high risk received postoperative radiation therapy. The inves- tigators reported that 82% of the patients receiving neoadjuvant chemotherapy were free of recurrence at 4 years compared with 67% of the patients treated with surgery alone (P = .007). It must be emphasized, however, that this trial was a nonran- domized comparison. The results of these two comparative trials of neoadjuvant chemotherapy before surgery are surprising, given the results of much more numerous comparative trials of neoadjuvant chemo- therapy before radiation therapy. When neoadjuvant chemo- therapy is administered before radiation therapy, there has been no survival difference in most trials (24-27), and there has been reduced survival in some trials (28,29). It is striking that the two comparative trials of neoadjuvant chemotherapy before surgery both demonstrated a survival advantage. Serious design flaws in both studies prevent definitive conclusions. In the study by Sardi et al. (22), a statistical advantage was found only in a sub— set of patients; all patients had received postoperative radiation therapy, and a large number of them had relatively small tumors (34 cm in diameter). The study by Namkoong et a1. (23) com- pared nonrandomly assigned patients. The group studied had received neoadjuvant chemotherapy before surgery. The control patients were treated during the same period of time and analyzed retrospectively. Such a study design is vulnerable to significant entry bias. It has been commonly noted that patients receiving neoad- juvant chemotherapy have a lower than expected incidence of pelvic lymph node metastases (8,]0-14.18.20.22). Table 6 lists the incidence of lymph node metastases in 10 assessable trials. Metastases were present in only 15% of the patients with stage IE or 11A disease, 21% of the patients with stage IIB disease, and 41% of the patients with stage III or IV disease. By com- parison. rates of pelvic lymph node metastases of 3l%—80% have been cited for patients with bulky stage [B or IIA disease treated with surgery alone (40—44). Bulky stage 18 cancer has been variably defined as a cancer with a diameter greater than 2, 3, 0r 4 cm. Nonetheless. these are the same variable criteria employed by investigators using neoadjuvant chemotherapy before surgery. It is remarkable that the cumulative incidence of lymph node metastases derived from these neoadjuvant chemo- therapy trials for patients with stage 1B or llA disease is less than half of the lowest number reported in the literature for patients treated with surgery alone. The frequency of pathologic complete responders among patients receiving neoadjuvant chemotherapy is shown in Table 7. Pathologic complete response rates of 9% for patients with stage [B or “A disease and 6% for patients with stage [11 or IV disease were achieved. There is a clear suggestion that neoad- juvant chemotherapy may be effective in eliminating micro- metastases. Only long-term. randomized studies, however, will be able to demonstrate improvement in survival with the use of neoadjuvant chemotherapy. Journal ofthc National Cancer Institute Monographs No. 21. 1996 Table 6. Frequency of lymph node metastases following neoadjuvant chemotherapy at time of radical surgery Surge" Authors and year .22 (reference No.) 1B. 11A IIB lll.1V Kim et al.. 1989 (8) 7/35 3/19 —— Benedetti Panici et al.. 1991 (I0) — 1/11 1/6 Benedetti Panicieta1.. 1991 (II) ()/8 7/311 8/23 Dottino etal.. 1991 (12) 0/10 4/7 5/11 Chang et al.. 1992(l4) 6/33 — — Fontanelli et al.. 1992 ( [5) 3/19 l/7 — Zanetta et al., 1993 ([8) 11/19 A — Sardi et al.. 1993 (22) 5/74 ~ — Bloss et al.. 1995 (20) NS NS 7/ll Eddy etal., 1995(2l) 3/31 — ~ Total(%) 35/230(15) 16/75 (21) 21/51 (41) *Each fraction represents the number of patients with histologically confirmed metastases to the pelvic lymph nodes divided by the number of patients undergo- ing lymphadenectomy. NS = not stated. On the basis of the observation that there is a group of patients with marked response to neoadjuvant chemotherapy, Sardi et al. (45) have proposed the use of neoadjuvant chemo— therapy prior to simple hysterectomy for select patients with stage IB disease. These patients must have initial tumor volume less than 40 cm}. Their initial biopsies must show high nuclear grade, high mitotic index, and prominent lymphoplasmo- monocytic infiltration. If following chemotherapy there is nega- tive ectocervical cytology, normal colposcopy and cervical endoscopy, and negative endocervical curettage, the patient would undergo a total hysterectomy, bilateral salpingo-oophor- ectomy, upper vaginectomy, and pelvic lymphadenectomy. Published Proposals for Future Research As mentioned above, Sardi et a1. (45) have proposed that patients with small stage IB disease who optimally respond to chemotherapy and who have favorable histologic features on Table 7. Frequency of pathologic complete response following neoadjuvant chemotherapy at time of radical surgery Stage* Authors and year ' " —* (reference No.) 18. [IA IIB 111, IV Friedlander et al.. 1984 (4) 0/7 0/90 0/14 Valle et al.. 1985 (5) — — 5/35 Rustin etal.. 1987(6) 1/1 NS NS Kirsten et al.. 1987 (7) 0/16 0/13 0/18 Benedetti Panici et al.. 1991 (ll) 4/6 1/11 0/9 Dottinoet al.. 1991 (I2) 2/10 1/7 1/10 Deppe et al.. 1991 (/3) 2/10 0/4 ()/3 Chang et al.. 1992 ( [4) 2/36 — — Fontanelli et al.. 1992 (I5) ()/19 2/8 — Sardi et al.. 1993 (22) 7/74 — — Bloss et al.. 1995 (20) i ()/19 0/11 Eddy et al.. 1995 (2]) 0/31 — — Total (f/t) 18/210 (9) 4/71 (6) 6/100 (6) *Each fraction represents the number of patients without residual cancer in the surgical specimen (including lymph nodes) divided by the number of patients undergoing surgery. NS = not stated. 97 Table 8. Ongoing randomized. prospective trials of neoadjuvant chemotherapy (NACT) before radical surgery (RS) reported in the literature Authors and year (reference No.) Disease stage Kim et al.. 1989(8) 18. ll Benedetti Panici et al.. 1991 (IU) Locally advanced Chang etal.. 1992(I4) lB(>-l cm). 11A Carlos do Valle et al.. 1992 ( I7) 111 Zanetta et al.. 1993 ( I8) Locally advanced Sardi et al.. 1993 (22) 18 (>2 cm) Eddy et al.. 1995+ (21) 18 (>3 cm) Location Design* Korea NACT and RS versus RS ltaly NACT and RS versus RT Taiwan NACT and RS versus RS Bra/.il NACT and RS and postoperative CT versus RT ltaly NACT and RS versus RT Argentina NACT and RS versus RS United States NACT and RS versus RS *RT : radiation therapy. CT = chemotherapy. (Awaiting activation their initial biopsies should undergo simple hysterectomy. It is conceivable that. should this program be effective. there may be a subgroup of patients who strongly desire future fertility who could be treated with neoadjuvant chemotherapy and cervical conization with pelvic lyrnphadenectomy. Table 8 describes randomized clinical trials employing in- travenous neoadjuvant chemotherapy before surgery; these studies have been reported either as under way or as approved and awaiting activation. Four (8,14,21,22) of these seven trials (8.I().l4.l7,l8,21.22) involve patients with bulky stage IB dis— ease and compare radical surgery with or without neoadjuvant chemotherapy. Two of these protocols (8J4) included patients with stage [1 disease. Bulky stage IB is defined as tumor di— ameter greater than 2 cm (22). greater than 3 cm (2]). or greater than 4 cm ([4). Considering the observation by Sardi et al. (22) that survival advantage could be demonstrated only when initial tumor volume was 60 cm3 or greater. it would seem prudent to limit entry to patients with tumor diameters greater than 3-4 cm. as suggested by Sardi et al. (22). Reduced objective response rates and operability in patients with stage 111 and [VA disease suggest restricting neoadjuvant chemotherapy to patients with a tumor diameter of less than 8 cm regardless of stage. Three other trials ([0. l8) involve using neoadjuvant chemotherapy and radical surgery for treatment of patients with locally advanced disease and comparing them with patients treated with radiation therapy alone. Park et al. (37) have activated a randomized trial in which treatments with intra—arterial chemotherapy and radical surgery. intravenous chemotherapy and radical surgery. and radiation therapy alone are compared in patients with stage 118 disease. It is hoped that this study will determine if the increased morbidity and difficulty in vascular access of intra—arterial chemotherapy can be compensated for by increased disease—free interval and survival. References (I) Horwich A. (‘ombined radiotherapy and chemotherapy in clinical oncol— ogy. London (UK): Edward Arnold. 1992. (2 Sardi Jli. di Paola GR. Cachau A. Orti/ ()C. Sanancs C. Giaroli A. et al. A possible new trend in the management of the carcinoma of the cervix uteri. Gynecol ()ncol 1986;25:139-49. (3) Benedetti Panici P. Scambia G. Greggi S. Di Roberto P. Baiocchi G. Mancuso S. Neoadjuvant chemotherapy and radical surgery in locally ad- \ anccd cervical carcinoma: a pilot study. ()bstet Gynecol 1988171 :344-8. (4) Friedlander ML. Atkinson K. Coppleson JV. Elliot P. Green D. Houghton R. et al. The integration of chemotherapy into the management of locally advanced ccr\ica| cancer: a pilot study. Gynecol ()ncol 1984:1911-7. 98 (5) (7 (8 (9 ([0 (II) (151 ([6) (I7) (I9) (20) Valle JC. Relcnder MR. Werneck C. Chu C. Figueiredo E, Neoadjuvant 84 adjuvant chemotherapy with adriamycin. bleotnycin & cisplatin (ABC) & modified radical hysterectomy in the cancer of cervix. Stage [11 (abstract). Proc ASCO 198514: 125. Rustin GJ. Newlands ES. Southcott BM. Singer A. Cisplatin. vincristine. methotrexate and bleornycin (POMB) as initial or palliative chemotherapy for carcinoma of the cervix, BrJ ()bstet Gynaecol 1987194: 1205-1 1. Kirsten F. Atkinson KH. Coppleson JV. Elliott PM. Green D. Houghton R. et al. Combination chemotherapy followed by surgery or radiotherapy in patients with locally advanced cervical cancer. BrJ Obstet Gynaecol 1987: 94:583-8. Kim DS. Moon H. Kim KT. Hwang YY. Cho SH. Kim SR. Two-year sur- vival: preoperative adjuvant chemotherapy in the treatment of cervical can- cer stages lb and 11 with bulky tumor. Gynecol Oncol 1989;33:225-30. Sardi .l. Sanancs C. Giaroli A. Maya G. di Paola G. Neoadjuvant chemo- therapy in locally advanced carcinoma of the cervix uteri. Gynecol ()ncol 1990238248633. Benedetti Panici P. Scarnbia G. Baiocchi G. Greggi S. Ragusa G. Gallo A. et al. Neoadjuvant chemotherapy and radical surgery in locally advanced cervical cancer. Prognostic factors for response and survival, Cancer 1991: 67:3729 Benedetti Panici P. Greggi S. Scarnbia G. Ragusa G. Baiocchi G. Battaglia F. et al. High—dose cisplatin and bleornycin neoadjuvant chemotherapy plus radical surgery in locally advanced cervical carcinoma: a preliminary report. Gynecol()nco1 I991 1411212-6. Dottino PR. Plaxe SC. Beddoe AM. Johnston C. Cohen CJ. Induction chemotherapy followed by radical surgery in cervical cancer. Gynecol Oncol 1991;4(117-11. Deppe G. Malviya VK. Han l. Christensen CW. Malone JM Jr. Kim Y. et al. A preliminary report of combination chemotherapy with cisplatin and mitomycin-C followed by radical hysterectomy or radiation therapy in patients with locally advanced cervical cancer. Gynecol ()ncol 1991;42: 178% 1. Chang HC. Lai CH. Chou PC. Tseng CJ. Chang TC. Hsueh S. et al. Neoadjuvant chemotherapy with cisplatin. vincristine. and bleornycin and radical surgery in early-stage bulky cervical carcinoma. Cancer Chemother‘ Pharmacol 1992;30:281-5. Fontanelli R. Spatti G. Raspagliesi F. Zunino F. Di Re 1". A preoperative single course of high~dose cisplatin and bleornycin with glutathione protec— tion in bulky stage 18/” carcinoma of the cervix [we comment citation in Medline], Ann ()ncol 1992:32117-21, Kuehnle H. Meerpohl HG. Eiermann W. AchterTath W. Neoadjuvant therapy for cervical cancer. Semin Oncol 1992:19( 1 Suppl 2):94—8. Carlos do Valle J. Chu C. Werneck Ribeiro C. Re/ende MC. Figueircdo 1i. Treatment of cancer of cervix stage 111 with adriamycin. bleornycirr and cisplatin neoadjuvant. modified radical hysterectomy and adriamycin. bleornycin and cisplatin adjuvant: eight years' follow—up (abstract). Proc ASCO 199211 1:229. Zanetta G. Landoni F. Colombo A. Pellegrino A. Maneo A. chentis C. Three—year results after neoadjuvant chemotherapy. radical surgery. and radiotherapy in locally advanced cervical carcinoma. Obstet Gynecol 1993;82:447-50. Shirnamoto T. Tateyama ll. Kohso T. Hamada M. Neoadjuvant chemo— therapy using a platinum. vincristine and peplomycin (CVP) combination in patients with carcinoma of the uterine cervix. Asia-Oceania J ()bstet Gynaecol 1993;19:261—9, Bloss JD. Ltrcci JA 3rd. DiSaia PJ. Manetta A. Schiano MA. Ramsinghani N. et al. A phase 11 trial of neoadjuvant chemotherapy prior to radical hysterectomy and/or radiation therapy in the management of advanced car'- cinorna of the uterine cervix. Gynecol Oncol 1995;59:105-10. Journal of the National Cancer Institute Monographs No. 21. 1996 (2/) Id CA. (24) (26) (27) (30) (3/) (32 Eddy GL. Manetta A. Alvarez RD. Williams L. Creasman WT. Neoad- juvant chemotherapy with vincristine and cisplatin followed by radical hysterectomy and pelvic lymphadenectomy for FIGO stage IB bulky cervi- cal cancer: a Gynecologic Oncology Group pilot study. Gynecol ()ncol 1995;57:412-6. Sardi J. Sananes C. Giaroli A. Bayo J. Rueda NG, Vighi S. et al. Results of a prospective randomized trial with neoadjuvant chemotherapy in stage IB. bulky. squamous carcinoma of the cervix [MT comment citations in Med~ line]. Gynecol Oncol I9931491156-65. Namkoong SE. Park JS. Kim JW. Bae SN. Han GT. Lee JM. el al. Com- parative study of the patients with locally advanced stages I and ll cervical cancer treated by radical surgery with and without preoperative adjuvant chemotherapy. Gynecol Oncol 199559: | 36-42. Chauvergne J. Rohart J. Heron JF. Fargeot P. Berlie J. David P. et al. Ran- domized phase III trial of neoadjuvant chemotherapy (CT) and radio- therapy (RT) vs RT in stage MB. III carcinoma of the cervix (CACX): a cooperative study of the French Oncology Centers (abstract). Proc ASC() 1988:7136. Tattersall MH. Ramirez C. Coppleson M, A randomized trial comparing platinunvbased chemotherapy followed by radiotherapy vs radiotherapy alone in patients with locally advanced cervical cancer, Int .l Gynecol Oncol l9‘)2:2:244-51. Cardenas J. Olguin A. Figueroa F. Becerra F. Huizar RM. Randomi/cd neoadjuvant chemotherapy in cervical carcinoma Stage [18. PEC + RT versus RT (abstract). Proc ASCO 1991;10:620. Sardi J. Early results of a randomized trial with neoadjuvant chemotherapy in squamous carcinoma cervix uteri (abstract). Proceedings of the 3rd meeting of the International Gynecologic Cancer Society. I99]. vol 40. Souhami L. Gil RA. Allan SE. Canary PC. Araujo CM. Pinto LH. et al. A randomized trial of chemotherapy followed by pelvic radiation therapy in stage IIIB carcinoma of the cervix. J Clin Oncol 199! 19:970-7. Tattersall MH. Lorvidhaya V. Vootiprux V. Cheirsilpa A. Wong F. A/har T. et al. Randomized trial of epirubicin and cisplatin chemotherapy fol- lowed by pelvic radiation in locally advanced cervical cancer. Cervical Cancer Study Group of the Asian Oceanian Clinical Oncology Association [rec comment citation in Medline]. J Clin Oncol 1995;13:444-5 l. Vermoken JB. The role of chemotherapy in squamous cell carcinoma of the uterine cervix: a review, lnt J Gynecol Cancer 1993:}: I 29-42. Thigpen JT. Vance R. Puneky L. Khansur T. Chemotherapy as a palliative treatment in carcinoma of the uterine cervix. Semin Oncol l‘)95:22(2 Suppl 3): 16-24. Lara PC. Garcia-Puche JL. Pedraza V. Cisplatin-ifosfamide as neoadjuvant chemotherapy in stage IIIB cervical uterine squamous cell carcinoma. Cait- cer Chemother Pharmacol [990:20 Suppl:S36-8. Journal ofthe National Cancer Institute Monographs No. 2 l. 1996 (33) (34 (35) (36) (37) (38) (39) (40) (4f) (44) Buxton EJ. Experience with bleomycin. ifosfamide. and cisplatin in pri- mary and recurrent cervical cancer. Semin Oncol 1992; 19(2 Suppl 5):9— |7:discussion 17-8. Nevin J. Bloch B. VanWijk L. Soeters R. Dehaeck K. Ball H. et al. Primary chemotherapy with bleomycin. ifosfamide and cisplatinum (BlP) followed by radiotherapy in the treatment of advanced cervical cancer. A pilot study. Eur] Gynaecol ()ncol l995: [6:30-5. Chambers SK. Flynn SD. Del Prete SA. Chambers JT. Schwartz PE. Bleomycin. vincristine. mitomycin—C. and cis—platinum in gynecologic squamous cell carcinomas: a high incidence of pulmonary toxicity. Gynecol Oncol 1989;32:303-9. Thigpen T. Vance RB. Balducci L. Blessing .l. Chemotherapy in the management of advanced or recurrent cervical and endometrial carcinoma. Cancer 198 l :480 Suppl):658—65. Park SY. Kim BG. Kim JH. Lee JH. Lee ED. Lee KH. et al. Phase III] study of neoadjuvant intraarterial chemotherapy with mitomycin-C. \incrisline. and cisplatin in patients with stage llb bulky cervical car— cinoma. Cancer |995176z8H-23. ltoh N. Sawairi M. Hanabayashi T. Mori H. Yamawaki Y. Tamaya T. Neoadjuvant intraarterial infusion chemotherapy with a combination of mitomycin-C. vincristine. and cisplatin for locally advanced cervical can— cer: a preliminary report. Gynecol Oncol l992:47:391—4. Kim DS. Moon H. Hwang YY. Cho SH. Preoperative adjuvant chemo» therapy in the treatment of cervical cancer stage lb. lla. and [lb with bulky tumor. Gynecol Oncol [98822932 | «32. Piver MS. Chung WS. Prognostic significance of cervical lesion size and pelvic node metastases in cervical carcinoma. Obstet Gynecol |975z46: 507—“). Van Nagel JR Jr. Donaldson ES. Parker JC. Van Dyke AH. Wood EC. The prognostic significance of cell type and lesion size in patients with cervical cancer treated by radical surgery. Gynecol Oncol I977:5: 141—51. Chung CK. Nahhas WA. Stryker JA. Curry SL. Abt AB. Mortel R. Analysis of factors contributing to treatment failures in stage IB and [IA carcinoma of the cervix. Am J ()bstet Gynecol 1980;138:550-6. Fuller AF Jr. Elliott N. Kosloff C. Lewis JL Jr. Lymph node metastases from carcinoma of the cervix. stage IB and HA: implications for prognosis and treatment. Gynecol ()ncol l982;|3:165-74. lnoue T, Prognostic significance of the depth of invasion relating to nodal metastases. parametrial extension. and cell types. A study of 628 cases with Stage [8. HA. and 11B cervical carcinoma. Cancer 1984;54:3035-42. Sardi J. Sananes C. Giaroli A. Gomez Rueda N. Di Paola G. Vighi S. et al. ls subradical surgical treatment for carcinoma of the cervix uteri stage IB logical? Gynecol ()ncol 19893223604. 99 Concomitant and Neoadjuvant Chemotherapy in Conjunction With Radiotherapy in the Management of Locally Advanced Cervical Cancer Martin H. N. Tattersall* Radiotherapy is standard treatment for women with locally advanced cervical cancer (stage IIB-IVA). Radiotherapy fails to control disease progression within the irradiated field in more than 40% of patients. The disease bulk is a major factor limiting the curative probability of pelvic radiotherapy. Chemotherapy has been integrated with pel- vic radiotherapy with the goal of improving local tumor con- trol and treating microscopic metastases outside the radiotherapy field. Simultaneous chemotherapy and radio- therapy, neoadjuvant chemotherapy followed by radio- therapy, and adjuvant chemotherapy after radiotherapy, have been investigated. Hydroxyurea during pelvic radiotherapy has been reported to be associated with an im- proved progression-free (P = .05) and survival (P = .066) advantage. Several randomized trials of concomitant radio- therapy with fluorouracil, mitomycin, or cisplatin have been completed, but no results are published. Neoadjuvant chemotherapy with cisplatin-containing regimens followed by radiotherapy has been studied in several randomized trials. Although tumor response after chemotherapy is reported in more than 50% of patients, no improvement in local disease control or in survival has been noted overall. In one large trial, neoadjuvant chemotherapy was inferior to standard pelvic radiotherapy in terms of local control and survival. Adjuvant chemotherapy after radiotherapy has been little studied, and no randomized trials have been reported. The results of trials in progress or completed but not yet reported will determine further clinical research directions. Chemotherapy during and after pelvic radiation therapy has been little studied. [Monogr Natl Cancer Inst 1996;21:101-3]. Primary pelvic radiotherapy is the standard treatment ap- proach in women with stage llB-IVA cervical cancer. Complete regression of local disease is obtained in about 80% of the women with stage llB disease, but with increased disease stage, and as tumor size increases, the probability of complete tumor regression falls. Complete response is achieved with radio- therapy in only about 50% of the women with stage III disease. The risk of disease recurrence or the development of metastatic disease is highest in the first 2-3 years after treatment, and less than 5% of recurrences occur after 5 years. The most common site of treatment failure after pelvic radiotherapy is within the Journal of the National Cancer Institute Monographs No. 21, 1996 pelvis, but relapse in the para-aortic and/or the supraclavicular lymph nodes is common; systemic disease also occurs, however, most often in association with progression in the pelvis or in lymph nodes. Despite developments in megavoltage radiation equipment, interstitial radiation therapy, and after-loading techniques, in- cluding high—dose-rate remote applicators, the results of radia- tion therapy in locally advanced cervical cancer have not improved recently. Hence, a variety of approaches have been in- vestigated with the goal of improving the treatment outcomes in women with locally advanced cervical cancer. Concomitant Chemotherapy and Radiotherapy Several chemical approaches have been investigated, with the goal of improved treatment outcomes in patients with locally ad- vanced cervical cancer. The use of chemicals that are hypoxic cell sensitizers or that enhance the antitumor action of ionizing irradiation has been one major field of endeavor. Hydroxyurea was found to perturb the cell cycling characteristics of sensitive tumor cells, killing those cells in S phase, and causing surviving cells to accumulate at the Gl-S boundary, a relatively radiosen- sitive portion of the mitotic cell cycle. These observations prompted a study (1) in tumor—bearing mice, and it was reported that hydroxyurea in conjunction with irradiation improved tumor responses compared with irradiation alone. Subsequently, Piver et al. (2) reported a small double—blind clinical trial in women with stage IIB and 1118 cervical cancer in which hydroxyurea or placebo was given during the radiation treat- ment. Encouraging results in this study prompted an expansion of the number of patients included in the trial, and in 1977, Piver et al. (3) reported the outcome in a total group of 148 women with stage 11B and IIIB cervical cancer, 75 of whom were surgically staged. Hydroxyurea (80 mg/kg) was given on the first day of radiation therapy and was continued every third day for 12 weeks. The results as reported do not present the out- come in the total population according to randomization. Sub- group analyses, however, of assessable patients with stage IIB disease without proof of para—aortic lymph node metastases reported that 20 of 27 patients who received hydroxyurea were *Correspondence to: Martin H. N. Tattersall, M.D., Department of Cancer Medicine, Blackburn Bldg, D06. University of Sydney. Sydney, New South Wales. 2006, Australia. l0] disease free at 2 years compared with 17 of 39 of the placebo group (P<.0]). In women with stage IIB disease who were not surgically staged, 12 of 18 were disease free at 2 years in the hydroxyurea group compared with nine of 24 in the placebo group (P = .05). The validity of these subgroup analyses and the conclusions of this report that hydroxyurea acts as a radiation sensitizer are questionable. The Gynecologic Oncology Group (GOG) has reported a trial of hydroxyurea at a dose of 80 mg/kg every third day during radiation therapy or for a period of 12 weeks versus placebo (4). One hundred ninety women with stage 111 and WA disease were entered in the study, but 26 patients were subsequently judged ineligible and 60 were judged inevaluable. Only 97 of the 190 women randomly assigned were judged assessable for disease progression and survival analysis. The reports of this trial do not present the outcome according to randomization, although in those assessable for progression and survival, a significant ad— vantage in terms of complete tumor regression, time to disease progression. and survival was noted. The development of hypoxic cell radiosensitizers in the 1970s led to the clinical evaluation of misonidazole as a radiation sen- sitizer in patients with a variety of cancer types. Leibel et a1. (5) have reported the results of a randomized trial of radiotherapy with or without misonidazole for patients with stage 1118 or lVA squamous cell carcinoma of the cervix. Misonidazole was ad— ministered orally 2-4 hours before each radiation fraction, and 30 doses of 400 mg/m2 each were given, for a total misoni— dazole dose of 12 g/m2. One hundred twenty—one patients were entered in the study. The accrual was terminated. however, prior to reaching the original sample size of 150, when interim analyses showed that the number ()f deaths observed among patients treated with misonidazole was inconsistent with the hypothesis that radiation therapy plus misonidazole improved survival by 15%—20% and so the study was unlikely to demon— strate a benefit in favor of the combined treatment with addi— tional accrual or follow—up. Tumor response rate was similar for the two treatment groups, as was the pattern of treatment failure. Toxicity was not significantly enhanced in the misonidazole— treated group, and there was no difference in survival or progression-free survival. The Medical Research Council of the UK. has similarly found no significant difference in tumor control, survival, or complications in 153 patients with stage III cervical cancer ran— domly assigned to receive radiation therapy and misonidazole compared with radiation therapy and placebo. It was concluded that misonidazole in the dosage used was an inadequate sen- sitizer of hypoxic cells (6). These disappointing results have prompted a re-evaluation of the pharmacokinetics of nitroimidazole compounds that have hypoxic cell sensitizing properties. The GOG has compared hydroxyurea versus misonidazole as radiosensitizing treatment in patients with surgically staged 118- [VA cervical cancer (and negative para-aortic nodes) (7). Three hundred eight patients were entered in a randomized study, and 14 (4.5%) were declared ineligible. The outcomes according to randomization in the remaining 294 patients were as follows: Hydroxyurea—treated patients had an improved progression-free survival (P = .05) and survival (P = .066) advantage. These trial 102 results underpin the GOG decision to accept radiotherapy and hydroxyurea as standard therapy against which other approaches are now compared. Rose (8) has reviewed the role of chemoradiation in advanced cervical cancer. A number ofdrugs were noted to have radiosen- sitizing properties, including hydroxyurea, fluorouracil, cis— platin, and mitomycin. A number of phase 1 and 11 trials of combined chemoradiation using these agents have been reported with, in some cases, very impressive outcomes. Unfortunately, no randomized trials of radiation therapy with agents other than hydroxyurea have been published, although several studies have completed accrual. The results of the completed studies will be extremely important in determining whether there is merit in chemosensitizing radiation. Even if the results are negative, the experience gained combining chemotherapy and radiotherapy may be advantageous in future trials in which chemotherapy may be given before and after pelvic radiotherapy. N eoadjuvant or Primary Chemotherapy in Locally Advanced Cervical Cancer During the past 15 years, many reports have documented tumor response following chemotherapy in patients with meta- static cervical cancer. With a variety of single agents, tumor response rates have ranged between 15% and 50%. It has been common, particularly in the last few years, to use cisplatin in combination with various other agents, and the accumulated ex— perience suggests that improved tumor response rates are seen compared with those that would have been expected with cisplatin treatment alone. In patients with metastatic cervical cancer, chemotherapy is palliative in intent, and few patients achieve a complete response; some patients, however, have very prolonged tumor response lasting Lip to several years. The antitumor activity, particularly of cisplatin—based combi- nation chemotherapy regimens, led to trials of primary chemotherapy in patients with locally advanced cervical cancer. An early report (9) indicated that tumor regression was seen in 50%—70% of patients and that subsequent pelvic radiotherapy was not complicated in most instances by this prior chemo— therapy. A report (10) of primary chemotherapy in locally ad— vanced cervical cancer suggests that tumors refractory to chemotherapy commonly were also relatively radioresistant, a result also reported in squamous cancers in the head and neck region (10). Concerns about delaying radiotherapy to complete certain cycles of chemotherapy led to a number of groups inves— tigating short courses of chemotherapy given over 2-6 weeks prior to pelvic radiotherapy, or in some cases, surgery. These short-course treatment programs have been associated with tumor response rates by 4 weeks in the range of 50% and ap— parently enhanced tumor control rates with subsequent radio- therapy ([0). Several randomized trials of primary chemotherapy in locally advanced cervical cancer have been published in the last few years. Many trials document tumor response following chemo- therapy in more than 50% of the patients, but none of the trials has reported improved tumor control rate, time to disease pro- gression, or survival. One of the larger trials undertaken in Asia showed a significant difference between the treatment arms, Journal ofthe National Cancer Institute Monographs No. 21, 1996 with standard radiotherapy having an improved local control rate and survival compared with the group treated with neoad— juvant or primary epirubicin and cisplatin (II). A number of trials of neoadjuvant or primary chemotherapy are still accruing patients. and some of these include short- course chemotherapy regimens that conceivably may have different biologic implications with regard to the effects of chemotherapy on tumor cell repopulation rates after drug—in— duced tumor shrinkage (/2). The overall treatment time—that is. time from starting chemotherapy until commencement and/or completion of radiotherapy——tnay be an important consideration in the optimal integration of primary chemotherapy and radio— therapy. At least two ongoing trials are starting radiotherapy within 30 days of initiating chemotherapy. and the results of these studies will clarify whether short—course neoadjuvant chemotherapy is an effective strategy. So far. none of the trials that have been reported has continued chemotherapy after corn- pletion of radiotherapy. and this strategy is worthy of investiga- tion in randomized trials. The reports of primary chemotherapy followed by radio- therapy have documented no major enhancement of radiation complications. but long—term follow-up is limited. Adjuvant Chemotherapy During and After Radiotherapy Several authors have advocated studies of this approach, but no randomized trials have been reported. The GOG is undertak— ing trials that compare radiotherapy and hydroxyurea versus radiotherapy plus cisplatin/fluorouracil. radiotherapy plus hy— droxyurea versus radiotherapy plus hydroxyurea plus cisplatin. and fluorouracil versus radiotherapy plus weekly cisplatin. Asian investigators (personal communication) are planning a trial of radiotherapy versus radiotherapy plus tluorouracil with or without adjuvant epirubicin/cisplatin and cisplatin/ifos- famide. The results of these trials will clearly be influential in directing future research. These trials may have the greatest chance of accruing adequate patient numbers in those areas of the world where locally advanced cervical cancer is a common disease (e.g.. Latin America and parts of Asia). Conclusion Radiotherapy remains the standard treatment approach in the treatment of women with locally advanced cervical cancer. Hydroxyurea given every third day for 12 weeks during radia- tion therapy has been shown to improve disease-free survival in a randomized trial. and the GOG has adopted this treatment as the standard against which new strategies are being compared. Journal ot'the National Cancer Institute Monographs No. 21. I996 Elsewhere. chemoradiation is not accepted as a standard ap- proach. Current trials of concomitant chemotherapy and radia- tion therapy in locally advanced cervical cancer are including a number of drugs that may have radiosensitizing properties and/or be active therapy for cervical cancer. The outcomes of these trials will be important in determining whether combined chemoradiotherapy becomes a standard approach. Primary chemotherapy in patients with locally advanced cervical cancer has yielded disappointing results. possibly because of altered tumor cell repopulation rates following tumor reduction caused by chemotherapy. Short—course primary chemotherapy regimens or concomitant chemoradiotherapy may overcome these ob- stacles and lead to improved outcome. The status of adjuvant chemotherapy after primary radiotherapy has not been inves- tigated. References (I) Piver MS. Howcs Ali. Suit Hl). Marshall N. Effect of hydroxyurea on the radiation response of C3” mouse mammary tumors. Cancer 1972;29:407— l2. (2) Piver MS. Barlow JJ. Vongtama V. Webster J. Hydroxyurea and radiation therapy in advanced cervical cancer. Am J Obstet Gynecol |974;|2():969- 72. (3) Piver MS, Barlow JJ. Vongtama V. Blumenson L. Hydroxyurea as a radia- tion sensiti/er in women with carcinoma of the uterine cervix. Am J Obstet Gynecol l977;12‘):379—83. (4) Hreschchyshyn MM. Aron BS. Boronow RL. Franklin EW 3rd. Shingleton HM. Blessing JA. Hydroxyurea or placebo combined with radiation to treat stages NIH and IV cervical cancer confined to the pelvis, Int J Radiat Oncol Biol Phys |9791513l7-22. (5) Leibel S. Bauer M. Wasscrlnan T. Marcial V. Rotman M. Hornback N. et al. Radiotherapy with or without misonidamle for patients with stage [118 or stage IVA squamous cell carcinoma of the uterine cervix: preliminary report of a Radiation Therapy Oncology Group randomized trial. Int J Radiat ()ncol Biol Phys |987zl3:541-9. Dische GS. Adams GE. Ash DV. Baker J. Barley V. Bleehen NM. et al. The Medical Research Council trial of misonida/ole in carcinoma of the uterine cervix. BrJ Radiol |9X~lz57z49l-9. Stehman FB. Bundy BN. Thomas G. Keys HM. d‘Ablaing G 3rd. Fowler WC. et al. Hydroxyurea versus misonidazole with radiation in cervical car- cinoma: long-term follow-up of a Gynecologic Oncology Group trial. J Clin ()ncol |9931||z|523-8. (8) Rose PG. Locally advanced cervical carcinoma: the role of chemoradia- tion. Semin ()ncol 19941391763. Vermorken JB. Neoadjuvant chemotherapy (NACT) in cervical cancer: still promising? leditorial] lsm' comment citation in Medlinel. Ann Oncol [99213:97-102. Symonds RP. Burnett RA. Haheshaw T. Kaye SB. Snee MP. Watson ER. The prognostic value of a response to chemotherapy given before radiotherapy in advanced cancer of the cervix. BrJ Cancer 1989;59:473—5. (ll) Tattersall MH. Lorvidhaya V. Vootiprux V. Cheirsilpa A. Wong F. Axhar T. et al. Randomized trial of epirubicin and cisplatin chemotherapy fol— lowed by pelvic radiation in locally advanced cervical cancer. Cervical Cancer Study Group of the Asian Oceanian Clinical Oncology Association [we comment citation in Medlinc]. J Clin Oncol l995:l3:444-Sl. (l2) Withers HR. Taylor JM. Maciejewski B. The hazard of accelerated tumor clonogen repoptilation during radiotherapy. Acta Oncol 1988227: I 3l—46. 16 t7 (9 (I0 103 Is There a Radiobiologic Basis for Improving the Treatment of Advanced Stage Cervical Cancer? Jacob C. Lindegaard, Jens Overgaard, Soren M. Benlzen, Dorie Pedersen * The success of radiotherapy in eradicating the primary tumor in patients with locally advanced cervical cancer is limited by normal tissue tolerance. Systematic recording of morbidity and treatment parameters is therefore very im- portant for radiobiologic treatment optimization and clinical decision making. There is substantial evidence that frac- tionation schedules employing large doses per fraction lead to a loss of therapeutic ratio. A similar argument could be used for high-dose-rate (HDR) brachytherapy that should also be administered in small dose fractions. However, HDR brachytherapy might convey some advantage to physical dose distribution that should be weighed against the radiobiologic advantages of Iow-dose-rate (LDR) continuous irradiation. Increasing overall treatment time reduces local control probability, whereas a shorter overall treatment time by accelerated fractionation may improve the thera- peutic ratio, at least in fast-growing tumors. Hypoxia and reduced oxygen delivery are associated with poor radiation response. Anemia should be compensated, if necessary. The role of hypoxic modification needs to be further explored. In the future, the therapeutic ratio may also be improved by the use of chemical and biologic response modifiers. Tumors are heterogeneous with respect to intrinsic radiosensitivity, proliferation parameters, and extent of hypoxia. Until a detailed prognostic profile can be obtained for each patient, optimal curative radiotherapy must aim for a sufficient dose, short overall treatment time, hypoxic modification, and [DR or low dose per fraction. [Monogr Natl Cancer Inst 1996;21:105-12] Introduction Radiotherapy is the only treatment option for controlling the primary tumor in most cases of locally advanced cancer of the uterine cervix (I). Nevertheless. the 5—year outlook in patients harboring International Federation of Gynecology and Ob— stetrics stages IIIB and IVA tumors is grim. In a Danish series of 293 patients with advanced cancer of the cervix (2). by 5 years half the patients had died after a recurrence. Among these. 48% had a local relapse without any clinical sign of distant dis— ease. The prognosis following a local relapse is dismal. with only 8% of the patients surviving more than 1 year after the recurrence (Fig. 1). Thus. intensified locoregional therapy is evi- Journal of the National Cancer Institute Monographs No. 21. I996 lOO '- ‘ ‘i \«« _V,\ \\ \d 80 “\(N ,— \J\<:\j \ \ 33 60 >— \ -:— Loca/ control Tn \ + Local re/apse .2 \ > t \ .J 1 (0 4c - \ \\ 20 l— o — \o l l l l l I O l 2 3 4 5 Observation time, years Fig. 1. Survival as a function of observation time in 293 patients with carcinoma of the uterine cervix. International Federation of Gynecology and Obstetrics stages NIH and IVA [data from (2.3)]. dently needed. However. this should be balanced against the radiotherapy-related morbidity. which may be considerable (3). A careful registration of morbidity is a prerequisite for the development of better radiotherapy schedules. However. many studies do not include a description of morbidity or do not use a recognized and validated scoring system (4). In addition. late complications are often reported as simple frequencies. a prac- tice that may seriously underestimate the true risk of complica— tions among long-term survivors (5). Recently. specific scoring systems (3.6. 7) have been developed for cancer of the uterine cervix. and these allow standardized recording and reporting of both early and late radiation damage. We hope that these and *A/flliulimt of'uurliorx: Danish Cancer Society. Department of Experimental Clinical Oncology, Aarhus University Hospital, Norrebrogade. Denmark. (‘1u‘rmprmtlmn‘0 to: Jens ()vergaard. M.D.. Department of Experimental Clini- cal Oncology. Aarhus University Hospital. Norrebrogade 44. DK-XOIX) Den- mark. l05 similar scoring systems will be used more frequently in the coming years. enabling a meaningful comparison of different treatment strategies. This paper reviews current radiobiologic knowledge that may improve the treatment of locally advanced cancer of the cervix by radiotherapy. Therapeutic Improvement by Reduction of Damage to Normal Tissue The incidence of damage to normal tissue increases with dose (1). A reanalysis of radiation morbidity in 442 consecutive patients with locally advanced carcinoma of the uterine cervix (3) revealed a relatively steep dose—response curve for severe late complications. Others have reported a lower incidence of severe late sequelae at comparable doses. probably reflecting not only different ways of recording. but also different treatment techniques. Thus. the amount of irradiated volume and concur- rent toxicity from other treatments, such as staging laparotomy. has a major impact on morbidity (8.9). Optimization of staging procedures and target volume may therefore reduce toxicity and/or allow an increased dose to the tumor. 3D dose planning. conformal external radiotherapy. and advanced intracavitary equipment with the use of a single stepping source are methods by which such optimization can be obtained. During the last decade. late-responding normal tissues have been shown to have a much larger capacity for repair of radia- tion damage compared with both acutely reacting normal tissues and most tumors (10). Late—responding tissues are therefore much more sensitive to large dose per fraction than tumors and acutely responding tissues. Several clinical studies (11-13). in— cluding both hypofractionation and hyperfractionation schedules. have confirmed this observation. Hypofractionation schedules should therefore not be used in curatively intended radiotherapy where a high total dose is needed to eradicate the tumor. The Dose-Rate Effect in Brachytherapy Cancer of the cervix was one of the first deep-seated malig- nant tumors to be cured by radiotherapy (1). The only treatment option available in the beginning of the era of radiotherapy was low-dose—rate (LDR) brachytherapy. Radiobiologie research in the last few decades has documented that this is indeed an effec- tive way to deliver a high dose to the tumor with an acceptable level of late normal tissue damage (14). When radiation is given at a high-dose rate (HDR). repair al- most exclusively takes place in the interval between fractions. Thus. the radiobiologic considerations in HDR brachytherapy are the same as in external-beam radiotherapy. However. as the dose rate is reduced. time is sufficient for biologic repair proces— ses to take place during irradiation (15). Thus. the biologic ef- fect depends not only on the repair capacity but also on the repair kinetics. Since late—responding tissues have the largest repair capacity. the dose-rate effect is expected to be most prominent in this type of tissue (10). A multivariate retrospec- tive study in patients with head and neck carcinoma by Mazeron et al. (16) showed that the risk of radionecrosis was significantly 1 06 higher for patients treated at a dose rate above 0.5 Gy per hour compared with lower dose rates. A prospective randomized phase [I] trial of brachytherapy in 204 patients with cervical cancer (17) compared the two dose rates. 0.38—0.73 Gy per hour. both in the LDR range. Dose rate had no significant influence on local tumor control and survival. However. the incidence of late complications almost doubled as the dose rate was increased (Fig. 2). The increasing use of HDR afterloading equipment has therefore caused concern and has resulted in a fierce debate be— tween advocates of LDR and HDR (14.18.19). Practical and economic motives have probably been the main reason for the introduction of HDR (l4). Radiobiologically. HDR intracavitary treatment of cervical cancer may be acceptably safe and effi- cient if it is administered with sufficiently low-dose fractions (14). Several nonrandomized clinical studies have shown that HDR treatment compares favorably with LDR treatment. How- ever. many of these studies have used historical control subjects. where differences in treatment planning and delivery may bias the comparison. Modern HDR treatment systems offer the pos- sibility of better anatomic dose distributions than traditional LDR equipment. where dose planning is much more difficult and the insertion of radioactive sources for a much longer time is required (20). The only randomized study comparing HDR with LDR brachytherapy. to date. is the study by Patel et al. (21). HDR was delivered in fractions of 9 or 9.5 Gy. with a dose rate of 130-160 cGy per minute at point A of the Manchester system. LDR was delivered in fractions of 35-375 Gy (55-65 cGy/hour at point A). The study comprised 482 patients. No statistically significant differences were found in the local control rate or in the incidence of severe complications. The study by Teshima et al. (22) was not strictly randomized and should therefore be regarded as a historical-control trial. Data were reported from a 60- :"j 0 38 Gy/h {:1 0. 73 Gy/li ~ 40— O<0 OOJ 30— l—l n l Local relapse Cumulative incidence at 2 years (0/0) Complications Fig. 2. Phase III randomized trial comparing two low dose rates in 204 patients with carcinoma of the cervix stage 18 or ll treated by brachytherapy and surgery ([7). Cumulative incidence of local relapse and late complications at 2 years. Journal ofthe National Cancer Institute Monographs No. 2 l. I996 total of 430 patients. The HDR and LDR regimens used yielded equivalent results concerning cause-specific survival but with a statistically significantly higher rate of moderate or severe com— plications after HDR ( l()C/() than after LDR (4%). Also. the non- randomized study by Sarkaria et al. (23) showed that HDR and LDR schedules. designed to be biologically equivalent using the linear—quadratic formula. may produce similar results as for tumor control and normal tissue complications. Many attempts have been made to retain the technical and clinical advantages of HDR without losing the radiobiologic benefits of LDR (20.24.25). The compromise currently in focus is the use of pulsed dose rate (PDR). ln PDR. the average dose rate is kept at a similar level as a continuous LDR schedule. but radiation is delivered with small and frequent HDR pulses (26). Because of the short interfraction interval in PDR. repair kinetics is crucial for the calculation of the therapeutic index. Brenner et al. (24) have suggested that PDR may produce fewer late effects than LDR. for a given level of tumor control if the repair half time is shorter for tumors than for late—responding normal tissues. Unfortunately. neither experimental nor clinical data support the existence of a systematic difference between the repair kinetics of tumor and late—responding normal tissues (27.28). A few studies have tested the supposed equivalence of PDR and continuous LDR therapy. PDR teletherapy was used in ad— vanced head and neck cancer but resulted in a very high rate of complications (29). Consequential late damage. self-sensitiza- tion. or increasing repair half time with larger dose per fraction have been suggested as possible explanations (25). Two studies (30.31) have brought evidence in support of the equivalence of PDR to continuous LDR irradiation. However. the dose-rate ef— fect is small in these experimental systems. and data from in vivo models of late-reacting tissues are lacking. Ideally. experimental animal testing and phase [-11 studies should have been done before advancing HDR and PDR into clinical use. HDR equipment is already being used in many in- stitutions in the treatment of advanced cancer of the uterine cer- vix. A significant increase in the incidence of late normal tissue damage apparently has not occurred ([4). However. a close and systematic registration of morbidity and tumor response is very important to knowing whether these initial observations hold true when evaluated systematically over a longer time span. From a radiobiologic point of View. HDR should be ad- ministered in small doses per fraction ([4). Importance of Overall Treatment Time Prolongation of the overall treatment time reduces the tumor response in radiotherapy of squamous cell carcinoma of the uterine cervix. This is illustrated by the experience gained with split-course radiotherapy. On the basis of the hypothesis that normal tissue would repopulate faster than tumors. a 2- to 3- week treatment-free interval was introduced. Total dose was ad— justed differently in patients with large and small tumors (2). In the large tumors. the split—course and the conventional regimens were isoeffective with respect to late effects. but split-course radiotherapy resulted in a significantly poorer local control (Fig. 3). In the smaller tumors. total dose was increased sufficiently to Journal ofthe National Cancer Institute Monographs No. 2 l. 1996 maintain the local control rate. but this was achieved at the ex— pense of an increase in late sequelae. These findings are consis— tent with the current hypotheses of rapid tumor repopulation and no appreciable repopulation in late-responding tissues. Tumor repopulation may also be a problem with neoadjuvant chemo- therapy (32). Several retrospective studies (33—35) have shown that prolongation of overall treatment time in a continuous radiotherapy schedule for advanced cancer of the cervix also reduces both local control and survival. A linear decrease of locoregional control and survival as a function of overall treat- ment time amounting to approximately 1% per day has been ob— served (34,35). Some care should be taken in interpreting the outcome of these studies since patients in a less favorable prog- nostic group with competing clinical problems are often treated in longer overall times (36). However. the split—course ex- perience (2) leaves no doubt that treatment prolongation should be avoided when possible. Overall treatment time may be reduced compared with a standard schedule either by increasing the dose per fraction or by giving more than five fractions of radiation per week. The former approach should be discouraged because of the high fractionation sensitivity of late—responding normal tissues. More than five fractions per week may be delivered by treating on weekend days or by treating with two fractions for all or some working days. Reduction of overall treatment time. so-called ac- celerated radiotherapy. is already being used in the clinic in an attempt to avoid the deleterious effect of tumor repopulation in squamous cell carcinoma of the head and neck. The best demonstration of the trade—off between dose and treatment time is from the recent trial of continuous hyperfractionated ac- celerated radiotherapy in squamous cell carcinoma of the head and neck (37). Accelerated schedules may run into problems with increased early toxicity because of less cellular renewal during therapy. A small total dose is therefore often employed in an attempt to trade overall treatment time with total dose. a principle that seems useful in squamous cell carcinoma of the head and neck. However. this again may cause failure in more radioresistant tumors. Therefore. a predictive assay discriminating fast—grow- ing from slow—growing tumors would be very important. Various proliferation markers may have a predictive potential in cancer of the uterine cervix (38,39). Yet. the most thoroughly tested parameter has been the potential doubling time of clonogenic tumor cells (Tpot). Tpot can be measured frotn a single tumor biopsy after injection with a thymidine analogue. The hypothesis is that Tpot could be an indication of the maxi- mal proliferation capacity of a tumor. Large studies in Europe and in the United States are currently testing this hypothesis. So far. the data are inconclusive. but this is likely to change in the coming few years (40). When giving two fractions a day. the interfraction interval should be as long as possible and not shorter than 6 hours (4]). The recent discovery that early-responding normal tissues may also have a long repair half time emphasizes the need for long intervals (28). This problem. together with the rapid dose ac— cumulation that may exceed the repopulation capacity of early- responding normal tissues. emphasizes that acutely responding tissues may be dose limiting in accelerated radiotherapy (42). 107 TOO I118 80 CRT (67%) 60 SCRT (49%) Local control IIIB Late severe rectosigmoid complications 40 40 SCRT (38%) L f o\0 20"— 20 r CRT (32%) >. .3; p <0.05 N.S t Cl , . , O [477 , . , I _| o t 2 3 4 5 1 2 3 4 6 CO 3); 100 ‘00 O W ( HE C: ' Late severe rsctosigmoid complications 0. 80 a in (CRT) 77% so — * (SCRT) 73% ' 60 ~ 50 - 40— 40,. SCRT (38%) 20- “B 20»— p=0.02 CRT (10%) ,. Local control N.S , Fl 0 I ' I I O L I ‘ I J 0 1 2 3 4 5 O 1 2 3 4 5 YEARS AFTER TREATMENT Fig. 3. Actuarial estimates of local control and severe late rcctosigmoid complications in patients with carcinoma of the uterine cervix treated by either continuous (CRT) or split»course radiotherapy (SCRT), Graphs show that increasing the overall treatment time (SCRT) results in a loss of local control with unchanged late mor- bidity (International Federation of Gynecology and Obstetrics stage ”[8. top panels) or an increase in late morbidity with unchanged local control (International Federation of Gynecology and Obstetrics stage [18. bottom panels). depending on whether total dose was increased [modified from (2)], Thus. accelerated fractionation schedules involve a delicate balance between total dose. rate of dose accumulation. and in- terfraction interval. Randomized studies of accelerated frac— tionation in cervical cancer have not been done. but the Gynecologic Oncology Group is planning to begin such a trial in the near future in the United States and in Canada. Hypoxia Histopathologic studies and invasive polarographic measure- ment of intratumoral partial oxygen pressure. p03 . have shown that many human tumors. including some carcinomas of the uterine cervix. contain regions with a low-oxygen tension (43,44). Biologically. there are probably two different mech- anisms behind this observation and these are called diffusion- limited and perfusion—limited hypoxia. Diffusion-limited hypoxia results from inadequate angiogenesis. This causes cells 108 to become chronically hypoxic at a distance of 100-180 um around tumor vessels as oxygen is consumed by cells nearer to the vessel (45). Perfusion-limited hypoxia results from intermit- tent closure of tumor blood vessels. leading to acute hypoxic conditions for all tumor cells situated downstream from the obstruction (46). Hypoxia activates several genes that may induce tumor dedif— ferentiation and increase the metastatic potential (47). Also. Graeber et al. (48) Showed that hypoxia provides a physiologic selective pressure for the expansion of cell variants that have lost their apoptotic potential. and in particular for cells acquiring p53 mutations. This may explain the strong prognostic value of hemoglobin level for both local control and survival in patients with advanced cancer of the cervix (49.50). Of greater impor— tance is probably the immense radioresistance of clonogenic hypoxic tumor cells observed experimentally (5/). The predic— tive value of intratumoral p0: in patients with advanced cervical Journal of the National Cancer Institute Monographs No. 2 I. (9% cancer emphasizes hypoxic radioresistance as a major obstacle to radiocurability (44). This is supported by a meta-analysis of clinical studies in cervical carcinomas where attempts have been made to overcome tumor hypoxia (Table 1). This shows a sig— nificant increase (P = .006) in local control of some 6% after this strategy. One way to reduce tumor hypoxia is by letting the patient breath high—oxygen-content gas mixtures or even irradiating patients in hyperbaric oxygen chambers. A hyperbaric oxygen trial in cervical cancer conducted by the Medical Research Council showed a benefit in the hyperbaric oxygen group for both local control and survival (52). Other studies have failed to show a statistically significant benefit from this strategy (51), partly because they have been too small to detect the difference. Also, the use of normobaric high-oxygen—content gas breathing has not shown a statistically significant advantage so far. Another strategy for overcoming hypoxia involves the use of oxygen-mimetic chemicals. These drugs are not metabolized and can sensitize hypoxic tumor cells to irradiation. Nitro- imidazoles are compounds with this property, and they have shown an effect in squamous cell carcinomas of the head and neck (53). However, disappointing results have been obtained with sensitizers in cancer of the uterine cervix (5]). A large trial with the analog pimonidazole showed an adverse effect of the drug (54), possibly because of drug-induced impairment in the blood supply to the tumor (55). Almost 10 years ago, it was shown that the vitamin B3 analogue, nicotinamide, could almost preferentially enhance radiation damage in several rodent tumor models by increasing their oxygen levels (56). Later, it was shown that the primary ef— fect of nicotinamide is to prevent transient fluctuations in tumor blood flow, thus reducing the development of perfusion—limited hypoxia (46,56). Clinical pharmacologic studies have shown that nicotinamide can be given in clinically effective doses (5]) and phase I—II studies are ongoing. Attempts are being made to overcome both the problem of diffusion- and perfusion-limited hypoxia, for example, by combining nicotinamide with hyper— thermia, perfluoro-chemicals, carbogen, or oxygen breathing. Experiments (51,56) show that this is a promising approach to the hypoxic problem. Because of the benefit already obtained by hypoxic modification, nicotinamide seems to be an interesting candidate for new protocols in cancer of the uterine cervix. Blood transfusion has also been shown to be of some clinical benefit. Thus, in a study by Bush et al. (57), the probability of local failure in patients with cervical cancer with an initial hemoglobin level of less than 7.5 mmol/L was reduced about 25% by transfusion during radiotherapy. Recently, erythro— poietin has been shown to be equally effective for raising hemoglobin levels in anemic patients with cervical cancer treated by radiotherapy (58). Carbon monoxide inhaled during smoking may substantially reduce tumor oxygenation, and smoking has been shown to reduce both local control and survival probability in patients with head and neck cancer (51,59). The data for patients with cervical cancer are less convincing. In a series of 151 patients with cervical carcinoma, Solberger and Sorbe (60) found smok- ing during radiotherapy to be a nonsignificant prognostic vari- able when tested in a multivariate analysis. Still, the experience from head and neck cancer suggests that patients should be ad- vised to stop smoking, at least during radiotherapy. Some new treatment strategies try to take advantage of hypoxia rather than attempting to reduce it. This can be ac- complished using hypoxia-specific cytotoxins (61) or bioreduc- tive drugs. An example of the latter is mitomycin C (62), which is currently being tested in randomized trials in head and neck cancer. During the last two decades, a biologic rationale for hyper- thermia as an anticancer treatment modality has been estab- lished (63). Blood supply to a tumor is often insufficient. This causes acidic waste products to build up, making the tumor cells more heat sensitive than surrounding normal cells (64,65). New tumor vessels can be formed, but they are fragile, tortuous, and do not contain a neuromuscular component (66). Many tumors are therefore not able to dissipate heat through vascular response to the same extent as most normal tissues. Thus, vascular stasis may occur during heating, leading to further acidification and heat accumulation in the tumor (67). A randomized multicenter trial in malignant melanoma has shown that the biologic ratio- nale for combining radiotherapy and hyperthermia is clinically relevant (68). Three smaller clinical studies (69-71) in advanced cancer of the cervix have also shown an increase in local control when radiotherapy is combined with hyperthermia. Unfortunately, present clinical hyperthermia equipment cannot deliver homo— geneous heating to a defined target volume, which probably is the main limitation to a more widespread clinical use right now. Biological and Chemical Response Modifiers Hematologic growth factor support of bone marrow function during antineoplastic therapy has been used routinely for several Table 1. Overview analysis of randomized trials of radiotherapy alone or combined with hypoxic or environmental modification in advanced cervical carcinoma* Local control, % No. of N0. of RT + RT Treatment trials patients modifier alone Odds ratio (95% Cl) Oxygen/HBO 5 819 73 62 1.65 (1.23222) Hypoxic sensitizer 6 956 59 61 0.92 (0.71-1.19) Transfusion l 135 84 69 2.27 (1.00-5.20) Hyperthermia 3 146 72 50 2.52 (1.27—5.01) Overall 15 2056 67 61 1.29 (1.08-1.55) *RT = radiation therapy; CI = confidence interval; HBO = hyperbaric oxygen. Journal of the National Cancer Institute Monographs No. 21, 1996 109 years. However, little is known about the possible use of growth factors and cytokines to reduce radiation—induced epithelial and stromal injury (72). Recent studies (72—74) show that epidermal growth factor. transforming growth factor-0t, and interleukin 1 may be strong stimulators of epithelial proliferation following acute radiation damage. As for stromal damage. increased levels of transforming growth factor—[3 have been found in patients showing excessive late radiation reactions (75). The concern has been raised that cytokines may force functional cells to divide. which precipitate latent radiation damage before regeneration from surviving cells has had sufficient time to occur (76). An obvious concern is whether growth factors would stimu- late tumor growth as well. So far. a complex pattern has emerged, with results pointing toward growth stimulation, growth inhibition, and radiosensitization (77.78). Regeneration of cell—mediated immunity may also be important for the tumor response. A recent paper (79) has presented a survival benefit through the use of a glycan sizofiram that apparently stimulates recovery of natural killer cell. T—cell. and macrophage function following radiotherapy for cancer of the uterine cervix. Many questions remain to be answered before cytokines can be intro- duced clinically. However, the improved understanding of the molecular basis for biologic response modification gives hope that new treatment options may emerge. Response modification can also be obtained by protaglandins. vascular modifiers. diet. and antibiotics (72). Yet their specific role in reducing normal tissue damage after radiotherapy for cancer of the uterine cervix remains to be established. The use of radioprotectors like WR—277l has gained renewed interest (80—82). An obvious prerequisite would be that these com— pounds are sufficiently normal tissue specific to avoid tumor protection (83). Also, a major concern with WR—277l is the in— duction of hypotension (80) that by a steal effect may increase tumor hypoxia and by that decrease radiosensitivity (55). Mit— suhashi et al. (82) did a retrospective analysis to investigate the radioprotective effects of WR-277l on the long—term outcome of radiation therapy for 83 patients with carcinoma of the uterine cervix treated with definitive radiation therapy from January 1978 through December 1984. Forty—six patients were treated with radiation therapy alone, whereas 37 patients were treated with radiation therapy plus WR—277l. No difference was seen in overall survival or pelvic control rate. Also, the in— cidence of late rectal and bladder complications was identical in the two groups. Intrinsic Radiosensitivity and Predictive Assays More than 20 years ago. it was shown that fibroblasts from patients with the rare genetic syndrome ataxia telangiectasia were hypersensitive to ionizing radiation in vitro. This fueled the notion that variability in clinical response to radiotherapy could depend on a genetically determined intrinsic cellular radiosensitivity, which again raised hope that predictive assays of intrinsic radiosensitivity could be developed to individualize dose prescription (39.84.85). Several studies (86-88) have shown a correlation between in vitro radiosensitivity of fibro- blasts and severity of late radiation morbidity. One problem may be that different predictive assays may have to be developed for 11() specific normal tissue reactions because of the apparent dis— sociation between the clinical sensitivities to various types of se- quelae (89). Also, the in vitro radiosensitivity of cervical carcinoma cells has been shown to be predictive of tumor response (85). However. some doubts remain about whether the assay can discriminate between normal and malignant cells (90). Other tumor biologic parameters such as apoptosis may have a predictive potential with respect to radiosensitivity of cervical cancer. but the results are conflicting (40.91). Other assays for radiation damage and for activity of oncogenes and tumor sup— pressor genes are being developed. It is to be hoped that these assays, or combinations of them, could predict the sensitivity profile of individual patients. It will then be possible to lower the dose to radiosensitive patients to reduce risk of severe com— plications and to increase the doses to the more resistant patients to achieve an improved tumor control probability without an in— creased complication rate. A recent analysis (84) suggests that this may lead to a therapeutic gain for some patients, Conclusion Radiotherapy with curative intention is the treatment of choice for locally advanced cervical cancer. The tumor response is dose dependent. but dose is limited by the tolerance of normal tissue. Knowledge of morbidity and treatment parameters is of the utmost importance for radiobiologic analysis and clinical decision making. The data obtained so far clearly show that large dose per fraction given at an HDR is radiobiologically dis— advantageous. Avoiding such strategies is therefore likely to im- prove the therapeutic ratio. In the future. use of growth factors, cytokines, etc.. may yield promise for relieving morbidity. Pre— vious studies have shown that extension of overall treatment time reduces curability. A radiobiologic rationale for improving the outcome by using accelerated schedules has recently been established. Hypoxia and reduced oxygen delivery are no doubt associated with poor response to radiotherapy. Anemia should be compensated, if necessary. However. the role of hypoxic modification needs to be further explored. Tumors are heterogeneous with respect to intrinsic radiosen- sitivity, proliferation parameters, and extent of hypoxia. These factors may influence the radiation response, and charac— terization of these parameters in the individual tumors and patients may be required to optimize treatment. Until a detailed prognostic profile for each patient can be obtained. optimal curative radiotherapy must aim for the following: sufficient dose (intrinsic radiosensitivity), short overall treatment time (reduce repopulation), hypoxic modification (eliminate hypoxic radio- resistant cells), and minimal late normal tissue damage (avoid large dose per fraction at HDR). The benefit from the radiobiologic knowledge is more likely to come from the proper application of treatment principles we already know than from introduction of new strategies. References (I) Hoskins WJ. Perez CA. Young RC, DeVita VT .lr. Hellman S. Rosenberg SA. editors. Cancer: principles and practice of oncology. 4th ed. Philadel- phia: Lippincott, 1993:1152—225. Journal of the National Cancer Institute Monographs No. 21. 1996 (2) (3 (4 Ln (6 t7) t8) (9 (/0) (II) \ \) (I3) (/4) (l5) (/6) (/7) (18) (I9) (20) (2/) (22) Pedersen D, Bentzen SM. Overgaard J. Continuous or split-course com- bined external and intracavitary radiotherapy of locally advanced car- cinoma of the uterine cervix. Acta ()ncol 1994;33:547-55. Pedersen D. Bent/.en SM. Overgaard J. Early and late radiotherapeutic morbidity in 442 consecutive patients with locally advanced carcinoma of the uterine cervix [.th comment citation in Medline]. Int J Radiat Oncol Biol Phys 1994;29:941-52. Sistnondi P. Sinistrero G. Zola P. Volpe T. Ferraris R. Castclli GL. et al. Complications of uterine cervix carcinoma treatments: the problem of a ttniform classification. Radiother ()ncol 1989;14:9-17. Bentzen SM. Vaeth M. Pedersen DE. Overgaard J. Why actuarial estimates should be used in reporting late normal-tissue effects of cancer treatment . . . now! [editorial] l.\'('t' comment citations in Medline]. lttt J Radiat Oncol Biol Phys 1995;32:1531-4. Chassagne D. Sistnondi P. Horiot JC. Sinistrero G. Bey P. Pernot M. et al. A glossary for reporting complications of treatment in gynecological can- cers. Radiother Oncol 1993126: 195-202. LENT/SOMA tables. Radiother Oncol 1995135: 17-60. Letschert JG. Lebesque JV. de Boer RW. Hart AA. Bartelink H. Dose- volume correlation in radiation—related late small-bowel complications: a Clinical study. Radiother Oncol 1990;18:307-20. Eifel PJ. Levenback C. Wharton JT. Oswald MJ. Tittle course and in- cidence of late complications in patients treated with radiation therapy for FIGO stage IB carcinoma of the uterine cervix [we comment citation in Medline]. lntJ Radiat Oncol Biol Phys 1995;32:1289-300. Thames HD. Bentlen SM. Turesson 1. Overgaard M. van den Bogaert W. Fractionation parameters for human tissues and tumors. lnt J Radiat Biol 1989;56:701-10. Horiot JC. Le Fur R. N'guyen T. Chenal C. Schraub S. Alfonsi S. et al. Hyperfractionated versus conventional fractionation in oropharyngeal car» cinoma: final analysis of a randomized trial of the EORTC cooperative group of radiotherapy [.wc comment citations in Medline]. Radiother Oncol 1992;25:23141. Overgaard M. Bentlen SM. Christensen J]. Madsen EH. The value of the NSD formula in equation of acute and late radiation complications in nor— mal tissue following 2 and 5 fractions per week in breast cancer patients with postmastectomy radiotherapy. Radiother Oncol 19871911 1. Singh K. Two regimes with the same TDF but differing morbidity used in the treatment of stage 111 carcinoma of the cervix. Br J Radiol 1978; 51:357—62. Scalliet P. Gerbaulet A. Dubray B. HDR versus LDR gynecologic brachytherapy revisited [sec comment citation in Medline]. Radiat Oncol 1993;28:118—26. Hall EJ. Brenner DJ. The dose—rate effect revisited: radiobiological con- siderations of importance in radiotherapy. Int J Radiat Oncol Biol Phys 1991;21:1403-14. Mazeron JJ. Simon JM. Le Pechoux C. Crook JM. Grimard L. Piedbois P. et al. Effect of dose rate on local control and complications in definitive ir- radiation of Tl—2 squamous cell carcinomas of mobile tongue and floor of the mouth with interstitial iridium-192 [we comment citation in Medline]. Radiother Oncol 1991 21:39—47. _ Lambin P. Gerbaulet A. Kramar A. Scalliet P. Haie-Meder C. Malaise E. et a1. Phase III trial cotnparing two low dose rates in brachytherapy of cer- vix carcinoma: report at two years. lnt J Radiat ()ncol Biol Phys 1993: 25:405-12. Eifel PJ. High'dose-rate brachytherapy for carcinoma of the cervix: high tech or high risk? [editorial] [we comment citations in Medline]. Int J Radiat Oncol Biol Phys 199224138347. Orton CG. HDR vs LDR for ca cervix: high risk or biased reporting? Int J Radiat Oncol Biol Phys 1992;24:387-8. Stilt JA. Fowler JF. Thomadsen BR. Buchler DA. Paliwal PB. Kinsella TJ. High dose rate intracavitary brachytherapy for carcinoma of the cervix: the Madison system: 1. Clinical and radiobiological considerations [we com— ment citation in Medline]. lntJ Radiat Oncol Biol Phys 1992;24:335-421. Patel FD. Sharma SC. Negi PS. Ghoshal S. Gupta BD. Low dose rate vs. high dose rate brachytherapy in the treatment of carcinoma of the uterine cervix: a clinical trial. Int J Radiat Oncol Biol Phys 1994;28:335—41. Teshima T. lnoue T. lkeda H. Miyata Y. Nishiyama K. lnoue T. et a1. High-dose rate and low—dose rate intracavitary therapy for carcinoma of the uterine cervix. Final results of Osaka University Hospital. Cancer 1993;72:2409-14. Sarkaria JN. Petereit DG. Stitt JA. Hartman T. Chappell R. Thomadsen BR. et al. A comparison of the efficacy and complication rates of low dose-rate versus high dose»rate brachytherapy in the treatment of uterine cervical carcinoma. lntJ Radiat Oncol Biol Phys 1994;30:75—82. Brenner DJ. Hall EJ. Huang Y. Sachs RK. Potential reduced late effects for pulsed brachytherapy compared with conventional LDR. Int J Radiat Oncol Biol Phys 1995;31:201'2. Journal of the National Cancer Institute Monographs No. 21. 1996 (25) (26) G E‘ (36) (3 7) (38) (19) (40 (4/) (42) (43) (44) (45) (46) (47) (48) (49) Fowler J. Mount M. Pulsed brachytherapy: the conditions for no sig— nificant loss of therapeutic ratio compared with traditional low dose rate brachytherapy.1ntJRadiatOncol Biol Phys 1992;23:661-9. Brenner DJ. Hall EJ. Conditions for the equivalence of continuous to pulsed low dose rate brachytherapy. Int J Radiat Oncol Biol Phys 1991: 202181—90. Fowler JF. Are half-times of repair reliably shottet‘ for tumors than for late normal—tissue effects [My comment citation in Medlinel'.’ Int J Radiat Oncol Biol Phys 1994;31:189-90. Bentzen SM. Ruifrok AC. Thames HD. Repair capacity and kinetics for human mucosa and epithelial tumors in the head and neck: clinical data on the effect ofchanging the time-interval between multiple fractions per day in radiotherapy. Radiother Oncol 1996;38:89- 101. Nguyen TD. Panis X. Froissart D. Legros M. Coninx P. Loirette M. Analysis of late complications after rapid hyperfractionated radiotherapy in advanced head and neck cancers. Int J Radiat Oncol Biol Phys 1988; 14123-5. Mason KA. Thames HD. Ochran TG. Ruifrok AC. Janjan N. Comparison of continuous and pulsed low dose rate brachytherapy: biological equivalence in vivo [see comment citation in Medline]. lntJ Radiat ()ncol Biol Phys 1994;28:667-71. Armour E. Wang ZH. Corry P. Martinez A. Equivalence of continuous and pulse simulated low dose rate irradiation in 9L gliosarcotna cells at 37 degrees and 41 degrees C. Int J Radiat Oncol Biol Phys 1991122: 1(19-1-1. Polish RA. Twiggs‘ LB. On the lack of demonstrated clinical benefit of neoadjuvant cisplatinum therapy for cervical cancer. lnt J Radiat Oncol Biol Phys 1993;27:975-9. Delaloye JF. Coucke PA. Pampallona S. De Grandi P. Effect of total treat- ment of time on event-free survival in carcinoma of the cervix. Gynecol Oncol 1996;60:42-8. Fyles A. Keane TJ. Barton M. Simtn J. The effect of treatment duration in the local control of cervix cancer [we comment citation in Medline]. Radiother Oncol 1992;25:273-9. Girinsky T. Rey A. Roche B. Haie C. Gerbaulet A. Randrianarivello H. et a1. Overall treatment time in advanced cervical carcinomas: a critical parameter in treatment outcome. Int J Radiat Oncol Biol Phys 1993; 27:1051—6. Bentzen SM. Time-dose relationships for human tumors: estimation from non-randomized studies. In: Beck-Bornholdt HP. editor. Current topics in clinical radiobiology of tumours. Medical Radiology. Berlin-Heidelberg: Springer-Verlag. 1993:1 1—26. Saunders M1. Dische S. Parmar MK. Harvey A. Gibson D. Randomised multicentre trials of CHART vs conventional radiotherapy in head and neck and non—small—cell lung cancer: an interim report. Br J Cancer 1996;73:1455—62. Tsang RW. Fyles AW. Kirkbride P. Levin W. Manchul LA. Milosevic MF. et al. Proliferation measurements with flow cytometry Tpot in cancer of the uterine cervix: correlation between two laboratories and preliminary clinical results [see comment citation in Medline]. Int J Radiat Oncol Biol Phys 1995;32:1319-29. Levine EL. Renehan A. Gossiel R. Davidson SE. Roberts SA. Chadwick C. et a1. Apoptosis. intrinsic radiosensitivity and prediction of radiotherapy response in cervical carcinoma. Radiother Oncol 1995;37:1—9. Begg AC. The clinical status of Tpot as a predictor? Or why no tempest in the Tpot! [editorial] [see comment citation in Medline]. Int J Radiat ()ncol Biol Phys 1995;32:1539—41. Rojas A. Joiner MC. The influence of dose per fraction on repair kinetics. Radiother Oncol 1989;14:329-361. Kaanders JH. van Daal WA. Hoogenraad WJ. van der Kogel AJ. Ac- celerated fractionation radiotherapy for laryngeal cancer. acute. and late toxicity. Int J Radiat Oncol Biol Phys 1992;24:497—503. Kolstad P. Vascularization. oxygen tension. and radiocurahility in cancer of the cervix. Oslo: Scandinavian University Books. 1963. Hockel M. Knoop C. Schlenger K, Vorndran B. Baussmann E. Mil/.e M. et a1. lntratumoral p02 predicts survival in advanced cancer of the uterine cervix. Radiother Oncol 1993;26:45-50. Tannoek 1F. The relation between cell proliferation and the vascular system in a transplanted mouse mammary tumour. BrJ Cancer 1968:22:258«73. Chaplin DJ. Olive PL. Durand RE. Intermittent blood flow in a murine tumor: radiobiological effects. Cancer Res 1987;47:597-601. Sutherland RM. Ausserer WA. Murphy BJ. Laderoute KR. Tumor hypoxia and heterogeneity: challenges and opportunities for the future. Setnin Radiat Oncol 1996;6:59-70. Graeber TG. Osmanian C. Jacks T. Housman DE. Koch CJ. Lowe SW. el al. Hypoxia-mediated selection of cells with diminished apoptotic potential in solid tumours. Nature 1996;379:88-91. Pedersen D. Sogaard H. Overgaard J. Bentzen SM. Prognostic value of pretreatment factors in patients with locally advanced carcinoma of the uterine cervix treated by radiotherapy alone. Acta Oncol 1995;34:787-95. lll (50) (51) (52) (53) (54) (55) (56) (57) (58) (59) (60) (6/) (62) (63) (64) (65) (66) (67) (68) (69) (70) (7/) 112 Girinsky T. Pejovic—Lenfant MH. Bourhis J. Campana F. Cosset JM. Petit C. et al. Prognostic value of hemoglobin concentrations and blood trans- fusions in advanced carcinoma of the cervix treated by radiation therapy: results of a retrospective study of 386 patients. Int J Radiat Oncol Biol Phys 1988;16:3742. Overgaard J. Horsman MR. Modification of hypoxia~induced radioresis— tance in tumors by the use of oxygen and sensitizers. Semin Radiat Oncol 1996:6210-21. Watson ER. Halnan KE. Dische S. Saunders Ml. Cade 1S. McEwen JB. et a1. Hyperbaric oxygen and radiotherapy: a Medical Research Council trial in carcinoma of the cervix. BrJ Radiol 1978;51:879—87. Overgaard J. Clinical evaluation of nitroimidazoles as modifiers of hypoxia in solid tumors. Oncol Res 199416150948. A trial of Ro 03-8799 (pimonidazole) in carcinoma of the uterine cervix: an interim report from the Medical Research Council Working Party on ad— vanced carcinoma of the cervix. Radiother Oncol 1993126293403. Chaplin DJ. Horsman MR. Tumor blood flow changes induced by chemi- cal modifiers of radiation response. Int J Radiat ()ncol Biol Phys 1992;22:459-62. Horsman MR. Nicotinamide and other hen/amide analogs as agents for overcoming hypoxic cell radiation resistance in tumors. A review. Acta Oncol 1995;34:571-87. Bush RS. Jenkin RD. Allt WE. Beale FA. Bean H. Dembo AJ. et a1. Definitive evidence for hypoxic cells influencing cure in cancer therapy. BrJ Cancer Suppl 1978;37:302-6. Dusenbery KE. McGuire WA. Holt PJ. Carson LF. Fowler JM. Twiggs LB. et al. Erythropoietin increases hemoglobin during radiation therapy for cervical cancer. Int] Radiat Oncol Biol Phys 199429107984. Browman GP. Wong G. Hodson l. Sathya J. Russell R. McAlpine L. et al. Influence ofcigarette smoking on the efficacy of radiation therapy in head and neck cancer [see comment citation in Medline]. N Engl J Med 1993;328:159-63. Solberger O. Sorbe B. Fever. haemoglobin and smoking as prognostic fac- tors during the treatment of cervical carcinoma by radiotherapy. Eur J Gynaecol Oncol 1990;] 1297-102. Brown JM. Siim BG. Hypoxia—specific cytotoxins in cancer therapy. Semin Radiat Oncol 1996;6:22~36. Sanorelli AC. Hodnick WF. Belcourt MF. Tomasz M. Haffty B. Fischer JJ. et al. Mitomycin C: a prototype bioreductivc agent. Oncol Res 1994; 6:501-8. Overgaard J. The current and potential role of hyperthermia in radio- therapy. IntJ Radiat Oncol Biol Phys 1989;16:535—49. Gerweck LE. Modification of cell lethality at elevated temperatures. The pH effect. Radiat Res 1977;70:224-35. Vaupel P. Kallinowski F. Okunieff P. Blood flow. oxygen and nutrient supply. and metabolic microenvironment of human tumors: a review. Can» cer Res 1989;49:6449-65. Folkman J. How is blood vessel growth regulated in normal and neoplastic tissue? G.H.A. Clowes Memorial Award lecture. Cancer Res 1986;46:467-774. Reinhold HS. Endrich B. Tumour microcirculation as a target for hyper— thermia. Int J Hyperthermia 1986;22111-37. Overgaard J. Gonzalez Gonzales D. Hulshof MC. Arcangeli G. Dahl O. Mella O. et al. Randomised trial of hyperthermia as adjuvant to radio- therapy for recurrent or metastatic malignant melanoma. European Society for Hypenhermic Oncology [you comment citations in Medlinel. Lancet 1995;345:5403. Datta NR. Bose AK. Kapoor HK. Thermoradiotherapy in the management of carcinoma cervix (1118): a controlled clinical study. Ind Med Gaz 1987;121:68-71. Overgaard J. van der Zee J. Vernon C. Thermoradiotherapy of malignant tumors. European randomized multicenter trials evaluating the effect of ad- juvant hyperthermia in radiotherapy. ln: Kogelnik HD. editor. Progress in radio—oncology V. Bologna: Monduzzi Editore. 1995:507-13. Sharma S. Singhal S. Sandhu AP. Ghoshal S. Gupta BD, Yadav NS. Local thermo—radiotherapy in carcinoma cervix: improved local control versus increased incidence of distant metastasis. Asia Oceania J Obstet Gynaecol 1991;17:5-12. (72) (73) (74) (75) (7o) (77) (78) (79) (80) (8/) (82) (83) (84) (85) (86) (87) (88) (89) Hendry JH. Biological response modifiers and normal tissue injury after ir— radiation. Semin Radiat Oncol 1994;4z 123—32. McKenna KJ. Ligato S. Kauffman GL Jr. Abt AB. Stryker JA. Comer RK. Epidermal growth factor enhances intestinal mitotic activity and DNA con— tent after acute abdominal radiation. Surgery 1994;! 15:626—32. Potten CS. Owen G. Hewitt D. Chadwick CA. Hendry H. Lord Bl. et al. Stimulation and inhibition of proliferation in the small intestinal crypts of the mouse after in vivo administration of growth factors. Gut 1995:3ozxo4~ 73. Canney PA. Dean S. Transforming growth factor beta: a promotor of late connective tissue injury following radiotherapy? BrJ Radio] 1990;63:620- 3. Michalowski A. On radiation damage to normal tissues and its treatment. 1. Growth factors. Acta Oncol 1990;29:1017—23. Leith JT. Harrigan P. Padfield G. Faulkner L. Michelson S. Modification of the growth rates and hypoxic fractions of xenografted A431 tumors by sialoadenectomy or exogenously supplied growth factor. Cancer Res 1991: 51:41 I 1—3. Laderoute KR. Ausserer WA. Knapp M. Grant TD. Sutherland RM. Epidermal growth factor modifies cell cycle control in A431 human squamous carcinoma cells damaged by ionizing radiation. Cancer Res 1994254140741. Miyazaki K. Mizutani H. Katabuchi H. Fukuma K. Fujisaki S. ()kamura H. Activated (HLA—DR+) T»lymphocyte subsets in cervical carcinoma and ef- fects of radiotherapy and immunotherapy with sizofiran on cell-mediated immunity and survival. Gynecol Oncol 1995;56:412-20. Capizzi RL. Oster W. Protection of normal tissues from the cytotoxic ef- fects of chemotherapy by amifostine: clinical experiences. Eur J Cancer 1995;31A. Suppl. 1:58-13. Wadler S. Beitler JJ. Rubin JS. Haynes H. McGill F. Rozenblit A. el al. Pilot clinical trial of cisplatin. radiation. and WR2721 in carcinoma of the uterine cervix: a New York Gynecologic Oncology Group study. J Clin Oncol 1993;11:1511-6. Mitsuhashi N. Takahashi l. Takahashi M. Hayakawa K. Niibe H. Clinical study of radioprotective effects of amifostine (YM-(18310. WR-2721) on long-term outcome for patients with cervical cancer. Int J Radiat ()ncol Biol Phys 1993;26:407-11. Denekamp J. Stewart FA. Rojas A. Is the outlook grey for WR-2721 as a clinical radioprotector'? Int J Radiat Oncol Biol Phys |983;9:1247-9. Tucker SL. Geara FB. Peters LJ. Brock WA. How much could the radiotherapy dose be altered for individual patients based on a predictive assay of normal tissue radiosensitivity? Radiother ()ncol 1996;38: 10313. West CM. Davidson SE, Roberts SA. Hunter RD. Intrinsic radiosensitivity and prediction of patient response to radiotherapy for carcinoma of the cer- vix. BrJ Cancer 1993;68:819-23. Burnet NG. Nyman J. Turesson 1. Wurm R. Yarnold JR. Peacock JH. Prediction of normal—tissue tolerance to radiotherapy from in-vitro cellular radiation sensitivity. Lancet 1992;339:1570-1. Johansen J. Bentzen SM. Overgaard J. Overgaard M. Evidence fora posi- tive correlation between in vitro radiosensitivity of normal human fibroblasts and the occurrence of subcutaneous fibrosis after radiotherapy. Int J Radiat Biol 1994;66:407-12. Geara FB. Peters LJ. Ang KK. Wike JL, Brock WA. Prospective coin- parison of in vitro normal cell radiosensitivity and normal tissue reactions in radiotherapy patients. Int J Radiat Oncol Biol Phys 1993;27:1 173-9. Bentzen SM. Overgaard J. Patient—to—patient variability in the expression . of radiation-induced normal—tissue injury. Semin Radiat Oncol 1994:4268- (90) (9/) 80. Stausbol-Gron B. Nielsen OS. Bentzen SM. Overgaard J. Selective assess- ment of in vitro radiosensitivity of tumour cells and fibroblasts from single tumour biopsies using immunocytochemical identification of colonies in the soft agar clonogenic assay. Radiother Oncol 1995;37:87-99. Wheeler JA. Stephens LC. Tomos C. Eifel PJ. Ang KK. Milas L. et a1. ASTRO Research Fellowship: apoptosis as a predictor of tumor response to radiation in stage IB cervical carcinoma. American Society for Thera- peutic Radiology and Oncology. Int J Radiat Oncol Biol Phys 1995: 31148793. Journal of the National Cancer Institute Monographs No. 21. 1996 Radiotherapy for the Treatment of Locally Recurrent Cervical Cancer Rachelle Lancian0* Following radical hysterectomy for stage I-IIA cervical car- cinoma, 10%-15% of women will have a recurrence. Sixty percent of these recurrences will be located in the pelvis alone. The treatment of a localized pelvic recurrence depends on a number of factors, but primarily the history of previous radiation therapy (RT) to the pelvis and the loca- tion of the recurrence. For patients without a history of prior treatment, RT with or without chemotherapy can lead to disease-free survival rates of 20%-50% and local control rates of 20%-60%, and should be considered first-line treat- ment. For those with a history of previous RT and sidewall or nodal recurrences, a combination of surgery and brachytherapy or intraoperative radiotherapy may improve survival for these women who were previously considered incurable. However, this requires special expertise and tech- nology not available at most centers. [Monogr Natl Cancer Inst 1996;21:113-5] Approximately 10%-15% of all patients will have recurrence following radical hysterectomy for stage I-IIA cervical car- cinoma (1). Sites of recurrence following radical hysterectomy include pelvis only in 60%, distant only in 20%, and combined sites in 20% (2). The most common sites of recurrence within the pelvis include vaginal vault, followed by parametrium/ sidewall and, less commonly, bladder/rectum. Pelvic recurrence rates increase with a tumor size of more than 4 cm, higher stage, adenocarcinoma histology, deep cervical stromal invasion, vas- cular space invasion, parametrial extension, involved margins, and, most important, involved pelvic lymph nodes (1—4). Post- operative pelvic external radiation therapy (RT) with or without vaginal brachytherapy has been used as an adjuvant treatment for patients with these high—risk features for pelvic recurrence following radical hysterectomy and has become the standard of care for patients with positive pelvic lymph nodes or involved margins. Therefore, when reviewing the results of treatment for localized pelvic recurrence following radical hysterectomy for cervical cancer, the most important factor that must be con- sidered is the history of previous adjuvant external pelvic RT. Other factors to be considered include the location of the pelvic recurrence (central versus sidewall), medical/surgical risk fac- tors, and the expertise/technology available. Locally Recurrent Cervical Cancer Following Hysterectomy: No Prior RT Eight series of RT alone and one series of concurrent chemo— therapy and RT have been reported for pelvic recurrence follow- Journal of the National Cancer Institute Monographs No. 21. 1996 ing radical hysterectomy (Table I) (2,5—12). Most patients presented with parametrial/sidewall with or without central recurrence. The median time to recurrence ranged from 12 to 24 months. Patients typically had negative lymph nodes at initial presentation and did not receive adjuvant pelvic RT. Radiation schemas varied by series but usually included external RT (55— 65 Gy), with a brachytherapy vaginal boost if the recurrence was confined to the vagina. The 5-year survival rates ranged from 16% to 47%, with pelvic control rates from 21% to 61%. The prognostic factors reported were variably associated with improved survival following RT for pelvic recurrence. These factors included a long interval from initial surgery and recur- rence, negative pelvic lymph nodes at initial presentation, central pelvic location, and nonpalpable or small-volume recur— rence. The most significant prognostic factors were related to recurrence location and size. Larson et al. (8) reported an 83% 5—year disease-free survival (DFS) rate for patients treated with RT for a central recurrence compared with a 33% DFS rate for a sidewall recurrence, with a median follow-up of 48 months. Job— sen et al. (9) described 100% pelvic control if the recurrence was confined to the vagina compared with 50% if the parametrium/sidewall was involved, with a minimum follow-up of 25 months. Ito et a1. (11) reported 48 patients with vaginal— only recurrence and found a significant improvement in the 5- year survival rate of 80% if the recurrence was nonpalpable compared with 8% for palpable disease. The improved outcome for central pelvic recurrence and small-volume disease may be related to the improved radiation dose delivery to the tumor with brachytherapy. Thomas et a1. (12) reported the only experience with combined fluorouracil and twice—daily external RT during chemotherapy. Although this was a high-risk group of 41 patients with 78% parametrial or sidewall involvement, the 5- year survival and DFS rates were 40% and 45%, respectively. Local tumor control was 52%, with a median follow—up of 57 months. Some series include the institutional experience with pelvic exenteration, which should not be compared with the RT series, since the patients treated with radical surgery have usual- ly failed adjuvant RT as their initial treatment and have central— only, small—volume disease. Locally Recurrent Cervical Cancer Following Hysterectomy: Prior RT RT has also been used as an adjunct to surgery (usually ex- enteration), either with brachytherapy or intraoperative electrons *Correspondence to: Rachelle Lanciano, M.D., Department of Radiation On— cology. Delaware County Memorial Hospital. 501 N. Lansdowne Ave, Drexel Hill. PA 19026—1186. Table l. Locally recurrent cervical cancer: no prior RT* 1nvestigator(s)/ No. of Parametrium/ 5»year DFS. LC, Medizm follow- affiliation Year patients sidewall. ’71 survival. (/1 0/! ‘/( Llp survivors Barber and O‘Neil (2)/MSK 1971 23 2- 17 — 7 All >5 y Deutsch and Parsons (5)/ 1974 38 16 16 21 All >2.5 y U. of Pittsburgh 46.5 mo Krebs et al. (6)/U. of Miatni 1982 22 V 18 14 ~ All >5 y Chen et a1. (7)/Okayama Ui, Japan 1985 40 0 40 — i All >3 y 775 100 0 — A (87% 25 y) 45 1 | 35 Larson et al. (AD/MDA 1988 6 0 v 83 50 mo 2 100 33 47 mo 15 60 53 48 mo Jobsen et al. (9)/U. of Leiden, 1989 4 (1 44 39 100 Minimum follow. The Netherlands 14 100 50 up, 25 mo 18 78 6| Potter et al. (l())/U. of Alabama 1990 I 28 34 2 36 v 74 mo Ito et al. (Il)/Keto U.. Japan 1991 28 0 8'1’ — i — 20 801‘ v 2 i 48 47 Thomas et al. (/2)/Toronto 1993 41 78 40 45 52 57 mo Bayview PMH *RT 2 radiation therapy; DFS = disease-free survival: LC 2 local control: MSK : D. Anderson Cancer Center; PMH = Princess Margaret Hospital. 'J'Ten-year survival (palpable versus nonpalpable). (IORT), for a localized recurrence when pelvic RT has been pre— viously delivered for the initial cervical cancer management (Table 2). Hockel et al. (13) have reported a unique experience with radical surgery and brachytherapy delivered to the pelvic sidewall. The principles of the combined operative and RT (CORT) technique included maximum debulking of the tumor from the sidewall; implantation of brachytherapy catheters on the residual tumor or tumor bed; muscle flaps to cover catheters; and reconstruction of the bladder, rectum, vulvoperineum, and vagina. The brachytherapy consisted of a fractionated high-dose rate, single—plane implant. The dose delivered was 30—60 Gy given twice weekly over 2.5—5 weeks. Hockel et al. (/3) reported 18 patients with sidewall disease treated with CORT, 83% of whom had received previous RT at initial presentation. They reported 56% pelvic control and a 39% DFS rate, with a median follow-up of 19 months for survivors. Their results with CORT were comparable with 14 patients from the same institu— tion with central recurrence treated with exenteration. CORT or exenteration was not recommended for recurrences of 5 cm or more in size, since the 4-year survival rate was 62% for recur— rences of less than 5 cm compared with 6% for larger lesions. Two consecutive reviews from the Memorial Sloan—Kettering Cancer Center (New York, NY) of larger numbers of patients Memorial Sloan—Kettering Cancer Center; MDA = The University of Texas M. treated with interstitial brachytherapy reported only 3% and 10% 5-year DFS rates for this same patient population (14.15). IORT is another technique to deliver RT to a localized vol- ume when re—treatment is necessary following optimal debulk- ing (Table 3). Two single institutional experiences have been recently reported, with local control rates of 48%—71% and 5- year DFS rates of 40%—43% (16,17). Both reports noted an im- provetnent in outcome for patients with microscopic residual disease prior to IORT, with a local control rate of 80% and 5— year DFS rate of 70%. A multi-institutional French experience (/8) reported a much lower local control rate of 21% and a 3— year survival rate of only 8%. There was no significant dif- ference between the percentage of patients with gross residual disease or previous RT to account for this reduced survival. Conclusions For patients with a localized pelvic recurrence and n0 pre— vious history of RT, high-dose external RT with or without brachytherapy results in a 5-year DFS rate of 40%-50% in recent series and should be considered first—line treatment. The use of chemotherapy was reported in only one large series but was associated with high local control and DFS rates. For Table 2.1nterstitial implant with or without surgery* Median survival/ lnvestigator(s)/ No. of Gross residual. Previous 57Year LC, median follow—up affiliation Year patients (71‘ RT, "/( DFS. '/r ‘Vr survivors Evans et al. (l4)/MSK 1971 70 — ~80 3 7 126/132 mo Nori et a1. (l5)/MSK 1981 75 — 75 10 — ~12 mol— Hockel et al. (I3)/U. of Main/,. Federal 1994 18 — 83 39 56 14 mo/l9 mo Republic of Germany (CORT) 114 *RT 2 radiation therapy: DFS = disease»free survival; LC 2 local control; MSK 2 Memorial Sloan—Kettering Cancer Center; CORT 2 combined operative and RT. Journal of the National Cancer Institute Monographs No, 21. 1996 Table 3. Locally recurrent cervical cancer: prior RT* Median survival/ lnvestigator(s)/ No. of Gross residual, Previous 5-Year LC. median follow-up affiliation Year patients ‘Yr RT. “/( DFS. We ‘71 survivors Garlon et al. (NU/Mayo Clinic 1993 21 52 4X 40 71 23/42 mo Stel/er et al. (17)/U. of Washington I995 22 55 68 43 (DSS) 4X 26 1110/— I996 70 57 7| l l rho/36 mo Mahe et al. (/8)/Nantes. Frances 8(3-y) 21 *RT = radiation therapy: DFS : disease—free survival; LC 2 local control; DSS = disease-specific survival. patients with a localized pelvic recurrence and a prior history of RT. comparable DFS rates are possible with a combined surgi- cal—RT approach. where RT is delivered either with brachy- therapy (CORT) or electrons (IORT). References (I) (2) (3 (4 (5 (6) 3' (8 Alvarez RD. Potter ME. Soong SJ. Gay FL. Hatch KD. Partridge Eli. et al. Rationale for using pathologic tumor dimensions and nodal status to sub- classify surgically treated stage IB cervical cancer patients. Gynecol ()ncol l99l;43:l()8»l2. Barber HR. O‘Neil WH. Recurrent cervical cancer after treatment by a primary surgical program. Obstet Gynecol |97|z372165-72. Monk BJ. Cha DS. Walker .lL. Burger RA. Rarnsinghani NS. Manetta A. et al. Extent of disease as an indication for pelvic radiation following radi— cal hysterectomy and bilateral pelvic lymph node dissection in the treat— ment of stage IB and [IA cervical carcinoma l.\'(‘(‘ comment citation in Medlinel. Gynecol ()ncol l994;54:4-9. Morris M. Early cervical carcinoma: are two treatments better than one? [editorial] [sec comment citation in Medlinel. Gynecol Oncol l‘)9~l:5-l:l—3. Deutsch M. Parsons JA. Radiotherapy for carcinoma of the cervix recur- rent after surgery. Cancer 1974;34:2051-5. Krebs HB. Helmkamp BF. Sevin BU. Poliakoff SR. Nadji M. Averette HE. Recurrent cancer of the cervix following radical hysterectomy and pelvic node dissection. Obstet Gynecol 1982;59:422-7. Chen NJ. Okuda H. Sekiba K. Recurrent carcinoma of the vagina follow- ing Okabayashi's radical hysterectomy for cervical carcinoma. Gynecol Oncol 1985;20:10—6. Larson DM. Copeland LJ. Stringer CA. Gershenson DM. Malone JM Jr. Edwards CL. Recurrent cervical carcinoma after radical hysterectomy. Gynecol Oncol 1988;30:381—7. Journal of the National Cancer Institute Monographs No. 2|. 19% (9 (10 (ll) (/3) (l4) (/5) (16) (/7) (I8) Jobsen JJ. Leer JW. Cleton FJ. Hermans J. Treatment of locoregional recurrence of carcinoma of the cervix by radiotherapy after primary surgery. Gynecol Oncol [9891332368-7l. Potter ME. Alvarez RD. Gay FL. Shingleton HM. Soong SJ. Hatch KD. Optimal therapy for pelvic recurrence after radical hysterectomy for early- stage cervical cancer. Gynecol ()ncol l99l);37:74-7. Ito H. Kutnagaya H. Shigematsu N. Nishiguchi l. Nakayama T. Hashimoto S. High dose rate intracavitary brachytherapy for recurrent cervical cancer of the vaginal stutnp following hysterectomy. Int J Radiat ()ncol Biol Phys I99l;2():‘)27-32. Thomas GM. Dembo A]. Myhr T. Black B. Pringle JF. Rawling G. Long- term results of concurrent radiation and chemotherapy for carcinoma of the cervix recurrent after surgery. Int J Gynecol Cancer W931}: 193-8. Hockel M. Baussmann E. Mil/e M. Knapstein PG. Are pelvic side~wall recurrences of cervical cancer biologically different from central relapses? Cancer |994174:648—55. Evans SR .lr. Hilaris BS. Barber HR. External vs. interstitial irradiation in unresectable recurrent cancer ofthe cervix. Cancer |97 | :28: 1284-8. Nori D. Hilaris BS. Kim HS. Clark DG, Kim WS. Jones WB. et al. Inter- stitial irradiation in recurrent gynecological cancer. Int J Radiat Oncol Biol Phys 1981;71 15l3—7. Garton GR. Gunderson LL. Webb M]. Wilson TO. Martenson JA .lr. Cha SS. et al. lntraoperative radiation therapy in gynecologic cancer: the Mayo Clinic experience. Gynecol Oncol [993:48z328—32. Stelzer KJ. Koh WJ. Greer BE. Cain .lM. Tamimi HK. Figge DC. et al. The use of intraoperative radiation therapy in radical salvage for recurrent cervical cancer: outcome and toxicity. Am J Obstet Gynecol 1995;172: 188l-X. Mahe MA. Gerard JP. Dubois JB. Roussel A. Bussieres E. Delannes M. et al. lntraoperative radiation therapy in recun'ent carcinoma of the uterine cervix: report of the French lntraoperative Group on 70 patients. Int J Radiat Oncol Biol Phys 1996;34:21-6. 115 Surgery for the Treatment of Locally Recurrent Disease ManuelA. Penalver, Giselle Barreau, Bernd-Uwe Sevin, Hervy E. Averen‘e * Background and Methods: Total pelvic exenteration is a sal- vage procedure done in the effort to eliminate completely pelvic cancer. Low colorectal anastomosis and continent uri- nary diversion are two new procedures that allow complete pelvic evisceration without the need for external appliances. From 1984 through 1994, 67 patients have undergone rec- tosigmoid colectomy and low-colorectal anastomosis. Sixteen patients underwent surgery as part of a total pelvic ex- enteration for recurrent cervical cancer, and 51 patients un- derwent surgery for either primary or recurrent ovarian carcinoma as part of an optimal debulking procedure. Be- tween 1988 and 1995, 55 patients have received continent urinary diversion with the Miami Pouch. Fifty-two patients underwent surgery for recurrent cervical cancer, two patients for advanced vulvar cancer, and one patient for a vesico-vaginal fistula. All of the patients with recurrent cer- vical cancer had previously received radiation therapy for gynecologic cancer. Results: Of the 16 patients with recur- rent cervical cancer who had a low colorectal anastomosis, 14 had a temporary colostomy. Of these 14 patients, eight had a colostomy takedown and have maintained fecal con- tinence. Of the 51 patients with ovarian cancer who had a low colorectal anastomosis, all achieved fecal continence. With the Miami Pouch, a urinary continence rate of 86% was obtained. Twenty-four (44%) patients had early com- plications, including ureteral obstruction, ureterocolonic anastomotic leak, reservoir cutaneous fistula, small bowel obstruction, and pyelonephritis. Nineteen (35%) patients had late complications, including ureteral reflux, urinary in- continence, difficult catheterizations, and reservoir stones. There was a perioperative mortality rate of 5%. Con- clusions: Low-colorectal anastomosis is an attractive alter- native to permanent colostomy, allowing all patients who had the protective colostomies taken down to achieve fecal continence. Continent urinary diversion with the Miami Pouch is also a worthwhile procedure because of its high continence rate. Although survival advantage for either pro- cedure has not been proven, the quality of life of patients un- dergoing such procedures has been substantially improved because of the avoidance of external appliances. This has been achieved with acceptable morbidity and mortality rates. [Monogr Natl Cancer Inst 1996;21:117-22] When pelvic cancer recurs, the only opportunity for total cure lies in the ability of the cancer to be completely resected. Be- cause of the natural history of many pelvic cancers. they may be Journal of the National Cancer Institute Monographs No. 21. 1996 locally advanced but still limited to the pelvis, therefore lending themselves to radical resection. Pelvic exenteration is a salvage procedure performed for central pelvic-recurrent gynecologic cancer. The majority of patients who undergo pelvic exentera— tion have a diagnosis of recurrent cervical cancer; however. the procedure has been done for recurrent endometrial as well as 10- cally advanced vulvar carcinoma. Due to its pattern of spread, recurrent ovarian carcinoma is rarely amenable to pelvic ex- enteration. This procedure offers hope to selected patients when no other possibility of cure exists. Furthermore. patients who survive this procedure can be rehabilitated to healthy living with a good quality of life. Recurrent cervical cancer has a l-year survival rate of 15% and a 5-year survival rate of less than 5%. A careful analysis of the patients with recurrent cervical cancer identifies a group of patients (approximately 25%) whose recurrence is in the central pelvis and. therefore. amenable to pelvic exenteration. Survival rates for these patients undergoing pelvic exenteration range from 32% to 62% in the literature (1—6). Recurrences of cervical cancer after radiation therapy are diagnosed by cytology, vaginal bleeding or discharge, hema- turia, or an increase in serum tumor antigen level. Occasionally, the first evidence of central recurrence is ureteral obstruction. A diagnosis of central recurrence is usually made by punch biopsy of the lesion. fine needle aspiration of a pelvic mass. cone biop- sy of the cervix. or biopsy of the rectum or bladder, Preoperative evaluation for pelvic exenteration includes a thorough search for extrapelvic disease. The triad of leg edema, ureteral obstruction or stricture, and pain is very suggestive of unresectable disease. Disease outside of the pelvis is a contrain— dication to the operation and. therefore, investigation must be done to rule out extrapelvic extension. Radiologic tests used in— clude pelvic sonogram, abdominal/pelvic computed tomography (CT) scan. and occasionally a magnetic resonance imaging scan. A chest x ray to evaluate the thorax is performed. A CT scan of the chest can be used when the radiograph is inconclusive. A bone scan may also be requested if there is suspicion of bony lesions. At surgery, a search for extrapelvic disease must be diligently performed. including running of the small and large intestines and biopsy of any suspicious omental. peritoneal. or retroperitoneal lesions. *Afiiliuriun of authors: Division of Gynecologic Oncology, University of Miami. Jackson Memorial Medical Center. FL. ("rirrespmulem‘1' m: Manuel A, Penalver. M.D.. Division of Gynecologic On- cology. University of Miami. Jackson Memorial Medical Center. East Tower Bldg. Suite 7007. Miami. FL 33136. ”7 Surgery begins with a para—aortic lymphadenectomy. Any metastatic disease to this area is a contraindication to the surgery. When possible. a pelvic lymphadenectomy is per- formed. Gross metastasis to this area also contraindicates the surgery. Performance of a pelvic and a para—aortic lympha- denectomy can be technically difficult due to the previous radia- tion therapy changes. On the basis of the preoperative examination under anesthesia and intraoperative findings. a decision is made as to whether to perform a supralevator or in- fralevator exenteration. Advantages of a supralevator exentera- tion include the preservation of a vaginal and an anorectal stump for elongation of the vagina and for low colorectal anastomosis to prevent a permanent colostomy. respectively. To perform a supralevator exenteration. a 2—cm free margin must be obtained to try to prevent local recurrence, When a supralevator exentera- tion is done. the vagina can be elongated using the sigmoid colon. the Martius bulbocavernosus flaps. or an omental wrap. Operative mortality associated with pelvic exenteration today should be no more than 5%. The risks depend. to a great extent. on the skill and experience of the surgeon: however. patient selection is also a very important determinant. As technology advances and surgical techniques are im- proved. more modifications to this type of radical surgery will be made. Two such modifications that have been introduced are low colorectal anastomosis and continent urinary diversion. These procedures improve the patient’s self—esteem and quality of life by eliminating the need for external appliances. yet do not compromise the rate of cure achieved with such extensive surgery. Low Colorectal Anastomosis Surgical Procedure Previously. a rectosigmoid resection resulted in the formation of a permanent colostomy: however. with recent advances in low colorectal anastomosis using the end—to—end anastomosis (EEA) stapler. reanastomosis is now used more frequently (7). Low colorectal anastomosis is defined as any anastomosis of the colon to the rectum located less than 6 cm above the anal verge. The EEA stapler permits safe performance of a low anastomosis at a level only 2-3 cm from the internal anal sphincter (8). The EEA stapler is introduced through the anus and the anvil of the stapler is introduced into the proximal segment of the descend— ing colon. These two segments are approximated and the stapler is fired to achieve the anastomosis. After the stapler is removed. the integrity of the donuts in the stapler must be established to ensure a complete anastomosis. To further confirm the anas— tomotic integrity. saline solution is instilled into the pelvis using an Asepto bulb syringe. and a small amount of air is insuffiated into the rectum. If there are any bubbles produced. the anas- tomotic leak must be identified and corrected with reinforcing sutures. In most cases. this can be achieved without a takedown and redoing the procedure. Results From 1984 through 1994 at the University of Miami School of Medicine. Division ofGynecologic Oncology. 67 patients un- H8 derwent rectosigmoid colectomy and reanastomosis. Sixteen patients underwent surgery as part of a total pelvic exenteration for recurrrent cervical cancer. and 51 underwent surgery for either primary or recurrent ovarian carcinoma as part of an op— timal debulking procedure. The In patients with recurrent cervi- cal cancer had previously received standard doses of radiation therapy for gynecologic cancer. whereas the other 5] patients with ovarian cancer had not received radiation therapy. All patients had the colorectal anastomosis performed with the EEA stapler (U.S. Surgical Corp, Norwalk. CT). The anastomosis was performed 2—3 cm above the levator ani muscle in all pa— tients. Of the 16 patients with cervical cancer who had a rectosig- moid resection. 14 received a temporary colostomy. Both pa— tients who did not receive a protective colostomy had an anastomotic dehiscence that required a colostomy. Of the l4 patients with a protective colostomy. eight had their colostomies taken down and have maintained fecal continence. Four patients developed recurrent cancer and did not have the colostomy taken down. and the other patients refused further surgery. Con— sequently. of the 16 patients with recurrent cervical cancer who underwent rectosigmoid resection. eight achieved fecal con— tinence with a low colorectal anastotnosis. Therefore. of those eight patients who had the temporary colostomy taken down and low colorectal anastomosis performed. all eight have achieved and maintained fecal continence. Discussion On the basis of the experience of the University of Miami School of Medicine. Division of Gynecologic Oncology. and other reports in the literature (9.]()). the rectosigmoid colectomy and low colorectal anastomosis can be performed effectively and safely in gynecologic oncology patients as part of a pelvic exenteration and debulking procedure. The EEA stapler is tech- nically simple to use and permits the safe reanastomosis of the colon and rectum. lts application to gynecologic cancer surgery is exemplary in that gynecologic cancers frequently involve the rectosigmoid. Furthermore. rectosigmoid colectomy and reanas- totnosis allows optimal cytoreduction and reduces the number of permanent and temporary c0l0stomies. One of the disadvantages of low colorectal anastomosis is fre- quent bowel movements. Fecal frequency is defined as more than four bowel movements per day. Although not all patients experience tenesmus and fecal frequency. it is bothersome and results in a diminished quality of life for those patients who do. Wheeless et al. (II) have shown a statistically significant dif- ference in fecal frequency between the end—to—end anastomosis and the J pouch reservoir. The creation of the J pouch reservoir in previously radiated patients appears to restore the rectal am— pulla and may be an effective method of decreasing tenesmus and fecal frequency. Low colorectal anastomosis is an attractive alternative to per- manent colostomy because it leaves the patient stoma free without the need for an external appliance. However. the benefits of palliative surgery are difficult to quantitate. Conse- quently. even though complete resection of pelvic disease can be accomplished with this operation. it should not be done in Journal ofthe National Cancer Institute Monographs No. 2 l. l996 those patients with unresectable disease in the upper abdomen For these patients. colostomy should be performed for rectosig- moid obstruction. Rectosigmoid colectomy and reanastomosis is a safe procedure that avoids colostomy and. although no sur— vival advantage has been proven. it is a worthwhile procedure that should be taught as part of gynecologic oncology training. Continent Urinary Diversion: the Miami Pouch Surgical Procedure For many years, gynecologic oncologists used a colon or ileoconduit for urinary diversion (/2). A conduit is an isolated intact segment of intestine to which the ureters are anastomosed that functions to carry the urine outside the body. The patient is incontinent and wears an external appliance. Disadvantages often encountered with conduit diversion include deterioration of renal function ([3), the problems inherent to the stoma. and the constant wearing of an external appliance. Since then. several investigators have developed different forms of urinary diversion procedures with the construction of continent urinary reservoirs. In 1975. Kock et al. (14) introduced the continent ileal reservoir. They found that intestinal segments prepared as a reservoir with the interruption of the circular fibers accorn— rnodate larger urine volumes at lower pressures than tubular in- testinal segments. This is the principle of detubularization of bowel to create a low—pressure reservoir. Following the work of Kock et al.. several techniques using the small and large intes- tine for continent urinary diversion have been reported (15—17). Since 1988, the Division of Gynecologic Oncology at the University of Miami School of Medicine has used a continent ileocolonic reservoir. the Miami Pouch. for urinary diversion during pelvic exenteration. The Miami Pouch is a colonic reser— voir with a tapered distal ileal segment and with purse-string su— tures around the ileocecal valve that has created continence. easy catheterization. and an acceptable incidence of reflux. Creating the Miami Pouch for continent urinary diversion is an extensive procedure with many ramifications that requires long—term patient acceptance and adaptation. All patients must be motivated and able to understand the implications of self- catheterization. The basic technique of the Miami Pouch centers around a differential in pressure between the ileal and colonic segments. If the pressure within the ileal segment is greater than the pressure within the colonic segment. all patients should be continent of urine. The technique used to decrease the colonic pressure is detubularization as described by Kock et al. The dis- tal ileum is transected 1() cm proximal to the ileocecal valve. The cecum and ascending colon are mobilized up to the right colic flexure and the transverse colon is transected just distal to the middle colic artery. The ascending and transverse colon are opened and anastomosed in an inverted U-shaped manner to create a reservoir. The ileal pressure is increased by tapering of the lumen and placing purse-string sutures at the level of the ileocecal valve. The ureters are reimplanted in a nontunneled and nonretluxing manner into the posterior wall of the colonic reservoir. This is done by bringing 2 cm of ureter inside the reservoir. spatulating the distal centimeter. and using 4-0 poly— glycolic acid sutures to anastomose it to the submucosal layer of the colon. The proximal centimeter of the ureter inside the reser— voir is left intact so that as the reservoir fills. it creates a tam— ponade effect that prevents reflux. The anterior wall of the reservoir is closed with absorbable sutures and then secured to the fascia of the anterior abdominal wall. Patient follow—up includes office visits every 3 months after the initial postoperative visit. In addition to the routine examina- tion. an interview is conducted to evaluate catheterization and continence. Serum electrolytes. serum creatinine, and a culture of the reservoir are obtained at each visit. Radiographs, such as an intravenous pyelogram and contrast reservoir studies. creatinine clearance. and urodynamic tests. are repeated every 6 months. Results From 1988 through 1995. a total of 55 patients have received this procedure as an alternative to other forms of urinary diver- sion at the time of exenteration. Fifty—two patients had surgery as part of a total pelvic exenteration for recurrent cervical can— cer. Two patients had surgery for advanced vulvar cancer and one patient had surgery for a vesico—vaginal fistula. The data ob- tained were analyzed in terms of complications resulting direct- ly from the surgery or from this form of urinary diversion. Early complications were defined as less than 6 weeks after surgery and late complications were defined as more than 6 weeks after surgery. Early complications included ureteral obstruction, ureterocolonic anastornotic leak. reservoir cutaneous fistula. small bowel obstruction. and pyelonephritis (Table l). A total of 24 (44%) patients had early complications. including a peri- operative mortality rate of 5% (three of 55 patients). The late complication rate was 35% and included ureteral reflux. urinary incontinence. difficult catheterizations. and reservoir stones (Table 2). Ureteral obstruction occurred in 7% (four of 55) of the patients. These patients were managed with percutaneous dilata- tion. One patient required a reoperation to relieve the obstruc- tion. Uterocolonic leaks occurred in 4% (two of 55 patients) of the patients. These patients were managed with percutaneous nephrostomy tubes: however. one patient required reoperation. Table l. Early complications of the Miami Pouch No. of patients Complication Ureter obstruction/stricture 4 Ureterocolonic leak 2 Reservoir fistula 3 Management Percutaneous dilatation Percutaneous nephrostomy Percutaneous nephrostomy Outcome (No. of patients) Resolution (2); partial resolution ( l ): reoperation ( 1) Resolution ( l ); reoperation ( 1) Resolution ( l )1 reoperation (2) Journal of the National Cancer Institute Monographs No. 2 l. l996 ll9 Table 2. Late complications of the Miami Pouch Complication No. of patients Incontinence 7 Difficult catheterization 5 Urinary stones 3 Scheduled catheterixations Scheduled catheterizations Management ()titcome (No. of patients) Resolution (5); reoperation (2) Resolution (5) Resolution (2); reoperation (I) Reservoir cutaneous fistulas occurred in 5% (three of 55) of the patients. Percutaneous nephrostomy resolved the problem in one patient; however. the other required reoperation. Difficulty with self—catheterization occurred in 9% (five of 55 patients) of the patients. These patients were managed with frequent scheduled catheterizations. patient education regarding position changes for self-catheterization. and ultrasound-guided needle emptying of the reservoir. All five patients had complete resolution of this difficulty. With the Miami Pouch. a continence rate of 86% has been achieved. Forty-seven of 55 patients remained continent with intermittent catheterization at 4-8-hour intervals. Five patients had incontinence. Urodynamic studies revealed higher ileal pressures than reservoir pressures in all continent patients. The incontinent patients had reservoir pressure greater than the tapered ileal pressure. Renal function data were available for 30 of the 55 patients. Of these 30 patients. two had a deterioration in renal function. Both patients had normal serum creatinine levels preoperatively (normal, 0.4-1.5 mg/IOO mL) and stabilized at 3.1 and 3.3 mg/IOO mL postoperatively. However. it was not associated with reflux, obstruction. or infection. It was attributed to amino- glycoside antimicrobial use postoperatively. In the other 28 patients. renal function remained unchanged at each follow-up visit. As of December 1995, 38 of the 55 patients were alive. Four— teen of the 17 deaths occurred from metastatic disease. Three patients died in the postoperative period. Two deaths were re— lated to the urinary diversion and the other death was attributed to disseminated intravascular coagulopathy. Surgical Modifications In an effort to reduce the incidence of complications with the Miami Pouch. a few modifications have been successfully at- tempted. Because of frequent stricture formation and obstruction at the ureterocolonic anastomosis, the colonic sulcus has been omitted from the surgical technique. Since it is radiated bowel. the less manipulation of the intestine at the time of surgery. the greater the possibility of reducing inflammation. edema. and consequent fibrosis. Also. the Malecot catheter has been discon— tinued because it leaves a large defect in a radiated intestinal segment. The role of the Malecot catheter was to provide an ac- cess site for saline flushes to eliminate the mucus produced by the colonic reservoir. thereby preventing stomal obstruction by a mucus plug. Since discontinuation of the Malecot catheter. how— ever. no patient has developed obstruction from a mucus plug. Another complication that was often encountered was difficulty with catheterizations. To correct this problem. the urinary reser- voir is now securely attached to the fascia of the anterior ab— dotninal wall. This allows a direct pathway for the catheter at l20 the time of self—catheterization. It also prevents kinking and twisting of the reservoir so that proper filling may occur. Discussion The goal of a suitable urinary reservoir is to construct a high— capacity. low—pressure reservoir that preserves the upper tract and achieves continence. The ileal or colon conduit should be the standard to which the risk and benefits of continent urinary diversion procedures are compared. The literature is replete with descriptions of new operations and suggested alternatives to standard and tested urologic procedures (18—20). The Miami Pouch was first used in patients with gynecologic cancer in 1988. The intention was to create a simple and more reliable technique to provide continence in previously radiated patients. With the Miami Pouch. a continence rate of 86% (47 of 55 patients) has been achieved. This high continence rate may be attributed to the reinforcement of the ileocecal valve with three purse—string sutures and the tapering ()f the distal ileal segment. in addition to the detubularization of the colonic segment. It is a well-recognized clinical phenomenon that the ileocecal valve is not reflux proof. Therefore. if the reservoir becomes distended. the ileocecal valve alone may not be sufficient to prevent uri- nary leakage. Consequently, these modifications of the Miami Pouch provide the necessary reinforcement of the ileocecal valve to achieve urinary continence. Deterioration of renal function has always been a concern of surgeons performing urinary diversion. However. the procedure of nontunnelled ureterocolonic anastomosis used with the Miami Pouch has shown good results. The incidence of renal failure. obstruction. and reflux has been similar to other forms of ureterointestinal anastomosis (21.22). The antireflux mechanism of this procedure is the low-pressure reservoir and the thickness of the muscular colonic wall that encircles the ureters when the reservoir is distended. With the use of this technique. reflux oc- curred in five (9%) patients. However. no evidence of renal deterioration occurred. Another postoperative complication that is often feared more than reflux is ureteral obstruction because it requires immediate treatment to prevent symptomatic urinary in- fection and renal failure. Ureteral obstruction occurred in four (7%) patients. Treatment of the anastomotic strictures consisted of percutaneous dilatation. and in one instance. reimplantation. A prolonged follow-up is required to properly assess the impact ofcontinent urinary diversion on renal function. Past experiences indicate that continence. ease of catheteriza— tion. and reflux are the areas that need improvement. Eight (15%) patients required revisional surgery for early or late com- plications. A conservative approach to the complication was at- tempted initially on all patients and reoperation was only Journal ofthe National Cancer Institute Monographs No. 21. 1996 attempted if this regimen failed. The conservative approach is recommended because of the high mortality rate of 2075—5076 that a second operation in the immediate postoperative period may carry (23.24). Reviews of other surgical procedures for continent urinary diversion show a comparable reoperation rate (25—27). It is anticipated that further refinements and modifica— tions in the surgical technique will reduce the need for surgical revision. Neovaginal Reconstruction The sexual rehabilitation of the patient after this extensive and disfiguring surgery is an important consideration. Re- construction after radical pelvic surgery requires large amounts of well-vascularized tissue that can be delivered by relatively few flaps. The two most frequently used flaps that fulfill this criteria are the gracilis muscle skin flap (28.29) and the rectus abdominis flap. Placement of these flaps in the pelvis reduces the risk of bowel obstruction and/or fistula due to the bowel ad- hering to a denuded pelvis during the postoperative period. The advantage of the rectus abdominis flap is that it closes in con- tinuity with a laparotomy incision. thus avoiding donor sites on the thighs. Restoration of surface quality is excellent with rectus abdotninis flaps and extensive dilators are not required to main— tain luminal dimension. Rectus abdominis flaps provide good vascularization for irradiated pelvic wounds. and the rate of in— fection after radical pelvic surgery has fallen dramatically with its use. The gracilis myocutaneous flap has been used extensively for vaginal reconstruction and pelvic floor management at pelvic exenteration. Advantages of gracilis flaps include restoration of a functional vagina and neovascularization to the pelvic floor. Disadvantages include pain in the donor sites of the thighs and the healing associated with a separate scar. At the University of Miami. Division of Gynecologic Oncology. most of the ex- periences with vaginal reconstruction have involved a sigmoid colon neovagina and the gracilis myocutaneous flaps. Although the sigmoid neovagina provides an adequate length of vagina. disadvantages include the excessive production of mucus. At the present time. data are being collected for a study of the rectus abdominis flaps for vaginal reconstruction. Conclusion Although the pelvic exenteration surgery is used for recurrent cervical cancer. it has also been used as primary treatment and as palliation. Patients who are sometimes considered for primary exenteration without having received prior radiation therapy in— clude patients whose disease extends into the bladder or rectum. These patients are considered for primary exenteration because of the difficulties of managing fistulas from the bladder and/or rectum associated with radiotherapy treatment. Although pelvic exenteration was originally described as a palliative procedure. it is rarely used for this purpose. A palliative exenteration has also been considered for patients who are symptomatic from pain or odor; however. they must understand the palliative intent of the operation. Similar to any surgical procedure. there is a learning curve where complications are less likely to occur with practice and Journal ofthe National Cancer Institute Monographs No. 21. I996 experience. Up to the present time. we are satisfied with the progress made with the Miami Pouch. It is a worthwhile proce— dure that can improve the quality of life for those patients under- going exenteration. In patients with gynecologic cancers, continent urinary diversion and low colorectal anastomosis allow patients to undergo radical surgery without the necessity of external appliances. Although a survival advantage for either procedure has not been proven. the quality of life of patients un— dergoing such procedures can be substantially improved. This has been achieved with acceptable morbidity and mortality rates. Consequently. teaching institutions should train their physicians in the surgical techniques necessary to carry out such extensive procedures. References (I) Krieger JS. Embree HK. Pelvic exenteration. Cleve Clin Q 1969;36:1-8. (2) Bricker EM. Pelvic exenteration. Advances in surgery. Vol. 4. Chicago: Yearbook. 1970. (3) Galante M. Hill EC. Pelvic exenteration: a critical analysis of a ten-year experience with the use of the team approach Am J Obstet Gynecol 1971; | 10: | 80. (J) Symmonds RE. Pratt JH. Webb MJ. Exenterative operations: experience with 198 patients. Am J Obstet Gynecol 1975;121:907-18. (a) Morley GW. Lindenauer SM. Pelvic exenterative therapy for gynecologic malignancy: an analysis of 70 cases. Cancer 1976;38tl Suppl):581~fi. to) Rutledge FN. Smith JP. Wharton JT. ()‘Quinn AG. Pelvic exenteration: analysis 012%) patients. Am J Obstet Gynecol l977:129:881—92. (7) Berek JS. Hacker NF. Lagasse LD. Rectosiginoid colectomy and reana— stomosis to facilitate resection of primary and recurrent gynecologic can— cer. Obstet Gynecol 1984164271521). (8) Wheeless CR Jr. Incidence of fecal incontinence after coloproctostomy below five centimeters in the rectum. Gynecol Oncol 1987;27:373-81. Hatch KD. Gelder MS. Soong SJ. Baker VV. Shingleton HM. Pelvic ex- enteration with low rectal anastomosis: survival. complications. and prog- nostic factors. Gynecol ()ncol 1990;38:462-7. (/0) Hatch KD. Shinglcton HM. Potter ME. Baker VV. Low rectal resection and anastomosis at the time of pelvic exenteration. Gynecol ()ncol 1988‘. 31:262-7. (II) Wheeless CR Jr. Hempling RE. Rectal J pouch reservoir to decrease the frequency of tenesmus and defecation in low coloproctostomy. Gynecol ()ncol 1989135: 136-8. (/2) Averette HE. Lichtinger M. Sevin BU. Girtanner RE. Pelvic exenteration: a 15-year experience in a general metropolitan hospital. Am J Obstet Gynecol 1984;150:179—84. (l3) McDougal WS. Kock MO. Accurate determination of renal function in patients with intestinal urinary diversions. J Urol 1987:135:| 175-8. Kock NG. Nilson Ali. Nilsson LO. Norlen LJ. Philipson BM. Urinary di\ersion via a continent ileal reservoir: clinical results in 12 patients. J l'rol 1982;128:469-75. Rowland RG. Mitchell ME. Bihrle R. The cecoileal continent urinary reservoir. WorIdJ Urol 1985:3185. (l6) Lockhan JL. An alternative method for continent supravesical diversion. Soc Ped Urol Newsletter 1987 Mar; p 18. Lockhart JL. Remodeled right colon: an alternative urinary reservoir. J Urol 1987;138:7304. Lockhan JL. Bejany DE. The antireflux ureteroileal reimplantation: an al« ternati\ e for urinary diversion. J Urol 1987;137:1467—70. (I9) Camey M. Le Duc A. L‘enterocystoplastie avec cystoprostatectomie totale pour cancer de la vessie. Ann Urol 1979;13:114. (20) Gilchrist RK. Merricks JW. Hamlin M11. Reiger 1T. Construction ofa sub— stitute bladder and urethra. Surg Gynecol Obstet 1950;90:752. (2/) ()rr JD. Shand JE. Walters DA. Kirkland lS. lleal conduit urinary diver- sion in children. An assessment of the long-term results. BrJ Urol 1981: 53:424-7. (22) Bejany DE. Politano V. Stapled and nonstapled tapered distal ileum for construction of a continent colonic urinary reservoir. J Urol 1988;140:491— 4. Lichtinger M. Small bowel complications after supravesical urinary diver— sion in pelvic exenteration. Meeting of Society of Gynecologic Oncol- ogists. Feb 1985. (9 (I4) (/5) (/7) (IX) 121 (24) Orr JW Jr‘ Shingleton HM. Hatch KD. Taylor PT. Partridge EE‘ Soong S]. (25) 136) Is.) Gastrointestinal complications associated with pelvic exenteration. Am J Obstet Gynecol 1983: |451325-32‘ Skinner DG. Lieskovsky G. Boyd SD‘ Continuing experience with the con- tinent ileal reservoir (Kock pouch) as an alternative to cutaneous urinary diversion: an update after 250 cases. J Urol l9871137: 1 I405. Lieskovsky G. Boyd SD. Skinner DG. Management of late complications ol'the Kock pouch form of urinary diversion] Urol 1987; | 37:1 I46. (27) (28) (29) Rowland RC. Mitchell ME. Bihrle Rt Alternative techniques for a con- tinent urinary reservoir. Urol Clin North Am l987;l4:797-8()4. Morrow CP. Lacey CG‘ Lucas WE. Reconstructive surgery in gynecologic cancer employing the gracilis myocutaneous pedicle graft. Gynecol ()ncol 1970:72176—87t Pursell SH. Day TG Jr. Tobin GR. Distally based rectus abdominis tlap for reconstruction in radical gynecologic procedures. Gynecol Oncol 1990.37: 234-8, Journal of the National Cancer Institute Monographs No. 21. 1996 Chemotherapy for Stage IVB or Recurrent Cancer of the Uterine Cervix George A. 0mura* Numerous drugs and drug combinations have been eval- uated, largely in phase I] trials, for the treatment of car- cinoma of the uterine cervix. There clearly is a need for effective chemotherapy for this disease, but the results to date have unfortunately not met that need. Cisplatin is ac- tive but overrated. Drug combinations have not shown a consistent advantage and require careful study, with the ini- tial focus on maximizing the complete response rate in women with no or minimal prior chemotherapy and with adequate sample size to have confidence about the results. The patient populations under treatment need to be careful- ly defined with respect to potential prognostic factors for response and survival, so that results will be more reproducible. The few large. randomized trials completed so far have failed to show a survival benefit with combination chemotherapy; in fact, there does not appear to have been a comparison of chemotherapy with best supportive care. Sys- temic therapy of cervical cancer remains experimental. [Monogr Natl Cancer Inst 1996;21:123-6] Chemotherapy of cancer of the uterine cervix is still quite un- satisfactory (l). Although some patients show notable responses to chemotherapy. the results tend to be inconsistent and not reproducible. The response rates are usually higher when chemotherapy is used as primary treatment than when it is used after surgery and/or radiotherapy. The disruption of the pelvic blood supply after these modalities. the likelihood that recurrent or persistent foci of cancer are more drug resistant. and the decreased tolerance for bone marrow—toxic agents exhibited by some patients after pelvic radiotherapy may contribute to the in— ferior results obtained when chemotherapy is given as a seconv dary treatment. With respect to recurrent cancers in pelvic versus extrapelvic sites. some investigators report no difference in the responsiveness of these cancers to drugs: however. in general. extrapelvic sites seem to be more responsive. Pelvic lesions are frequently more difficult to measure and follow. Obstructive uropathy is common and may compromise the use of nephrotoxic or renally excreted drugs. There has not been a systematic study of chemotherapy in newly diagnosed stage IVB cases. separate from recurrent cases. so inferences must be made from reports of neoadjuvant chemotherapy in earlier stages and from the experience with ex— trapelvic recurrences. Stage lVB cases might be more respon- sive. for the reasons noted above, but they also could have a worse outcome because of their potentially more aggressive biology. In one recent Gynecologic Oncology Group (GOG) chemotherapy trial (2). only 7% of cases were newly diagnosed Journal of the National Cancer Institute Monographs No.21. I996 stage NB: the response rate (23%) was the same as for recur— rent cases. The present review includes several tabulations of drugs and drug combinations that have been evaluated during the past quarter century in cervix cancer. The listings are intended to be representative. rather than encyclopedic. The 95% confidence intervals have been included to underscore the imprecision of high response rates in small trials. Cisplatin has been cited repeatedly as the most active agent in the treatment of squamous cell cervix cancer (Table 1). The ex— perience with this drug is very extensive. but the results are ac— tually rather modest. Response rates in the 18%-21% range for the 50—mg/m2 dose have underscored the need to look else- where. In a comparison of 50 mg/m2 with 100 mg/mz. there was a dose—response relationship. with the response rate increasing from 21% to 31%. but there was only a minimal increase in complete response rate from 10% to 13% as well as no sig- nificant improvement in response duration. progression—free in- terval. or survival (3). Carboplatin and iproplatin. analogues of cisplatin with dif- ferent toxicity. have had only 15% and 11% response rates. respectively. Although they have been regarded as inferior to cisplatin in treating cervix cancer. the 95%) confidence limits for carboplatin overlap at least one of the large trials with cisplatin. Other “standard" drugs (Table 2) have given variable results. Only ifosfamide and anthracyclines have reproducibly had response rates above 15%. in one study. single-agent epirubicin gave a 45% response rate (eight complete responses [CR] + 10 partial responses [PR] in 40 assessable cases). but the response criteria were somewhat ambiguous. A confirmation of the vincristine results (29%) is needed. Vinblastine apparently has not been studied first line. With regard to drugs that are still investigational (Table 3). mitolactol has been notable. with a 29% response rate. Vinde- sine. a desacetyl derivative of Vinblastine. seems to be active. al- though no complete responses were reported. lrinotecan has some activity (24% response rate) and warrants a confirmatory trial. Several cisplatin combinations (Table 4) have yielded a wide range of response rates. Most regimens have given inconsistent results. Two regimens are currently popular, BIP (bleomycin. ifosfamide. and cisplatin) and MVAC (methotrexate. Vinblas— tine. doxorubicin. and cisplatin). BlP is being compared with *Afliliu/iuns‘ of’autlmr: Departments of Medicine and Gynecology. University of Alabama at Birmingham. and BioCryst Pharmaceuticals. 1nc.. Birmingham, AL. ('urrmlmmlem‘r to: George A. Omura. M.D.. BioCryst Pharmaceuticals. lnc.. 2190 Parkway Lake Dr.. Birmingham. AL 35244. Table l. Single-agent platinum in the treatment of cervix cancer (no priorchemotherapy) Regimen Response rate* 21 CR + 31PR/16613lz24-38) 7 11 CR+21 PR/128125: 18-33) Cisplatin. 50 mg/m‘ 15 CR + 16 PR/I50121; 15-28) Cisplatin. 50 mg/m2 19 CR + 37 PR/320118: 13»22) Cisplatin. 51) mg/m2 W 9 CR + 16 PR/137 (19; 12-25) lproplatin. 270 mg/m‘ 7 CR + 12 PR/17711127-17) Carboplatin. 400 mg/m 10 CR + 17 PR/l75 (15:9-20) Carhoplatin. 400 mg/m 2 CR + 4 PR/41 ( 15: 6-29) Cisplatin. 100 mg/mz Cisplatin. 21) mg/m‘ X 5 u 1.) Median survival. mo Reference 7 J Clin Oncol 3: 1079 6 l “ 7.1 “ 6.3 Gynecol Oncol 32:198 8 Gynecol Oncol 60: I 20 5.5 JClin Oncol 7:1462 6.2 .. e Gynecol Oncol 392332 *CR 2 complete response; PR = partial response. Overall response rates and 95% confidence intervals are in parentheses. Table 2. Other standard drugs used in the treatment ofcervix cancer (minimal or no prior chemotherapy) Agent Cyclophosphamide Cyclophosphamide Response rate”: 15/76 (20; 11—31)) 2/40(5: 1-17) Cyclophosphamide 2 PR/30 (7; 1-22) lfosfamide 6 CR + 6 PR/30 (40; 23—59) lfosfamide 1 CR + 9 PR/30133: 17—53) lfosfamide 2 CR + 6 PR/Sl (16; 7-29) Lomusline 1 CR + 1 PR/58 (310»12) Metholrexate 1 CR + l PR/2311313-2-l) Methotrexate 3 CR + 2 PR/33 (15; 5-32) Bleomycin 2 CR + 8 PR/32 (31; 1651)) Bleomycin 3 PR/231l3;3-34) llexamethylmelamine Hexamethylmelamine Vincristine 8 PR/34124:11~4|) 0/14 (0; 0-23) 9 PR/31 (29: 14448) Fluorouracil 20/208 1 10:6»15) Mitoxantrone 0/1710:()»20) Etoposide 0/17 (0; ()-20) Mitomycin 3 CR + 3 PR/52(12:-1-23) Paclitaxel 2 CR + 7 PR/52 (17: 8-30) Doxorubicin 2 CR + 4 PR/38 116:6-311 Survival 1 J. 'c 1 C . mo Reference Am J Obstet Gynecol 97:800 Gynecol Oncol 5:109 Cancer Treat Rep 65:901 Cancer Cheniother Pharmacol 182281) Cancer Treat Rep 702727 Am J Obstet Gynecol 168:805 Cancer Treat Rep 62:833 Cancer Chemother Rep 57:429 Clin Oncol 1:70 Cancer 40:2027 Cancer Chemother Rep 572439 Cancer Chemother Rep 57:497 Cancer Clin Trials 3:345 Proc Am Assoc Cancer Res 4:29 Gynecol ()ncol 5: 109 Cancer Treat Rep 71): 1 239 Cancer Treat Rep 63:2089 Gynecol Oncol 571376 .1 Clin ()ncol 14:792 Cancer Treat Rep 62: 1435 2 CR + 3 PR/27 (19:6-38) 8 CR + 10 PR/40 (~15; 29-62) Epiruhicin Epiruhicin 0c 'JI Acta Oncol 30:325 M Cancer 63: | 279 *CR = complete response: PR 2 partial response. Overall response rates and 95% confidence intervals are in parentheses. ifosfamide and cisplatin in an ongoing randomized GOG trial. Eligible patients are required. among other things. to have ade— quate pulmonary function. This implies an additional element of patient selection. The 24-hour high-dose administration of ifos- famide requires careful attention to such factors as serum a1— bumin and creatinine levels to avoid excessive toxicity. Similarly. MVAC appears to be quite active. with a 73% (19/26) response rate; however. patients must be carefully selected with respect to renal function. and some attention must be paid to their cardiac status. Doxorubicin combinations (Table 5) have given inconsistent results. The high response rate with the addition of vincristine. fluorouracil, and cyclophosphamide (58%) does not seem to have been replicated. Second-line chemotherapy has generally been unrewarding. A review (4) indicated a 27% (8/30) response rate for cisplatin. Table 3. Some investigational drugs used in the treatment ofcervix cancer (minimal or no prior chemotherapy) Agent Response rate* Mitolactol 1 CR + 15 PR/55129; 18-43) Dichloromethotrexate 3 PR/3618; 2-22) Echinomycin l PR/19 (5; 0-26) Menogaril 0/22 (0: 0—15) Vindesine 6 PR/20130: 12-54) Didemnin ()/27(():1)—13) lrinotecan 5 CR + 8 PR/55124: 13:37) Semustine 1 CR +2 PR/6215; 1-14) Median survival. mo Reference 53 JClin Oncol 7:1892 — Cancer Treat Rep 71:1295 — Invest New Drugs 10:25 — Gynecol Oncol 522229 7 Cancer Treat Rep 70: 1455 6.2 Gynecol Oncol 452303 9 Proc ASCO 1 1:224 5.4 Cancer Treat Rep 622833 *CR 2 complete response; PR = partial response. Overall response rates and 95% confidence intervals are in parentheses, 124 Journal of the National Cancer Institute Monographs No. 21. 1996 Table 4. Representative cisplatin cotnhinations used in the treatment ofcervix cancer (no prior chemotherapy) Reginien* Survival. mo Response rate?“ Reference 8 PR/25(32:15-5~1) 5 Am J Clin Onco1112149 Cisplatin + BLEO Cisplatin + BLEO + CY + DOX 15 PR/27 (562 35-75) 7.7 Cisplatin + BLEO + VLB 6 CR + 16 PR/33 (67; 48—82) v Gynecol Oncol 16:275 Cisplatin + BLEO +1FX 10 CR + 24 PR/49 (692 55-82) 10.2 J Natl Cancer Inst 81 2359 Cisplatin + BLEO +1FX 3 PR/20 ( 15‘. 3-38) 9 GynecolOncol-162203 Cisplatin + BLEO + MMC Cisplatin + BLEO + MMC + VCR 3 CR + 4 PR/44(1617»30) r Gynccol Oncol 48:1 1 6CR+14PRI39(51;35-68) * Arch Gynecol 2402247 Cisplatin + BLEO + MMC + VCR 4 CR + X PR/5-1(22; 12-36) 6.9 .1 Clin Oncol 5: | 791 Cisplatin + MMC 2 CR +11 PR/51(25214-40) 7 “ Cisplatin + CY + DOX 5 CR + 1 PR/28 (21: 8—41) 10 Gynecol Oncol 29:32 Cisplatin + CY + DOX 3 PR/30110:2-27) — GynecolOnco126z225 Cisplatin + 5-FU 4 CR + 8 PR/24 (50: 29-71 ) — GynecolOnco129z76 Cisplatin + 5-FU 8 CR +6 PR/52 (27: 1641) 8.6 Gynecol Oncol 372354 Cisplatin + MTX 5 CR + 16 PR/37157; 39-73) 11 Med Pediatr Oncol 14:17 Cisplatin + MTX + 5-FU + BLEO 6 CR + | PR/23 (30'. 13-53) 12 GynecolOnco153216| Cisplatin + VLB + DOX 6 CR + 13 PR/26 (73'. 52—88) 11.5 GynecoIOnco1572235 Cisplatin + Mitolactol 14 CR + 17 PR/140 (21'. 15-29) 7.3 Gynecol Oncol 602120 Cisplatin +1FX 19 CR + 28 PR/141 (322 24-42) 8.3 “ *BLEO : bleomycin: CY = cyclophosphamide: DOX = doxorubicin: VLB = vinhlastine; IFX = ifosfamide; MMC : mitomycin C; VCR : vincristine: 5-FU : fluorouracilz MTX = methotrexate. ‘rCR : complete response: PR 2 partial response. Overall response rates and 95% confidence intervals are in parentheses. Table 5. Doxorubicin combinations used in the treatment ofcervix cancer (without platinum. minimal or no prior drugs) Regimen* DOX DOX + VCR + 5-FU DOX + VCR + 5—FU + CY DOX + Methyl-CCNU DOX + MTX DOX + MTX DOX + MTX + BLEO Response rate'1' 2 CR + 4 PR/38 ( 162 6-31) 1 CR + 2 PR/31 (10; 2-26) 4 CR +14 PR/31 (58: 39—75) 9 CR + 5 PR/3l (45; 276-1) 5 PR/2-1(212 7-42) 4 CR + 10 PR/33 (42; 25-61) 1 CR + 5 PR/25 (24: 9-45) Survival. mo Reference 5.9: Cancer Treat Rep 62: 1-135 7: Cancer Treat Rep 65:901 — Cancer 4922437 8.5§ Am J Obstet Gynecol 1322545 6: Obstet Gynecol 552-188 11$ Clin Oncol 1:70 8: Chemioter 4: 143 *DOX = doxorubicin; VCR = vincristine; 5-FU =tluorouraci12 methyl-CCNU = semustine: MTX 2 methotrexate: BLEO = bleomycin. TCR : complete response; PR 2 partial response. Overall response rates and 95%) confidence intervals are in parentheses. :Median survival. §Mean survival. a 14% (3/22) response rate for teniposide. and an 11% (3/27) response rate for ifosfamide, but less than 10% for mitox- antrone, echinomycin, aminothiadiazole. amsacrine. diaziquone. and maytansine; there were no responses to Yoshi 864. eso— rubicin, vinblastine. PALA (i.e.. N-phosphonoacetyl—L—aspar- tate). N-methylformamide. and spirogermanium in women who had prior chemotherapy. Teniposide should be studied first line. Since there are so few drugs with reproducible activity, the paucity of phase 111 trials in advanced cervix cancer (Table 6) should not be surprising. Unfortunately, none ofthese trials with either doxorubicin combinations or platinum regimens has shown a difference in overall survival. Parenthetically, one trial. with a total of 25 patients. reported a survival advantage for cisplatin plus methotrexate over hydroxyurea (5), but preran- domization was used. performance status was not accounted for. and there may have been other imbalances to account for the differences in outcome. In a large-scale, randomized trial with 438 eligible patients. the GOG recently compared the following three regimens: cisplatin alone (50 mg/m2 every 3 weeks), cisplatin plus ifos— famide (5 g/mz). and cisplatin plus mitolactol (180 mg/m2 daily for 5 days). The response rate was significantly higher with cisplatin plus ifosfamide compared with cisplatin alone (31% versus 18%; 32% versus 18% if eligible but untreated cases are Table 6. Phase 111 trials in advanced cervix cancer Study‘i" DOXi-VCRoriCY 7 W Cisplatin. 50 mg/m‘ vcrsus 100 mg/m‘ versus 20 mg/m‘ X 5 Cisplatin. rapid versus prolonged Carboplatin versus iproplatin Cisplatin i IFX or i mitolactol Survival. mo Reference 5.9, 5.5. 7.3 Cancer Treat Rep 62:14.35 71. 7.0. 6.1 JClin Oncol 321079 6.2. 6.4 Gynecol Oncol 322198 6.2, 5.5 J Clin Oncol 711462 8.0. 8.3, 7.3 Gynecol Oncol 602120 *DOX = doxorubicin: VCR = vincristine: CY = cyclophosphamide: IFX 7- ifosfamide. Journal ofthe National Cancer Institute Monographs No. 21. 1996 excluded). and the progression-free survival improved (4.6 months versus 3.2 months [median]: P : .003). but the overall survival did not (8.3 months versus 8.0 months [median] for the combination or cisplatin alone) and there was significantly more toxicity with cisplatin plus ifosfamide. Mitolactol is an active drug. having shown a 29% response rate as a single agent; how- ever. when combined with cisplatin. there was no improvement in response (21%) or survival (7.3 months lmedianl) compared with cisplatin alone (2). In that study. ifosfamide was given over a 24—hour period. rather than fractionated over a period of several days. Early in the trial. there were episodes of encephalopathy in association with hypoalbuminentia or marginal renal function. so the eligi— bility requirements were tightened to require a serum albumin of at least 3 g/dL and normal renal function. Also. patients with bilateral hydronephrosis were excluded. Even with the modifications just mentioned. there was significantly more central nervous system toxicity associated with the cisplatin— plus—ifosfamide arm than with cisplatin alone (P<.01). There were also more hematologic and renal toxic effects with the ifosfamide combination. The response rate by site of disease (31% for extrapelvic ver— sus 18% for pelvic lesions) in that trial supports the notion most investigators have that extrapelvic sites are more responsive to chemotherapy. Although there was a significant advantage for cisplatin plus ifosfamide in progression-free survival. the incre— ment in median progression—free survival was less than 11/: months (4.6 months versus 3.2 months). This may be the first evidence of an improvement in progression—free survival with a drug combination in cervix cancer; however. unfortunately. it did not translate into a survival benefit. Prognostic factors for response to chemotherapy of squamous cell cervix cancer have not been studied extensively. 1n the study just referred to. increasing age (P = .025) and better per- formance status (P<.001) were significantly favorable for surv vival in a multivariate analysis (2). 1f tnore attention were paid to prognostic factors. then the results might be tnore consistent from one trial to another. Adenocarcinoma is said to make up 1()‘/r—18% of invasive cervix cancer. Only a handful of studies have reported specifi- cally on chemotherapy of non-squamous cases. There is no good evidence that these non—squamous cases respond to drugs dif— ferently than the squamous cases. Parenthetically. small-cell. undifferentiated carcinoma arises in the cervix in rare instances and has been treated in the same way as small—cell lung cancer. In the absence of data to the contrary. it would seem reasonable to conduct chemotherapy trials in “non-small-cell cervix car— cinoma" rather than to continue the current practice of evaluat— ing squamous cases separately. Despite great general interest in agents such as interleukins. interferons. monoclonal antibodies. and tumor necrosis factor in the treatment of cancer. little has been reported so far with 126 respect to cervix cancer. A combination of interferon alfa and 13-ci.s'-retinoic acid yielded 13 responses among 26 cases with no prior treatment of any sort (6). A confirmation is needed. Thus. several cytotoxic drugs have antitutnor activity in cer— vix cancer. but as yet none have had a reproducible major im— pact. Cisplatin combinations have not shown a consistent advantage over single—agent therapy: a recent trial (2) showed significant improvements in response rate and progression-free survival with the use of cisplatin plus ifosfamide. but there was significantly tnore toxicity and there was no survival benefit. A subset of patients probably exists who can benefit frotn currently available drug regimens: this suggestion is based on the impressive results obtained in some phase 11 trials and the occasional spectacular response in individual cases. Part of the challenge for the next few years will be to determine if such a subset can be reliably identified. Eligible. consenting patients should be enrolled in clinical tri- als employing new drugs and/or new strategies. Since there is as yet no evidence of curative potential for chemotherapy in cervix cancer and no established survival benefit. and uncertainty ex— ists as to how often response translates into symptom relief (“palliation"). nonprotocol therapy should not be encouraged. Nevertheless. for a patient who is ineligible or unwilling to par- ticipate in a study but who wants treatment. there may still be an indication for chemotherapy—what might be called "psycho— logical support" or hope (7). When such a patient insists on treatment and will seek quack remedies rather than hospice if orthodox chemotherapy is not offered. single-agent cisplatin or carboplatin may be justified. with due attention being paid to contraindications and toxic side effects. An interval response as- sessment and a finite period of treatment are indicated. Objec— tive benefit is possible. but not likely. References (l) Omura GA. Chemotherapy for cervix cancer. Semin ()ncol 1904;21:54—62. (2) Omura GA. Blessing J. Vaccarcllo 1.. Berman M. Mulch I). Clarke—Pearson D. et al. A randomized trial of cisplatin versus cisplatin + mitolactol \ersus cisplatin + ifosfamide in advanced squamous carcinoma of the cervix by tlte GynecologicOncology Group. Gynecol Oncol 1996;60: 120. (3) Bonomi P. Blessing JA. Steltman HF. DiSaia Pl. Walton L. Major Fl. Ratt- domi/ed trial of three cisplatin dose schedules in squamous-cell carcinoma of the cervix: a Gynecologic Oncology Group study. .1 Clin ()neol 1985;31 107985. (4) Park RC. Thigpen JT. Chemotherapy in advanced and recurrent cervical cancer. A review Cancer l‘)93;7l(~l Suppl): 14-16-50. (5) Be/woda WR. Nissenbaum M. Derman Dl’. Treatment of metastatic and recurrent cervix cancer with chemotherapy: a randomised trial comparing hydroxyurea with cisdiamminedichloro—platinum plus methotrcxate. Med PediatrOncol 1986:14: 17-1). (6) Lippman SM. Kavanagh JJ. Paredes-lispino/a M. Delgadillo-Madrueno F. Paredes-Casillas P. Hong WK. et al. l3»t~i,v-Retinoic acid plus interferon (1- 2a: highly active systemic therapy for squamous cell carcinoma of the cer- vix Ivor comment citation in Medlincl. J Natl Cancer Inst 1992;84:241—5. Omura GA. Indications for cancer chemotherapy Heller]. N l€ng1 J Med 1982;307:826. 17 Journal ofthe National Cancer Institute Monographs No. 21. 1996 Radiation Palliation of Cervical Cancer William J. Spanos, Jr.. Thomas J. Pajak, Bahman Emami, Philip Rubin, Jay S. Cooper, Anthony H. Russell, James D. Cori< Radiation is a useful modality for palliation of local-regional disease in patients with cervical cancer who require pallia- tion because of distant metastases, extensive local—regional disease, medical consideration, or patient concerns. Two radiation schedules have been reported on for the treatment of advanced pelvic disease including cervical cancer. The large single-dose schedule consisted of 10-Gy fractions repeated at monthly intervals to a maximum of 30 Gy. This schedule has produced good palliative results with symp- tomatic improvement in approximately 50% of patients and objective response in 35%-80%. However, severe late toxicity was shown to be as high as 42% (actuarial). The second schedule tested by the Radiation Therapy Oncology Group consisted of 3.7-Gy fractions given twice a day for 2 days (14.8 Gy) repeated after 2-4 weeks for a maximum of 44.4 Gy. There were 284 patients accrued, and the subgroup of 61 cervical cancer patients is analyzed in this article. The subjective response (50%-100% complete response) and ob- jective response (53%) were similar to those observed with the large single-fraction schedule. The late toxicity was sig- nificantly lower (7%—actuarial). For patients who may sur- vive 6 months or longer, this second schedule is preferable. [Monogr Natl Cancer Inst 1996;21:127-30] Cervical cancer presents most frequently in early stages when palliation is not a consideration. However, approximately 5%— IS‘Vr of patients present with advanced stage disease (Interna— tional Federation of Gynecology and Obstetrics stage [VA or lVB). depending on the region of the United States (I). In addi— tion. there are patients with earlier stages of disease who are not candidates for curative approaches because of underlying medi— cal conditions. poor general condition. logistics. or patient con— cerns. The incidence of these patients is difficult to determine but probably adds approximately 5%-10% to the number of patients presenting for consideration of palliation. Palliation as a concept is poorly defined. In broad terms. it usually refers to alleviation of adverse symptoms in a patient for whom life expectancy is limited or in whom systemic progres— sion of disease is a dominant process. However. the concept of palliation may include pre-emptive treatment for the purpose of maintaining an existing quality of life in the face of a disease site that has a high likelihood of causing a symptomatic progres- sion. Palliation can be local—regional or systemic in nature and includes surgery, radiation therapy. or chemotherapy. For pallia- tion to be an acceptable consideration. the palliative modality must have a reasonable chance of success. The acceptable pal- liative success rate will vary among investigators, but it should not fall below 50% unless side effects are negligible or there is a Journal of the National Cancer Institute Monographs No. 2 1. I996 clear understanding of the goals between patient and physician. In some circumstances. no treatment other than supportive care is the best palliative measure. Palliation decisions in cervical cancer must take into account the likely pattern of spread. Cervical cancer has a propensity to progress locally and regionally and to produce local—regional symptoms before distant metastases produce significant symp— toms. lt is in this context that pelvic radiation therapy is a con- sideration for palliation. Radiation therapy palliation schemes are quite varied. As a general consideration. higher doses per fraction place the patient at higher risk of developing late complications in sensitive. nor- mal tissue such as the bowel or the urinary bladder. Smaller doses per fraction with prolonged treatment schedule tend to produce more prolonged acute symptoms and obviously lengthen the time that the patient is under treatment. The goal for radiation therapy palliation in the pelvis is to provide good quality symptomatic relief in as compact a schedule as is pos- sible without producing significant acute symptoms and without placing patients at significant risk of developing severe com- plications over their anticipated life span. Radiation Therapy Palliation Schedules Daily, High Dose per Fraction The most common general palliative schedule uses 2.5-3.0 Gy per fraction daily to a total dose of 30—325 Gy over a 2- to 3-week treatment course. This schedule is used most frequently for bone palliation. Although this schedule or a variation of it is undoubtedly used in palliation of advanced cervical cancer. neither the effectiveness nor the side effects have been reported in the literature. Experience at the University of Louisville with this schedule in areas that include significant bowel volume (such as the pel- vis) has been associated with significant acute bowel symp- toms. i‘Afliliulionx (giant/mm: W. J. Spanos. Jr.. University of Louisville School of Medicine. KY: T. J. Pajak. Radiation Therapy Oncology Group (RTOG) Head— quarters. Philadelphia. PA; B. Emami. Washington University School of Medicine. St. Louis. MO; P. Rubin. University of Rochester Medical Center. NY: J. S. Cooper. New York University. NY: A. H. Russell. Radiation Oncol- ogy Center. Sacramento. CA: J. D. Cox. The University of Texas M, D. Ander« son Cancer Center. Houston. (‘orrex/mnzlcm'e Io: William J. Spanos. Jr.. M.D.. Department of Radiation Oncology. University of Louisville School of Medicine. 529 South Jackson St.. Louisville. KY 40202, 127 Large Single Dose—Monthly Interval This schedule consists of 10 Gy per fraction delivered at monthly intervals to a maximum dose of 30 Gy. It has been spe- cifically targeted at the palliation of advanced pelvic cancers. The effectiveness and toxicity of this schedule have been re- ported by several institutions and in a prospective trial conducted by the Radiation Therapy Oncology Group (RTOG) (2-5). This schedule has the advantage of minimal time investment by the patient and low acute toxicity. It also provides flexibility to allow the physician to modify the total radiation dose if the patient‘s condition deteriorates or progression of systemic dis— ease decreases the value of pelvic palliation. Symptomatic palliation. The symptomatic palliation for this schedule is shown in Table l. The rate of symptomatic relief for the different institutions was approximately 50%. except for the experience at Manitoba (4). where there was a 100% relief of symptoms. Objective palliation. The objective response to this schedule is shown in Table 2. Subjective and objective palliation with this schedule was good. There was fair consistency in subjective response among the various studies. There was less consistency in objective response. probably as a result of a differing mix of tumor bulk and the difficulty in assessing tumor response in the pelvis. The RTOG data represent a prospective trial that included the putative radiosensitizer misonidazole. The patient distribution consisted of a lower percentage of patients with cervical cancer than the single-institution reports and a higher percentage of patients with colorectal tumors. Because an objective response in nongynecologic pelvic tumors is more difficult to quantify. this may partially account for the lower response rates for the RTOG study. The remaining studies demonstrated a 33%—5()% complete response rate for patients completing treatment. Toxicity. The acute and late toxic effects of the reporting in— stitutions are summarized in Table 3. As expected with a large dose per fraction. acute toxicity was small. However. late corn- plications were relatively frequent except for The University of Texas M. D. Anderson Cancer Center experience. It is often difv ficult to determine cause and effect for bowel or urinary bladder damage when viable tumor is present in the affected area. The University of Texas M. D. Anderson Cancer Center experience concluded that the damaged areas were a result of persistent cancer. The RTOG assigned all damage to the radiation unless a clear tumor cause—effect relationship could be demonstrated. Either of these approaches may underestimate or overestimate true incidence and conclusions. All of the late complication rates listed in Table 3 were crude percentages. Since this patient population had a short average Table l. Symptomatic palliation with large single-doscimonthly interval schedule No. of patients Bleeding Pain Reference (No. with cervix) (’/¢ response”‘) (‘/( response?) The University of Texas M. D. 86 (77) 45 ( |()()) 45(100) Anderson Cancer Center (2) University of North Carolina (3) 42 (32) 44 5() Manitoba (4) 27 ( 14) l()() l()() Radiation Therapy Oncology 46 (5) 58') Group (5) *Values in parentheses : ‘7} response for patients completing all three courses. 'l‘Overall symptomatic 9; response—not subdivided by W response. bleeding or pain. life span, the true complication rate was underestimated by crude percentages. The RTOG used an actuarial life-table method to look at the late complication incidence and demon- strated a projected 42% risk of late complications by 12 months. This information led to the premature closure of the protocol. Similarly. the investigators at the University of North Carolina recognized a higher than expected complication rate in patients surviving more than 12 months and recommended use of this schedule only in patients who had an expected survival of less than 12 months. Accelerated Split Course This schedule consists of 3.7 Gy per fraction given twice a day to a total dose of 44.4 Gy with a 2— to 4—week rest after 14.8 Gy and 29.6 Gy. The schedule was developed in response to the disappointingly high late complication rates seen following the large single-fraction schedule tested by the RTOG. The aims of the revised schedule were to retain the compact schedule (each segment of the new schedule occupies 2 days instead of l). to maintain response rate. and to significantly decrease late corn- plications. RTOG tested this schedule extensively as a replace— ment palliative schedule for advanced pelvic cancers (6—8). A total of 284 patients were entered into the protocol. The RTOG study achieved the goals presented above. The study group had a mixture of tumor histologies and primary sites. There were 6] patients with cervical cancer. For the pur— pose of analysis of palliation of cervical cancer. the subgroup of cervical cancer patients is analyzed separately below. For the purpose of analysis of toxicity. the entire patient population is analyzed. since toxicity should not be dependent on primary site or histology. The characteristics of the group are given in Table 4. Table 2. Objective response vi ith large single-dose—monthly interval schedule No. of patients Response* Reference (No. with cervix) Complete, ‘71 Partial. ‘/r Overall. 0/r The University of Texas M. D. Anderson Cancer Center (2) 86 (77) lo (50) (14 (50) 80 ( IOO) UniversityofNorthCarolinaU) 42(32) I788) l9(27) 36(65) Manitoba (4) 27(14) 33 26 59 Radiation Therapy Oncology Group (5) 4b (5) I3 ( lo) 24 (27) 37 (42) ‘E‘Values in parentheses : 5? response for patients completing 30 Gy. 128 Journal ofthe National Cancer Institute Monographs No. 2 1. I996 Table 3. Acute toxicity and late complications with large single-dosc—monthly interval schedule No. of Acute Late Reference patients toxicity complications, % :1: The University ofTexas M. D. 86 Minimal None‘?‘ Anderson Cancer Center (2) University of North Carolina (3) 42 Minimal [2 Manitoba (4) 27 Not stated I 1 Radiation Therapy Oncology Group (5) 46 4% 2t) ”‘Grade 3 or 4 using Radiation Therapy Oncology Group complications scor~ ing (5), ‘i'There were 15 t 17% ) grade 3 or 4 bowel injury attributed to tumor. Table 4. Characteristics of 61 patients with cervical cancer. Radiation Therapy Oncology Group study No. ofcase No. of attaly/able patients receiving Characteristic case patients (% ) three courses (% ) Total 61 34 Age. y S50 9(15) 6(18) 51-60 IltIX) 5(l5) 61-70 18(30) I()(2‘)) 271 23(38) 13(38) Karnol‘sky’s performance status 50-70 25 (41) 15 (44) 80— 100 36 (5‘)) l‘) (56) Disease extent Pelvis only 37(61) 25 (74) Outside pelvis 24 (3‘)) ‘) (26) Symptomatic palliation. For the 61 patients with cervical cancer. the rate of complete clearance of symptoms is sum- marized in Table 5. The symptomatic response rate was good and compared favorably with the results obtained with the large single-fraction regimen. Objective palliation. Objective palliation for the 61 patients with cervical cancer is summarized in Table 6. The overall and complete response rates were slightly lower than those reported with the large single—fraction regimen. but higher than those reported by RTOG (5). University of Louisville. There were 18 patients with cervi— cal cancer treated for palliation with the accelerated split-course Table 5. Rate of complete clearance of symptoms for the 61 patients with cervical cancer No. of patients presenting ‘rl totally ‘rl totally Symptom with symptom cleared cleared. lull dose Dysuria ‘9 67 50 Pain 36 3 l 37 Ulceration 14 50 86 Bleeding 4‘) 76 100 ()bstipation 7 43 100 'l‘enesmUs 4 100 1()() Journal ofthe National Cancer institute Monographs No. 21. 1996 Table 6. Objective palliation for the 61 patients with cervical cancer Response No. of patients % response % response. full course Complete ‘) 15 24 Partial | 4 23 2‘) Minor ‘) I5 21 ()\ erall‘i‘ 23 38 53 ”‘Complete response + partial response, technique. The subjective response rate was 83%. and the objec- tive response rate was 88% (27% complete response). Toxicity. Acute toxicity consisted of diarrhea or cystitis. The acute toxicity was 3% overall. it was not noted for patients with longer (4—week) rest intervals but was 6%) for patients with short (2-week) rest intervals. Late complications (grade 3 or 4) oc— curred in 7% (actuarial). Varying the rest interval had no effect on the incidence of late complications. Survival. The survival of the cervical cancer patients treated for palliation in the RTOG protocol testing the accelerated split- course schedule is shown in Figs. 1 and 2. The median survival was 5.5 months. For the subgroup of patients receiving the full dose. the median survival was 8.5 months. There was a tail on the survival curve of approximately 5% at 2 years. This tail rep— resents a few patients whose disease was limited to the pelvis and who had a complete response. The number of patients with cervical cancer was too small to characterize the long-term sur— vivors. 1n the overall population of the study group. 33% of complete responders survived more than 24 months. Discussion Both the large—fraction radiation schedule and the accelerated split-course radiation schedule offer substantial objective pallia— tion and symptomatic palliation. Both schedules are flexible and provide for patient and/or physician modification of the total dose. should the patient’s condition deteriorate or systemic dis— ease progress such that pelvic palliation is no longer a quality— of-life factor. in addition. both are compact and require little 100 75 p E R’ C 50 E N T 25 O O 6 12 18 24 MONTHS FROM DATF 0F RANDOMlZATlON Fig. l. Surviuil for all cet'\ic;1| cancer patients: Radiation Therapy Oncology (lrotlp 8502. 12‘) 100‘ 75 P E R C 50 E N T 25 O O 6 12 18 24 MONTHS FROM DATE OF RANDOMIZATlON Fig. 2. Survival for cervical cancer patients completing three courses: Radiation Therapy Oncology Group 8502. commitment of patient time. Moreover, both are associated with minimal acute toxicity. There is a concern over the potential late complications from the large single-fraction schedule noted by RTOG. the Univer- sity of North Carolina. and Manitoba reports. The investigators from The University of Texas M. D. Anderson Cancer Center did not observe the high incidence of late complications from radiation therapy. but they did see a high incidence of tumor-re- lated complications (17%). The accelerated split-course sched— ule clearly demonstrated a lower late complication rate with the use of RTOG observation criteria. 130 From the above considerations. radiation palliation of cervical cancer is effective and can be accomplished with little expendi- ture of time or toxicity. If the patient‘s expected life span is less than 6 months. either schedule may be considered. If the patient‘s expected life span is 6 months or more. the accelerated split-course schedule lowers her risk of developing late complications. References (I) Pellersson F. Annual report on the results of treatment in gynecological can» cer. Vol 22. Stockholm. Sweden: International Federation of Gynecology and Obstetrics. 1995. Boulware RJ. (‘aderao JB. Delclos L. Wharton JT. Peters LJ. Whole pelvis megavollage irradiation with single doses of 1000 rad to palliate advanced gynecologic cancers. Int J Radiat ()ncol Biol Phys 1979:52333 —>+ In hypoxlc cells and Intrinsic radioresistance l of oxlc cells I Immunosuppresslon Normal Tissue Sensitivity —> tumor Fig. 1. Current model for the progression of cervical cancer that synthesizes the recent studies on the viral. immunologic. genetic. and microem‘ironmental factors that are highly associated with the disease and its control. HPV = human papillomavirus‘. ampl. : amplification; metaloprot. : metalloproteinase; cone, : concentra— tion; XRT = radiation therapy. has recently been demonstrated with the use of embryonic fibroblasts derived from wild—type and p53—deficient transgenic mice (9). Hypoxia induced apoptosis in oncogenically trans— formed cells and selected for cells with defects in apoptosis. such as those with p53 mutations (9). Such apoptotically resis- tant hypoxic cells may possess decreased sensitivity to antican- cer therapy and may possess an increased ability to metastasize. The metastatic potential of cervical cancer cells could be further enhanced by other extrinsic factors. such as fever and im— munosuppression (as will be discussed later in this article). lead— ing to distant failure. In addition to its potential role as a selective agent. hypoxia will also decrease the effectiveness of radiotherapy and some chemotherapeutic agents by acting as a dose-modifying factor. However. one of the most important questions remaining is whether HPV or cervical tumorigenesis alters the intrinsic capacity of cervical cells to repair radiation- or chemotherapy— induced damage. Are those cells that have diminished apoptotic programs more intrinsically resistant to DNA-damaging agents? Obviously. the radiotherapist must always contend with the balance between increasing overall treatment dose and normal tissue complications. but only knowledge of the intrinsic cellular radioresistance of both tumor cells and normal cells will allow us to address whether it will be possible to tailor the treatment for each individual. The more we understand the biology of cervical neoplasia. the greater will be our ability to choose the appropriate course of treatment. This model will be employed to identify potential avenues for future research in four areas. I32 Molecular Genetics Considerable advances in our understanding of the patho- genesis of cervical cancer have been made. and the role of HPVs in the development of cervical cancer has been elucidated in some detail (3,4). HighArisk viral HPV DNA (types l6. l8. 3]. 32. 33. 35. and 45) is found in more than 90% of cervical cancers (3.10). The viral DNA usually is integrated into the cel- lular DNA. and the cells express viral E6 and E7 gene products (II). The continued expression of E6 and E7 is necessary to im- mortalize human foreskin or cervical keratinocytes. but it is not sufficient to produce malignant clones (12.13). High-risk E6 protein binds to the p53 protein and promotes its degradation (14). High—risk E7 protein binds to the protein of the retinoblas- toma susceptibility gene (15—I7). Most cervical cancers with HPV DNA have a normal p53 gene. but a small percentage of the HPV—negative cervical cancers have mutant p53 genes (18- 20). Therefore. by alteration of the retinoblastoma susceptibility gene or p53 gene function (in high—risk HPV—positive cells) or by mutation in the p53 gene (in HPV-positive or -negative cells). regulations of both DNA replication and cell cycle are al- tered. The prolongation and further modification of these altera— tions by as yet to be identified mutagens (e.g.. products of cigarette smoke) can result in the development of cancer. Our understanding of the molecular genetics of the develop- ment of cervical cancer points to many potential opportunities for therapeutic intervention. including the following: blocking the transcription of DNA for E6 and E7. blocking the translation of messenger RNA (mRNA) for E6 and E7. inhibiting the bind— Journal ofthe National Cancer Institute Monographs No. 21. I996 ing of the E6 protein to p53 protein. restoring p53 gene function by gene therapy with the wild—type p53 gene, restoring the retinoblastoma susceptibility gene function by gene therapy with the wild—type retinoblastoma gene. developing a vaccine to prevent high-risk HPV infection. and developing a vaccine to treat recurrent or persistent cancers. Considerable progress has been made in several of these areas. Plasmids with antisense RNA to HPV18 E6/E7 sequences have been shown to inhibit the expression of E6 and E7 and to decrease cell growth in HPV-positive cell lines in vitro (2l—23). Nuclease—resistant phosphorthioate oligodeoxynucleotides com- plementary to mRNA of the HPV 16 E6- and E7-containing cer- vical cancer cell line (CaSki cells) have been shown to exhibit potential antiproliferative activity (24,25). Cationic liposome— mediated delivery of antisense oligonucleotides targeted to HPV16 E7 mRNA inhibits proliferation of CaSki cells (26). More recent research (27) has shown that antisense HPV18 E6 and E7 sequences as well as sense complementary DNA (cDNA) for the human wild-type retinoblastoma susceptibility gene decreased in vitro and in vivo growth of transfected HPV18 cells. Growth suppression of cell lines expressing HPV 1 8 was also found by overexpression of a plasmid—contain— ing cDNA for human wild—type p53 gene or by recombinant adenovirus containing wild-type p53 cDNA (28.29). Another approach (30) included the use of ribozyme-mediated in vitro cleavage of transcripts arising from the major transforming genes of human HPV16. Use of vaccines to prevent HPV16 in- fection or to destroy established lesions has been reviewed (3/). There is an obvious need to study the effects of HPV infec- tion and p53 gene mutations on intrinsic tumor cell sensitivity to radiation therapy (and chemotherapy). In cell lines from head and neck cancers. p53 mutations “predispose to increase radioresistance" (surviving fractions at 2 Gy for the group with mutations versus tumors without p53 mutations were 0.568 and 0.507, respectively; P = .28) when compared with wild—type p53 cells (32). Increased resistance to radiation in bone marrow and spleen cells in transgenic mice with p53 mutations has been reported (33). and p53 mutations in lymphoma cells tended to make mutant cells more radioresistant than wild—type cells (34). Conflicting results with regard to chemotherapy sensitivity have been noted in other tumor systems with p53 mutations. Lowe et al. (35,36) reported that p53-deficient tumors had increased resistance to y radiation. fluorouracil, etoposide. and doxorubi— cin compared with tumors expressing the p53 gene. However. an enhanced sensitivity to a wide range of chemotherapeutic agents has also been noted in other p53-deficient cells (37.38). The few clinical studies that have tested for an association be— tween p53 mutations and prognosis in cancer of the cervix have shown variable results. Crook and Vousden (39) demonstrated an association between p53 mutation and metastatic lesions: Tsuda and Hirohashi (40) reported an association between p53 mutation and undifferentiated cervical cancer or clinical stage IV disease. and Tsuda et al. (41) noted a poorer prognosis in patients with adenocarcinoma of the cervix and p53 mutations. However. Oka et a1. (42) reported no significant correlation with outcome of mutant p53 in patients with stage III cervical cancer treated with radiation. Similarly, the presence of p53 mutations was not found to influence prognosis in patients with cervical cancers of stages [B to IIB treated with surgery (43). However. a significant negative influence on outcome has been reported for other solid tumors. including breast cancer (44-46). non-small— eell lung cancer (47). prostate cancer (48). gastric cancer (49). colorectal cancer (50), pancreatic cancer (5]). and astrocytomas (52) associated with p53 mutations. Preliminary evidence sug— gests that the site of mutations within the p53 gene may corre- late with other prognostic factors as well as with response to systemic therapy or radiation therapy (46). Possible correlations with the detection of various types of HPV DNA present in the tumor and treatment outcome in cervi— cal cancer are summarized in Table 1. The discrepancies be- tween the results of the various studies may be due to the relatively small numbers of patients in each trial. differing selec— tion criteria for patient inclusion, differing staging techniques. differing sensitivities of the techniques used to detect the HPV DNA (polymerase chain reaction versus Southern blot hy— bridization). differing types of HPV tested. and differing treat- ments (radical surgery versus radiation therapy versus combination therapy). Further studies are warranted to clarify the prognostic implications of the influence of HPV infection (and type) and/or p53 mutations in patients with carcinoma of the cervix. Hypoxia A review of 16 clinical studies in the literature (1969-1989) showed an adverse effect of anemia on outcome for patients with carcinoma of the cervix treated with radiation (59). Most of the studies showed an impairment of local control and survival in patients with moderate anemia (hemoglobin level, 10-12 g/l()() mL) (59). A clinical trial reported by Bush (60) showed that patients who were transfused to maintain hemoglobin of 12.5 g/IOO mL or had initial hemoglobin of 12.5 g/100 mL or Table I. Prognostic implications of human papillomavirus (HPV) in carcinoma of the cervix* Author, y (reference No.) % HPV positive King et al.. 1989153) 45 RS. XRT. or both Walker et al.. 1989 (54) 64 NA Riou et al. 1990 (55) 8-1 R5 i XRT or XRT alone Higgins et al.. 1991 (56) 80 NA Rose et al.. 1991 (57) 69 RS i XRT Burnett et al.. 1992 (58) 100 RS Treatment Prognostic significance None For stage IB: HPV18. worse outcome HPV negative: higher relapse. DM HPV negative: high fatality None Infection with multiple HPV types or HPV18 increased recurrence *RS : radical surgery. XRT : radiation therapy. NA = not available. and DM = distant metastases. Journal of the National Cancer Institute Monographs No. 21. 1996 133 more had improved control versus patients with a hemoglobin level during treatment of between l0 and 12.5 g/l()0 mL. The adverse effect of anemia on local control following radiation therapy in patients with cervical cancer has not been directly re— lated to tumor hypoxia. However. in agreement with fundamen— tal radiobiologic principles, a strong correlation between tumor hypoxia and outcome has been demonstrated more recently by direct tumor oxygen measurements (6]). The role of hypoxia and its modification for direct radiosensitization in tumors have been reviewed (62) and will not be discussed further here. Table 2 summarizes the potential therapeutic interventions that have been proposed to improve the treatment outcome of potentially hypoxic tumors. An alternative therapeutic approach exploits the presence of hypoxia by treatment with specific hypoxic cell cytotoxins. From a theoretical point of view. it may be better to kill hypoxic cells in tumors rather than to radiosensitize Or oxygenate them (63). The sensitized cells still need to be killed. whereas a direct hypoxic cell toxin will eliminate these cells. Such a premise as- sumes. for maximum benefit. that tumors “rehypoxiate” (i.e.. become hypoxic again following each radiation fraction) during fractionated treatments so that repeated administration of a hypoxic cell toxin will have the maximum gain. Laboratory tri- als have been reported on one such hypoxic cell toxin. tirapaz- amine. This drug is activated by cellular reduction to an oxidizing radical. whereas in the presence of oxygen it is “back- oxidized" to the parent drug (64). Laboratory studies have clear— ly demonstrated the selective killing of hypoxic cells with this drug. The hypoxic cytotoxicity ratio (the concentration of drug in air divided by the concentration of drug in hypoxia to produce the same level of killing) for this drug has been shown to be be— tween 50 and 200 for various human tumor cell lines (64). When tirapazamine is used in combination with radiation (which selectively kills aerobic cells). one can obtain a relative- ly constant level of killing over a broad range of oxygen con- centrations (lO’z-IO} pM 03). In combination with fractionated radiation in mouse tumors. a greater than additive response was seen with tirapazamine without an enhancement of early skin reactions or late reactions (64). Dose escalation trials in humans with weekly and up to three- times—a-week doses of tirapazamine in combination with daily fractionated radiation therapy have been completed at Stanford University and the Joint Center for Radiation Therapy at Har— vard University. A Radiation Therapy Oncology Group (RTOG) phase II trial with radiation therapy in patients with glioblas— toma multiforme has already completed accrual. and a phase II trial with radiation therapy in patients with head and neck can- cer is ongoing at Stanford University. In addition. there is a clinical trial utilizing this drug with radiation in the treatment of brain metastases from melanoma (von Roemling R: personal communication. 1996). A phase I trial of the oral preparation of tirapazamine has been initiated at Stanford University. The side effects of tirapazamine to date have been minimal and include muscle cramping and diarrhea. No serious late effects have thus far been reported. Successful completion of these phase I trials. coupled with the direct demonstration of hypoxic tumor cells in patients with cervical cancer, suggests that phase II trials of tirapazamine 134 Table 2. Tumor hypoxia: potential therapeutic interventions ' Increase oxygenation of tumor ’l‘ransfusions; erythropoietin Ilyperbaric oxygen Perfluorocarbons Nicotinamide and carbogen Stop smoking llyperthermia ('1) ° llypoxic cell sensitizers ' Use of hypoxic-specific cytotoxins Bioreductivc activation SR 4233 (tirapazamine) should be seriously considered in the treatment of hypoxic tumors of the cervix. Noninvasive (e.g.. magnetic resonance spectroscopy) or minimally invasive (e.g.. nitroimidazole-bind— ing assays of biopsy specimens) methods to determine which tumors are “hypoxic“ should be verified (65). Monitoring tumors during treatment to determine the effect of radiation therapy on the hypoxic cell fraction should also be considered. In addition. the relationship between tumor hypoxia. p53 gene alterations. and apoptosis and the patient‘s prognosis should be determined (9.66). It has been proposed that hypoxic stress can result in apoptosis in wild—type p53—containing tumor cell lines. Mutant p53 cell lines do not undergo apoptosis in response to hypoxia. and persistent hypoxic stress can result in the selection of a p53 (—/—) genotype (66). If hypoxia selects cells that are p53 (—/—). then one would expect that in hypoxic tumors p53 (—/—) tumor cells would predominate. These tumors would be ex- pected to have an increased level of apoptosis (as a result of hypoxia-induced killing of p53 wild-type cells) and a possible increased radioresistance (assuming that the selected p53 mutant cells are of increased radioresistance). The results of a study by Levine et a1. (67) are consistent with this concept, demonstrating poorer survival for patients whose tumors had higher than average levels of apoptosis. Intrinsic Sensitivity Using a wide variety of assay systems. investigators (68—70) have demonstrated a wide range of intrinsic radiosensitivities between tumors of varying clinical responsiveness and within tumors of the same histopathologic type and origin. Similarly. there exists a wide range of intrinsic radiosensitivities for cells from normal tissues (e.g.. fibroblasts, lymphocytes). suggesting a broad range of genetic susceptibilities to radiation therapy and chemotherapy (7/). In particular, differences exist in the radiosensitivity of patients who are homozygotes for ataxia— telangiectasia compared to heterozygotes for ataxia—telangiec- tasia and possibly for a subset of patients with autoimmune dis- eases. A pretreatment knowledge of intrinsic tumor cell radiosensitivity would permit selection of radioresistant tumors for alternative treatments. including surgery. adjuvant chemo- therapy. higher dose radiation therapy. or radiosensitizers. The very radiosensitive tumors could be selected for less aggressive radiation therapy to minimize complications. Similarly, if the normal tissue sensitivity were known prior to institution of radiation therapy. one could exclude the patients with very Journal of the National Cancer Institute Monographs No. 21. 1996 radiosensitive nortnal tissues from standard radiotherapy regi- mens and select patients with very radioresistant normal tissues for more aggressive radiation therapy treatment (if they had tumors of advanced stage) (69). Pioneering work using the Courtenay—Mills clonogenic assay performed by West (69) has shown a correlation between the in— trinsic radiosensitivity in cervical cancer and outcome. In 88 patients with carcinoma of the cervix treated with radiation therapy in vitro. tumor cell survival following a dose of 2 Gy (SFz) was significantly predictive of outcome. Patients with radiosensitive tumors (SF3<0.40) had significantly higher local control (P = .001) and survival (P = .002) than patients with more radioresistant tumors (SF3>0.40). In addition. tumors taken from patients with recurrent disease had higher SF: values (0.54 i 0.17, mean i standard deviation) than tumors taken from patients who were alive and well (0.38 i 0.16; P<.01 ) (72). Studies on lymphocytes from 98 patients obtained prior to treatment with radiation for cervical cancer were also assessed. The intrinsic radiosensitivity of the lymphocytes was sig- nificantly correlated with the development of complications. The lymphocytes in 13 patients who developed either acute or late reactions (minimum follow—up of 26 months) were. on average. significantly more radiosensitive than the lymphocytes in 82 individuals who showed no adverse reactions to treatment (69). In addition. a study of 135 patients with stage 111 car— cinoma of the cervix showed a highly significant correlation be- tween 5—year survival rate and gastrointestinal complications. Patients with severe diarrhea or proctitis had a 5-year survival rate of 66%; those with either diarrhea or proctitis had a 5—year survival rate of 41%; those with neither complication had a 5— year survival rate of 17% [reported in (68)]. This finding sug- gests that intrinsic cellular radiosensitivity may be. in part. genetically determined in both malignant and normal tissues ofa given patient. It is now well established that there are inherited genetic sus- ceptibilities to radiation damage (7]). Patients who are homo- zygotes for ataxia-telangiectasia have a high risk of developing lymphomas during childhood and fatal complications following standard—dose radiation therapy. Fibroblasts from patients who are homozygotes for ataxia-telangiectasia have been shown to be 2.5—3.0 times more sensitive to radiation than normal con- trols. Ataxia-telangiectasia heterozygotes may constitute ap- proximately 1% of the general population and have a twofold to fourfold increased risk ()f developing cancer (73). Approximate- ly 9%—18% of all breast cancer patients are ataxia-telangiectasia heterozygotes (73). Fibroblasts from these heterozygotes are 1.2-fold to 1.3-fold more radiation sensitive than normal (7]). and this observation may account for increased reactions to radiation in these patients. Microchip—based sequencing technol- ogy now exists to allow rapid resequeneing of genes (74). and such techniques can be employed to detect ataxia-telangiectasia heterozygotes prior to treatment decisions. Some patients with autoimmune diseases (e.g.. dermatomy— ositis, scleroderma) may have an associated risk of developing cancer. Severe complications following radiation therapy have been reported in these patients (75,76). For example. of 158 consecutive patients treated with radiation therapy for cancer of the cervix at the Cancer Control Agency of British Columbia. Journal ofthe National Cancer Institute Monographs No. 21. 1996 the only treatment—related death in a 6—year review of their ex— perience was in a patient with systemic lupus erythematosis. This patient died of pelvic necrosis with extensive fibrosis and fistulization (77). Not all studies. however. have confirmed an increased risk of acute or late reactions in patients with autoim— mune diseases (78), and a way to better identify which patients may be at increased risk would be of significant value. It has been noted that between 39% and 55% of patients with autoim- mune disease have autoantibodies against the Ku protein (Chu G: personal communication. 1995). Genes encoding the Ku protein and its catalytic subunit DNA-PK have been shown to be required for the repair of DNA double-strand breaks induced by ionizing radiation (79,80), Therefore. patients with autoim— mune disease may have mutations in Ku. altered Ku function. and enhanced radiation sensitivity. Patients can be screened for Ku antibody levels and x—ray sensitivity, Peripheral blood cells of patients with autoimmune disease have previously been shown to have intrinsic increased radiosensitivity (8/). Such as- says may permit us to identify patients with autoimmune dis— eases who are at increased risk of developing radiation sequelae. Much work remains to be done in the area of predictive as- says. Confirmation of the ability of SF: or other measures of in— trinsic radiosensitivity to predict therapeutic outcome needs to be obtained in larger. multi-institutional trials. The development of more rapid (hours to days). inexpensive. reproducible assay systems to predict intrinsic tumor radiosensitivity is needed so that treatment decisions based on these assays can be made prior to the institution of therapy. The elimination of the potential problem of contamination of cultures with normal cells (e.g.. by histopathologic staining for specific tumor markers in the cells cultured) needs to be incorporated into these assays (82). Similarly. the development of rapid. inexpensive assay systems for normal tissue sensitivity warrants further attention. Addi- tional studies are needed on the relationship between normal cell intrinsic radiosensitivity and complications from frac— tionated radiotherapy. Unanswered questions relate to the pos— sible correlation between sensitivities of different normal cells (e.g.. fibroblasts versus lymphocytes) and the potential need for assays to be tailored to the type of damage (e.g.. fibroblast as- says for potential risks of fibrosis). Further work also is required to establish the relationship between normal cell sensitivity and tumor cell sensitivity. Finally, rapid. cost—effective screening as— says for genetic mutations likely to affect radiosensitivity should be developed and tested. Distant Metastases A wide range of intrinsic (tumor cell—related) and extrinsic (host—related) factors influences the development of tumor metastases [see Fidler (83) for reviewl. lntrinsic factors (includ- ing the HPV status. c—myc amplification. and matrix metal— loproteinase levels) and extrinsic factors (including fever and immunosuppression) have all been demonstrated to play a po- tential role in the frequency of metastases in carcinoma of the cervix. The risk of distant metastases among 94 patients with car- cinoma of the cervix of stages [B to “B was significantly in- creased if their tumors were HPV negative (11—fold). in those 135 whose tumors had a threefold or more amplification of c—myc (threefold to 10-fold increase), and in those whose tumors had a threefold or more amplification of c-myc (56—fold increase) (84). Studies on matrix metalloproteinases have demonstrated their involvement in tumor invasion. The 72—kd type IV collagenase cleaves type IV collagen of basement membranes (85). A recent study (86) of 34 patients with stage l squamous cell carcinoma of the cervix treated with radical hysterectomy (plus postopera— tive radiation therapy in 21 patients) quantitated the percentage of tumor cells that stained with an anti—72—kilodalton metal— loproteinase antibody. A statistically significant correlation was demonstrated between a higher percentage of tumor cell staining and 1) lymphovascular space invasion, 2) lymph node involve— ment, 3) number of positive lymph nodes. and 4) disease—free survival (independent of tumor size, lymph node involvement, or vascular invasion) (86). Laboratory investigations have been undertaken to develop matrix metalloproteinase inhibitors. Transfection of invasive and metastatic rat cells with cDNA for the metalloproteinase in- hibitor/TIMP-Z resulted in a partial suppression of the formation of lung colonies (87). The matrix metalloproteinase inhibitor BB-94 (batimastat) decreased tumor burden and prolonged sur- vival of mice bearing human ovarian carcinoma xenografts (88). Further studies on batimastat showed that it significantly reduced both the mean number of liver tumors and their cross- sectional area in a liver-invasive tumor model in nude mice while having no direct cytotoxic effects on the in vitro growth of the tumor cell line (89). Ongoing clinical trials are evaluating the efficacy of batimastat in patients with malignant ascites (from ovarian cancers). malignant pleural effusions, and bone metastases (90). The deleterious influence of fever during radiation therapy on survival in patients with carcinoma of the cervix has long been recognized, but its mechanism of action is unclear (9]). A Cox proportional hazards model analysis of 398 patients with in- vasive cervical cancer investigated the influence of fever during intracavitary radiation therapy while controlling for the in- fluence of other factors (e.g.. stage, age, and hemoglobin level). Increasing maximum temperatures noted during intracavitary radiation therapy were significantly associated with both a decrease in patient survival (P = .014) and an increase in the fre— quency of distant metastases as the initial sites of treatment failure (P = .038). In patients with maximum temperatures of 101.0 °F or higher. a 12% rate of metastases was noted com- pared with patients with maximum temperatures of less than 101.0 ”F. of whom 6.3% developed distant metastases (9]). Investigation of the potential adverse effect of systemic hy- perthermia has been performed in numerous laboratory studies utilizing murine tumors as well as in trials in spontaneous canine tumors. Such studies have demonstrated increased distant metastases associated with systemic hyperthermia (92). For ex- ample, an analysis of 64 dogs with localized soft-tissue sar— comas randomly allocated to receive local hyperthermia plus radiation therapy or the same regimen with the addition of whole-body hyperthermia (40.5 ”C, 30 minutes) showed no dif— ference in local control. However. the time to distant metastases was shorter in the dogs receiving the whole-body hyperthermia. 136 The hazards ratio for metastatic disease was 25-fold higher with the adjuvant use of whole-body hyperthermia compared with local hyperthermia and radiation therapy (93). The concurrent administration ofchemotherapy with systemic hyperthermia has been shown in several murine systems to reduce or eliminate the adverse effect of systemic hyperthermia on distant metastases. Oda et a1. (94) reported their results on Lewis lung carcinoma—bearing mice treated on day 5 with sys- tetnic hyperthermia (42 °C, 30 minutes) with or without simul— taneous chemotherapy. The mice were killed at 17 days, and their lungs were assayed for the number of metastatic foci. The lung involvement was graded from 0 (none) to 4 (>70 colonies) or 5 (total replacement of the lung with tumor). This study demonstrated an increase in lung metastases as a result of whole—body hyperthermia (average grade equaled 2.4 versus 1.6 for controls: P<.01). This increase could be prevented by the concurrent administration ofcisplatin (3.0 mg/kg) or by the con- cun‘ent administration of mitomycin C (1.0 mg/kg). When ad- ministered without concurrent whole-body hyperthermia. these drugs, at the doses employed, had no effect on the number of lung metastases (94). It therefore is suggested both in the murine tumor trials and in clinical trials. using whole—body hyperthermia in conjunction with chemotherapy, that chemotherapy may suppress or prevent the deleterious effect of whole-body hyperthermia on distant metastases. This hypothesis supports the possible use of concur— rent chemotherapy during intracavitary radiation therapy in patients at risk for fever (92). Patients in whom immunity is severely or chronically im— paired have an increased risk of developing cancer (95, 96). Such patients include organ transplant recipients and human im— munodeficiency virus (HIV)—positive patients. HIV-infected women with cervical cancer present with more advanced disease and have more rapid disease progression with higher death rates than HIV-negative patients (97). The patient‘s immune status has been shown to have a significant impact on subsequent dis- ease. Only patients with CD4 (helper-inducer) cell counts of more than 500/mm3 had a prolonged survival (97). These studies on tumor metastases suggest several avenues for future research. Genetic techniques to overcome c—myc amplification or antisense techniques to inhibit the expression of genes associated with metastatic potential are all approaches that need to be considered. The use of metalloproteinase in— hibitors to prevent metastases in cervical cancers that demonstrate a high percentage of cells expressing metal- loproteinase activity by cell staining should also be considered as a therapeutic adjuvant. It would seem prudent to prevent or suppress fever during brachytherapy and to consider adjuvant chemotherapy trials in patients with fever or with other high-risk factors for metastases (i.e., advanced stage of disease). Finally, the development of minimally immunosuppressive therapies would be of value. especially in patients with cervical cancer who are immunocompromised. It should also be noted that a higher rate of distant metastases has been reported in patients with cervical cancer who have locoregional failures (98). The potential selection factors that lead to local failure and that may concurrently lead to a greater metastatic potential require further investigation. Every attempt Journal ofthe National Cancer Institute Monographs N0. 21, 1996 should be made to optimize the initial therapy to prevent such locoregional failures. Conclusions Our simplified model suggests many new directions for radiobiologic research in carcinoma of the cervix. Customized therapy should be developed and should be based on the follow- ing: HPV status and p53 mutation status of the individual tumors. a consideration of tumor hypoxia, an evaluation of the intrinsic tumor and normal tissue radiosensitivities, and an evaluation of risk factors for tumor metastases. References (I) (4 (5 (6) (/0) (I3) (/4) (l5) (l6) (l7) (/8) Parkin DM. Pisani P. Ferlay J. Estimates of the worldwide incidence of eighteen major cancers in 1985. Int J Cancer 1993;54:594-606. Pisani P. Parkin DM. Ferlay J. Estimates of the worldwide mortality from eighteen major cancers in 1985. Implications for prevention and projec- tions of future burden. IntJ Cancer 1993;55:891—903. Iur Hausen H. Papillomaviruses in anogenital cancer as a model to under» stand the role of viruses in human cancers. Cancer Res 1989;49:4677-8 l. /.ur Hausen H. Human papillomaviruses in the pathogenesis of anogenital cancer. Virology 1991;184:9—13. Livingstone LR. White A. Sprouse J. Livanos E. Jacks T. Tlsty TD. A1- tered cell cycle arrest and gene amplification potential accompany loss of wiId~type p53. Cell 1992;70:923-35. Yin Y. Tainsky MA. Bischoff FZ. Strong LC. Wahl GM. Wild-type p53 restores cell cycle control and inhibits gene amplification in cells with mutant p53 alleles. Cell 1992;70:937-48. Tainsky MA. Bischoff FZ. Strong LC. Genomic instability due to germline p53 mutations drives preneoplastic progression toward cancer in human cells. Cancer Metastasis Rev 1995;14:43—8. Mullokandov MR. Kholodilov NG. Atkin NB. Burk RD. Johnson AB. Klinger HP. Genomic alterations in cervical carcinoma: losses of chromosome heterozygosity and human papilloma virus tumor status. Can- cer Res 1996;56:197-205. Graeber TG. Osmanian C. Jacks T. Housman DE. Koch CJ. Lowe SW. et a1. Hypoxia-mediated selection of cells with diminished apoptotic potential in solid tumours [we comment citations in Medlinel. Nature 1996;379:88— 91. Bosch FX. Manos MM. Munoz N. Sherman M. Jansen AM. Peto J. et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International biological study on cervical cancer (IBSCC) Study Group [.\‘(’(‘ comment citation in Medlinel. J Natl Cancer Inst 1995;87:796-802. Schwarz E. Frese UK. Gissmann L. Mayer W. Roggenbuck B. Stremlau A. et al. Structure and transcription of human papillomavirus sequences in cervical carcinoma cells. Nature 1985;314:111-4. Munger K. Phelps WC. Bubb V. Howley PM. Schlegel R. The E6 and E7 genes of the human papillomavirus type 16 together are necessary and suf- ficient for transformation of primary human keratinocytes. J Virol 1989:63:4417~21. Pecoraro G. Lee M. Morgan D. Defendi V. Evolution of in vitro transfor- mation and tumorigenesis of HPV16 and HPVIX immortalized primary cervical epithelial cells. Am J Pathol 1991;138:1-8. Scheffner M. Werness BA. Huibregtse JM. Levine AJ. Howley PM. The E6 oncoprotein encoded by human papillomavirus types 16 and 18 promotes the degradation of p53. Cell 1990:63:1 129-36. Munger K. Werness BA. Dyson N. Phelps WC. Harlow E. Howley PM. Complex formation of human papillomavirus E7 proteins with the retinoblastoma tumor suppressor gene product. EMBOJ 1989;814099—105. Dyson N. Howley PM. Munger K. Harlow E. The human papilloma virus4 16E7 oncoprotein is able to bind to the retinoblastoma gene product. Science 1989;243:93-7. Heck DV. Yee CL. Howley PM. Munger K. Efficiency of binding the retinoblastoma protein correlates with the transforming capacity of the E7 oncoproteins of the human papillomaviruses. Proc Natl Acad Sci U S A 1992;89:4442-6. Crook T. Wrcde D. Tidy JA. Mason WP. Evans DJ. Vousden KH. Clonal p53 mutation in primary cervical cancer: association with human-papil- lomavirus-negative tumours [we comment citations in Medline]. Lancet 1992;339:1070-3. Journal of the National Cancer Institute Monographs No. 21. 1996 (/9) (20) (26) (27) (28) (29) (33) (34) (35) (36) (37) (38) (39 (40) (4/) (43) (44) Kaelbling M. Burk RD. Atkin NB, Johnson AB. Klinger HP. Loss of heterozygosity on chromosome 17p and mutant p53 in HPV-negative cer» vical carcinomas. Lancet 1992;340:140—2. Busby—Earle RM. Steel CM. Williams AR. Cohen B. Bird CC. p53 muta- tions in cervical carcinogenesis—low frequency and lack of correlation with human papillomavirus status. BrJ Cancer 1994;69:732—7. von Kncbel Doeberitz M. Oltersdorf T. Schwartz. E. Gissmann L. Correla» tion of modified human papilloma virus early gene expression with altered growth properties in C4-1 cervical carcinoma cells. Cancer Res 1988;48: 3780—6. Steele C. Sacks PG. Adler-Storthz K. Shillitoe EJ. Effect on cancer cells of plasmids that express antisense RNA of human papillomavirus type 18. Cancer Res 1992;52:4706-1 1. Steele C. Cowsen LM. Shillitoe EJ. Effects of human papillomavirus type 18»specific antisense oligonucleotides on the transformed phenotype of human carcinoma cell lines. Cancer Res 1993:53tI()Supp1);233()»7. Storey A. Oates D. Banks L. Crawford L. Crook T. Antisense phos— phorothioate oligonucleotides have both specific and non—specific effects on cells containing human papillomavirus type 16. Nucleic Acids Res 1991;19:4109—14. Tan TM. Ting RC. In vitro and in vivo inhibition of human papillomavirus type 16 E6 and E7 genes. Cancer Res 1995;55:4599-605. Lappalainen K. Unti A. Jaaskelainen I. Syrianen K. Syrjanen S. Cationic Iiposomes mediated delivery of antisense oligonucleotides targeted to HPV 16 E7 mRNA in CaSki cells. Antiviral Res 1994;23:119-30. Hu G. Liu W. Hanania EG. Fu S. Wang T. Deisseroth AB. Suppression of tumorigenesis by transcription units expressing the antisense E6 and E7 messenger RNA (mRNA) for the transforming proteins of the human papilloma virus and the sense mRNA for the retinoblastoma gene in cervi- cal carcinoma cells. Cancer Gene Ther 1995;2z19-32. Zheng PS. Iwasaka T. Ouchida M. Fukuda K. Yokoyama M. Sugimari H. Growth suppression of a cervical cancer cell line (TMCC-I ) by the human wi1d«type p53 gene. Gynecol Oncol 1996;60:24550. Hamada K. Zhang WW. Alemany R. Wolf J. Roth JA. Mitchell MF. Growth inhibition of human cervical cancer cells with the recombinant adenovirus p53 in vitro. Gynecol Oncol 1996;60:373-9. Lu D. Chatterjee S. Brar I). Wong KK Jr. Ribozyme-mediated in vitro cleavage of transcripts arising from the major transforming genes of human papillomavirus type 16. Cancer Gene Ther 1994;1:267~77. Khan SA. Cervical cancer. human papillomavirus and vaccines. Clin Oncol (R Coll Radiol) 1993538600. Brachman DG. Beckett M. Graves D. Haraf I). Vokes E. Weichselbaum RR. p53 mutation does not correlate with radiosensitivity in 24 head and neck cancer cell lines [we comment citation in Medline]. Cancer Res 1993;53:3667-9. Lee JM. Bernstein A. p53 mutations increase resistance to ionizing radia- tion. Proc Natl Acad Sci U S A 1993;90:5742-6. Fan S. eI-Deiry WS. Bae I. Freeman J. Jondle D. Bhatia K. et a1. p53 gene mutations are associated with decreased sensitivity of human lymphoma cells to DNA damaging agents. Cancer Res 1994;54:5824—30. Lowe SW. Ruley HE. Jacks T. Housman DE. p53-dependent apoptosis modulates the cytotoxicity of anticancer agents. Cell 1993;74:957—67. Lowe SE. Bodis S. McClatchcy A. Remington L. Ruley HE. Fisher DE. et a1. p53 status and the efficacy of cancer therapy in vivo. Science 1994; 266:807-10. Petty RD. Cree IA. Sutherland LA. Hunter EM. Lane DP. Preece PE. et a1. Expression of the p53 tumour suppressor gene product is a determinant of chemosensitivity. Biochem Biophys Res Commun 1994;199:264-70. Hawkins DS. Demers GW. Galloway DA. Inactivation of p53 enhances sensitivity to multiple chemotherapeutic agents. Cancer Res 1996;56:892— 8. Crook T. Vousden KH. Properties of p53 mutations detected in primary and secondary cervical cancers suggest mechanisms of metastasis and in- volvement ofenvironmental carcinogens. EMBOJ 199221 123935—40. Tsuda H. Hirohashi S. Frequent occurrence of p53 gene mutations in uterine cancers at advanced clinical stage and with aggressive histological phenotypes. Jpn J Cancer Res 1992;83:1184—91. Tsuda H. Jiko K. Tsugane S. Yajima M. Yamada T. Tanemura K. et a1. Prognostic value of p53 protein accumulation in cancer cell nuclei in adenocarcinoma of the uterine cervix. Jpn J Cancer Res 1995;86:1049-53. Oka K. Nakano T. Arai T. p53CMI expression is not associated with prog- nosis in uterine cervical carcinoma. Cancer 1993;72:160—4. Kain]. C. Kohlberger P. Gitsch G. Sliutz G. Breitenecker G. Reinthaller A. Mutant p53 in patients with invasive cervical cancer stages 1B to 11B. Gynecol ()ncol 1995;57:212-4. Davidoff AM. Hemdon JE 2d. Glover NS. Kerns BJ. Pence JC. Iglehart JD. et a1. Relationship between p53 overexpression and established prog— nostic factors in breast cancer. Surgery 199111 10125964. (45) (46) (47) (48) (49) (50 (5/) (53) (53 ) (54) (55) (59) (60) (6/) (62) (63) (64) (65) (66) (67) (68) (69) (7(1) (7/) (72) I38 Thorlacius S Borresen AL. Eyfiord JE. Somatic p.53 mutations in human breast carcinoma in an Icelandic population: a prognostic factor. Cancer Res 1993153: 1637—41. Bergh J. Norberg T. Sjogren S Lindgren A. Holmberg L. Complete se- quencing of the p53 gene provides prognostic information in breast cancer patients. particularly in relation to adjuvant systemic therapy and radiotherapy. Nat Med 1995:l :1029-34. Hon'o Y. Takahashi T. Kuroishi T. Hibi K. Suyama M. Niimi T. et al. Prognostic significance of p53 mutations and 3p deletions in primary resected non-small cell lung cancer. Cancer Res 1993;53:1-4. Visakorpi T. Kallioniemi OP. Hcikkinen A. Koivula T. lsola J. Small sttb— group of aggressive. highly proliferative prostatic carcinomas defined by p53 accumulation. J Natl Cancer Inst 1992;84:883-7. Manin HM. I-‘ilipe MI. Morris RW. Lane DP. Silvestre F. p53 expression and prognosis in gastric carcinoma. Int J Cancer 1992;50:859-62. Sun XF. Carstensen JM. Zhang H. Stal O. Wingren S. Hatschek T. Nor" denskjold B. Prognostic significance of cytoplasmic p53 oncoprotein in colorectal adenocarcinoma. Lancet 19921340: 136973. Nakamori S. Yashima K. Murakami Y. Ishikawa ()_ ()Itigashi H. llnaoka S. et al. Association of p.53 gene imitations with short survival in pancreatic adcnocarcinoma. Jpn J Cancer Res 1995186: I 74-81. Jaros E. Perry RH. Adam L. Kelly PJ. Crawford PJ. Kalbag RM. et al. Prognostic implications of p53 protein. epidermal growth factor receptor. and Ki»67 labelling in brain tumours. BrJ Cancer 1992;66:373—85. King LA. Tase T. Twiggs LB. ()kagaki T. Savage JE. Adcock LL. et al. Prognostic significance of the presence of human papillomavirus DNA in patients with invasive carcinoma of the cervix. Cancer 1989;63:897-900. Walker J. Bloss JD. Liao SY. Bcrman M. Bergen S. Wilc/ynski SP. Human papillomavirus genotype as a prognostic indicator in carcinoma of the uterine cervix. ()bstet Gynecol 1989;74:781-5. Riou G. Favre M. Jeannel D. Bourhis J. Le Dottssal V. 0th G. Association between poor prognosis in early-stage invasive cervical carcinomas and non-detection of HPV DNA Iver comment citation in Medline]. Lancet 1990:33511 171-4. Higgins GD. Davy M. Roder I). U/eIin DM. Phillips GE. Burrell C]. In— creased age and mortality associated with cervical carcinomas negative for human papillomavirus RNA. Lancet 1991;338:910—3. Rose BR. Thompson CH. Cossarl YE. Elliot PE. Tattersall MH. Papil- lomavirus DNA and prognosis in cervical cancer. [letter; we comment citation in Medlinel. Lancet 1991;337:489. Burnett AF. Barnes WA. Johnson JC. Grendys E. Willett GD. Barter JF. et al. Prognostic significance of polymerase chain reaction detected human papillomavirus of tumors and lymph nodes in surgically treated stage IB cervical cancer. Gynecol Oncol 1992;47:343-7. Dische RS. Radiotherapy and anemia—the clinical experience. Radiother ()ncol 1991:20 Suppl:35—40. Bush RS. The significance of anemia in clinical radiation therapy. Int J Radiat ()ncol Biol Phys 1986;12:2047-50. Hockel M. Schlenger K. Mil/.e M. Schaffer U. Vaupel P. Hypoxia and radiation response in human tumors. Setnin Rad ()ncol 1996;6z3-9. Overgaard J. Horsman MR. Modification of hypoxia—induced radioresis— tance in tumors by the use of oxygen and sensiti/ers. Semin Rad ()ncol |996;6:I()»21. Brown JM. Giaecia AJ. Tttmour hypoxia: the picture has changed in the 1990s. lntJ Radiat Biol 1994;65:95-102. Brown JM. SR 4233 (tirapa/amine): a new anticancer drug exploiting hypoxia in solid tumours. BrJ Cancer 1993:6711 163-70. Stone HB. Brown JM. Phillips TL. Sutherland RM. Oxygen in human tumors: correlations between methods of measurement and response to therapy. Summary of a workshop held November 1920. 1992. at the Na« tional Cancer Institute. Bethesda. Maryland. Radiat Res 1993;136:422-34. Giaccia AJ. Hypoxic stress proteins: survival of the fittest. Setnin Radiat Oncol 1996.6:46-58. Levine EL. Renehan A. Gossiel R. Davidson SE. Roberts SA. Chadwick C. et al. Apoptosis. intrinsic radiosensitivity and prediction of radiotherapy response in cervical carcinoma. Radiother Oncol 1995137: 1-9. Peters LJ. Inherent radiosensitivity of tumor and normal tissue cells as a predictor of human tumor response. Radiother ()ncol 1990;17:177-90. West CM. Invited review: intrinsic radiosensitivity as a predictor of patient response to radiotherapy. BrJ Radiol 1995582827737. Deschavanne PJ. Fertil B. A review of human cell radiosensitivity in vitro, Int J Radiat ()ncol Biol Phys 1996;34:251—66. Busch D. Genetic susceptibility to radiation and chemotherapy injury: diagnosis and management. Int J Radiat Oncol Biol Phys 1994;30:997- 1002. West CM. Davidson SB. Robetts SA. Hunter RD. Intrinsic radiosensitivity and prediction of patient response to radiotherapy for carcinoma of the cer- vix. BrJ Cancer 1993;68:819-23. (73) (74) ( 75) (76) (77) (78) (79) (80) (8/) (82) (83') (84) (85 ) (86) (87) (88) (89) ( 90 (9/) (92) (93) (94) (95) (96 (97) (98 Swift M. Morrell D. Massey RB. Chase CL. Incidence of cancer in 161 families affected by ataxia-telangiectasia [we comment citations in Med- line]. N Engl J Med 1991;325:1831—6. Fodor SP. Read JL. Pii‘rung MC. Stryer L. Lu AT. Solas D. Light—directed. spatially addressable parallel chemical synthesis. Science 1991;251:767- 73. Fleck R. McNeese MD. Ellerbroek NA. Hunter TA. Holmes FA. Conse- quences of breast irradiation in patients with pre-existing collagen vascular disease. lntJ Radiat Oncol Biol Phys l989:17:829»33. Abu»Shakra M. Lee P. Exaggerated fibrosis in patients with systemic sclerosis (sclerodertna) following radiation therapy. J Rheumatol 1993; 20: 1601—3. Olivotto IA. Fairey RN. Gillies JH. Stein H. Fatal outcome of pelvic radiotherapy for carcinoma of the cervix in a patient with systemic lupus erythetnatosis. Clin Radiol I989;40:83—4. Ross JG. Hussey DH. Mayr NA. Davis CS. Acute and late reactions to radiation therapy in patients with collagen vascular diseases. Cancer 1993;71:3744-52. Smidcr V. Ralhmell WK. Licbcr MR. Chu G. Restoration of X~ray resis- tance and V(D)J recombination in mutant cells by Ku cI)NA. Science 1994;266:2889 I. Kirchgessner CU. Patil CK. Evans JW. Cuomo CA. Fried LM. Carter T. el al. DNA-dependent kinase (p350) as a candidate gene for the murine SCID defect. Science 1995;26711 178—83. Harris G. Cramp WA. Edwards JC. George AM. Sabovljev SA. Hart L. et al. Radiosensitivity of peripheral blood lymphocytes in autoimmune dis— ease. Int J Radiat Biol Relat Stud Phys Chem Med 1985;47:689-99. Lawton PA. Hodgkiss RJ. Eyden BP. Joiner MC. Growth of fibroblasts as a potential confounding factor in soft agar clonogenic assays for tumour cell radiosensitivity. Radiother Oncol 1994;32:218-25. Fidler IJ. 7th Jan Waldenstrom Lecture. The biology of human cancer metastases. Acta Oncol 1991 :301668-75. Riou G. Le MG. Favre M. Jeannel D. Bourhis J. 0th G. Human papil- Iomavirus-negative status and c»myc gene overexpression: independent prognostic indicators of distant metastasis for early-stage invasive cervical cancers. J Natl Cancer Inst 1992;84:1525-6. A/navoorian S. Murphy AN. Stctler-Stevenson WG. Liotta LA. Molecular aspects of tumor cell invasion and metastases. Cancer 1993;71:1368—83. Gar/.ctti GG. Ciavaltini A. Lucarini G. Goteri G. Romanini C. Biagini G. The 72-kDa metalloproteinase immunostaining in cervical carcinoma: relationship with lymph nodal involvement. Gynecol Oncol 1996;60:271- 6. DeCIerck YA. Perez N. Shitnada H. Boone TC. Langley KE. Taylor SM. Inhibition of invasion and metastases in cells lransfected with an inhibitor of metalloproteinases. Cancer Res 1992;52:701-8. Davies B. Brown PD. East N. Crimtnin MJ. Balkwill FR. A synthetic matrix metalloproteimtse inhibitor decreases tumor burden and prolongs survival of mice bearing human ovarian carcinoma xenografls [published erratum appears in Cancer Res 1993;53:36521 Cancer Res 19935322087— 91. Watson SA. Morris TM. Robinson G. Crimmin MJ. Brown PD. Hardcastle JD. Inhibition of organ invasion by the matrix tnetalloproteinase inhibitor batimastal (BB-94) in two human colon carcinoma metastasis models. Cancer Res 1995;55:3629-33. Brown PD. Clinical trials of a low molecular weight matrix metallopro» teinase inhibitor in cancer. Ann N Y Acad Sci 1994;732:217-21. Kapp [)S. Lawrence R. Temperature elevation during brachytherapy for carcinoma of the uterine cervix: adverse effect on survival and enhance- ment ofdistant metastasis. Int .1 Radiat Oncol Biol Phys 1984;10:2281—92. Kapp DS. Thermal dose response. systemic hyperthermia. and metastases: old friends revisited [editorial]. Int J Radiat ()ncol Biol Phys 1996;35:189‘94. Thrall DE. Prescott DM. Samulski TV. Rosner GL. Denman I)L. Legor— reta RL. et al. Radiation plus local hyperthermia versus radiation plus the combination of local and whole-body hyperthermia in canine sarcomas. Int J Radiat ()ncol Biol Phys 1996;34:1087-96. Oda M. Koga S. Maeta M. Effects of total-body hyperthermia on inelas— tases frotn experimental mouse tumors. Cancer Res 1985145: 1532—5. Porrcco R. Penn 1. Droegemueller W. Greer B. Makowski E. Gynecologic malignancies in immunosuppressed organ homograft recipients. ()bstet Gynecol 1975;45:359-64. Penn 1. Why do immunosuppressed patients develop cancer? Crit Rev Oncog 1989:1127-52. Maiman M. Fruchter RG. Guy L. Cuthill S. Levine P. Serur E. Human im~ munodcficiency virus infection and invasive cervical carcinoma. Cancer 1993:71 1402—6. Fagttndes H. Perez CA. Grigsby PW. Locketl MA. Distant metastases after irradiation alone in carcinoma of the uterine cervix [we comment citation in Medline]. lntJ Radiat ()ncol Biol Phys 1992;24:197-204. Journal of the National Cancer Institute Monographs No. 21. I996 Note We thank Dr. Martin M. Brown for his discussions on tumor hypoxia and tirapazamine, Dr. Gilbert Chu for his discussions on genetic effects on DNA Journal of the National Cancer Institute Monographs N0. 21. I996 repair with particular reference to the Ku mutations. and Dr. Donald Thrall for permission to use his results on whole»body hyperthermia and metastases in his dog trial prior to their final publication. We also thank Darci Mclntyre for her secretarial assistance in preparing this manuscript Human Papillomavirus Biology Douglas R. Lowy, John T. Schiller* Human papillomaviruses (HPVs) induce papillomas of the skin and mucous membranes. and some HPVs are now asso- ciated with the development of malignant epithelial tumors (1- 3). These cancers include anogenital cancers. especially cancer of the cervix (3,4). A wealth of epidemiologic and molecular biologic data now points to an etiologic link between HPV in— fection and most cervical cancers. More than 70 distinct HPV genotypes (types) have been described (23). Among these. approximately 20 are considered genital/mucosal types because they mainly infect the genitalia and mucosal sites. The genital/mucosal group of HPVs is further classified into high-risk and low—risk types according to the fre- quency with which they are found in cervical cancer (3.4). The genomes of HPVs are a circular. double-stranded DNA around 8 kilobases in length (5). As with most viruses. HPV genomes encode nonstructural proteins as well as those that form the viral capsid (Fig. l). The viral genome is divided into three functional regions: the upstream regulatory region. the early region. and the late region. The upstream regulatory region contains binding sites for cel- lular and viral regulatory proteins. It serves to regulate the ex- pression of several viral genes and also contains the origin of viral DNA replication. The genes in the early region encode the nonstructural viral proteins El—E8, which direct viral DNA replication and seem to be responsible for the altered growth properties induced directly by the virus. El binds to the upstream regulatory region and is required for viral DNA replication (6,7). E2 regulates viral transcription via its binding to E2-binding sites in the upstream regulatory region and elsewhere in the genome. It also par- ticipates in viral DNA replication (8). E5. E6. and E7 are the viral oncoproteins (9.10). The major and minor virion structural proteins L1 and L2. respectively. are encoded by the late genes. L2 is required for encapsidation of the viral DNA (11,12). Ll can form properly E7 E1 E5 L2 -— .— E6 E2 L1 _ — E4 - URR Early Region Late Region I l l l Fig. l. Organi/ation of human papillomavirus type lo (HPVlo) genome. The genes in the early region (El-E8) encode the nonstructural viral proteins. includ- ing the E6 and E7 \iral oncoproteins. which are preferentially retained and ex- pressed in cervical cancers. The late region encodes the structural viral proteins Ll and L2. L1 is the major capsid protein: it can form the capsid without L2. The major neutralizing antibodies are directed against type-specific conforma— tional epitopes on LI. The upstream regulatory region (URR) contains many of the DNA regulatory elements that participate in replication of the viral DNA and transcription of the viral genes. Journal of the National Cancer Institute Monographs No. 2|. I996 folded. empty viral capsids in the absence of L2 (13). The major neutralizing antibodies are directed against conformational epitopes of Ll. Although Ll is the most conserved protein among HPVs. most neutralizing Ll epitopes are type specific (12,14). It is therefore likely that most protective immunity against HPV is type specific. HPV infection and replication are limited to stratified squamous epithelia. with no viremic phase (15). Most of the virus replication takes place in suprabasal cells that are undergo— ing terminal differentiation. a strategy that seems designed to evade immune surveillance. In general. the nonstructural (early) proteins are expressed in the lower and upper levels of the epithelium. whereas the structural (late) proteins are expressed only in the upper layers. Benign lesions usually express all the viral genes. whereas only a subset of the viral genes. mainly the E6 and E7 oncogenes. is expressed in cervical cancers. It is un- common for the structural viral proteins to be expressed in severe dysplasias or invasive tumors. Thus. progression to can— cer is an aberration. from the viral point of view. that interferes with virus production (3,4,9). Cervical infection with HPV is transmitted sexually in adults. This infection is extremely common in sexually active women under 30 years of age. with prevalence rates ()f 2()%-40‘7r as as— certained by the most sensitive techniques (16). Most HPV in- fections. even those with high—risk types. are self-limited. and the proportion of women positive for HPV DNA decreases with age. Most infected women do not develop cytologic abnor- malities. Reliable epidemiologic evidence based on direct identifica— tion of HPV DNA in cervical samples. however. has now shown that infection by high—risk HPV types is by far the most impor- tant risk factor (10—fold to 2()()—fold compared with controls) for the development of high—grade squamous intraepithelial lesions and for invasive cervical cancer (3.17—20). The apparent lack of a strong association between cervical HPV infection and precan- cerous and malignant lesions of the cervix. noted in some studies carried out in the 1980s. probably resulted from methodologic difficulties in HPV DNA detection. which led to false—positive and false—negative misclassification of the HPV status of lesions (21.22). Overcoming these technical problems led to the consistent findings of HPV as the major risk factor. Indeed. in many recent studies (3). risk factors traditionally associated with high-grade lesions and with cervical cancer. such as number of sexual partners or smoking. appear to have been surrogate markers for HPV infection. 'E:/l[’1111111()ll ofuul/Iurx: Laboratory of Cellular Oncology, Division of Basic Sciences. National Cancer Institute. Bethesda. MD. (‘4u‘rtntlmmlwu't' IU.‘ Douglas R, Lowy. M.D.. National Institutes of Health. Bldg. 36. Rm. 1D32. Bethesda, MD 20892. l4l HPV infection of the cervix generally precedes the onset of cancer by many years. The pathogenetic importance of HPV in- fection preceding the development of cancer is underscored by a prospective. nested case—control study (23). based on serologic diagnosis of past infection with HPVI6. Infection with HPVI6 conferred an excess risk (odds ratio 2 I32) for subsequent development of cervical carcinoma, and odds ratios were higher for patients with more than 5 years between sampling and diag— nosis (23). Close to 95% of cervical cancers have been found to contain HPV DNA. The relative frequency with which the various high- risk HPV types are found in cervical cancer is remarkably similar in most regions of the world (24). The high—risk HPVI6 is found in 5()%-70% of cervical cancers. with high-risk types 18. 31. 33. and 45 accounting for most of the remaining HPV- positive tumors. The low—risk HPV6 and HPVI l are often found in benign lesions, but almost never in cervical tumors. The apparent difference in oncogenic potential between high— risk and low-risk HPVs is reflected in their in vitro biological and biochemical activities (3.4.9.25). High-risk HPVs can im- mortalize human keratinoeytes in culture. whereas low—risk HPVs are ineffective in this assay. Further analysis has shown that keratinocyte immortalization requires two viral genes. E6 and E7. from a high-risk HPV type. These are the two viral genes that are preferentially retained and expressed in cervical cancers and in cell lines derived from them. Continued expres- sion of the E6 and E7 oncogenes is required for cell prolifera— tion and viability in these cell lines. Two important biochemical properties of the high-risk E6 and E7 oncoproteins are that they each bind and inactivate a protein of a cellular tumor suppressor gene: p53 for E6 and pr for E7. The low-risk E6 and E7 proteins lack these functions or carry them out less efficiently. The long interval between HPV infection and the develop- ment of cervical cancer suggests that factors other than viral in- fection per se are required for tumor development. Some aspects involve the virus. Persistence of viral infection is one corn— ponent (26). While the viral DNA in benign lesions is episomal. it is integrated into the host genome in most cervical cancers. Viral expression is usually limited to E6 and E7. Other changes associated with progression may involve E6- and E7-induced chromosome instability and the deregulation of cellular growth control genes. Progression is more common in long—term renal transplant patients on immunosuppressive therapy or in women who are human immunodeficiency virus positive (3). indicating the role of impaired host defense mechanisms such as cellular immune function. As in the clinical situation. keratinocytes that have been recently immortalized in vitro by high-risk HPV are not fully transformed. The addition of an activated ras oncogene can. however. render these cells tumorigenic in nude mice (27). In summary. a combination of molecular. epidemiologic. and experimental evidence has now established a close link between cervical infection with certain HPV types and the development of cervical cancer. As with other tumors. cervical cancer repre- sents the outcome of a multistep process. Infection with high- risk HPV represents a critical step in this process. but it is not by itself sufficient to produce cervical cancer. The key role of HPV infection in the pathogenesis of cervical cancer. however. makes I42 efforts to prevent and treat genital HPV infection a potentially attractive target for the prevention and treatment of this cancer. References (I) Lowy DR. Human papillomaviruses and epithelial cancer. In: Mukhtar II. editor. Skin cancer: mechanisms and human relevance. Boca Raton (FL): CRC Press. 1994:35I-8. (2) /.ur Hausen H. De Villiers EM. Human papillormiviruses. Annu Rev Microbiol 1994;48:427-47. (3) Human papillomaviruses. IARC Monogr Eval Carcinog Risks Hum. Volume 64. I995. (4) xur Hausen H. Molecular pathogenesis of cancer of the cervix and its causation by specific human papillomavirus types. Curr Top Microbiol Im- munol 1994;186: [31-56. Turek LP. The structure. function and regulation of papillornaviral genes in infection and cervical cancer. Adv Virus Res |994;44:3()5—56. Ustav M. Stenlund A. Transient replication of BPVAI requires two viral polypeptides encoded by the El and E2 open reading frames. EMBO J 1991;10:449-57. (7) Chiang CM. Ustav M. Stenlund A. Ho TF. Broker TR. Chow LT. Viral El and E2 proteins support replication of homologous and heterologous papil— lomavirus origins. Proc Natl Acad Sci U S A 19918925799803. (8) Ushikai M. Lace MJ. Yamakawa Y. Kono M. Anson J. lshiji T. et al. trans activation by the full-length E2 proteins of human papillomavirus type I6 and bovine papillomavirus type I in vitro and in vivo: cooperation with ac- tivation domains ofcellular transcription factors. J Virol I994;68:6655—66. Werness BA. Munger K. Howley PM. Role of the human papillomavirus oncoproteins in transformation and carcinogenic progression. In: DeVita VT Jr. Hellman S Rosenberg SA. editors. Important advances in oncology I99l. Philadelphia: Lippincott. 1991348. (I()) Straight SW. Hinklc PM. Jewers RJ. McCance DJ. The ES oncoprotein of human papillomavirus type I6 transforms fibroblasts and effects the downregulation of the epidermal growth factor receptor in keratinocytes. J Virol 1993;67:452I-32. Zhou J. Stenlel DJ. Sun XY. Fra/er IH. Synthesis and assettibly of infec- tious bovine papillomavirus particles in vitro. J Gen Virol l993:74(Pt 4):763—8. Roden RB. Greenstone HL. Kirnbauer R. Booy FP. Jessie J. Lowy DR. et al. In vitro generation and type-specific neutralization of a human papil- lomavirus type 16 virion pseudotypc. J Virol. In press. Kirnbauer R. Booy F. Cheng N. Lowy DR. Schiller JT. Papillomavirus LI major capsid protein self-assembles into virus»like particles that are highly immunogenic. Proc Natl Acad Sci U S A I992;X9: | 2I8()—4. (I4) Christensen ND. Kirnbauer R. Schiller JT. Ghirn SJ. Sehlegel R. Jenson AB. et al. Human papillomavirus types 6 and l I have antigenieally distinct strongly immunogenic conformationally dependent neutralizing epitopes. Virology I994;2()5:329—35. (I5) Taichman LB. LaPona RF. The expression of papillomaviruses in human epithelial cells. In: Sal/.man NP. Howley PM. editors. The PUPOI'II\‘II‘I(I(I('.' Volume 2. The papillomaviruses. New York: Plenum Press. I987: I()9»39. (I6) Schiffman MH. Recent progress in defining the epidemiology of human papillomavirus infection and cervical neoplasia. J Natl Cancer Inst I992; 84394-8. Lorincl. AT. Reid R. Jenson AB. Greenberg MD. Lancaster W. Kurman RJ. Human papillomavirus infection of the cervix: relative risk associa- tions of 15 common anogenital types. Obstet Gynecol I992z79z328-37. Meijer CJ. van den Brule AJ, Snijders PJ. Helmerhorst T. Kenemans P. Walboomers JM. Detection of human papillomavirus in cervical scrapes by the polymerase chain reaction in relation to cytology: possible implica- tions for cervical cancer screening. In: Munoz N. Bosch FX. Shah KV. Meheus A. editors. The epidemiology of human papillomavirus and cervi~ cal cancer. Lyon: International Agency for Research on Cancer. 1992:27I» X I. (I9) Schiffman MH. Bauer HM. Hoover RN. Glass AG. (‘adell DM. Rush BB. et al. Epidemiologic evidence showing that human papillomavirus infec- tion causes most cervical intraepithelial neoplasia [we comment citations in Medline]. J Natl Cancer Inst 1993;85:958-64. (20) Munol. N. Bosch FX. de Sanjose S. Shah KV. The role of HPV in the etiology ofcervical cancer. Mutat Res I994z305z293—30l. Franco EL. The sexually transmitted disease model for cervical cancer: in— coherent epidemiology findings and the role of misclassification of human papillomavirus infection. Epidemiology I99] :2:98- 106. (22) Schiffman MH. Schatzkin A. Test reliability is critically important to molecular epidemiology: an example from studies of human papil- LI. 3 (9 (II (I2 (I3) (I7) ([8) (2]) Journal of the National Cancer Institute Monographs No. 21. I996 (24) Iomavirus infection and cervical neoplasia [we comment citation in Med— linel. Cancer Res I994;54(7 Suppl): l944s—7s. Lehtinen M. Dillner .l. Knekt P. Luostarinen T. Aromaa A, Kirnbauer R. ct al, Serologically diagnosed infection with human papillomavirus type lo and risk for subsequent development of cervical carcinoma: nested case— control stttdy [we comment citation in Medlinel. BM] |9<)(»:312:537~9. Bosch FX. Manos MM. Munol. N. Sherman M. Jansen AM. Peto I. et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International Biological Study on Cervical Cancer (IBSCC) Study Group [sou comment citation in Medlinel, J Natl Cancer Inst 1995;87:706-802. Journal of the National Cancer Institute Monographs No. 21. I996 (25) Vousdcn K. Interactions of human papillomavirus transforming proteins (26 (27 ) with the products oftumor suppressor genes. FASEBJ 1993;72872-9. Ho GY. Burk RD. Klein S. Kadish AS. Chang CJ. Palan P. et al. Persistent genital human papillomavirus infection as a risk factor for persistent cervi- cal dysplasia. I Natl Cancer Inst 1995;87:1365-7]. DiPaoIo IA. Woodworth CI). Popescu NC. Notario V. Doniger J. Induc— tion of human cervical squamous cell carcinoma by sequential transfection with human papillomavirus 16 DNA and viral Harvey ras. Oncogene 1989:4395—9. Human Papillomavirus Immunology and Vaccine Prospects Michael/l. Steller, John T. Schiller* As a result of several recent advances in molecular biology, the association between human papillomavirus (HPV) infec- tion and cervical cancer has been firmly established and the oncogenic potential of certain HPV types has been clearly demonstrated. These observations provide the impetus for the development of novel vaccines to prevent or treat HPV- associated cervical cancer. Because there is no effective cul- turing system to propagate HPV, traditional approaches for studying HPV and developing vaccines have been hampered. However, recent studies using recombinant subunit prepara- tions in animals have yielded promising results and en- courage their investigation in human trials. Strategies currently under investigation focus on the induction of effec- tive humoral immune responses for prophylaxis against sub- sequent HPV infection; for the treatment of existing HPV infections, techniques to improve cell-mediated immunity by enhancing viral antigen recognition are being studied. Small-scale human trials using several different vaccine ap- proaches should be completed within the next few years and field trials of the most promising 0ne(s) could begin within a decade. The development of successful therapeutic and/0r prophylactic vaccines offers an attractive alternative to ex- isting screening and treatment programs for cervical cancer and may result in a substantial reduction in the worldwide morbidity from this disease. [Monogr Natl Cancer Inst 1996;21:145-8] On the basis of extensive epidemiologic data. human papil- lomaviruses (HPVs) are strongly associated with a spectrum of anogenital neoplasms. including condyloma, cervical dysplasia. and cervical cancer. Such studies indicate that nearly 90% of in- vasive cervical tumors harbor HPV DNA. and genotype 16 can be detected in approximately 50% of squamous cell carcinomas of the cervix (I). More than 70 different HPV genotypes have been identified. Mucosotropic HPVs are grouped into low—risk or high—risk categories on the basis of each genotype's associa- tion with a benign or malignant disease process (2). Low—risk serotypes 6 and l 1. for instance. are commonly detected in con- dyloma accuminata (venereal warts) but are virtually never found in cervical cancer. In contrast. high-risk serotypes l6 and 18 can be detected in roughly 70% of all invasive cervical tumors. Other high-risk serotypes include, among others, HPV- 31. -33. and -45. Recent research has shown that the E6 and E7 genes of HPV—l6 and -I8 encode for oncoproteins that can im- mortalize human keratinocytes. This potential appears to be limited to high-risk types. since E6 and E7 from HPV-6 or -11 are nonimmortalizing (3,4). Continued expression of these Journal ofthe National Cancer Institute Monographs No. 2]. I996 nucleoproteins appears necessary to maintain proliferation of cervical cancer—derived cells in vitro (5). Several lines of evidence suggest the importance of the host‘s immune response in the pathogenesis of HPV—associated cervi— cal lesions. For example. more than 60% of HPV-positive. mild- ly dysplastic lesions resolve spontaneously (6) and such factors as human immunodeficiency virus infection and exposure to im- munosuppressive agents have been strongly associated with HPV—induced carcinogenesis (7.8). Also. the prevalence of geni— tal HPV infections peaks soon after the onset of sexual activity in women and declines thereafter. suggesting that long-term protection is generally induced (9). These observations form the basis for the development of vaccine strategies to combat HPV infection. Both prophylactic and therapeutic HPV vaccine strategies are being developed. Prophylactic vaccines would likely act by inducing antivirion antibodies that would prevent HPV infection. while therapeutic vaccines would eliminate pre- existing infections through cell-mediated mechanisms. Therapeutic Vaccines Because oncoproteins E6 and E7 are selectively retained and expressed in cervical tumors associated with HPV—l6 ([0). they are attractive targets for novel immunotherapies. One therapeu— tic vaccine strategy is based on the generation of a T—cell—medi— ated immune response to target cells engineered to overexpress viral proteins. The genes for E6 and/or E7 can be introduced into target cells by transfection of naked DNA or infection via recombinant viral vectors. and robust cellular immune responses can be elicited. Presentation to T cells of endogenously syn- thesized viral antigens involves degradation of the viral protein into small epitope peptides that are transported via the endoplas— mic reticulum to the cell surface in association with class I major histocompatibility complex (MHC) molecules (11.12) (Fig. 1). Cytolytic T cells (CTLs) recognize antigens by the binding of their clonotypic T-cell receptors to the processed en— dogenous peptides associated with class I molecules. While this strategy has been shown to elicit desirable immune responses. it also has its limitations. For example. E6 and E7 are viral on- cogenes that might promote tumorigenesis. This problem might be overcome by using mutant forms of E6 or E7. but these mutations may compromise immunogenicity in some cases. In *Al'fi/iulions afoul/mm: M, A. Steller (Section of Gynecologic Oncology. Surgery Branch). J. T. Schiller (Laboratory of Cellular Oncology). National Cancer Institute, Bethesda, MD. (‘m'n'x/mmlem'e Io: Michael A. Steller. M.l).. National Institutes of Health. Bldg. I0. Rm. 28—42. Bethesda. MD 208924502. 145 Virus as V :l: l Rig? (I/V" Cleavage refEs“, - " _ z. ‘n Endoplasmic "‘ ‘b Reticulum Peptide ....... Transport \ Peptide MHC . / Transport Golgi 3 L.) Apparatus O K o / \ / / Class | 2" Viral Peptide Class | Peptide 2 Cervical Cancer Intravenous of Cells and lL»2 / f ( Machine ( t ‘ (- Isolates [I ,, Blood Return 335.21%? C /"HPV Epitope Specific CTL /’ f l (t. ‘ : Lymphocytes \. ) (I (3/ Peripheral Lymphocytes and HPV Peptide with IL-2 Fig. l. ()ncoproteins encoded by viral DNA undergo proteolytic degradation during nortnal cellular processing. Cleaved protein fragments. or peptides. are transported into the endoplastnic reticulum. then shuttled to the cell surface after specifically binding to major histocompatibility complex (MHC) class I immune molecules. This multitnolecular complex is then recognized by cytotoxic T lyin- phocytes. with subsequent destruction ofthe virally infected cell. addition. Uene transfer will likely be tnost efficient using viral vectors. and these vectors themselves might cause substantial morbidity. especially in immunosuppressed patients. Another viable approach in therapeutic vaccine development is the use of peptide epitopes of E6 or E7. MHC class I- restricted CTL epitope peptides have been identified for both human and mouse model systems. In vivo experiments in mice have demonstrated that HPV epitope peptides in the presence of nonspecific adjuvants are capable of eliciting an immune re— sponse (13). In humans. recent studies (14,15) in both normal volunteers and patients with cervical cancer have demonstrated that peptide epitopes of HPV-I6 E7 can induce CTL responses that specifically lyse tumor cells that harbor HPV—16. These fragments of viral proteins could elicit cellular immune respon— ses after binding to empty cell surface class 1 complexes. In con— trast to the transfer of transforming viral genes. peptide immunogens are likely to be safe and relatively simple to ad- minister via subcutaneous injection. Alternatively. it is conceiv— able to immunize an HPV-16+ tumor—bearing patient by stimulating the patient's own lymphocytes in vitro with the im— munologically relevant peptides and then expanding this lym- phocyte population before transfusing them back into the patient (Fig. 2). Because the binding of these epitopes is restricted by I46 Fig. 2. Schematic diagram of adoptive immunotherapy for the treatment of cer» vical cancer. Lymphocytes obtained by leukapheresis are cocultured with im- munologically relevant human papillomavirus (HPV) peptide epitope. Clones of HPV—specific cytolytic T lymphocytes (CTLs) are expanded using interleukin 2 (IL-2). then transfused back into the patient. the allelic expression of MHC class I molecules. it will be necessary to identify a library of peptides to overcome the ex— treme diversity of human leukocyte antigen (HLA) haplotypes. lnitial efforts to develop epitope vaccines have focused primari— ly on class I HLA—A2-restricted models. since this allele is the one most commonly expressed (16). A third type of immunogen for a therapeutic vaccine is the entire Eb or E7 viral protein. which could be cheaply generated in bacteria. While soluble proteins generally favor an antibody rather than a cell—mediated response. the use of recently devel— oped adjuvants that specifically promote cell-mediated respon— ses may make it possible to develop effective therapeutic vaccines that are based on soluble viral proteins. Table 1 shows several clinical trials currently ongoing to test various thera- peutic vaccine strategies in patients with advanced cervical can- cer. A critical consideration for all therapeutic vaccine strategies aimed at eliciting an effective cellular immune response is whether the infected cell‘s epitope—presenting machinery is in— tact. Cell-surface HLA class I molecules are frequently down- regulated in cervical cancers ([7,18). and mutation or deletion of the [fig—microglobulin gene has also been shown to impair CTL recognition of malignant cells (/9). Therefore. successful development of therapeutic cellular immune responses might Journal ofthe National Cancer Institute Monographs No. 2|. I996 Table 1. Therapeutic human papillomavirus-l(u-hased clinical trials open for accrua|* Institution University of Leiden (The Netherlands) University onueensland (Australia) Cytel University of Wales (UK) and National Cancer Institute (NCI) Cantab NCI NCl Cytel Corporate collaborator Commonwealth Serum Laboratories Vaccine construct Eo and E7 epitopes (HI..A~A2) Mutated E7 protein Mutated E6 and E7 genes (vaccinia virus vector) E6 and E7 epitopes loaded onto antigen-presenting cells (HLA-A2) Lipidated E6. E7. and helper epitopes tHLA—AZ) *All protocols are phase 1 studies and only accrue patients with advanced disease. HLA : human leukocyte antigen. eventually focus on patients with preinvasive cervical lesions, where HLA class I antigen expression remains relatively pre— served (18). Prophylactic Vaccines Prophylactic vaccines prevent viral infections by generating antivirion antibodies that neutralize the virus before it enters the target cell. Certain prophylactic vaccines have been enormously effective in preventing subsequent infection by other human viruses, including hepatitis B. measles. mumps. and polio. In contrast to the oral poliovirus vaccine. which is an attenuated form of the poliovirus. development of an attenuated HPV vac— cine has been precluded because there is currently no effective culturing system to propagate HPV. Because of the paucity of available tissue. the use of inactivated virus or crude viral ex— tracts from infected humans has been impractical and has the theoretic disadvantage of exposing normal subjects to HPV DNA-encoding viral oncogenes. Therefore. prophylactic vac— cine development for HPV has focused on recombinant subunit preparations consisting of the L1 and L2 virion structural proteins. A similar strategy was used in the highly successful prophylactic vaccination program for hepatitis B virus. A study (20) in both animals and humans has confirmed the production of protective antibodies against the surface antigen of hepatitis B virus after immunization with recombinant hepatitis B vac- cine and demonstrated its efficacy in preventing the subsequent transmission of this virus. It has been shown that the papillomavirus Ll and L2 capsid proteins. including those of the major high-risk types. coas- semble into papillomavirus-like particles (VLPs) when ex— pressed at high levels in eukaryotic cells (2]). These VLPs are morphologically indistinguishable from authentic virions. ex- cept that they lack the viral DNA (22) (Fig. 3). This approach represents a major advance in HPV prophylactic vaccine development for several reasons: 1) previous attempts to generate recombinant HPV capsid proteins in bacteria yielded proteins that lacked critical conformational epitopes and there- fore were not effective in eliciting neutralizing antibodies; 2) there is no risk of administering potentially oncogenic viral DNA since the VLPs do not contain the E6 or E7 genes; and 3) VLPs are relatively easy and inexpensive to produce and purify. VLPs have proven to be effective in generating type-specific protection from viral challenge in animal papillomavirus models (23,24). Journal of the National Cancer Institute Monographs No. 21. 1996 Important considerations involving a prophylactic HPV vac- cine for cervical cancer include safety. stability. cost—effective- ness. activity against all or at least most of the oncogenic HPV subtypes. and the capacity to generate a robust antibody re— sponse on genital mucosal surfaces. Since numerous epidemi— ologic studies strongly suggest that HPV infection is in large part sexually transmitted (25), a successful prophylactic vaccine will only be effective when it is administered to women before they acquire HPV infection through sexual activity. By dimin— ishing transmission rates. vaccine programs targeting young men as well as young women may enhance the overall effective- ness. even though males rarely manifest diseases related to high- risk genital HPV infections. Limited analysis suggests that divergent variants of at least HPV-16 are serologically cross-reactive (26). However. im— munity in VLP-based vaccines is likely to be primarily genotype specific (27). Therefore. a multivalent vaccine should be con- sidered for clinical trials. Since 80% of HPV—associated cervical cancers contain either HPV—16. -18. —31. or -45 (1). it would be desirable to include VLPs of at least these four genotypes in a vaccine. Conclusion Although cervical cytology screening programs have drasti- cally reduced the incidence of cervical cancer in developed Fig. 3. Transmission electron micrograph of human papillomavirus-16 virus-like particles (VLPs). Conformationally correct VLPs self-assemble when the genes encoding the Ll and/or L2 capsid proteins are expressed in eukaryotic cells. The VLPs are morphologically indistinguishable from authentic \ irions. 147 countries. the cost of this service in the United States alone is nearly $6 billion annually (28). Furthermore, in developing countries, cervical cancer remains a leading cause of cancer- related death in women. In the future, the development of a therapeutic and/or prophylactic vaccine for cervical cancer may offer an attractive and cost—effective immunologic approach to reduce the need for expensive prevention programs involving screening and surveillance and substantially reduce the world- wide morbidity from this disease. References (I) (2 (3 (4 (5 (7 (8 (9 (I0) (/1) (I2) I48 Bosch FX. Manos MM. Munoz N. Sherman M. Jansen AM, Peto J. et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International Biological Study on Cervical Cancer (IBSCC) Study Group [see comment citation in Medlincl. J Natl Cancer Inst 1995: 87:796-802. de Villiers EM. Heterogeneity of the human papillotnavirus group. J Virol 1989;63:4898—903. Pecoraro G. Morgan 1). Defendi V. Differential effects of human papil- lomavirus type 6. 16. and 18 DNAs on immortalization and transformation of human cervical epithelial cells. Proc Natl Acad Sci U S A 1989;86:563— 7. Barbosa MS. Vass WC. Lowy DR. Schiller JT. In vitro biological ac- tivities of the E6 and E7 genes vary among human papillomaviruses of dif- ferent oncogenic potential. J Virol 1991;65:292-8. Crook T. Morgenstern JP. Crawford L. Banks L. Continued expression of HPV—l6 E7 protein is required for maintenance of the transformed phenotype of cells co—transfortned by HPV—16 plus EJ-ras. EMBO J 1989;825139 Nasiell K, Roger V. Nasiell M. Behavior of mild cervical dysplasia during long—term follow-up. Obstet Gynecol 1986;67:665-9. Maiman M. Fruchter RG. Serur E. Remy JC. Fetter G. Boyce J. Human immunodeficiency virus infection and cervical neoplasia. Gynecol Oncol 1990;38:377-82. Shamanin V. Glover M. Rausch C. Proby C. Leigh 1M. zur Hausen H. et al. Specific types of human papillomavirus found in benign proliferations and carcinomas of the skin in immunosuppressed patients. Cancer Res 1994;54:4610-3. Schiffman MH. Epidemiology of cervical human papillomaviruses. In: zut' Hausen H, editor. Human pathogenic papillomaviruses. Heidelberg: Springer—Verlag. 1994. Schwarz E. Freese UK. Gissmann L. Mayer W. Roggenbuck B. Stretnlau A. et al. Structure and transcription of human papillomavirus sequences in cervical careinotna cells. Nature 1985;314:111—4. Townsend AR. Gotch FM. Davey J. Cytotoxic T cells recognize fragments ofthe influenza nucleoprotein. Cell 1985;42:457-67. Maryanski JL. Pala P. Corradin G. Jordan BR. Cerotttini JC. H—2—t'estricted cytolytic T cells specific for HLA can recognize a synthetic HLA peptide. Nature 1986;324:578—9. (I31 2 L (/5) (/6) (/7) (IX) (/9) (20) (24) (25) (26) R. 3c Feltkamp MC. Smits HL. Vierboom MP. Minnaar RP. de Jongh BM. Drijfhout JW. et al. Vaccination with cytotoxic T lymphocyte epitope~con- taining peptide protects against a tumor induced by human papillomavirus type 16-transformed cells. EurJ lmmunol 1993;23:2242-9. Ressing ME. Sette A. Brandt RM, Ruppert J. Wentwoith PA. Hartman M. et al. Human CTL epitopes encoded by human papillomavirus type 16 E6 and E7 identified through in vivo and in vitro imtnunogenicity studies of HLA-A*()2()I—binding peptides. J Immunol 1995;154:5934-43. Alexander M. Salgaller M. Celis E. Settc A. Barnes W. Rosenberg SA. et al. Generation of tumor specific cytolytic 'I‘-lymphocytes from peripheral blood of cervical cancer patients by in vitro stimulation with a synthetic HPV-16. E7 epitope. Am J Obstet Gynecol. In press. Lee TD. Distribution of HLA antigens in North American caucasians. North American blacks and orientals. In: Lee J. editor. The HLA system: a new approach. New York: Springer-Verlag. 1990:141—78. Keating PJ. Cromme FV. Duggan-Keen M. Snijders PJ. Walbootners JM. Hunter RD. et al. Frequency of down—regulation of individual HLA—A and -B alleles in cervical carcinomas in relation to TAP—l expression. Br J Cancer 1995;72:405-1 1. Hilders CG. Houbicrs JG. Krul EJ. Fleuren GJ. The expression of his- tocompatibility-related leukocyte antigens in the pathway to cervical car— cinoma. Am J Clin Pathol 1994;1(11 15-12. Restifo NP. Marincola FM. Kawakami Y. Tauhcnberger J. Yannelli JR. Rosenberg SA. Loss of functional beta Z-tnicroglobulin in metastatic melanomas from five patients receiving immunotherapy. J Natl Cancer Inst 1996;88:100—8. Stevens CE. Taylor PE. Tong MJ. Toy PT. Vyas GN. Nair PV. et al. Yeast-recombinant hepatitis B vaccine. Efficacy with hepatitis B immune globulin in prevention of perinatal hepatitis B virus transmission. JAMA 1987;257:2612—6. Kirnbauer R. Booy F. Cheng N. Lowy DR. Schiller JT. Papillomavirus LI major capsid protein self-assembles into virus-like particles that are highly immunogenic. Proc Natl Acad Sci U S A 1992;89:12180-4. Kirnbauer R. Taub J. Greenstone H. Roden R. Durst M. Gissmann L. et al. Efficient self-assembly of human papillomavirtts type 16 L1 and Ll—LZ into virus»like particles. J Virol 1993;67:6929—36. Suzich JA. Ghim SJ. Palmer-Hill FJ. White WI. Tamura JK. Bell JA. et al. Systemic immunization with papillotnavirus L1 protein completely prevents the development of viral mucosal papillomas. Proc Natl Acad Sci U S A 1995:92111553-7. Breitburd F. Kirnbauer R. Hubben NL. Nonnenmacher B. Trin-Dinh-Des- marquet C. Orth G. et al. Immunization with viruslike particles from cot- tontail rabbit papillomavirus (CRPV) can protect against experimental CRPV infectionJ Virol 1995;69:3959-63. Brinton LA. Epidemiology of cervical cancer—overview. IARC Sci Publ 199211 19:3—23. Cheng G. Icenogle JP. Kirnbauer R. Hubbert NL. St. Louis ME. Han C. et al. Divergent human papillomavirus type 16 variants are serologically cross»reactive. J Infect Dis 1995:17221584-7. Roden RB. Hubbert NL. Kirnbauer R. Christensen ND. Lowy DR. Schiller JT. Assessment of the serological relatedness of genital human papil- lomaviruses by hemagglutination inhibition. J Virol 1996;70:3298-301. Kurman RJ. Henson DE. Herbst AL. Noller KL. Schiffman MH. Interim guidelines for tnanagetnent of abnormal cervical cytology. The 199?. Na— tional Cancer Institute Workshop. JAMA 1994;271:1866—9. Journal of the National Cancer Institute Monographs No. 21. 1996 New Publication Nllllll the National Ntlllttl Institute .......... NN ............................................................................... Nate N1 N 4th Edition, 1996 International incidence and mortality rates 5-year survival rates Known cancer risk factors Risks for major cancers Call the National Cancer Institute's Cancer Information Service at l-800-4-CANCER (1—800-422-6237). Persons with TTY equipment, dial 1—800-332-8615. Is There a Radiobiologic Basis for Improving the Treatment of Advanced Stage Cervical Cancer? Jacob C. Lindegaard, Jens Overgaard, Soren M. Bentzen. Dorte Pedersen Radiotherapy for the Treatment of Locally Recurrent Cervical Cancer Rachelle Lanciano Surgery for the Treatment of Locally Recurrent Disease Manuel A. Penalver, Giselle Barrcau, Bernd-Uwe Sevin. Hervy E. Averette Chemotherapy for Stage lVB or Recurrent Cancer of the Uterine Cervix George A. Omura Radiation Palliation of Cervical Cancer William J. Spanos. Jr.. Thomas J. Pajak, Bahman Emami. Philip Rubin. Jay S. Cooper. Anthony H. Russell. James D. Cox New Directions for Radiation Biology Research in Cancer of the Uterine Cervix Daniel S. Kapp. Amato J. Giaccia Human Papillomavirus Biology Douglas R. Lowy, John T. Schiller Human Papillomavirus Immunology and Vaccine Prospects Michael A. Steller, John T. Schiller NIH Publication No. 96-4152 US. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service National Institutes of Health [05 If“ M] 145