key: cord-328214-2azb8789
authors: Piper-Vallillo, Andrew J.; Mooradian, Meghan J.; Meador, Catherine B.; Yeap, Beow Y.; Peterson, Jennifer; Sakhi, Mustafa; Do, Andrew; Zubiri, Leyre; Stevens, Sara; Vaughn, Jeanne; Goodwin, Kelly; Gavralidis, Alexander; Willers, Henning; Miller, Adam; Farago, Anna; Piotrowska, Zofia; Lin, Jessica J.; Dagogo-Jack, Ibiayi; Lennes, Inga T.; Sequist, Lecia V.; Temel, Jennifer S.; Heist, Rebecca S.; Digumarthy, Subba; Reynolds, Kerry L.; Gainor, Justin F.
title: COVID-19 Infection in a Lung Cancer Patient Population: Incidence, Presentation and Alternative Diagnostic Considerations
date: 2020-11-12
journal: JTO Clin Res Rep
DOI: 10.1016/j.jtocrr.2020.100124
sha: 
doc_id: 328214
cord_uid: 2azb8789

Introduction Lung cancer is associated with severe COVID-19 infections. Symptom overlap between COVID-19 and lung cancer may complicate diagnostic evaluation. We aimed to investigate the incidence, symptoms, differential diagnosis, and outcomes of COVID-19 in lung cancer patients. Methods To determine an at-risk population for COVID-19, we retrospectively identified lung cancer patients receiving longitudinal care within a single institution in the 12 months (4/1/19 – 3/31/20) immediately preceding the COVID-19 pandemic, including an “active therapy population” treated within the last 60 days of this period. Among patients subsequently referred for COVID-19 testing, we compared symptoms, laboratory/radiographic findings, and outcomes of positive versus negative patients. Results Between 4/1/2019-3/31/2020, 696 patients received longitudinal care, including 406 (58%) in the active therapy population. Among 55 patients referred for COVID-19 testing, 24 (44%) were COVID-19 positive, representing a cumulative incidence of 3.4% (longitudinal population) and 1.5% (active therapy population). Compared to COVID-19-negative patients, COVID-19-positive patients were more likely to have a supplemental oxygen requirement (11% vs. 54%, p=0.005) and to have typical COVID-19 pneumonia imaging findings (5 vs. 56%, p=0.001). Otherwise, there were no significant differences in presenting symptoms. Among COVID-negative patients, alternative etiologies included treatment-related toxicity (26%), atypical pneumonia (22%), and disease progression (22%). Sixteen COVID-19-postive patients (67%) required hospitalization, and 7 (29%) died from COVID-related complications. Conclusions COVID-19 was infrequent in this lung cancer population, but these patients experienced high rates of morbidity and mortality. Oncologists should maintain a low threshold for COVID-19 testing in lung cancer patients presenting with acute symptoms.

Since the onset of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) outbreak in December 2019, 1 evidence has accumulated indicating that cancer patients are particularly susceptible to complications from COVID-19, the infection caused by SARS-CoV-2, with higher rates of intensive care unit (ICU) admission, need for mechanical ventilation, and death. [2] [3] [4] [5] [6] Moreover, patients with a history of cancer or active cancer not on cancer therapy also appear to be at higher risk for severe illness or mortality due to COVID-19 compared to the general population. 4, 7 Patients with lung cancer may be particularly vulnerable to complications from COVID-19. In an initial retrospective analysis of COVID-19 outcomes among 105 cancer patients in Wuhan, China, patients with lung cancer had the second highest mortality rates due to COVID-19, behind only hematological malignancies. 7 More recent studies have confirmed high rates of hospitalization and death within thoracic oncology populations affected by COVID-19. For example, in the Thoracic Cancers International COVID-19 Collaboration (TERAVOLT) registry, 8 which pooled data from 200 patients across 42 institutions and eight countries, 76% of thoracic oncology patients with COVID-19 required hospitalization, and 33% died.

Importantly, recent data from Luo and colleagues suggest that patient-specific factors, such as smoking status and chronic obstructive pulmonary disease (COPD) rather than diseasespecific factors (e.g., prior surgery, systemic therapy) are the major determinants of COVID-19 infection severity among patients with lung cancer. 9, 10 J o u r n a l P r e -p r o o f While registry studies have been valuable in elucidating risk factors and clinical outcomes for severe COVID-19 infection among lung cancer patients, such studies have not captured the incidence of COVID-19 infection within this patient population. Furthermore, given the potential for symptoms of lung cancer and/or toxicities from lung cancer therapies to mimic COVID-19 infection, it is imperative to characterize the frequency of alternative diagnoses among patients with lung cancer presenting with respiratory symptoms during this pandemic. These data may guide diagnostic algorithms and clinical workflow while ensuring prompt treatment initiation for COVID-19 and minimizing transmission.

As of June 1, 2020, 100,805 patients have been diagnosed with COVID-19 in the Commonwealth of Massachusetts, and Boston became an early epicenter in the United States. 11 In this study, we retrospectively reviewed rates of COVID-19 infection among patients with primary lung malignancies treated at Massachusetts General Hospital, a tertiary academic medical center in Boston during the COVID-19 pandemic. To differentiate characteristics of COVID-19 from common complications of lung cancer, we evaluated all urgent clinic encounters from the onset through the first three months of the COVID-19 surge in Massachusetts.

To determine the lung cancer population at risk of COVID-19 in our center, we retrospectively identified all patients with lung cancer receiving longitudinal care in the Center for Thoracic Cancers at the Massachusetts General Hospital (MGH) between April 1, 2019 and March 31, J o u r n a l P r e -p r o o f 2020 using institutional databases and manual chart review. We included in the longitudinal population all patients for whom MGH was the primary provider of lung cancer therapy (systemic therapy, radiation and surgery) during the year prior to the pandemic. The electronic medical record (EMR) was manually reviewed to extract data on demographics, baseline clinical characteristics, prior and current cancer therapies, including surgical resection, radiation therapy (RT) and/or systemic therapy. To ensure adequate clinical follow-up, we excluded patients that had not received cancer therapy at our institution over the study period (e.g., second opinion consultations or patients who ultimately received therapy at an outside institution), those with non-lung primary thoracic malignancies (e.g., thymoma, thymic carcinoma), patients never treated with lung-cancer directed therapy, and those who died prior to March 2020. All clinical, radiographic, and outcome data were compiled under an ongoing institutional review board-approved protocol at our institution with an appropriate waiver of consent.

Within the "longitudinal population" of patients who were receiving cancer therapy at MGH, the "active therapy" population was defined as those patients treated with systemic therapy (chemotherapy, immune checkpoint inhibition, combination chemotherapy plus immune checkpoint inhibition, tyrosine kinase inhibition, or any other anti-neoplastic therapy, including investigational agents) within the 60 days prior to and including March 31, 2020. Treatment modifications employed to reduce the risk of COVID-19 exposure were assessed within the "active therapy" population by manual chart review of encounters between March 1, 2020 to COVID-19 testing location, date(s) and results were extracted. Clinical data (presenting symptoms, need for supplemental oxygen above baseline, laboratory values including complete blood count, creatinine and liver function tests) on the day of testing were captured.

Data pertaining to clinical management were also collected, including need for hospitalization and treatments administered. Among patients who ultimately tested negative for COVID-19, alternative diagnoses were based upon the treating clinicians' assessment. Such cases were also reviewed independently by CM, MM, APV and JG to ensure consensus. All data were collected under an institutional review board (IRB) approved protocol.

J o u r n a l P r e -p r o o f

All cases of COVID-19 infection in this series were confirmed using one of three assays: MGH SARS-CoV-2 assay, 12 a real-time polymerase chain reaction (RT-PCR) test intended for the qualitative detection of nucleic acid from SARS-CoV-2; Cepheid Xpert Xpress SARS-CoV-2 nucleic acid test; or Roche COBAS SARS-CoV-2 nucleic acid test. The MGH SARS-CoV-2 test was employed early in the pandemic, then phased out with FDA-clearance and implenation of the Cepheid and Roche tests in mid-March 2020. Patients referred for preprocedural COVID-19 testing were excluded from this analysis.

Patients with suspected COVID-19 infection underwent radiographic imaging at the discretion of the treating clinicians. Imaging findings from chest plain films and/or computed tomography scan were retrospectively reviewed by a dedicated thoracic radiologist (SD) who was blinded to COVID-19 testing results. Images were graded per Radiological Society of North America guidelines, which designate findings as either negative, indeterminate, atypical or typical for COVID-19 infection. 13 The presence of lung cancer, post-radiation and post-surgical changes were also noted.

Comparisons of presenting symptoms, laboratory and radiographic findings between COVID-19 positive and negative groups were made using Fisher's exact test and Wilcoxon rank-sum test. Imaging findings and oxygen supplementation were dichotomized into binary outcomes and analyzed using Fisher's exact test. Rates of hospitalization and mortality amongst patients J o u r n a l P r e -p r o o f who received any versus no systemic therapy or radiation for lung cancer within 30 days of COVID-19 positivity were compared using Fisher's exact test. All p-values were based on a two-sided hypothesis and computed using Stata 12.1 (StataCorp).

Between April 1, 2019 and March 31, 2020, 1895 patients were seen by medical oncologists in the Center for Thoracic Cancers (Figure 1 ). Within that twelve-month period, 696 patients with lung cancer received at least one form of cancer therapy ("longitudinal population"), including systemic therapy (90%), radiation therapy (30%) or surgical resection (8%; Table 1 ). Among patients in the longitudinal population ever treated with systemic therapy (90%), the most recent systemic therapy was a tyrosine kinase inhibitor (37%), chemotherapy (24%), immune checkpoint inhibitor monotherapy (19%), and chemotherapy plus immune checkpoint inhibitor (11%). In total, 58% (n=406) of patients in the longitudinal population received systemic therapy within 60 days of March 31, 2020 and were defined as the "active therapy" population ( Table 1) .

Fifty-seven patients (14%) in the active therapy population had a documented treatment modification as a result of the COVID-19 pandemic ( Table 1) . The most common treatment modification was a delay of 1 week in the administration of a patient's ongoing systemic therapy regimen, which occurred in 47 (82%) of these patients. Immune checkpoint inhibitor monotherapy was the most frequently delayed treatment type (53%).Most patients who J o u r n a l P r e -p r o o f experienced a treatment delay were receiving palliative systemic therapy (77%) and had disease control (86%) on the most recent surveillance imaging. As of the data cutoff, 33 (70%) had resumed therapy after a median duration of 38 days (interquartile range, "IQR," 35) treatment hold.

Fifty-five patients with lung cancer were referred for evaluation of possible COVID-19 infection due to concerning symptoms or imaging findings (Figure 1) . The most common symptoms within this group were cough (75%), fatigue (51%), dyspnea (45%) and fever (45%; Table 2 ).

testing via nasopharyngeal RT-PCR swab. COVID-19 testing was performed in the MGH Center for Thoracic Cancers outpatient clinic (79%), ambulatory COVID-19 testing centers (14%) or in the emergency department/hospital (7%). Four patients were referred for evaluation but did not undergo COVID-19 testing. Ultimately, these four patients were found to have symptoms attributable to progressive disease (n =1), treatment-related adverse effects (n =2) and in one case an atypical pneumonia with rapid resolution of symptoms after a course of azithromycin, though confirmatory respiratory testing was not performed.

Among the patients tested for COVID-19, twenty-four (47%) were positive by SARS-CoV-2 RT-PCR. Across the longitudinal and active therapy populations, this represents an estimated cumulative incidence of 3.4% and 1.5%, respectively. Eighteen (75%) were diagnosed with COVID-19 by their initial RT-PCR test, while six (25%) initially tested negative and were diagnosed in subsequent testing (median tests for entire cohort, 1; range, 1-4 tests). In those J o u r n a l P r e -p r o o f initially testing negative, the median time from first negative RT-PCR to positive was 10 days (range, 1 -63 days). In two of the six cases, a clear alternative diagnosis was made at the time of initial COVID-19 negative test (liver abscess and post-obstructive pneumonia), which were performed more than 30 days prior to eventual COVID-19 positivity in both cases. By contrast, among the other four patients who were COVID-19 negative on initial testing, subsequent testing was performed and found to be positive within 1 -11 days. For these six cases, symptoms and additional work-up at the time of presentation are presented in Supplementary Table 1 .

Of the 51 patients tested for COVID-19, 27 (53%) were negative by RT-PCR testing. In this COVID-negative cohort, 15 patients (56%) had a single COVID-19 test, 7 (26%) had two tests and 5 (18%) underwent testing three times, all of which were negative. Additional work-up for the COVID-19 negative group included computed tomography (CT) or plain film imaging of the chest (81%) and laboratory evaluation (63%). After medical evaluation, 15 patients (56%) were found to have a definite alternate etiology for their symptoms, and 10 (37%) a likely alternative etiology (Figure 2) . The most common alternative etiologies included treatment-related toxicity (n=7), atypical pneumonia (n=6), and disease progression (n=6).

Among the six patients diagnosed with an atypical pneumonia, two underwent more than one COVID-19 test, two underwent additional viral respiratory testing for common pathogens (influenza, respiratory syncytial virus) and two were empirically treated with azithromycin for potential bronchitis. Notably, the limited testing for respiratory viruses other than SARS-CoV Table 2 . With the exception of a greater need for supplemental oxygen initiation among those with COVID-19 infection (54% versus 11%, p=0.005), there were no significant differences in presenting symptoms or lab abnormalities between COVID-19-positive and COVID-19-negative patients.

Radiographic imaging was obtained in 38 patients (75%), with 45% evaluated by plain film and 55% by CT imaging. Among patients with RT-PCR-confirmed COVID-19 infection ( Table 2) , imaging findings were "typical" for COVID-19 infection 13 in nine (56%), "negative" in four (25%), "indeterminate" in two (12%), and "atypical" in one (6%). RT-PCR-confirmed COVID-19 cases J o u r n a l P r e -p r o o f were significantly more likely than RT-PCR negative cases to have findings typical for COVID-19 pneumonia based on imaging classification (56% versus 5%; p=0.001).

Five patients were referred for COVID-19 testing based primarily on radiographic findings on surveillance imaging, and two were found to be COVID-19 positive. Of these positive cases, retrospective chart review revealed that one patient had low-grade fever and altered mental status at the time of surveillance imaging, and one reported chronic dyspnea as well as new dry cough and diarrhea. In the three patients who tested negative, alternative etiologies for radiographic findings were treatment-related pneumonitis, atypical pneumonia, and etiology unknown (Figure 3) .

COVID-19 positive patients (n = 24) were predominantly current or former smokers (87%), white (84%) and older (median age 75 years , range 57-87) ( Table 3) Sixteen patients (67%) required hospital admission and two patients (8%) received ICU-level care ( Table 4) . During COVID-19 course, supplemental oxygen was required in 54% of cases ( Table 2 ) with maximal level of support being nasal cannula (34%), non-rebreather (13%), BiPAP (4%) or mechanical ventilation (4%). No patients received high flow nasal canula as their maximal level of supplemental oxygen therapy. More than half of patients (58%) received antibiotic therapy (Supplemental Table 2 ). One patient received remdesivir and tocilizumab.

Hydroxychloroquine and corticosteroids were used in 21% and 4%, respectively. No patients received renal replacement therapy or extracorporeal membrane oxygenation.

At the time of hospital admission, six of 16 (38%) patients had a documented code status of do not resuscitate/do not intubate (DNR/DNI). Of the ten (62%) patients who were "full code" on admission, three transitioned to DNR/DNI and four to comfort measures only (CMO). Both patients who received ICU-level care were "full code" on admission, with one patient transitioning to comfort measures only (CMO).

Eight COVID-19-positive patients (33%) died during the study period with seven deaths (29%) attributed to COVID-19 complications, including acute respiratory distress syndrome (n=5) ( Table 4) . One COVID-19-positive patient died from disease progression in the central nervous system. Of the seven patients who died from COVID-19 infection, the median age was 75 years (range 63-87 years). Six patients (85%) had a documented ECOG PS of 3 at last medical oncology assessment prior to COVID-19 infection, and 85% were active or former smokers. Three of seven patients had documented progression of disease on the most recent oncologic staging assessment. Hospitalization rates and mortality were slightly higher among COVID-19-positive patients on active therapy within the preceding 30 days, but this was not statistically significant (Supplemental Table 3 In this study, we retrospectively estimated the incidence of COVID-19 among a thoracic oncology population within a tertiary academic medical center in one of the early epicenters in the United States. We found that roughly 3.4% of lung cancer patients overall and 1.5% receiving active therapy tested positive for COVID-19 within the first three months of the pandemic. Our cumulative incidence of COVID-19 infection is comparable to other, larger series that have reported a prevalence of 1-3% in patients with active or previous malignancy. 2, At our center, based upon earlier reports of increased COVID-19 complications among oncology populations in China 2, 3 and Italy, 15 we proactively pursued strategies to minimize COVID-19 infection risk, including restrictions on all visitors, implementation of a universal face mask policy for providers and patients alike, transition of all non-essential visits to a virtual format, and, in some cases, modifications in systemic therapy. Reassuringly, in our active therapy population, 86% of patients were able to continue active therapy. Furthermore, for those who experienced a delay in systemic therapy, the majority had successfully resumed treatment by the time of data cutoff. As the pandemic continues, guidelines on optimal methods to safely delivery cancer treatment will be crucial. 16 Future studies are also needed to understand the impact of these modifications on cancer-specific outcomes.

In contrast to other recent COVID-19 registry series, our dataset was uniquely poised to highlight the diagnostic challenges faced by thoracic oncologists because we collected data on all patients referred for COVID-19 testing. Overall, we found that 44% of those referred for COVID-19 testing were COVID-19 positive. Importantly, 25% of COVID-19 RT-PCR-positive cases required two or more tests to confirm infection, underscoring the continued diagnostic constraints of RT-PCR testing. Our study also highlights the need for thoracic oncologists to maintain a broad differential diagnosis for patients presenting with acute symptoms, as nearly 50% of patients referred for COVID-19 testing in our cohort were found to have an alternative etiology. These alternative diagnoses included treatment-related complications, progressive disease, atypical pneumonia, pulmonary embolism, congestive heart failure, and COPD flare.

The presenting signs and symptoms of COVID-19 infection among patients with lung cancer were difficult to distinguish from these other etiologies.

While clinical symptoms were largely poor predictors of COVID-19 status in this series, radiographic imaging of the chest provided diagnostic clarity in a subset of cases.

Radiographic changes consistent with lung cancer and prior treatment (radiation and surgery) were common in both COVID-19 positive and negative cohorts, but acute imaging findings "typical" for COVID-19 in the COVID-19 pneumonia classification system were significantly more common in patients with RT-PCR-confirmed COVID-19 compared to negative patients (56% versus 5%, respectively; p=0.001). These findings underscore the diagnostic challenges faced by thoracic oncologists and suggests that clinicians should have a low threshold for COVID-19 testing and radiographic imaging in patients presenting with new respiratory symptoms.

Similar to previously reported data, 3, 6, 8 we found that more than half of our lung cancer patients infected with COVID-19 required hospitalization and nearly one-third died. Patients who died were commonly older than 70, had ECOG PS 3 and had disease progression on recent scans. Findings from larger cohorts, such as the CCC19 consortium, have also indicated age and poor PS are risk factors for mortality/poor outcomes in COVID-19 positive oncology patients. 5 Of note, no specific systemic therapies have to date been linked with increased risk of severe COVID-19 infection in NSCLC. 8, 9 In our study, hospitalization and mortality rates were slightly higher among patients on active therapy within 30 days prior to COVID-19 positivity, but such comparisons are limited by a small sample size (Supplemental Table 3 ).

In our hospitalized cohort, 62% of patients had a documented code status of "full code" on admission. All seven patients who died from COVID-19 (or their appointed health care proxies) elected to de-escalate care during hospitalization, transitioning code status to CMO or DNR/DNI. Whether ICU-level care would have improved clinical outcomes in our cohort remains unknown, but the incidence of elective de-escalation of care indicates that metrics of ICU utilization and rates of mechanical ventilation alone may underestimate the severity of COVID-19 infection in an oncology patient population. These findings emphasize the ongoing importance of critical illness conversations in all patients with advanced malignancies.

Our study has several limitations. First, this was a single institution, retrospective study with a small sample size and limited duration of follow-up. Second, we may have underestimated the true incidence of COVID-19 in our study population due to false negative RT-PCR testing, an inability to capture patients who were diagnosed outside of our hospital system, and excluding patients who underwent testing for COVID-19 as a pre-procedural protocol. It should be noted, however, that our study period was relatively early in the pandemic, and pre-procedural asymptomatic testing hadn't been implemented at the start of our study period. To minimize the number of patients who may have been diagnosed at other institutions, we restricted our study population to those receiving longitudinal care at MGH and also manually reviewed all external hospital records available to us through our EMR system. In defining the "at-risk" population receiving care at our instituion, we did not include patients with active disease who had not received any lung cancer therapy within the preceding year (e.g., patients receiving best supportive care). The impact of COVID-19 on this population remains an important area for future investigation. Another limitation of this analysis is that 22% of patients in the COVID-J o u r n a l P r e -p r o o f unable to perform comprehensive viral testing to determine a definitive organism due to restrictions on viral respiratory testing (except for COVID-19) at our institution to conserve resources during the surge. Finally, our study was conducted prior to FDA-approval of remdesivir and corticosteroids as therapy for COVID-19. While the efficacy of these therapies

has not yet been proven in patients with lung cancer, it is possible that clinical outcomes would have been improved by guideline-directed administration in hospitalized patients.

In summary, COVID-19 infections were identified in a relatively small proportion of lung cancer patients during the initial wave of the pandemic in our institution, but these patients experienced high rates of morbidity and mortality. There is a critical need for ongoing investigation into optimal patient triage strategies, indications for testing, test modalities and ways to mitigate exposure of patients while still ensuring the safe administration of cancerdirected therapies during the COVID-19 pandemic. cases (100%). b. First site listed if patient underwent RT to multiple sites in 12 months; 47 patients underwent RT to multiple sites.

J o u r n a l P r e -p r o o f 

A Novel Coronavirus from Patients with Pneumonia in China

Cancer patients in SARS-CoV-2 infection: a nationwide analysis in China

SARS-CoV-2 Transmission in Patients With Cancer at a Tertiary Care Hospital in Wuhan, China

COVID-19 and Cancer: Lessons From a Pooled Meta-Analysis

Clinical impact of COVID-19 on patients with cancer (CCC19): a cohort study

Case Fatality Rate of Cancer Patients with COVID-19 in a New York Hospital System

Patients with Cancer Appear More Vulnerable to SARS-CoV-2: A Multicenter Study during the COVID-19 Outbreak

COVID-19 in patients with thoracic malignancies (TERAVOLT): first results of an international, registry-based, cohort study. The Lancet Oncology

Impact of PD-1 blockade on severity of COVID-19 in patients with lung cancers

COVID-19 in patients with lung cancer

Massachusetts General Hospital SARS-CoV-2 Test EUA Summary

Radiological Society of North America Expert Consensus Statement on Reporting Chest CT Findings Related to COVID-19. Endorsed by the Society of Thoracic Radiology, the American College of Radiology, and RSNA

Clinical characteristics of COVID-19-infected cancer patients: a retrospective case study in three hospitals within Wuhan

The experience on coronavirus disease 2019 and cancer from an oncology hub institution in Milan, Lombardy Region

KDIGO clinical practice guidelines for acute kidney injury

Patients not on active therapy, n = 290 c