key: cord-313474-1gux1gsi
authors: nan
title: Physicians Abstracts
date: 2015-03-20
journal: Bone Marrow Transplant
DOI: 10.1038/bmt.2015.27
sha: 
doc_id: 313474
cord_uid: 1gux1gsi

nan

Introduction: The incidence of cGvHD after allogeneic SCT is higher when peripheral blood stem cells are used as stem cell source. There is a strong need for preventing cGvHD after ASCT without increasing the risk of relapse. Materials (or patients) and methods: We performed a multicenter, multinational, open-label, randomized study comparing anti-lymphocyte globulin (ATG-Fresenius s ) 10 mg/kg on day -3, -2 and -1 with no ATG in patients with AML (n ¼ 110) or ALL (n ¼ 45) in 1 st complete remission (CR; n ¼ 139) or 2 nd CR (n ¼ 16) who received peripheral blood stem cells from their HLA-identical sibling after standard TBI (12 Gy)/Ccclophosphamide (120 mg/kg) or busulfan (16 mg/ kg)/Cy (120 mg/kg) based myeloablative conditioning regimen. Standard GvHD prophylaxis consisted of cyclosporine A and a short course of MTX (10 mg/m 2 on day þ 1, þ 3, þ 6 and þ 11). The primary study aim was to compare the cumulative incidence of cGvHD at 2 years after ASCT. Results: Out of 161 randomized patients from 27 centers and 4 nations 6 were withdrawn before conditioning and ASCT due to leukemia progression, or cancellation of the donor. 155 patients were analyzed for safety and efficacy; 83 were randomized to ATG and 72 to non-ATG. The treatment groups were comparable regarding recipient and donor age and sex, CMV serostatus, disease (AML vs ALL), 1 st or 2 nd CR. The median time to leukocyte (41.0x10e9/l) and platelet (4 20x 10e9/l) engraftment was significantly delayed in the ATG group (18 vs. 15 days, Po 0.001 and 20 vs 13 days, Po0.001). The incidence of acute GvHD grade I-IV was 25% for the ATG arm and 35% for the non-ATG arm (P ¼ 0.32 ) and for severe grade III/IV acute GvHD 2% and 8%, respectively (P ¼ 0.2). The cumulative incidence of cGvHD at 2 years was 32% (95% CI 22-47%) in the ATG and 69% (95% CI 51-74%) in the non-ATG arm (Po0.0001). Regarding limited and extensive cGvHD the CI at 2 years was 26% vs 53% (P ¼ 0.002) and 19% vs 53% (Po0.001), respectively. There was no difference in infectious complications, CMV and EBV reactivation between both arms. The cumulative incidence of therapy related mortality at 2 years was 14% (95% CI 8-24%) for the ATG arm and 12% (95% CI 6-22%) for the non-ATG arm (P ¼ 0.60), resulting in 2 year relapse-free and overall survival of 59% (95%CI 48-69%) and 74% (95% CI 63-82%) for the ATG group and of 65% (95% CI 51-75%) and 78% (95% CI 66-86%) for the non-ATG group (P ¼ 0.46 and P ¼ 0.21, respectively). Chronic GvHD free survival at 2 years was 50% for the ATG and 23% for the non ATG arm (Po 0.001). A composite endpoint cGvHD and relapse-free survival at 2 years was in favor for the ATG treated patients (30% vs. 17%, P ¼ 0.005). Conclusion: This is the first randomized cGvHD prevention study providing evidence that ATG-Fresenius s (3x 10 mg/kg) is highly effective in preventing limited and extensive cGvHD in HLA-identical sibling PBSC transplantation when used within a myeloablative preparative conditioning regimen. The use of ATG-Fresenius did not result in an obvious increase of infectious complications and relapse, resulting in similar overall survival rates, but improved cGVHD/relapse free survival.

Introduction: The establishment of large transplant registries and introduction of novel statistical techniques have paved the way for large scale data analysis. Nevertheless, contemporary tools for risk prediction of transplant related mortality (TRM) following allogeneic (allo) hematopoietic stem cell transplantation (HCT) are of limited clinical use, owing to a sub-optimal predictive accuracy. Apart from inherent procedural uncertainty, methodological factors impeding prediction might be attributed to the statistical methodology used, number and quality of features collected, or simply the population size. Using an in-silico approach (i.e. iterative computerized simulations), based on machine learning (ML) algorithms, we aimed to define prediction limiting factors of day 100 TRM and rank variable contribution. ML is a field of artificial intelligence dealing with the construction and study of systems that can learn from data, rather than follow explicitly programmed instructions. Commonly applied in complex data scenarios, such as financial settings, it may be suitable for outcome prediction of HCT. Materials (or patients) and methods: Study population was a cohort of 28,236 adult acute leukemia allo-HCT recipients from the EBMT-ALWP. Twenty four variables were analyzed, including recipient, donor and transplant characteristics. Study design involved two phases. The first, focused on development and comparison of several ML based prediction models of day 100 TRM. In the second phase, a repetitive computerized simulation was applied. Factors necessary for optimal prediction were explored: algorithm type, size of data set, number of included variables, and performance in specific subpopulations. Models were assessed and compared on the basis of the area under the receiver operating characteristic curve (AUC). Results: Six ML based prediction models for day 100 TRM were developed on the entire dataset. Optimal AUCs ranged from 0.65-0.68 . Depending on the algorithm used for prediction model development, the in-silico experimental system yielded the following results: Predictive performance plateaued on a population size ranging from n ¼ 5647-8471; A range of 6-12 ranked variables, selected by a separate feature selection algorithm, were necessary for optimal prediction; Disease status and donor type were consistently top ranking variables. Predictive performance of models developed for specific subpopulations, ranged from 0.59 to 0.67 for patients in second complete remission and patients receiving reduced intensity conditioning respectively. Conclusion: We present a novel experimental system for assessment of prediction boundaries in HCT. The present approach has clinical implications. We show that predictive performance of day 100 TRM is unlikely to improve with the data routinely gathered on HCT recipients, as different algorithms reach approximately the same performance. In addition, an exhaustive search for variable importance, reveal that few variables "carry the weight" with regard to predictive influence. Predictive performance converged when sampling more than 5000 patients, reflecting the importance of large registry studies. Overall, it seems we have reached a point of predictive saturation. Improving predictive performance will likely require additional types of input like genetic, biologic and procedural factors. Disclosure of Interest: None declared.

Introduction: Allogeneic stem cell transplantation (SCT) can provide long-term disease control in selected patients with relapsed refractory NHL. Restricted availability of a matched sibling donor limits its use especially for patients with rapidly progressing disease in whom unrelated donor search cannot be awaited. Because data of alternative donor transplants in NHL is sparse, we aimed to compare outcome of haploidentical and cord blood transplants with conventional relatedand matched unrelated donor transplats for NHL. Materials (or patients) and methods: Information of patients with mantle cell lymphoma (MCL), DLBCL, T-cell lymphoma (TCL) and follicular lymphoma (FL) who received an SCT from a sibling donor (SIB), 10/10 matched unrelated donor (MUD), haplo-identical donor (HAPLO) or cord blood (CORD) between 2007 and 2012 was downloaded from the EBMT database. Results: 2798 patients with NHL were identified in the EBMT database meeting the inclusion criteria. 2065 received a transplant from a SIB, 447 from a MUD, 167 from CORD (18 MCL, 36 DLBCL, 43 FL, 70 TCL) and 119 from a HAPLO donor (16 MCL, 30 DLBCL, 22 FL, 51 TCL) . Active disease (Po0.01) and Karnofsky Index (KI) below 80% was also more common in the HAPLO group (P ¼ 0.02). Other variables were balanced. Median follow-up after SCT for all patients was 27 months (CI 25 to 29). Relapse incidence after conventional transplants (SIB, MUD) and alternative donor transplants (HAPLO, CORD) was not significantly different within the whole group (HAPLO: HR 1.2 95% CI 0.9-1.8 P ¼ 0.23; CORD: HR 1.1 95% CI 0.7-1.4, P ¼ 0.74 ) and across all studied disease entities. In contrast, NRM incidence was not significantly different between SIB and MUD, but significantly higher with alternative donor transplants (HAPLO: HR 1.8 95% CI Po0.001 ; CORD: HR 1.9 95% CI Po0.001) . With the exception of FL where MUD in addition to HAPLO and CORD transplants had a significantly higher NRM incidence than SIB transplants. Because patients who received a HAPLO transplant had more commonly active disease at transplant and worse KI, we performed multivariate modeling of relapse-and NRM incidences adjusting for the above mentioned covariates. No different relapse incidences between donor groups was observed. NRM incidence in contrast, was significantly higher in MUD (reference SIB, HR , P ¼ 0.033) and CORD (reference SIB, HR , P ¼ 0.001) but not in HAPLO transplants. Most interestingly, acute GVHD incidence was significantly increased in MUD compared to SIB (P ¼ 0.003) transplants but not in HAPLO (P ¼ 0.08) or CORD (P ¼ 0.34) transplants. Multivariate adjustment for diagnosis (MCL, DLBCL, FL, TCL), remission prior to SCT, KI (KIo90% vs. 490%) and conditioning intensity (RIC vs. MAC) did not reveal worse OS for HAPLO but a worse OS for CORD (HR Po0.0003 Introduction: Retrospective studies in MDS/sAML suggest that reducing the intensity of the conditioning regimen prior to allogeneic stem cell transplantation reduces the risk of nonrelapse mortality but is associated with a higher risk of relapse, but prospective randomized studies for MDS are lacking so far. Materials (or patients) and methods: Within the Chronic Malignancies Working Party (CMWP) of EBMT, we performed a prospective randomized trial comparing a busulfan based (Busulfan 8 mg/kg orally or equivalent dosis intravenously (iv) plus fludarabin 180 mg/m 2 ) reduced intensity conditioning regimen (RIC) and a standard myeloablative busulfan (Busulfan 16 mg/kg orally or equivalent dosis iv plus cyclophosphamide 120 mg/kg) based regimen (MAC) in patients with MDS or sAML (o20% blasts). Between May 2004 and December 2012, 129 patients were included from 18 centers and 7 nations and 127 could be analysed. Major inclusion criteria were: MDS (according to FAB: RA, RARS, RAEB, RAEB-t), CMML, and sAML, blasts less than 20%, matched related or unrelated donor (HLA 8/8, 1 mismatch was allowed), age 18 -60 years (for unrelated) and 18 -65 years (for HLA-identical sibling). Included patients were stratified according related vs unrelated donor and blast countoor4than 5%. The primary endpoint of the study was 1 year non-relapse mortality.The median age of the patients was 51.4 years (r.19-64y) . Donors were HLA-identical sibling (n ¼ 33), matched unrelated (n ¼ 74) and mismatched related or unrelated (n ¼ 20). The patients were well distributed in both arms regarding age, gender, IPSS risk profile, number of blasts at transplantation, donor source and mismatch donor. Results: Graft failure occurred in 1 after MAC and 2 after RIC. Median time to leukocyte (4 1.0 x 10e9/L) occurred after 14days after MAC and 15days after RIC and and platelet (420x10e9/L) engraftment occurred after 14 days after MAC and 15 days after RIC, respectively. Acute GvHD II-IV was noted in 28% after RIC and 31% after MAC. Chronic GvHD was seen in 64% after RIC and 63% after MAC. The cumulative incidence (CI) of non-relapse mortality (NRM) after 1 year was 17% (95% including 12 pediatric/adolescent cases o18 years), 74 with myeloablative conditioning (67%; 34 TBI and 40 non-TBIbased), 64 from HLA-identical siblings (58%), and 9 using ex vivo T-cell depletion. In addition to the above matching criteria, cases and controls were also balanced for other factors such as donor gender and gender mismatch, CMV serostatus, in vivo T-cell depletion and Karnofsky's performance status. Compared to HIV-neg, HIV-pts had lower rates of neutrophil engraftment (92.6% vs 97.5%, P ¼ 0.03), higher incidence of grade III-IV acute GVHD (21% vs 13%, P ¼ 0.05), higher NRM at day 100 (17% vs 11%, P ¼ 0.04) and 2 years (33% vs 23%, P ¼ 0.04), and similar incidence of relapse (29% vs 25%, P ¼ 0.96). Overall, HIV-pts had poorer PFS (39% vs 52%; P ¼ 0.03; HR 1.42 [1.04-1.94] ) and OS (47% vs 59%; P ¼ 0.004; HR 1.60 [1.16-2.22 ]) at 2 years than HIV-neg cases.Outcomes within HIV-pts were comparable for myeloablative vs reduced intensity conditioning and among peripheral blood, bone marrow and cord blood stem cell sources (data not shown). Finally, while HIV-pts' outcomes were comparable for allo-HCT from HLA-identical siblings and alternative donors (OS: 48% vs 40%; P ¼ 0.38) , the use of such alternative donors in HIV-pts was less common than in HIV-neg HCT recipients (42% vs 50%, P ¼ 0.02).

Conclusion: This study showed that the outcome of allo-HCT is poorer in HIV-pts than in the general population, primarily driven by higher NRM, and in keeping with their inferior overall life expectancy despite HAART. Even so, allo-HCT is feasible in HIV-pts with hematologic indications, with a 47% OS at 2 years. In view of the current reduced use of alternative donors despite comparable results to matched sibling donors, HIV-pts requiring an allo-HCT should be granted access to donor search and consideration for transplantation at the same level as HIV-neg counterparts. These data are key to inform allo-HCT strategies in HIV-pts, including its investigational use to eradicate HIV infection. Disclosure of Interest: None declared.

Introduction: HAPLO-HSCT is a therapeutic option for patients with high risk hematologic neoplasms with the advantages of quick availability, easy programation and logistics, and a committed donor. It has shown promissing results in patients diagnosed with relapsed or refractory Hodgkin Lymphoma (HL) at least comparable to allogeneic transplant from siblings or unrelated donors (Burroughs LM et al. Biol Blood Marrow Transplant 2008; 14:1279 -1287 . Materials (or patients) and methods: We retrospectively evaluate the results of HAPLO-HSCT with IV Busulfan (BUX) based RIC regimens (Fludarabine 30 mg/m2 x5 days (-6 to -2), Cyclophosphamide14,5 mg/kg x2 days (-6 to -5) , BUX 3,2 mg/ kg x 1 (BUX1)or 2 days (BUX2) on days -3 to -2) and GvHD prophylaxis based on PTCY (50 mg/kg on days þ 3 and þ 4) and a calcineurin inhibitor plus mycophenolate from day þ 5 performed in GETH centers to patients diagnosed with relapsed or refractory HL. Results: From March-2009, 43 HAPLO-HSCT have been performed in patients diagnosed with relapsed or refractory HL in 11 GETH centers. Median age was 31 years (17-53), 67% were males and all were in advanced phases of their disease, after a median of 4 prior treatment lines (2) (3) (4) (5) (6) (7) (8) . Autologous HSCT was previously employed in 79%, and allogeneic HSCT in 7%. Five patients (11.5%) have received more than 2 prior transplants. Disease status at HAPLO-HSCT evaluated by PET was complete remission in 14 (32%) and persistent disease in 29 (68%). Bone marrow was employed in 11 (26%) and peripheral blood in 32 (74%), without T-cell depletion in all cases. The haploidentical donor was patients mother (20), father (3), siblings (19) or daughter (1) . The RIC regimens employed were BUX1 in 14 (32.5%) and BUX2 in 29 patients (67.5%). Median neutrophils engraftment was day þ 18 (13-44) and platelets 420 K was day þ 26 (13-150). Graft failure with autologous reconstitution happened only in 1 patient (2.5%). The day þ 100 cumulative incidence (CI) of non-relapse mortality (NRM) was 7% (3/43) and 16% (7/43) at 1 year posttransplant. The day þ 100 CI of grade II-IV acute GVHD was 43%, and grade III-IV was 14.5%. Chronic GVHD CI was 26.5% at 1 year, being extensive in 6%. After a median follow-up for survivors of 13 months , the event-free survival (EFS) was 59.5% and overall survival (OS) was 84%. The 1-year CI of relapse or progression was 25%. Factors related with better 1year EFS were CR prior to HAPLO-HSCT (93% vs 45%; P ¼ 0.017) and receiving less than 4 treatment lines prior to HAPLO-HSCT (100% vs 51.5%; P ¼ 0.018). No significant differences were seen when comparing BUX1 against BUX2 in terms of NRM, EFS or OS. Conclusion: HAPLO-HSCT with PTCY and BUX based RIC conditioning in relapsed or refractory HL patients, renders long-lasting remissions with acceptable toxicity and GVHD, obtaining better results in those transplanted in CR and with less than 4 treatment lines prior to HAPLO-HSCT. Disclosure of Interest: None declared.

Introduction: B-lineage acute lymphoblastic leukemia (ALL) is the most common childhood cancer. Although this disease can be successfully treated in 80% of patients, prognosis for primary refractory or relapsed disease is very poor. Even after allogeneic stem cell transplantation (SCT), relapse rates are considerable and correlate significantly with persistent minimal residual disease (MRD) prior to and after SCT. MRD constellations represent favorable effector-target ratios and thus might be optimally suited for immunotherapeutic intervention with therapeutic antibodies. Materials (or patients) and methods: We developed an Fcoptimized CD19 antibody (4G7SDIE) and produced it in pharmaceutical quality. 4G7SDIE mediates markedly enhanced antibody-dependent cellular cytotoxicity (ADCC) through its improved capability to recruit FcgRIIIa bearing effector cells including NK cells and gd T cells. 4G7SDIE was applied on compassionate use to MRD-positive pediatric patients with relapsed or refractory ALL (CR1 n ¼ 3, ZCR2, n ¼ 11). Results: Side effects such as headache and fever were negligible. In all patients complete CD20 þ B-cell depletion was observed during therapy. After discontinuation of 4G7SDIE therapy B cell counts recovered rapidly to normal levels. In 9/14 patients MRD was reduced byZ1 log or fell below MRD-detection threshold of 10 -4 over the course of treatment. 2/9 responders were receiving additional treatment. 6 patients relapsed, 1 patient died of CNS chemotherapy associated toxicity and 1 patient died of late posttransplant sepsis. 6 patients have been in sustained remission for 264-1115 days (median follow-up 720 days). In initial cytotoxicity screenings, performed in an allogeneic setting, significantly increased lysis of ALL blasts by PBMC after adding 4G7SDIE was observed. NK cells and gd T cells were identified as main effector cell populations. CD19 expression on patient blasts was confirmed by quantitative flow cytometry (mean 1.71x10 4 molecules/ cell, ± 0.54x10 4 ). Cytotoxicity assays using patient PBMC on autologous blasts confirmed sustained functionality of patient effector cells over the course of 4G7SDIE treatment. Cytotoxicity assays were performed using PBMC from transplanted patients obtained at different time points of 4G7SDIE treatment. Lysis of autologous ALL blasts was increased when 4G7SDIE or autologous patient serum taken after antibody infusion was added. After infusion of 20 mg/m 2 -40 mg/m 2 4G7SDIE serum half-life was 20 h -43 h. Serum levels of 4G7SDIE remained above saturating concentrations ofZ700 ng/ml (EC 50 ¼ 65 ng/ml) until the following application in the bi-weekly treatment cycle. Notably, in 2/2 analyzed patients under 4G7SDIE therapy, a down-modulation of CD19 surface expression on the leukemic blasts was observed. In vitro antigenic shift assays on patient blasts showed considerable but very heterogeneous shift of CD19 surface expression. Furthermore, a positive correlation between CD19 surface expression levels and 4G7SDIE mediated lysis was observed. These observations hint at in vivo tumor escape mechanisms and moreover indicate selective pressure exerted by immunotherapy with 4G7SDIE, underlining its therapeutic potential, but also delineating possible limitations. Conclusion: Promising anti-leukemic effects of the 4G7SDIE antibody have been observed in vitro and in vivo. We are currently preparing a clinical phase I/II trial. Disclosure of Interest: None declared. Introduction: Allogeneic HCT with myeloablative conditioning is considered a standard of care for adults with high risk acute lymphoblastic leukemia (ALL). However, with improving results of conventional-dose chemotherapy and the introduction of novel agents on one hand, and the improvement in transplantation techniques on the other, the indications for alloHCT require re-evaluation, taking into account patient-and procedure-related factors associated with the risk of nonrelapse mortality (NRM). The aim of this study was to analyze the results of myeloablative alloHCT treatment for patients with ALL according to recipient age and donor type. Materials (or patients) and methods: 4859 patients treated with alloHCT in first complete remission during the period 1993-2012 were included. The outcomes were analyzed for the periods 1993-2002, 2003-2007 and 2008-2012 , in various age groups, separately for HLA matched sibling (MSD, n ¼ 2681) and unrelated donors (URD, n ¼ 2178). Results: For MSD-alloHCT recipients treated during the period 2008-2012, the following two-year probabilities of OS were obtained: 76% for the 18-25 years old (y.o.) group, 69% for both the 26-35 y.o. and 36-45 y.o. groups and 60% for the 46-55 y.o. group. The incidence rates of NRM were 12%, 11%, 15% and 24%, respectively for those same age groups. In comparison with the 1993-2007 period, significant improvements were observed for all age groups except for the 26-35 y.o. patients. The improved survival rates were a consequence of reduced NRM and a tendency towards a reduced risk of relapse. Among URD-alloHCT recipients, the OS was 66% (18-25 y.o.) , 70% (26-35 y.o.) , 61% (36-45 y.o.) , and 62% (46-55 y.o.) , while the respective incidence rates of NRM were 21%, 20%, 21% and 19%, The improvement of OS over time was documented for 36-45 y.o. (P ¼ 0.005) and 46-55 y.o. (0.0007 ) patients due to the reduced incidence of NRM with no significant effect on relapse. In a multivariate analysis adjusted for disease-, patient-, donorand procedure-related factors, transplantations performed for the period 2008-2012, when compared to 1993-2007 , were associated with significantly reduced risks of the overall mortality (HR ¼ 0.72, P ¼ 0.0003), treatment failure (either relapse or NRM, HR ¼ 0.77, P ¼ 0.002), and NRM (HR ¼ 0.73, P ¼ 0.01) and showed a trend towards reduced risk of relapse (HR ¼ 0.81, P ¼ 0.06). The overall mortality was reduced for transplants with TBI-based compared to chemotherapy-based conditioning (HR ¼ 0.71, P ¼ 0.005) as a result of reduced risk of relapse (HR ¼ 0.55, P ¼ 0.00004). Type of donor (MSD vs. MUD) had no significant effect on survival (HR ¼ 0.9, P ¼ 0.24). Conclusion: Results of alloHCT for adults with ALL improved significantly over time in all age groups, mainly due to the reduction of NRM. Importantly, results obtained with matched unrelated transplants were comparable to sibling transplants. Total body irradiation should still be considered as the preferable type of myeloablative conditioning for ALL. Disclosure of Interest: None declared.

CMV seronegativity is associated with a 20% decrease in 1-year survival in patients undergoing reduced intensity sibling-donor transplantation for treatment of myeloid malignancy D. J. Lewis 1,* , C. Holmes 1 , K. Peggs 2 , A. Peniket 3 , M. Nikolousis 4 , S. Nagra 5 , G. Pratt 1,4 , C. Craddock 1,5 , R. Malladi 1,5 , P. Moss 1, 5 1 School of Cancer Sciences, University of Birmingham, Birmingham, 2 University College Hospital, London, 3 Oxford University Hospital, Oxford, 4 Birmingham Heartlands Hospital, 5 Centre for Clinical Haematology, University Hospital Birmingham, Birmingham, United Kingdom Introduction: Reduced intensity (RIC) allogeneic stem cell transplantation is a highly effective treatment for acute myeloid leukaemia and the immunological 'graft versus leukaemia' effect is believed to be a major mechanism of disease control. Cytomegalovirus reactivation has been suggested to reduce the rate of disease relapse in acute myeloid leukaemia (AML). We investigated the influence of CMV serostatus on the clinical outcome of patients who underwent T cell depleted RIC allografts for myeloid malignancy, and went on to examine reconstitution of lymphoid subsets. Materials (or patients) and methods: We studied patients who underwent RIC allografts for AML (n ¼ 272) and MDS (n ¼ 82) from four UK centres, with fludarabine and melphalan conditioning þ /-alemtuzumab. Overall survival was calculated. Relapse rate and non-relapse mortality were calculated with competing risk analyses. Results: The median overall survival was 2.17 years. The relapse rate of the entire cohort was 23% at 1 year with a nonrelapse mortality of 22%. The overall survival for the 'CMV at risk' group was 64.7% at 1 year compared to 57.9% for CMV (-/-). This difference was most marked in patients transplanted from sibling donors (n ¼ 140); the overall survival at 1 year was 79% in CMV-at-risk patients (n ¼ 99) compared to only 59% in the CMV seronegative group (n ¼ 41) (P ¼ 0.027). This 20% increment in survival was due to a 37% reduction in the rate of disease relapse in patients that were CMV-at-risk (1 year relapse rate of 21% versus 33%). There was a non-significant trend towards improved overall survival in those that experienced CMV reactivation amongst the CMV-seropositive patients (1 year OS 83% versus 61%, P ¼ 0.19) , mainly due to a reduction in the rate of relapse (1 year relapse rate 17% versus 28%). Because of this large difference in relapse risk, we went on to examine the effects of CMV serostatus and alemtuzumab use on reconstitution of lymphocyte subsets at 3 months post transplant. Alemtuzumab led to 5-fold decrease in the T cell count at 3 months compared to transplants in which T cell depletion was not used. However, within this alemtuzumab group, positive CMV serostatus in the donor or recipient was associated with a relative 7 fold and 35 fold increase in the CD3 þ and CD8 þ T cell count compared to CMV seronegative pairs. Indeed, CD8 þ T cells were virtually undetectable at this time point in CMV seronegative transplant/donor pairs. Conclusion: CMV seropositivity is markedly beneficial for patients who undergo a sibling donor reduced intensity allograft for myeloid malignancy and in whom alemtuzumab is used for conditioning. This effect is most likely to be due to the profound influence of chronic viral replication on boosting T cell immune reconstitution in the early post transplant period. CMV seronegative patients with AML should be considered at risk of impaired survival for certain subgroups of stem cell transplant. Disclosure of Interest: None declared. Introduction: Allogeneic stem cell transplantation is the only curative option for patients with high risk acute myeloid leukemia (AML) and for those experiencing relapse. Either matched sibling donor (MSD) or unrelated donor (UD) is indicated. Materials (or patients) and methods: With the aim to compare the outcomes of both strategies we have retrospectively analysed 1554 adults with AML receiving either MSD (n ¼ 961) or UD (n ¼ 593) in EBMT centers from 2000-2012. For UD, 481 were 10/10 HLA matched and 112 were 9/10. Median follow up was 28 (range 3-157) months. There were statistical differences between the 2 groups. Compared to MSD recipients, UD transplants were older (49 vs 52 years, P ¼ 0.001), were performed more recently (2009 vs 2006, P ¼ 0.001) , had longer interval between diagnosis to transplant (10 vs 9 months, P ¼ 0.001), had more often secondary AML (13% vs 19%, P ¼ 0.002) and were transplanted with higher proportion of CMV negative donors (38% vs 37%, P ¼ 0.001). Peripheral blood stem cells (PBSC) was used as graft source in 90% of patients in both groups, P ¼ 0.14. Conditioning regimen was more frequently myeloablative for patients transplanted with a MSD (61% vs 46%, P ¼ 0.001). MSD received more often busulfan and cyclophosphamide as MAC (16%) or a TBI based regimen (18%). For UD, Bu-fludarabine was the most frequent conditioning used (19%). Results: Cumulative incidence (CI) of neutrophil engraftment was similar (93% vs 92% for MSD vs UD, respectively, P ¼ 0.07).

Grade II-IV acute GVHD was 26% vs 30% (P ¼ 0.11 ) and CI of chronic GVHD was 25% vs 24% (P ¼ 0.9) for MSD and UD, respectively. For MSD and UD respectively, CI of relapse (RI) was 57% vs 49%, P ¼ 0.001; CI of non-relapse mortality (NRM) was 23% vs 24%, P ¼ 0.24. Probability of leukemia-free survival (LFS) at 2 years was 21% vs 26%, P ¼ 0.001, and overall survival (OS) was 26% vs 33% P ¼ 0.004, respectively. Chronic GVHD as time-dependent variable was associated with lower RI (HR 0.78 , P ¼ 0.05), higher NRM (HR 1.71 , P ¼ 0.001), and higher OS (HR 0.69, P ¼ 0.001). According to HLA-match for MSD, 10/10 UD and 9/10 UD, CI of relapse (RI) was 57% vs 50% vs 45%%, P ¼ 0.001; CI NRM was 23% vs 23% vs 29%%, P ¼ 0.26 and probability of LFS at 2 years was 21% vs 27% vs 25%, P ¼ 0.003, respectively. In a multivariate analysis adjusted for differences between the 2 groups, UD was associated with lower RI (HR 0.76, P ¼ 0.001) and higher LFS (HR 0.83, P ¼ 0.001) compared to MSD. When analyzing according to HLA-match, there was no differences for patients transplanted with an UD 9/10 or a 10/10 for RI (HR 0.77, P ¼ 0.10) , and LFS (HR 0.92, P ¼ 0.53) . The other factors independently associated with better outcomes were the interval between diagnosis and transplant (RI HR 0.62, Po0.001) , and LFS (HR 0.67, Po0.001) . Conclusion: Unrelated donor transplant was associated with better LFS due to lower RI compared to MSD for high-risk patients with AML transplanted in first relapse. There were not differences for the UD 9/10 match probably due the graftversus-leukemia effect in the setting of patients transplanted with active disease. Disclosure of Interest: None declared. Introduction: Allogeneic reduced intensity transplantations (RICT) in elderly patients (pts) with AML in CR1 has become a commonly used treatment modality. However, no prospective studies to date have been reported to support use of this strategy. The aim of this prospective, multicenter study is to compare outcomes of patients receiving RICT with patients being treated with conventional chemotherapy. In 2012 an amendment allowed also the use of matched unrelated donors. The study is currently ongoing in 8 countries with a total of 250 pts included. Only pts included as per the original protocol are accounted for in this interim analysis which focuses on safety. Materials (or patients) and methods: The study was designed by the Transatlantic Leukemia Group (TRALG) consisting of centers associated with the Swedish and Canadian BMT Groups, and centers from Germany, Norway, Finland and New Zealand. Patients should have intermediate or high risk AML in CR1 and not being eligible for a myeloablative transplant due to age or comorbidities, and have at least one potential sibling donor. Date of inclusion was defined as the date of HLA-typing of the first sibling, and pts were allocated to the donor (D) or no-donor group (noD) based on HLA match. Overall survival at 3 yrs is the primary endpoint; secondary endpoints include RFS, GvHD, non-relapse (NRM) and transplant related (TRM) mortality. Conditioning consisted in the vast majority of patients of fludarabine 150 mg/m 2 IV and busulphane 8 mg/ kg PO or 6.4 mg/kg IV. GvHD prophylaxis was CyA/Methotrexate or CyA/MMF. The study started to accrue patients in 2004. Results were analyzed on an intent to treat basis. Results: From 2004 pts were included in Sweden (n ¼ 62), Canada (n ¼ 56), Germany (5) , Norway (13), New Zealand (5) and Finland (5) . 12 patients were excluded from analysis; 3 favorable risk, 9 protocol violations and 134 pts; 60 females, 74 males, median age 62 (51-74) yrs, 21 sAML, were allocated to the D group (n ¼ 75), or the noD group (n ¼ 59). In the D group, 16 pts (21%) did not reach transplant, 15 of these died (12 after relapse, 3 other causes). Cytogenetic risk groups were categorized as per ELN criteria. The distribution of total disease risk categories was 25% adverse, 66% intermediate, 9% unknown. Median follow-up of surviving pts at Dec 6 th 2014 was 4.6 (2.6 -10.1 ) yrs after inclusion. Ninety-one pts relapsed (75%). Kaplan-Meier estimates of OS and RFS in the whole group at 3 yrs were 45% and 37% respectively. For adverse and intermediate risk pts OS and RFS at 3 yrs were 36% and 21%, and 51% and 44%, respectively. In the transplanted pts acute GvHD grade 0-1 occurred in 73% of pts, grade 2-4 in 20%. Extensive GvHD occurred in 42%, limited in 15%. 89 pts (66%) have died, 17 without and 72 after AML relapse. In the noD group 5 (8%) pts died from non-relapse causes. In the D group, 3 pts died before transplant, while the TRM was 9/59 (15%) with causes of death: 6 GvHD, 3 other. Conclusion: Selected patients with AML in CR1 tolerate RICT or standard management with low mortality. Disease control remains a major issue. This multicenter prospective protocol will continue to accrue patients until relevant conclusions can be drawn comparing a RICT to standard treatment. The current rapid inclusion of pts and participation of new sites in Australia, Greece and Estonia will help to complete the study. Disclosure of Interest: None declared. Introduction: The Wilms' tumor gene 1 (WT1) is overexpressed in 480% of acute myeloid leukemias (AML) and myelodysplastic syndromes (MDS) and proved to be a good marker for minimal residual disease (MRD) monitoring.

Although allogeneic haematopoietic stem cell transplantation (allo-HSCT) is the most effective treatment for AML/MDS, disease recurrence remains a major problem. After allo-HSCT, quantitative WT1 monitoring can represent a useful tool to detect MRD and tailor immunotherapeutic strategies. Materials (or patients) and methods: We included in this retrospective analysis 111 pts with AML/MDS (99 and 12 respectively) overexpressing WT1 and allotransplanted in our Institution from 12/2007 to 01/2014. Twenty two pts were transplanted from a matched related donor, 26 from a matched unrelated donor, 57 from a mismatched related donor and 6 from cord blood units. 101/111 pts received a reduced toxicity conditioning treosulfan-based. At the time of transplant, 73/111 pts (66%) were in CR, while 38/111 (34%) had active disease. Median follow up (FU) of our cohort was S10 599 days (range, 55-2538 days). WT1 MRD monitoring was performed by RQ-PCR and considered positive for values4250 copy numbers of WT1 per 10e4 ABL [1] . After allo-HSCT, detection of positive WT1 was followed by immunomodulatory therapeutic interventions according to the time from transplant, the presence of active graft-versus-host disease (GvHD) and the general clinical conditions: tapering and/or discontinuation of immunosuppressive drugs (IS), donor lymphocytes infusions (DLI), administration of hypomethylating agents. Median time to disease relapse was calculated from the time of detection of WT1 value above the threshold.

Results: At day 30 post allo-HSCT, 109 out of 111 pts (98%) were in CR. Forty-five out of 109 (41%) CR pts had WT1 levels persistently negative during follow up (FU) and 23/45 remained in CR at the last FU. Sixty-four out of 109 pts (59%) had at least one increase of BM WT1 levels above the threshold during observation and were evaluated for preemptive treatment. In 45/64 (70%) IS was tapered until suspension and/or DLI þ /-azacytidine was administered.

Median time from first WT1 positive value to treatment was 30 days (range, 15-45 days). In 16 out of these 45 pts (35%) WT1 level normalized, whereas 29 pts (65%) progressed to overt disease. All the 16 that normalized WT1 remained in CR. No grade III or IV acute GvHD was observed, while severe chronic GvHD occurred in 2/45 pts (4%). Median time to disease relapse for the 45 treated pts was 116 days (range, 20-951 days). In 11/64 pts (17%) with a post allo-HSCT WT1 positivity, the presence of an active form of acute or chronic GvHD prevented from applying further immunotherapy strategies. Among these pts, 5 remained in CR, 6 relapsed. Median time to disease recurrence for these 11 pts with GVHD was 156 days (range, 32-204 days). Finally in 8/64 pts (13%) concurrent clinical issues (e.g. active infection) did not allow any attempt to prevent relapse and 4/8 relapsed. Median time to disease relapse for untreated pts was 35 days (range, 15-82 days).

Conclusion: In our series, pre-emptive immune-modulatory maneuvers targeting WT1 levels proved to be feasible and safe and of potential clinical benefit to postpone overt relapse in high risk AML/MDS pts post allo-HSCT. Introduction: Allogeneic stem cell transplantation (SCT) with both myeloablative (MAC) and reduced intensity conditioning (RIC) is effective therapy in AML. Several studies have shown that leukemia-free survival (LFS) is similar after SCT with MAC and RIC. However, there is paucity of data on the long term outcome (beyond 10 years) following RIC due to the relative recent introduction of this approach. The ALWP of EBMT published in Leukemia 2005 the largest study until that time, comparing the outcome of AML patients (pts) given RIC (n ¼ 315) and MAC (n ¼ 407). The median follow-up was 13 months. In multivariate analysis, non-relapse mortality (NRM) was significantly lower and relapse rate was significantly higher after RIC resulting in similar LFS. Materials (or patients) and methods: In order to better predict long-term outcome and late events we have now updated SCT outcomes in a larger cohort of pts, age Z50 years (n ¼ 1423) after SCT from matched sibling donors, in the years 1997-2005 with a median follow up of 8.3 years (0.1-17) . Results: 722 pts were given RIC and 701 MAC regimens. The median age at SCT was 57 (50-75) and 54 (50-72) years, respectively (Po0.001). 21% of RIC recipients had advanced disease at SCT compared to 25% of MAC recipients. The percentage of pts in CR1 and CR2/ later CR was 62% and 17% compared to 63% and 12%, respectively (P ¼ 0.02). RIC recipients were more likely to receive PBSC (92% Vs 73%, Po0.001 ) and in vivo T-cell depletion (33% Vs 12%, Po0.001). 16% and 20% had poor-risk cytogenetics, respectively (P ¼ 0.19 1.4, Po0.001 ) and poor cytogenetics (HR 1.7, Po0.005) . The conditioning regimen did not predict 10-year LFS. NRM rates were higher after MAC than RIC throughout the late post SCT course, while relapse rates were only mildly decreased at the late phases. In pts surviving leukemia-free 2 years after SCT the subsequent NRM was 15% and 9%, respectively (P ¼ 0.03). Subsequent relapse rates were 14% and 9% (P ¼ 0.12) and LFS was 71% and 73%, respectively (P ¼ 0.76). In pts surviving leukemia-free 5 years after SCT, subsequent NRM was 9% and 4%, respectively (P ¼ 0.06). Subsequent relapse rates were 5% and 6% (P ¼ 0.53), and LFS was 86% and 90%, respectively (P ¼ 0.27). Conclusion: LFS remains similar after RIC and MAC even 10 years after SCT. The trend for excess NRM with MAC remains throughout the late course, however excess relapse after RIC is more obvious in the early phases. Pts remaining leukemia-free 5 years after SCT can expect excellent subsequent outcome with both regimens. Long-term follow-up studies (beyond 10 years) are of significant importance when assessing SCT outcomes. Disclosure of Interest: None declared.

myeloablative Busulfan based strategies in transplantation particularly of children with non-malignant diseases. A reduction in short term mucosal and hepatic (SOS) toxicity and the absence of long term pulmonary toxicity have been demonstrated. It is unclear, however, whether this reduction of toxicity is accompanied by an equal or inferior myeloablative capacity compared to Busulfan based myeloablative regimens. Materials (or patients) and methods: We performed a retrospective analysis of consecutive patients with nonmalignant diseases transplanted in 4 German pediatric transplant centers (Hamburg, Hannover, Munich and Ulm) with a Treo based regimen (Treo and Fludarabin (Flu) ± Thiotepa (TT)±Serotherapy) in the period between January 1st 2000 and June 30th 2013. Results: We identified 153 patients with inborn errors including primary immunodeficiencies, hemoglobinopathies, hemophagocytic lymphohistiocytosis, mucopolysaccharidosis and osteopetrosis. The median age at transplantation was 4.8 years (0.1-22) . All pts received Treo/Flu. TT was added in 102 of these pts, serotherapy in 139 pts. The OS after 2 years was 89%, the EFS 81%. The incidence of aGvHD 1II-1IV was 19% and 5.4% for cGvHD 11-13. Primary engraftment of donor cells was present in 97% of patients. However, mixed chimerism at any time point was found in 59% and disease recurrence in 11%. An additional cellular therapy (ACT: stem cell boost, DLI, or 2nd transplant) was applied in 28% of patients. ACT was more often needed after transplantations from MMFDs and MUDs than from MSD/MFDs (P ¼ 0.08). There was a significant difference in patients who received Alemtuzumab as serotherapy early (d-13 to d-4) versus late (d-3 or later) during conditioning, with 32% ACT in the early and 52% ACT in the latter group (P ¼ 0.02). Cell source (BM or PBSC) and addition of TT did not affect the cumulative incidence (CI) of ACT. Conclusion: There is an excellent overall survival and EFS with Treo based conditioning in the transplantation of patients with non-malignant diseases. Despite a good primary engraftment, a high rate of mixed chimerism, disease recurrence and need for additional cellular therapy was observed. Interestingly, early administration of serotherapy was correlated with a higher probability of stable donor engraftment and has to be considered as a relevant factor for transplant outcome. A randomized controlled prospective study comparing conditioning regimens with Treo against Busulfan in nonmalignant diseases is needed. Disclosure of Interest: None declared.

Long-term outcomes of reduced-intensity conditioning allogeneic stem cell transplantation for adult high-risk acute lymphoblastic leukemia in first complete remission S. Lee 1,* , K. Introduction: Reduced-intensity conditioning (RIC) allogeneic stem cell transplantation (SCT) has emerged as an option designed to lower nonrelapse mortality (NRM) for older patients and those with comorbid conditions. However, the role of RIC-SCT in adult acute lymphoblastic leukemia (ALL) remains unclear because the interpretation of transplantation outcome is mainly limited by the small sample size, short follow-up duration, various regimens for conditioning and graft-versus-host disease (GVHD) prophylaxis, and the heterogeneity of the criteria used to select patients for RIC-SCT. Previously, we conducted a phase 2 trial of RIC-SCT in adults with high-risk ALL and showed the potential role of this strategy, especially in patients in first complete remission (CR1). Here, we report on the updated results of RIC-SCT by analyzing 93 consecutive adult high-risk ALL transplants in CR1. Materials (or patients) and methods: During the period between 2000 and 2012, 93 consecutive patients in CR1 (median age, 51 years [range, 15-65 years]) were given an identical RIC regimen consisting of fludarabine (150 mg/m 2 in total) and melphalan (140 mg/m 2 in total). The indications for RIC-SCT were advanced age (Z50 years; n ¼ 53; 57.0%) and comorbid conditions (n ¼ 40; 43.0%). Graft sources were peripheral blood stem cells (n ¼ 91; 53 8/8-matched sibling donor, 15 8/8-matched unrelated donor, 23 7/8-matched unrelated donor) and bone marrow (n ¼ 2; 1 8/8-matched unrelated donor, 1 7/8-matched unrelated donor). The median time-to-transplantation was 161 days (range, 106-291 days). GVHD prophylaxis was attempted by administering calcineurin inhibitors (cyclosporine for sibling donor transplants, tacrolimus for unrelated donor transplants) plus methotrexate. Antithymocyte globulin (2.5 mg/kg in total) was administered to the patients who received allele-mismatched unrelated donor grafts. If residual leukemia was detected in the absence of GVHD at 3 months after transplantation, calcineurin inhibitors were rapidly discontinued. Results: Fifty-two patients developed grade II to IV acute GVHD (42 grade II, 6 grade III, 4 grade IV). The cumulative incidence of acute GVHD at 100 days was 55.9%. Of the 87 patients who survived for at least 100 days with sustained engraftment after transplantation, 61 developed chronic GVHD (30 limited, 31 extensive), resulting in a 5-year cumulative incidence of 65. 6% . After a median follow-up of 60 months (range, 24-174 months), the cumulative incidence of relapse (CIR) and NRM at 5 years were 28.1% and 22.9%, respectively, and the 5-year disease-free survival (DFS) and overall survival (OS) rates were 54.4% and 60.4%, respectively. Within the cohort of Ph-negative ALL transplants (n ¼ 43), the 5-year CIR, NRM, DFS, and OS rates were 26.4%, 15.5%, 61.7%, and 67.4%, respectively. In a subgroup of Ph-positive ALL transplants (n ¼ 50), the 5-year CIR, NRM, DFS, and OS rates were 29.7%, 28.9%, 48.6%, and 55.0% respectively, and for these patients, minimal residual disease kinetics (early-stable molecular response vs late molecular response vs poor molecular response) and the presence of chronic GVHD were closely related to CIR and DFS. Conclusion: Our data suggest that RIC can be considered as a reasonable choice for providing a long-term disease control for adult high-risk ALL patients in CR1. Disclosure of Interest: None declared.

Substitution of TBI by intravenous Busulfan for elderly AML/MDS patients within the FLAMSA-RIC protocol is feasible and yields comparable results M. Schleuning 1,* , D. Judith 1 , I. Burlakova 1 , H. Baurmann 1 , R. Schwerdtfeger 1 , G. Stuhler 1 1 Centre for hematopoietic cell transplantation, DKD Helios Klinik, Wiesbaden, Germany Introduction: The FLAMSA-RIC protocol is a highly effective conditioning protocol for high risk myeloid leukaemia patients (pts). However, many of these pts are beyond the age of 60. In this population the use of total body irradiation (TBI) might be toxic, even with the reduced dose of 400 cGy, as described in the original FLAMSA-RIC protocol. In an attempt to further reduce toxicity for elderly (460y) or comorbid pts we substituted TBI by intravenous Busulfan (ivBu; 8 x 0.8 mg/kg BW) within the FLAMSA-RIC protocol. Materials (or patients) and methods: Retrospective study to analyze the results of ivBu in comparison to those achieved in pts receiving the classical FLAMSA-RIC protocol with TBI during the same time period. Results: From November 2006 to October 2012 173 pts with high-risk AML or MDS received an allogeneic stem cell transplant after FLAMSA-RIC conditioning. Eighty-three pts (median age 47y) received TBI and ninety pts (median age S12 64y) received ivBu. In the TBI group 76 pts suffered from AML and 7 pts from MDS and in the ivBu group diagnoses were AML in 74 and MDS in 16 pts. Unfavourable cytogenetics or molecular genetics were found in 64% of TBI and 50% of ivBu pts, respectively. In the TBI group 43% were transplanted with active disease as compared to 71% in the ivBu group. In both groups stem cell grafts from unrelated donors were used in approximately 75% of pts. All pts engrafted. After a median follow-up of 5.5 y for surviving pts the probability of leukaemia-free survival at 5 y after transplant is 40% in the TBI group and 35% in the ivBu group (P ¼ 0.133). Twentyseven relapses occurred in the TBI and 21 in the ivBu group. Non-relapse mortality (NRM) for the TBI cohort was 10% and for the ivBu group 24% at day þ 100 (P ¼ 0.052) and 33% and 41% during the entire observation period (P ¼ 0.004). The difference in NRM is probably owing to the older age of the pts in the ivBu group and was due to more infectious complications. No significant difference in survival was observed in MUD or sibling transplants in either group. Conclusion: In conclusion, substitution of TBI by ivBu is feasible with no enhanced relapse rates observed and should be further evaluated in prospective clinical trials also for younger pts. Introduction: To ascertain the therapeutic potential of non-TBI-based conditioning for CD34 þ HPC-selected, T celldepleted allografts, we conducted a trial comparing our standard regimen, arm (A) 1375 cGy HFTBI þ thiotepa,5 mg/ kg/day x 2 days þ cyclophosphamide 60 mg/kg/day x 2 days vs. arm (B) Busulfex 0.8 mg/kg/6 h x12 (dose adjusted) þ melphalan 70 mg/kg/day x 2 þ fludarabine 25 mg/m 2 /day x5 and arm (C) Clofarabine 20 mg/m 2 /day x 5 þ melphalan 70 mg/m 2 /day x 2 þ thiotepa 5 mg/kg/day x2, as preparation for T-cell depleted CD34 þ PBSC transplants from GCSFmobilized leukocytes fractionated with the CliniMACS CD34 þ reagent system. Materials (or patients) and methods: Primary endpoints were engraftment, GVHD, transplant-related mortality (TRM) and 2 yr OS and DFS (Confer Table) . Stratification of pts to arms A (standard), B or C was based on the patient's disease, disease stage and clinical factors such as age, prior therapy or comorbidities enhancing risks of TBI. Arm B was the non-TBI arm predominantly used for myeloid and Arm C for lymphoid malignancies. Prior to transplant, recipients of HLA-matched or non-identical transplants received rabbit thymoglobulin at 2.5 mg/kg/day x2 or 3 days respectively, to prevent graft failure. No GVHD drug prophylaxis was given post transplant.

Results: A total of 215 consecutive patients, accrued between 5/13/2010 and 11/20/2014, were analyzed (84 in arm A, 103 in arm B, 28 in arm C). These pts have been followed for a median of 19.7 months. Donors were related or unrelated and HLA-matched for 73% of the patients and 1-2 HLA alleles disparate for 27%. Median age for the entire group was 47.8 years, with older pts predominating in the non-TBI groups (medians arm A, 31.2 yrs; arm B, 58.8 yrs; arm C, 24 yrs). The CD34 þ PBSC transplant provided a mean dose of 9.7 x 10 6 CD34 þ progenitors/Kg (range 1.4 -89.7 ) and 4.5x10 3 CD3 þ T-cells/Kg (range 0.6-25.3 ). All pts engrafted; but 2 pts (1.9%) in arm B experienced late graft failure, one of whom was reconstituted after a secondary graft. Overall the incidence of grade II-IV acute GVHD was 6%, 6% and 7% for arm A, B and C respectively. TRM at 1 year was 9% in Arm A, and 15% in Arm B and 16% in Arm C. Two year OS and DFS for each arm are: arm A -62.2% and 56%; arm B -68% and 59% and 48% and 48% for arm C. For the 115 pts who received standard risk transplants (i.e., pts with high risk forms of AML, ALL or NHL in 1 o CR, AML in 2 o CR, MDS RA/RCMD, CML in 1 o CP or MM in CR1), 2 year OS and DFS are: arm A -60% and 59%; arm B -74% and 65%; arm C -50% and 50%, with relapse rates at 2 yrs of arm A -23%, arm B -10.4%, and arm C -11.5%. cumulative incidence of relapse (CIR) stratified by risk group reveals that the probability for relapse is significantly higher for the pts with high risk disease (p o0.0005), whereas the cumulative incidence of non-relapse mortality (NRM) is comparable (p ¼ 0.36)( Figure 1 ). The median time to relapse has not been reached in either group. The estimates of relapse at 1 year were 23.2% (95% CI, 16.5-30.0% ) for high risk pts and 11.6% (95% CI, 6.7 -16.6%) for the other group. The estimated 2-year CIR was 31.4% (95% CI, 23.5-39.4%) for the group at high risk and 14.6% (95% CI, As measured from alloSCT prior to DLI BM progression and focal progression patterns were highly similar with cumulative incidences of BM progression of 23%±13% and 23%±13%, and of focal progression of 19% ± 12% and 21% ± 12%, after 12 and 24 months post-transplant, respectively. In contrast, as measured from DLI BM progression and focal progression patterns showed strong dissimilarity: at 12 and 24 months after DLI cumulative incidences of BM progression were 14%±13% and 17%±14% respectively, whereas focal progression was 48% ± 19% and 62% ± 18%, respectively, illustrating a potent immunological response in BM with only limited effect of DLI on focal lesions. Conclusion: Disease progression patterns of multiple myeloma after TCD-RIC alloSCT diverged from initially similar BM and focal progression patterns in the absence of alloimmune responses towards disease control in BM with focal progression after DLI. This finding illustrates failure of donor lymphocytes to target extra-medullary/focal disease in multiple myeloma. Disclosure of Interest: None declared.

Engineered T cells modified to express a CS-1-specific chimeric antigen receptor (CAR) confer anti-myeloma activity in vitro and in pre-clinical in vivo models Introduction: Adoptive immunotherapy with T cells that were modified by gene-transfer to express tumor-targeting chimeric antigen receptors (CARs) has therapeutic potential in advanced B-cell malignancies. We are pursuing the glycoprotein CS1 (SLAMF7/CD319) as candidate target for CAR T cells in Multiple Myeloma (MM), due to its restricted high level expression on malignant plasma cells in a significant proportion of MM patients. Here, we evaluated the anti-MM function of T cells that we modified with CS1-specific CARs in vitro and pre-clinical in vivo models. Materials (or patients) and methods: We constructed two CS1-specific CARs with antigen-binding domains derived from the huLuc63 and Luc90 mAbs that target distinct CS1 epitopes, each comprising a signaling module of CD3zeta and a CD28 co-stimulatory domain, and encoded both constructs in lentiviral vectors for gene-transfer.

Results: CD8 þ and CD4 þ T-cell lines expressing the huLuc63 and Luc90 CS1-CARs could be readily generated from healthy donors and MM patients (n ¼ 5/3), and propagated and expanded in vitro with similar kinetics as T cells expressing a CD19-specific CAR that we included as a reference. In functional experiments, CD8 þ T cells expressing either of the CS1-CARs conferred specific high level lysis of MM lines (MM1.S, NCI-H929 and OPM-2), primary MM, K562 that had been stably transfected with CS1, but not native CS1-negative K562. We also detected high level production of IFNg and IL-2 (CD44CD8), and productive proliferation after stimulation with CS1 þ target cells, with significantly superior anti-MM reactivity mediated by the huLuc63 compared to the Luc90 CS1-CAR construct. We confirmed the anti-MM efficacy of CS1-CAR modified T cells using a xenograft model in immunodeficient mice (NSG/MM1.S). Mice were inoculated with fireflyluciferase labeled MM1.S myeloma by tail vein injection and 14 days later, when mice presented with disseminated disease, administered a single dose (i.v.) of a cell product consisting of equal proportions of CD8 þ and CD4 þ T cells modified with either the optimal CS1-CAR, the CD19-specific CAR or mock transduced. We observed rapid and durable complete rejection of established MM from bone marrow and resolution of extramedullar MM manifestations in all of the mice treated with CS1-CAR T cells (n ¼ 4), whereas mice treated with CD19-CAR T cells, or control T cells had to sacrificed due to progressive disease (n ¼ 4/4). Of interest, in this in vivo model, we observed similarly effective anti-MM responses mediated by CS1-CAR T cells that had been derived from healthy donors and MM patients. Conclusion: Our data suggest the potential of T cells expressing CS1-specific CARs to confer anti-MM activity in clinical settings. The experience with anti-CS1 mAb huLuc63, that as single agent has only minute anti-MM activity, indicates targeting this molecule will be safe and not be associated with toxicity to normal tissues. We observed stronger anti-MM reactivity with CS1-CARs targeting a proximal (huLuc63) rather than a distal (Luc90) epitope on CS1 protein, in line with our previous observation that the targeted epitope on a given antigen affects tumor recognition of CAR T cells. Experiments to analyze the function of our CS1-CARs against panels of primary MM in our NSG model are ongoing to inform our efforts of clinically translating CAR T-cell therapy in this entity. 4;14) , t(14;16), del17p by FISH and/or del13q by karyotyping]. All pts had to achieve at least a partial response from preceding salvage chemotherapy (n ¼ 32) or second salvage auto HSCT (n ¼ 12). Pts underwent allo TCD HSCT with busulfan (0.8 mg/kg x 10 doses), melphalan (70 mg/m 2 x 2 days), fludarabine (25 mg/m 2 x 5 days) and rabbit ATG (2.5 mg/ kg x 2 days). TCD was performed by positive CD34 selection (Isolex) followed by rosetting with sheep erythrocytes for the initial 13pts (2008-09) and by CD34 þ enrichment by the Miltenyi Device in 31pts thereafter, achievingo10 4 CD3 þ /kg for all grafts. None of these pts received immuno suppressive therapy post TCD HSCT. Pts with 10/10 HLA matched donors were also eligible to receive low doses of DLI (5x10e 5 -1x10e 6 CD3 þ /kg) no earlier than 5mos post allo HSCT.

Results: 44 pts with a median follow up of 24.8 mos (range: 11.1-81.2 mos) of survivors are reported, median age 56 years (range 32-69). All pts engrafted promptly (median d þ 10, range þ 9 -þ 12).TRM (grade II-IV) at 12mos is 18% (95% CI: 8% -31%). Acute GvHD was 2% (95% CI: 0% -11%) and chronic GvHD was not observed in any pt. The overall survival (OS) and progression-free survival (PFS) with their 95% confidence intervals (CI) are shown in Table 1 . Factors associated with worse outcome were disease status and number of previous treatments prior to TCD HSCT. (1) . In this study, we have analyzed the molecular consequences of del(8)(p21), an abnormality we and others have previously shown to have an adverse impact on survival of MM patients (2) (3) (4) . Materials (or patients) and methods: In a cohort of 140 patients that were diagnosed with MM between 2001 and 2012, we have investigated the clinical impact of del(8)(p21) on time to progression (TTP) and overall survival (OS). Moreover, response rate of 84 patients to 1 st line bortezomib treatment was investigated. We have also analyzed the expression profiles of genes located near the 8p21 region in patients with and without del(8)(p21). Additionally, we have analyzed the in vitro response of primary MM cells with and without the deletion to bortezomib-mediated killing and sensitization to TRAIL/APO2L-triggered apoptosis in an attempt to understand why MM patients carrying 8p21 deletion respond poorly to bortezomib treatment. Results: We found that MM patients carrying del(8)(p21) deletion had significantly shorter TTP compared to patients without the deletion (P ¼ 0.011) and most importantly these patients had significantly shorter OS compared to patients without the deletion (P ¼ 0.001). In a cohort of 84 patients, we observed that patients with del(8)(p21) (n ¼ 24) responded poorly to bortezomib, 50% showing no response while 90% of patients without the deletion (n ¼ 60) responded to bortezomib treatment. In vitro analysis revealed that MM cells from patients with del(8)(p21) show higher resistance to bortezomib treatment possibly due to upregulated expression of genes such as PTK2B, CCDC25, RHOBTB2, NFkB, MYC and BCL2 while showing downregulated levels of TP53 and SCARA3 when compared to MM cells without the deletion. Furthermore, we have observed that MM cells with del(8)(p21) express higher levels of the decoy death receptor, TRAIL-R4 and fail to upregulate the pro-apoptotic death receptors TRAIL-R1 and TRAIL-R2 that are located in the 8p21 region. As a result, MM cells with del(8)(p21) were largely resistant to bortezomib and TRAIL/APO2L-mediated apoptosis. Conclusion: Substantiating the clinical outcome of the patients, our data provides a potential explanation regarding the poor response of MM patients with del(8)(p21) to bortezomib treatment. Furthermore, our clinical evaluation suggests that including immunomodulatory agents such as lenalidomide in the treatment regimen may help to overcome this negative effect, providing an alternative thought in planning treatments of patients with del(8)(p21). Introduction: Autologous stem cell transplantation is the standard treatment in patients with multiple myeloma (MM). However, there is discrepancy over the optimal mobilization regimen. Therefore a randomized study was conducted to compare cellular composition of the collected grafts as well as early hematopoietic and immune recovery in MM patients receiving G-CSF with or without low-dose cyclophosphamide for mobilization of blood grafts after induction with lenalidomide, bortezomib and dexamethasone. Materials (or patients) and methods: Thirty patients with MM were included into this prospective multicenter study. There were 16 males and 14 females with a median age of 62 years (range 43-70). Fourteen patients were mobilized with cyclophosphamide plus G-CSF (arm A) whereas sixteen patients were mobilized with G-CSF alone (arm B). Melphalan 200 mg/m 2 was used as high-dose therapy and patients having graft CD34 þ cell countso3 x 10 6 /kg (measured before freezing) were scheduled to receive G-CSF after the graft infusion. Cryopreserved graft samples were analyzed with a flow cytometry for T and B cells (CD3/CD8/CD45/CD19) as well as for NK cells (CD3/CD16 þ CD56/CD45). Also CD34 þ cell subclasses were analyzed (CD34/CD38/CD133/CD45). Complete blood counts were evaluated on day þ 15 and one month post-transplant and a flow cytometry for blood lymphocyte subsets (T, B, NK) was performed one month after the graft infusion. Results: The blood grafts in arm A contained significantly higher amounts of CD34 þ cells and the grafts of the arm B contained significantly higher proportion of primitive CD34 þ CD133 þ CD38cells and T, NK and B lymphocytes ( Table 1) . The median amount of infused CD34 þ cells was comparable between the arms (3.9 Â 10 6 /kg in group A vs. 3.1 Â 10 6 /kg in group B, P ¼ 0.056). The number of platelets was slightly lower in the group B at d þ 15 (P ¼ 0.094) otherwise the course of early hematological and immune recovery was comparable between the groups.

The use of G-CSF alone instead of a combination with cyclophosphamide seems to enrich the blood grafts with significantly higher number of T and B lymphocytes and a higher proportion of more primitive stem cells. The hematological and immune recovery was comparable between the arms. The possible effects of graft composition in long-term patient outcomes will be further evaluated in the ongoing GOA study (Graft and Outcome in Autologous stem cell transplantation). Disclosure of Interest: None declared. Introduction: PET is a useful tool that allows deeper assessment of response beyond that measured by M protein levels. It has been reported to predict outcome following both ASCT. To be able to integrate PET-CT negativity to internationally accepted response criteria the cut-off level needs to be validated by independent investigators. This prospective study was initiated to elucidate the prognostic role of PET-CT in the ASCT setting utilizing the cut-off found in our patients in Ankara University (3.35 ) comparing with those initially reported by Barlogie et al (3.6 ) and Zamagni et al (4.2) Materials (or patients) and methods: 85 consecutive patients diagnosed and transplanted in Ankara University with pre-and post-ASCT FDG-PET-CT imaging were included. Patients were: Median age 56.6 þ /-8.8 (M/F: 45/40), ISS I/II/III: 37/33/15, renal impairment (8,2%), bone involvement (94,1%), del13q (40.9%), t (4;14) and/or p53(26.5%), LDH high(10,5%), induction with Bortezomib (72,9%). Pre-ASCT clinical response 3 VGPR: 52.9%, post-ASCT clinical response 3 VGPR:77.5%. Overall Survival (OS): median: 33 months (4.2-141 months). PASW statistics for Windows program was used for statistical analysis. Results: As reported previously ROC analysis revealed 3.35 as a significant cut-off level (P ¼ 0.005; OS). PET-CR was defined FDG uptake less than 4.2 or 3.35 depending on the analysis. Post-ASCT PET (44.2) was predictive for PFS (P ¼ 0.05) but not OS (P ¼ 0.096) . However PET (43.35 ) was predictive for OS (P ¼ 0.037) but not PFS. Depending on the cut-off more (SUV 3 4.2: 43/64) or less (SUV 3 3.35 : 29/64) patients met the criteria for PET-negativity (or remission) following ASCT. Expert PET assessment resulted with PET-CR 39/85 similar to the SUV 3.35 frequency. As shown in figure patients to converted to CR after ASCT (positive/negative group) displayed a better PFS than those who had reached CR prior to ASCT. This analysis was significant if cut-off was 3.35 but not 4.2. Expert assessment was also able to differentiate patients with better prognostic features. 0.0-0.49 ). 23 leukaphereses were analized for MRD: the median plasma cells value was 0,03% (0,00-0,7%); in 6 pts PCs wereo0,01% (cut off for MRD negativity). Conclusion: Mobilization with Cy-BOR-Dx þ G-CSF and BORbased ASCT is safe and effective in elderly MM patients. This schedule allows the collection of an adequate dose of CD34 þ cells, with a very low rate of mobilization failure (2%), also in elderly MM pts. A low rate of clonal PCs contamination in the harvest was also observed. This approach allows to perform ASCT in most elderly pts, achieving high response rate and promising outcome with a short term treatment:20 weeks compared to the non-ASCT programs (54 weeks in the VMP program). Disclosure of Interest: None declared. MICA expression levels were investigated in n ¼ 180 gut biopsies with SYBR GREEN s qRT-PCR. Histological grades of the gastrointestinal GvHD (GI GvHD) were determined by the pathology department at the University Clinic, Regensburg and severity was grouped by assigning an apoptotic score (0 ¼ absence of apoptosis, 3 ¼ maximum apoptosis). A Protein Biochip Array (Evidence Investigator s , RANDOX) was utilised for measuring MICA serum levels and evaluated in n ¼ 129 samples from allo-HSCT patients collected at pretransplantation, day-7, day þ 14, day þ 28 and 3 months post transplantation.

Results: Our analysis showed that the Methionine allele in rs1051792 was associated with an increased risk of relapse (P ¼ 0.029). The same allele was also found to be associated with a reduced overall survival (P ¼ 0.041) which was more severe for non-T cell-depleted allo-HSCT (P ¼ 0.001). Vice, versa, the presence of the Valine allele was associated with the development of aGvHD (P ¼ 0.044).

In the gut, MICA expression was investigated in patients treated with low doses of steroids (r 20 mg/kg), as high dose steroid treatment strongly suppressed MICA expression. Higher levels of MICA were associated with an apoptotic score ¼ 0 (no apoptosis) (P ¼ 0.044) and the absence of active GI GvHD (P ¼ 0.046). Increased soluble MICA levels at 3 months post-transplantation were significantly associated with aGvHD (P ¼ 0.0123).

Conclusion: MICA molecules have been shown to play prominent roles in immune processes and therefore are also potential aGVHD biomarkers. In this study, we showed that the Methionine (MICA-129Met) allele was associated with the incidence of relapse while the Valine (MICA-129Val) allele was associated with an increased risk aGvHD. A low overall survival for patients who did not have had the T cell depletion treatment was also associated with the presence of the Methionine (MICA-129Met) allele. In the gut of patients treated with low doses of steroid, MICA gene expression levels were higher with the absence of GvHD. This may indicate that the isoforms are able to meditate NK-cell and T cell inactivation, and down-regulate NKG2D with high levels of soluble MICA contributing to the development of aGvHD. Eleven patients needed GvHD treatment: 4 pts received ganciclovir iv (GCV 5 mg/Kg/12 h/14 days), 7 pts valganciclovir per os (VGCV 900 mg/12 h/14 days). Both GCV and VGCV were effective in control clinical manifestations of GvHD in a median of 13 days (range 7-52) and resulted in a significant reduction in numbers of circulating TK-cells, without reduction of CD3 þ TK-negative lymphocytes resulting in no effect on long-term immune reconstitution. In five patients additional concomitant treatment with low-dose steroid (prednisone o0.5 mg/kg per day for a median of 2 weeks) was given. A pt who presented severe gut and liver GvHD and one pt who received at transplantation an high dose of unmanipulated lymphocytes (5.4x10 5 /Kg) -were succesfully treated with a combined therapy of prednisone and cyclosporine or rapamicine in association with GCV. One patient developed a severe classic de novo c-GVHD, with sclerodermatous lichenoid skin and mouth features plus moderate dry-eye symptoms that was successfully treated with VGCV and a transient course of mycophenolate mofetil (2 g per day) over a 2 months period. No cases of quiescent or progressive c-GvHD was observed after a median follow-up of 679 days (range 139/4035). Conclusion: In our long-lasting clinical application of haploidentical TK-cells, an effective induction of immune reconstitution and a complete control of GvHD, provided a long-term immunosoppressive therapy free survival in absence of GvHD related deaths or longterm complications. Introduction: Despite major improvements in allogeneic hematopoietic cell transplantation (allo-HCT) over the last decades, severe corticosteroid-refractory acute and chronic graft-versus-host-disease (GvHD) still remains a life-threatening complication characterized by high mortality rates (40-70%). Since preclinical and early clinical evidence indicated anti-inflammatory effects of ruxolitinib, we collected the outcome data from multiple Stem Cell Transplant Centers using the JAK1/2 inhibitor ruxolitinib as salvage treatment in patients suffering from corticosteroid-refractory GvHD. Materials (or patients) and methods: A total of 13 Stem Cell Transplant Centers in Germany, France, Switzerland and United States reported outcome data from 52 patients who received ruxolitinib for corticosteroid-refractory GvHD (skin, mucosa, intestine, liver, lung, musculoskeletal) between 01/ 2012 and 12/2014. Patients were classified as having acute (n ¼ 32) or chronic (n ¼ 20) GvHD. The median number of previous GvHD-therapies was 4 for acute GvHD (range: 1-7) and 3 for chronic GvHD (range: 2-10).

Results: The overall response rate was 84.3% (27/32) in acute GvHD comprising 10 CRs (31.2%) and 17 PRs (53.1%). In chronic GvHD the overall response rate was 80% (16/20). Clinical improvement was rapid with a median time to response of 1.5 (1-4) weeks and 1 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) week after initiation of ruxolitinib treatment in acute and chronic GvHD, respectively. All responders were in persistent remission at last follow-up. The median follow-up was 18 (2-58) and 13 (2-70.5) weeks for acute and chronic GvHD patients, respectively. Non-responders (acute GvHD: 5/32, chronic GvHD: 4/20) received other salvage therapies. Cytopenias (anemia, leukopenia or thrombocytopenia) and CMV reactivation were observed during the time of ruxolitinib treatment in both acute (18/32, 56.2% and respectively 11/32, 34.3% ) and chronic (4/20, 20% and respectively 3/20, 15%) GvHD patients, sometimes however cytopenias already preceded ruxolitinib treatment. Ruxolitinib treatment was stopped or reduced in 2 patients because of cytopenia. CMV was controlled by antiviral therapy in all patients. 3 ruxolitinib responders died, one because of leukemia relapse 4 weeks after ruxolitinib was stopped, and two from GvHD progression.

In one of the patients who died, ruxolitinib was stopped due to impossibility of oral drug application. Conclusion: Overall, these data collected in multiple centers using different strategies for GvHD prophylaxis and treatment suggest that ruxolitinib is a very promising agent in the treatment of corticosteroid-refractory acute or chronic GvHD and may be successfully used to treat a major subset of patients beyond 2nd line of GvHD treatment. A prospective randomized multicentre clinical trial testing therapeutic JAK1/ 2 inhibition as salvage treatment in GvHD is planned to verify the efficacy of ruxolitinib and to identify potential biomarkers that may be predictive for response. Disclosure of Interest: None declared.

Transplant-associated renal microangiopathy is associated with a high risk of refractory acute GVHD and characterized by a specific biomarker signature Introduction: There is increasing evidence that endothelial damage is involved in the pathogenesis of steroid-refractory graft-versus-host disease (refGVHD). Recently, serum soluble ST2 (suppressor of tumorigenicity, IL-33 receptor), an independent risk factor of cardiovascular death, has been identified as a risk factor of refractory GVHD. However, the pathomechanism of endothelial cell dysfunction which is associated with mortality from GVHD is yet poorly characterized. Renal transplant-associated microangiopathy (TMA) is another endothelial complication of alloSCT, and its association with severe GVHD and with biomarkers of endothelial damage (ST2, sCD141 (soluble thrombomodulin)), and endothelial function (VEGF) is investigated in this study. Materials (or patients) and methods: Evidence for renal TMA was studied in a cohort of 508 patients who underwent alloSCT between 2002 and 2013 at our institution and who have provided informed consent for this observational study. Criteria to diagnose renal TMA included an otherwise unexplained 50% rise in creatinine and lactate dehydrogenase (LDH) levels (or a pre-existing LDH above 400 U/L), a 50% drop in platelet counts (or a pre-existing platelet count below 50/nl) and at least 4% schistocytes. Cytokines were measured in sera taken prior to alloSCT and on the indicated days thereafter and stored at minus 80C. Statistical analyses were performed using SPSS19 and included a cumulative incidence analysis of causespecific hazards and the non-parametrical median test of independent probes. Results: Both renal TMA and refGVHD were rare complications after alloSCT but were significantly associated with each other (TMA only 18/508 (3.5%), refGVHD only 26/508 (5.1%), both 20/508 (3.9%), chi 2 0.000). Median time intervals from alloSCT to renal TMA and refGVHD were 1.93 (0.23-60.16 ) months and 1.23 (0.16-12.4 ) months respectively. In the overlap group, GVHD onset usually occurred before renal TMA (- 0.69 months, -55.4 to 11.2) NRM rates were significantly increased in all three cohorts but approached 100% in patients with both complications. Serum sTM levels as well as soluble ST2 levels increased between transplantation and day 100/day 200 after alloSCT in all cohorts, refGVHD, TMA and both. In contrast, VEGF levels (day 100) were significantly lower specifically in patients with TMA with or without refGVHD, but not in patients with refGVHD without TMA.

Conclusion: This study identifies renal TMA as an endothelial cell dysfunction associated with extremely high mortality rates in the context of GVHD. The absence or presence of renal TMA defines two separate subsets of refGVHD with a different prognosis. Biomarkers of endothelial damage or vulnerability, such as sTM, ST2, and VEGF can help to dissect TMA and refGVHD, and might be useful to identify and guide management of patients with endothelial dysfunction who are at high risk of fatal complications after alloSCT. Disclosure of Interest: None declared. Introduction: SNPs of the key cytokines and chemokines involved in the pathogenesis of aGvHD have become an object of major interest recently. Here we present SNPs rs3774937 CC/CT/TT and rs3774959 AA/AG/GG of the NF-kB1 gene in association with aGvHD. Materials (or patients) and methods: In our single-center study we analyzed 70 patients allografted for the following hematological malignancies: AML (37%), ALL (14%), CLL (13%), MDS (13%), CML (6%), NHL (6%), IMF (3%), CMML (4%), and other hematological disorders (4%) between 2009-2014. The median age of the study group was 51 (21-62) years. Patients were allografted after myeloablative (11%), non-myeloablative (14%) and reduced (intensity/toxicity) conditionings (74%). GvHD prophylaxis was done by solo cyclosporine-A (80%), cyclosporine-A with mycophenolate mofetil (19%) and cyclosporine-A with short-methotrexate (1%). ''In vivo'' T-depletion with thymoglobuline was used in 64% of recipients. Patients were allografted from HLA identical donors (related 46%) with median age 37 (18-63) years. The female donor/male recipient combination represented 14% of all pairs. The grafts contained median 4.6 (2.8-8.0 ) x106 CD34 þ cells/kg and median 6.8 (1.9-27.1) x108 MNC /kg. SNPs analysis was done from genomic DNA isolated from EDTA-treated peripheral blood. Genotyping was performed with Sequenom MassARRAY platform using allele-specific MALDI-TOF mass spectrometry assay (Sequenom, San Diego, CA, USA). Primers were designed using the Sequenom SNP Assay Design software version 3.0 for iPLEX reactions. Univariate analysis was performed to find significant difference in aGvHD among the patient groups with different NF-kB1 profiles. The asymptotic Pearson's chi-square test was used in cross tabulation with significance level set to 0.05. Results : Out of 70 patients 46 patients (66%) are still alive, 24 patients (34%) died (12 died of transplant-related complications). The median post-transplant follow-up was 1.6 (0.1-4.5) years. AGvHD developed in 21 patients (30%), grade III-IV in 7 of them (10%). Both NFkB1 SNPs were completely correlated in the sense that knowing the genotype of one fully determined the genotype of the second one (5 patients did not follow this correlation and were excluded from the analysis). Consequently, we defined a common predictor NF-kB1 which codes the information carried by both NF-kB1 SNPs in the following way: all patients carrying the rs3774937 CC genotype were also positive for rs3774959 AA allele and were marked as NF-kB1 ¼ I, all patients carrying the rs3774937 CT genotype were also positive for rs3774959 AG allele and were marked as NF-kB1 ¼ II and finally all patients carrying the rs3774937 TT genotype were also positive for rs3774959 GG allele and were marked as NF-kB1 ¼ III. The NF-kB1 profile was found to be significantly associated (P ¼ 0.002) with aGvHD in the following way: Patients in the NF-kB1 ¼ I group are more probable to suffer from aGvHD than in the NF-kB1 ¼ III group.

Conclusion: This is the first report showing the association of NF-kB1 gene SNPs with aGvHD. The transcription factor NF-kB has been implicated in the regulation of cellular stress and inflammatory signals. According to our pilot data patients with inherited genetic abnormalities of the NF-kB1 gene may be prone to aGvHD. Patients.The two groups -control/treatment -were well balanced in terms of diagnosis (P ¼ 0.9), and disease phase (P ¼ 0.4) : the most frequent diagnosis was AML (n ¼ 75), followed by ALL (n ¼ 38) and MDS (n ¼ 10). The median age for control/treatment was 46 years (1-69) vs 38 yers (0, , (P ¼ 0.06); the proportion of patients over 50 years was 51% in the control and 49% in treatment group (P ¼ 0.8). The donor type in the control/treatment arms was as follows: HLA identical siblings n ¼ 36/n ¼ 34, unrelated cord blood (CB) n ¼ 6/n ¼ 2, unrelated donor (UD) n ¼ 36/n ¼ 42, and haploidentical family donors (HAPLO) 6/7 (P ¼ 0.4). Skin biopsies.A skin biopsy before randomization, was not mandatory: it was performed in 38 patients. GvHD was diagnosed as proven, probable and possible respectively in S23 21%, 32% and 29%of the patients.These different reports were equally distributed in treatment and controls (P ¼ 0.7). Results: The cumulative incidence of acute GvHD grade II (primary end point), was 50% in controls and 35% in treatment patients (P ¼ 0.02). This difference was maintained in different subgroups. The CI of transplant related mortality (TRM) was 18% (control) vs 27% (treatment) (P ¼ 0.1), despite a non significant younger median age in treatment patients. Excess mortality in the treatment arm was due to an excess incidence of infections. Actuarial 1 year survival was 80% (control) vs 78% (treatment) (P ¼ 0.1). Cases of death in the control/treatment groups were as follows: GvHD 10% -13%; infection, 4%4 8%; interstitial pneumonia, 2% -0%; toxicity, 1% -2%; leukemia relapse 19% -18% (P ¼ 0.3). There was no significant difference in TRM among different Centers (P ¼ 0.5) Progression of GvHD and skin biopsies.The proportion of patients progressing to GvHD grade II þ was simlar in proven, probable, possible GvHD (62%, 50%, 45% ) (P ¼ 0.7). However the proportion of GvHD related deaths was 37% for proven GvHD, and 8% for probable/possible GvHD. Introduction: Graft versus host disease (GvHD) is a common and severe complication after allogeneic stem cell transplantation. During pregnancy, the placenta and the fetal membranes function as an immunological barrier, protecting the fetus from the mother's immune system. We have isolated stromal cells from the decidual layer of term placentas. Decidual stromal cells (DSCs) are of maternal origin and strongly inhibit the alloreactivity of T-cells in vitro. The effect is mainly contact dependent, and decreases the production of several key cytokines involved in the cytokine storm promoting the continuation of GvHD. Materials (or patients) and methods: To investigate the effect of DSCs on acute GvHD we enrolled 34 patients diagnosed with acute GvHD and clinically non-responsive to standard therapy. The protocol was modified after 17 patients, the DSCs were then thawed and infused in infusion solution with albumin instead of AB-plasma and given repeatedly and earlier upon diagnosis. This led to the formation of two treatment groups (group 1 n ¼ 17 and group 2 n ¼ 17), which were compared to matched historical controls (n ¼ 54). We also performed a retrospectively corrected analysis of steroid refractivity.

Results: Group 1 received a median of 1 infusion on day 11 after standard treatment compared to group 2, who received a median of 2 infusions (P ¼ o0,05) on day 7 (ns). No adverse events related to the treatment were observed. At 4 weeks after treatment, 60% of the patients in group 1 hade responded to the treatment. In contrast, all patients in group 2 responded (P ¼ o0,05). All patients in the treatment groups received fungal prophylaxis. The overall cumulative survival (OS) at one year was 28% for the controls as compared to 47% for group 1 and 80% for group 2 (P ¼ o0.001). When the groups were corrected for steroid-refractivity, the OS was 3% for the controls, 38% for group 1 and 70% for group 2(P ¼ o0.001). In the steroid-refractory control group (n ¼ 32), the risk of dying from GvHD at one year was 81%, whereas no patients in the steroid-refractory group 2 died from GvHD(n ¼ 13, P ¼ o0.001).

In the last 19 months, no patients have died from acute GvHD at our center.

In conclusion, DSCs might be an effective treatment of acute GvHD. The infusion should be prepared in albumin, given as early as possible and in repeated doses.

To confirm these striking findings, we will extend the followup time and enroll more patients in our study. Disclosure of Interest: None declared.

Oral session: Stem cell source and donor type O041 Unmanipulated haploidentical stem cell transplantation after reduced intensity or ablative conditioning regimen for the treatment of acute leukemia-a report from the acute leukemia working party of the EBMT M. Introduction: Haploidentical hematopoietic stem cell transplantation(Haplo-HSCT)is feasible option for patients with acute leukemia(AL)at high risk of relapse who do not have HLA-matched related or unrelated donors. Haplo-HSCT was associated with severe acute graft-versus-host disease (aGVHD) in unmanipulated transplants and a high incidence of graft rejection in T-cell depleted transplants because of the high frequency of T cells that recognized major class I or II HLA disparities between donor and recipient. Two approaches were developed to overcome these problems:megadose of T-cell depleted hematopoietic progenitor cells without any posttransplant immunosuppression and T-cell replete grafts with innovative pharmacological prophylaxis of aGVHD.Posttransplant cyclophosphamide(PTCY)is regarded as a GVHDspecific immunosuppressant in adults but its feasibility is unknown in children. Purpose: to evaluate the feasibility and outcome of haplo-HSCT in children and adolescents with acute leukemia in active disease depending on conditioning regimens and methods of harvesting haplo-graft. Primary end points: overall survival (OS), transplant-related mortality (TRM). Secondary end points: engraftment rate, aGVHD, cGVHD, relapse rate. Materials (or patients) and methods: 56 patients(range from 1-21y.o. median 9 y.o.)with AL(ALL-32pts, AML-24pts)in progressive disease (PD)(cytoreduction chemotherapy (CTx) prior conditioning regimen-15pts, without-41pts) were analysed.MAC þ ATG regimen based on GIAC protocol received 20pts, MAC þ PTCY 50 mg/kg on D þ 3, þ 4-5pts, RIC þ ATG regimen based on Flu-18pts, RIC þ PTCY 50 mg/kg on D þ 3, þ 4-13pts. All pts received prophylaxis of aGVHD based on CsA-30pts, Tac-8pts, Tac þ Sir-18pts. G-CSF-primed T-cell replete BM was used as a graft source in 33pts (median CD34 þ cells 4,7x10 6 /kg), G-CSF mobilized peripheral blood (CD34 þ selected by CliniMACS, Miltenyi Biotec) and G-CSFprimed BM-23pts (median CD34 þ cells 11,3x10 6 /kg).

[O042]

Results: 3-year OS was 33,3%. Ptso9y.o. had significantly higher OS vs pts49y.o. 46,7% and 18,5% respectively (P ¼ 0,01). 3-year OS in pts receiving CTx prior conditioning regimen was 50% vs 26,8% in pts with leukemic burden(P ¼ 0,02). Significantly difference were observed in 3year OS in pts transplanted G-CSF-primed T-cell replete BM 45,5% vs 13% in pts after G-CSF mobilized peripheral blood and G-CSF-primed BM (P ¼ 0,03). Pts receiving RIC þ PTCY had 3-year OS 61,5%, RIC-11,1%, MAC-35%, MAC þ PTCY-20% (P ¼ 0,06). TRM after haplo-HSCT in pts with AL in PD was 38%. Engraftment was sustained in 82,1% pts. Full donor chimerism was achieved in 73,2% pts on D þ 30. Median ANC engraftment (40,5x10 9 /L) D þ 19,PLT recovery(420x10 9 /L) D þ 17. Cumulative incidence of grade 2-4 aGVHD was 27,3%,cGVHD-23,9%. Cumulative incidence of relapse was 39,3%.

Conclusion: Haplo-HSCT G-CSF-primed unmanipulated BM is an effective method of achieving remission with good sustained engraftment rate in children and adolescents with resistant disease. RIC regimen followed by T-cell replete haplo-HSCT with PTCY on D þ 3, þ 4 was associated with good OS, low incidences of GVHD and TRM. Before any definitive conclusions can drawn, a randomized study is required. Disclosure of Interest: None declared.

The detection of donor specific anti-HLA antibodies in recipients of unmanipulated haploidentical blood and marrow transplantation is predictive of poor graft function Introduction: Our previous study suggest that choosing young, male, non-inherited maternal antigen-mismatched donors is reasonable following unmanipulated haploidentical blood and marrow transplantation (HBMT). Recently, a correlation between the presence of donor-specific anti-HLA antibodies (DSA) and graft failure has been demonstrated in haploidentical transplant settings. In our protocol, approximately 99% patients can achieve sustained, full donor chimerism. However, poor graft function (PGF) remains one of complications after unmanipulated HBMT. Therefore, we determined the effect of DSA on primary PGF in order to provide further evidence for donor selection. Materials (or patients) and methods: Three hundreds and fourty-five patients with hematological diseases receiving HBMT were enrolled in this prospective study. The median age of the patients was 26 years (range, 2-58 years). These patients were randomly selected as training group (n ¼ 173) and validation group (n ¼ 172). Peripheral blood serum were collected pre-conditioning regimens. DSA were determined using the Luminex-based assay.

Results: In all 345 patients, the percentages of DSA positive cases were 11.3% (39/345). The incidence of DSA in female patients was higher than that of male cases (16% vs. 8%, P ¼ 0.020). In the training set, a cutoff value of DSA (MFI ¼ 2000) were developed. Multivariate analysis showed that the presence of DSA (patients with MFIZ2000 vs. cases with MFIZ2000) was associated with primary PGF. In the validation set, the association of DSA with primary PGF following transplantation was also confirmed. The association of PGF with inferior overall survival (OS) was demonstrated both in the training group and in the validation group. In all 345 patients, the median time to platelet recovery in DSA positive (MFIZ2000) patients was slower than that of DSA negative ones (25 days vs. 18 days, P ¼ 0.004). The incidence of primary PGF and primary graft failure was 5.5% and 0.9%, respectively. DSA positive patients experenced higher incidence of primary PGF (31% vs. 3 Introduction: Transplacental trafficking of maternal and fetal cells during pregnancy establishes long-term, reciprocal microchimerism in both mother and child because of exposure of the two immune systems to the non-self alloantigens (Maloney et al., J Clin Invest. 1999 ). Studies show the immune system in the mother is capable of being sensitized by paternal histocompatibility antigens. For example, antibodies directed against paternal HLA-antigens (van Rood et al., Nature. 1958 ) and T lymphocytes directed against paternal major (van Kampen et al., Hum Immunol. 2001 ) and minor histocompatibility antigens (Verdijk et al., Blood. 2004) are frequently detected in multiparous women. We previously demonstrated mother/child immune interactions positively influenced the outcome of mother to child HLA haploidentical T cell-depleted hematopoietic transplantation. In a series of adult and pediatric patients we demonstrated mother donors conferred protection against leukemia relapse and improved transplant related mortality (TRM), which was largely due to infection, and improved survival (Stern et al., Blood 2008). Materials (or patients) and methods: The Kaplan-Meier method evaluated leukemia-free survival. Cumulative incidence estimates were used for relapse and TRM, as they are competing risks. Multivariate analysis assessed the impact of diverse variables on transplantation outcomes. Results: We analyzed the outcomes of 238 adult acute leukemia patients after T cell-depleted haploidentical transplantation. When compared with transplantation from all other family members, transplantation from mother donors was associated with significantly lower TRM (largely infectious) (27% vs 50% from all other donors, P ¼ 0.01). Multivariate analyses demonstrated transplantation from mother donors was an independent factor predicting improved survival (hazard ratio 0.41, 95% confidence interval 0.12 to 0.95, P ¼ 0.03). In an attempt to elucidate the mechanism, we analyzed donor T cell repertoires that were specific for CMV antigens presented by recipient antigen-presenting cells (by ELISPOT and by limiting dilution cloning). Unlike all other donor/recipient pairs, mothers possessed CMV-specific CD8 cell clones that killed child's and father's CMV-pulsed dendritic cells (DCs). Such clones were nonalloreactive as they did not kill the child's or father's non-CMV-S26 pulsed DCs. Mothers also possessed CD4 T cell clones that produced IFN-gamma in response to child's and father's CMVloaded DCs. Such clones were non-alloreactive as they did not respond to child's or father's non-CMV-pulsed APCs. Thus, mothers possessed a T cell repertoire that recognized CMV antigens also when presented by the unshared, father's, HLA haplotype. In fact, they showed twice as many T cells that recognized CMV antigens presented by the child's APCs than all other donor/recipient pairs (Po0.05).

Conclusion: Therefore, pregnancy resulted in the generation of an additional T cell repertoire that specifically recognized pathogen antigens presented by the unshared paternal HLA haplotype antigens on the child's APCs. Apparently, upon mother to child T cell-depleted hematopoietic transplantation, such repertoire expands over time and helps reduce infectious mortality. Further studies are needed to elucidate the mechanisms underlying mother T cell selection/education by paternal HLA haplotype antigens on the child's APCs.

Disclosure of Interest: None declared.

Uni-directional and bi-directional non-permissive HLA-DPB1 T cell epitope group mismatches have similar risk associations in 10/10 matched unrelated donor HCT , was analyzed after separating uni-directional from bi-directional non-permissive mismatches. Non-permissive mismatches were defined as unidirectional HvG when the donor but not the patient carried an HLA-DPB1 allele from a TCE group not present in the patient, and vice versa as uni-directional GvH. Bi-directional nonpermissive mismatches were present when none of the HLA-DPB1 alleles in patient and donor were from the same TCE group. The associations with clinical endpoints of overall survival (OS), transplant related mortality (TRM), relapse, acute GvHD (aGvHD) and chronic GvHD (cGvHD) were studied using multivariate proportional hazards methods.

Results: The number of transplants with permissive or nonpermissive uni-directional HvG, uni-directional GvH and bidirectional mismatches was 1537, 527, 529 and 143, respectively. In the TRM analysis, non-permissive uni-directional HvG (HR 1.32, P ¼ 0.001) and GvH mismatches (HR 1.28, P ¼ 0.005) had significantly higher relative risks (RR) of TRM compared to the permissive group. The bi-directional group had similar RR (HR 1.35 , P ¼ 0.05). In pairwise comparisons, there were no statistical differences between the uni-and the bi-directional non-permissive groups for any of the outcomes tested. Introduction: There are several alternative sources of donor stem cells available for patients (pts) who need an allo-SCT and especially for those who lack a HLA-matched donor. Outcomes of mismatched-unrelated-donor (MM-URD) transplant have recently improved, and a comparison between matched and MM-URD sources in a uniform cohort of pts has not been performed after RIC regimens. Materials (or patients) and methods: Pts, aged Z50 year, who underwent fully matched or MM-URD RIC PBSCT or BMT from 2000-2012 were included in the study. All donors were HLA-matched (10/10) or mismatched at one or two-loci (9/10 or 8/10). The Kaplan-Meier-estimator, the cumulative incidence function and Cox proportional hazards regression models were used where appropriate. Results: In total 3197 pts receiving matched or mismatched RIC-URD allo-SCT were included in the study (AML 2947, ALL 250). 2370 10/10 HLA-matched pts were compared with recipients receiving 9/10 (n ¼ 712) or 8/10 (n ¼ 115) after RIC MM-URD allo-SCT. Median age of 10/10, 9/10 and 8/10 HLAmatched recipients were 60 years. Higher female donor to male recipients were in 8/10 cohorts (20%) compared to 10/10 (12%) and 9/10 (14%) group (P ¼ 0.02). More pts with CR2 and advanced-disease were among 9/10 and 8/10 cohort compared to 10/10 matched donor recipients (CR2 24, 24 and 20%; advanced disease 26, 25, 24% respectively; P ¼ 0.04). Also higher percentage of pts with secondary leukemia were in 8/10 cohorts compared 10/10 and 9/10 matched donor (35, 27, 25%; P ¼ 0.07). Percentages of engraftment (97%, 95%, 97%, P ¼ 0.16) were no different between the 3 groups. Acute GVHD grade II-IV was 27%, 33%, 33%, and grade III-IV 10, 13, and 10%, respectively for 10/10, 9/10 and 8/10 matched donors, respectively (P ¼ 0.003 and 0.03). In univariate analysis, 2-year survival rate was significantly higher for pts receiving 10/10 donor RIC-URD allo-SCT in CR1 compared to 9/10 or 8/10 MM-URD (OS: 55%, 46%, 46%, P ¼ 0.01; LFS: 51%, 43%, 45%, P ¼ 0.03, respectively). However, among the CR2 and advanced disease groups there were no differences in outcome between fully matched or MM-URD (9/ 10 or 8/10) donor (CR2 þ : OS 49%, 43%, 48%, P ¼ 0.30; LFS 42%, 37%, 45%, P ¼ 0.36; advanced-disease: OS 38%, 31%, 31%, P ¼ 0.39; LFS 32%, 26%, 28%, P ¼ 0.50, respectively). There was no difference in RI between the 3 groups and NRM was higher after fully matched donor compared to MM-URD (9/10 or 8/10) only in pts with CR1 diseases (P ¼ 0.02). Multivariate analysis showed higher NRM after 9/10 (HR 1.33, P ¼ 0.001) compared to fully matched donor and no difference in NRM between 9/10 vs. 8/10 MM-URD. There was no difference in RI between 10/10 vs. 9/10 or 9/10 vs. 8/10 MM-URD donor. OS and LFS were superior after fully matched donor vs. 9/10 MM-URD (OS: HR 1.24, P ¼ 0.001; LFS: HR 1.19 , P ¼ 0.002). However, there was no difference in adjusted OS and LFS between 9/10 vs. 8/10 MM-URD RIC allo-SCT. Chronic GVHD rate was not different between matched or MM-URD allo-SCT groups. Conclusion: Despite the limitations of a retrospective registrybased study, our analysis shows no significant outcome difference between 9/10 and 8/10 MM-URD allo-SCT after RIC regimen in patient aged Z50 year. In the absence of prospective data, we conclude that MM-URD RIC allo-SCT is a therapeutic option for acute leukemia pts not having fully matched donor. Disclosure of Interest: None declared. Introduction: The use of minors as HSC donors is medically and legally accepted and is increasing. However, there is a lack of understanding of the physical and psychosocial effects of pediatric HSC donation. The goal of this investigation was to longitudinally investigate HRQoL in this group. Materials (or patients) and methods: Participants were related pediatric donors (n ¼ 105) who donated at domestic U.S. centers between 4/10 and 5/13. Data were collected from donors and their parents via structured telephone interviews at pre-donation, and 4 weeks and 1 year post-donation. A healthy age/gender matched pediatric sample was generated from existing data for normative comparisons. Interviews gathered socio-demographics, psychosocial characteristics, and multidimensional HRQoL using the well-validated Pediatric Quality of Life Inventory (PedsQL) which produces a total score and physical, emotional, social, school and psychosocial subscores. T-tests were used to compare HRQoL from donor self-report, parental proxy-report, and the normative sample across the three assessment points. Mixed logistic models were used to examine the effects of pre-donation variables on post-donation HRQoL. Results: Donors were 5-17 yrs (median ¼ 11 yrs) and all but one were sibling donors. Most parental respondents were mothers (74%; median age 40 yrs), 87% were married, and 36% had at least a bachelor's degree. Donor vs Proxy. Across all HRQoL domains except emotional functioning and at all three assessment time points, donor self-reported HRQoL was significantly lower than that reported by parental proxies. Donor vs Norm. At pre-donation, as compared to the normative sample, donors reported significant HRQoL deficits across multiple subdomains and in total HRQoL (t ¼ -2.76,Po.01). At 4 weeks post-donation, donors reported deficits in physical functioning (t ¼ - 5.99, Po.001 ) and total HRQoL (t ¼ -2.82,Po.01). At 1 year post-donation, donors reported deficits in physical (t ¼ -2.51,Po.05) and school functioning (t ¼ -2.12,Po.05). Donors at HRQoL risk. Across the three assessment time points, 21%, 19%, and 17% of donors respectively had self-reported PedsQL total scores below the standard cutoff indicating significant clinical risk of poor HRQoL -scores below the cutoff are similar to those of chronically ill children. Sixteen percent, 13%, and 5% were below the cutoff at only one, two, or all three assessments respectively. The youngest donors (5-7 yrs) were at significantly greater risk of being below the cutoff than were their older counterparts with 37%, 39% and 82% of this group below cutoff at each of the three assessments respectively. In multivariable analyses, the pre-donation factor most strongly and consistently associated with below cutoff PedsQL scores at 4 weeks and 1 year post-donation was pre-donation donor self-reported PedsQL score (likelihood ratio: 9.17,Po.01; 15.54,Po.001). Conclusion: These findings suggest that there may be significant HRQoL deficits among pediatric HSC donors and that in this particular context, parents are not able to accurately report those deficits. These findings also indicate that research to identify predictors of poor HRQoL and the development of interventions to screen and address HRQoL deficits are urgently needed. Disclosure of Interest: None declared. Introduction: In allo HCT patients (pts) with disseminated AdV disease, mortality is reported to be up to 80%. Antiviral treatment (tx) usually consists of IV cidofovir (CDV), which has a significant risk of nephrotoxicity. BCV is an orally-available, lipid-conjugate of CDV with no evidence of nephrotoxicity in clinical trials. The pilot portion of the Phase 3 AdVise (CMX001-304) study was initiated in March 2014 to enroll B100 allo HCT and other immunocompromised AdV pts with, or at risk of progression to, disseminated AdV disease, to guide the final study design. As of 10NOV2014, 73 subjects have been enrolled and entered into the database, including 60 allo HCT pts (48 with disseminated AdV disease), 7 solid organ transplant pts and 6 ''other'' pts. Preliminary safety and virologic results for the 48 allo HCT pts with disseminated disease are described. Materials (or patients) and methods: All subjects receive open-label BCV 100 mg (Z50 kg) or 2 mg/kg (o50 kg) twiceweekly for 12 wks, extendable up to 24 wks for pts at high-risk of relapse, and are followed for 24 wks post-tx. AdV DNA viral load (VL) in plasma is measured using a quantitative PCR test (limit of detection [LOD] 2 log 10 c/mL). Results: Baseline (BL) characteristics for the 48 subjects are: median (range) age 12 (0.7, 69) y, 65% o18 y; 69% male; median (range) plasma AdV VL 4.6 (o LOD to 7.6 ) log 10 c/mL (n ¼ 45); 42% AdV positive by qualitative PCR in respiratory secretions, 60% in urine, 65% in stool; 29% with CMV in plasma, 6% EBV in plasma and 42% BKV in urine; 42% received prior IV CDV. As of 25NOV2014, 5 subjects had completed tx and 21 had discontinued tx prematurely. The most common reasons for tx discontinuation were death (n ¼ 11) and adverse event ([AE] n ¼ 4). Median (range) tx duration was 38 (1, 141) days (n ¼ 45). Virologic response in CDV-naïve and exposed subjects with detectable plasma AdV VL at BL are summarized in the table. In subjects with positive AdV PCR at BL, 65% (13/ 20) cleared AdV in respiratory secretions, 55% (16/29) in urine and 48% (15/31) in stool. Through 08DEC2014, 40% (19/48) of allo HCT subjects with disseminated AdV disease had died, with a median 71-day observation period for living subjects. No death was attributed to BCV. AEs leading to permanent tx discontinuation attributed to BCV were vomiting and abdominal pain in 1 subject, and acute GVHD in 1 subject.

Median (range) Change in AdV VL from BL (log 10 c/mL) Median (range) Time to Minimum On-tx (days) Proportion ?3 log 10 Reduction in AdV VL or to Undetectable at Nadir Minimum On-tx Last On-tx CDV-naive (n=23) -2.0 (-5.1, 0.5) - 1.8 (-5.1 , +2.1) 15 (3, 106) 57% (13/23) CDV-exposed (n=18) - 1.5 (-5.4 , +0.6) - 1.2 (-5.4 , +0.6) 15 (4, 77) 72% (13/18)

S29 Conclusion: The observed mortality rate was 40% for allo HCT pts with disseminated AdV disease in AdVise, which is lower than literature rates reported for this pt population (50-80%; Ison 2006, Sandkovsky 2014). BCV showed potent virologic activity in CDV-naïve and exposed pts with no new safety concerns. These preliminary data support expansion of the pilot portion to a definitive Phase 3 study. Introduction: The guidelines for immunization of hematopoietic stem cell transplant (HSCT) recipients recommend 3 doses of anti-pneumococcal conjugate vaccine (PCV) from 3-6 months after transplant, followed by a dose of polysaccharide 23-valent (PPV23) vaccine at 12 months in case of no chronic graft-versus-host disease (GVHD), or an additional PCV dose in case of GVHD. However, due to lack of long-term data, there is no recommendation for boosts after 12 months. Our goal was to assess the retainment of anti-pneumococcal antibodies in allogeneic HSCT recipients vaccinated 10 years ago. Materials (or patients) and methods: In 2009, the IDWP published the results of the IDWP01 trial that compared the immune response assessed one month after 3 doses of PCV7, started either at 3, or at 9 months after myeloablative HSCT 1 . Additionally, all patients received 1 dose of PPV23 at 12 or 18 months after transplant. All surviving patients had been assessed for anti-pneumococcal antibodies against the vaccine-serotypes 24 months after HSCT. This study was the basis of the current guidelines for anti-pneumococcal immunization after allogeneic HSCT. The present study included 30 surviving patients from the IDPW01, who were assessed for antibody levels against the 7 PCV7-antigens and against 2 of the PPV23antigens (pn1 and pn5), between 8.3 and 11 years after transplant, i.e. 6 to 9 years after the last assessment in the initial study. The mean age was 39 y (18-55), and 18/30 had acute leukemia. Only 7 had chronic GVHD (limited: 6, extensive: 1) and 2 had suffered a leukemia relapse. Eleven (37%) had received an additional dose of PPV23 at a mean time of 6.5 years (2-11 years) after transplant, according to local procedure.

The rates of persistent responses to all 7 antigens of PCV7 were 65.5% for an Ab cut-off of 0.15 mg/mL, and 40% for a cut-off of 0.50 mg/mL. Compared to the response rate at 24 months after transplant, these rates were not significantly decreased but showed important serotype-specific variability. Similar findings were observed for pn1 and pn5 antibody levels. Neither the recipient or donor age, donor type, source of stem cells, GVHD, nor the administration of an additional dose of PPV23 (given to 11/30 patients) influenced the maintenance of the response. The timing of the initial vaccination was the only parameter influencing the long-term response; patients who were vaccinated lately after HSCT (from 9 months) had a significantly better maintenance of the response than patients vaccinated early (from 3 months) after transplant. The 3 patients who were not responders at 24 months and who received an additional dose of PPV23 at 39, 40 and 72 months after transplant, respectively, did not respond.

Conclusion: In long-term HSCT survivors without severe chronic GVHD vaccinated against S pneumoniae according to the current guidelines, the specific immunity is not fully maintained a decade later. Patients, who received an additional PPV23 dose after 24 months post-transplant, do not seem to benefit from this boost. Boosts with PCV should be explored. So far, the optimal schedule of anti-pneumococcal vaccination in HSCT recipients after 12 months remains to be established.

References : 

Early cytomegalovirus reactivation -a potential factor for early robust T cell reconstitution and possibly a prognostic factor for aGvHD after HSCT P. R Many previous studies have shown that aGvHD puts patients at risk of CMV reactivation, most likely due to more intensive immunosuppression. However, recent studies and case reports also show that that CMV-R could be a risk factor for aGvHD.

Here, we studied the effect of CMV-R on T cell reconstitution in patients with and without aGvHD. Materials (or patients) and methods: 106 CMV R þ /D þ patients transplanted 2005-2013 in our institution were included in this study. All the patient samples were monitored for the CMV viral load (CMVpp65 expressing cells/400,000 leukocytes) and T cell reconstitution (CD3, CD4, CD8 and all available HLA specific CMV tetramers for each patient) within the first 100 days after HSCT. Patients were subdivided into five groups: group 1: No aGvHD but CMV-R (no-aGvHD-CMV-R), group 2: aGvHD after CMV-R (aGvHD-after-CMV-R), group 3: aGvHD before CMV-R (aGvHD-before-CMV-R), group 4: aGvHD but no CMV-R (aGvHD-no-CMV-R) and group 5: no aGvHD-and no CMV-R (no-aGvHD-no-CMV-R).

Results: The characteristics for onset of CMV reactivation and aGvHD in the different subgroups are provided in Table 1 . CD3, CD4, CD8 and CMV specific T cells were analyzed on day 50±10days after HSCT. In order to investigate the potential influence of CMV-R in the absence of aGvHD on T cell reconstitution, we compared the T cell numbers in the groups CMV-R þ /-subsequent aGvHD (i.e. group 1 þ 2) with group 5 (no-aGvHD-no-CMV-R). We found significantly more CD3 (P ¼ 0.021) and CD8 T cells (P ¼ 0.0057) in groups 1 þ 2 compared to group 5. There were no differences in T cells between the groups with CMV-R þ /subsequent aGvHD (i.e. groups 1 þ 2) compared to the groups with aGvHD þ /-subsequent CMV-R (i.e. groups 3 þ 4). Moreover, there were no differences in T cell numbers between the groups with aGvHD þ /-subsequent CMV-R (i.e. groups 3 þ 4) compared to group 5. To study the impact of CMV-R on T cell reconstitution in patients with subsequent aGvHD we compared the T cell numbers in group 2 (aGvHD-after-CMV-R) with group 1 (no-aGvHD-after-CMV-R). There were significantly more CD4 T cells (P ¼ 0.0041) and a trend for more CD3, CD8 and CMV-CTLs in group 2 (aGvHD-after-CMV-R) compared to group 1 (no-aGvHD-after-CMV-R). Subsequently, we compared the potential influences of CMV-R and of aGvHD on T cell reconstitution. We found significantly more CD3 (P ¼ 0.0138) and CD8 T cells (P ¼ 0.0125) in group 2 (aGvHD-after-CMV-R) compared to group 4 (aGvHD-no-CMV-R). Moreover, we studied the overall potential influence of CMV-R in the presence of aGvHD on T cell reconstitution. We found a trend for more T cells (CD3, CD4, CD8 and CMV-CTLs) in group 2 (aGvHD-after-CMV-R) compared to group 3 (aGvHDbefore-CMV-R).

In conclusion, patients with aGvHD after CMV-R had considerably more CD4 T-cells on day 50 compared to patients with CMV-R but no aGvHD and significantly more CD3 and CD8 T cells compared to patients with aGvHD but no-CMV-R. These results suggest that early CMV-R enhances overall T cell reconstitution which could be a potential risk factor for developing aGvHD after HSCT. Single and double CBT were used in 289 and 159 cases, respectively. TBI was part of the conditioning regimen in 263 cases (59%). In vivo T-cell depletion by ATG was used in 61% of patients. At least one PCR with a viral load 43 log/mL of blood was sufficient to define HHV-6 reactivation after the graft. The impact of HHV-6 reactivation on CBT outcomes has been studied as a time-dependent variable. 

In multivariate analysis, HHV-6 was independently associated with graft failure (HR: 1.44 Conclusion: Our study confirms that HHV-6 reactivation is a risk factor for graft failure in CBT recipients after myeloablative conditioning regimen. This result has to be confirmed prospectively and in the setting of reduced-intensity conditioning CBT. This paves the way also to test prospectively the indication of ganciclovir or foscarnet use as anti-HHV-6 prophylaxis in CBT recipients. Disclosure of Interest: None declared.

Introduction: Candida is the second more frequent cause of invasive fungal infection in haematological immunocompromised hosts, especially in the patients who undergo an haematopoietic stem cell transplantation (HSCT). The aim of this study was to analyse retrospectively the outcome of patients with Candida infections acquired in the first 100 days after allogeneic HSCT. Materials (or patients) Materials (or patients) and methods: In prospective study 173 alloHSCT recipients were included from Dec 2012 to Jul 2013. The median age was 34 y.o., males -54%. Most of pts had high-risk acute leukemia (70%). AlloHSCT from MUD were performed in 57%, MRD -24%, haplo -11%, MMUD -8%, predominantly with RIC (80%). EORTC/MSG 2008 criteria for diagnosis and response to therapy were used. Since 2011 active diagnostic strategy, including bronchoscopy with BAL, in pts with CT-scan lung lesions before alloHSCT has been introduced to the routine practice. ''Active IA'' is the IA diagnosed just before HSCT.

Results: Incidence of IA before alloHSCT was 22,5% (n ¼ 39/ 173). According to EORTC/MSG 2008 criteria 92% of pts had probable IA and 8% proven IA. The main sites of infection were lungs -95%, central nervous system -3%, and colon -3%, other localizations were observed mostly in a combination with lung involvement: sinuses -5%, spleen -3%, and liver -3%. Antifungal therapy before alloHSCT was administrated in 69% pts (Voriconazole -95%, other -5%) with the median duration of therapy -2 months. Complete response to antifungal therapy was registered in 10 (26%) pts, partial response or stabilization in 17 (43%), and ''active IA'' in 12 (31%) pts. After alloHSCT all pts received antifungal therapy with Voriconazole (first line -31%, continuation of treatment -43%, and secondary prophylaxis -26%). Median length of treatment was 166 days (37-394). Cumulative incidence of relapse or progression of IA after alloHSCT was 12,4% (n ¼ 6).

Relapse of underlying disease was the main risk factor for the relapse or progression of IA after alloHSCT (6% vs 33%, P ¼ 0,007). Progression of IA after alloHSCT was treated with Voriconazole 600 mg per day (n ¼ 1) and combination Vori þ Caspo (n ¼ 3). Relapse of the IA after alloHSCT was treated with Voriconazole 400 mg per day (n ¼ 2). No toxicity of the antifungal treatment was registered. Complete response was achieved in 4 pts, and stabilization -2. 12-weeks overall survival (OS) after the start of antifungal therapy was 67%. Two pts died with the progression of the underlying disease. 100days OS after alloHSCT was 70%, 1-year OS after alloHSCT was 57%. There was no significant difference in OS in pts with or without IA before alloHSCT.

Conclusion: Incidence of the IA before alloHSCT was 22,5%. Cumulative incidence of relapse or progression of IA after alloHSCT in pts with proven and probable IA before alloHSCT was 12,4%. Relapse of the underlying disease was the main risk factor for relapse or progression of IA after alloHSCT. Secondary prophylaxis with Voriconazole should be used in pts with IA before alloHSCT. Relapse or progression of IA after alloHSCT didn't impair OS. IA is not a contraindication for alloHSCT. Disclosure of Interest: None declared. PT. 12 patients had acute kidney injury which was managed conservatively without the need for renal replacement therapy. 3/6 survivors have normal LFT's, 2 patients have a residual mild increase in transaminases due to cGvHD, whereas 1 patient has a moderate increase in LFT's due to cGvHD. Conclusion: Busulphan based conditioning were the most important risk factor for VOD. Myelofibrosis had a strong trend towards causing VOD (P-0.06). Early intervention with defibrotide along with supportive management was able to completely resolve VOD in most of the cases and the 100 day mortality was only 3/13 (23%). Only 1 death was directly attributed to VOD and 1 death each due to sepsis and biopsy proven drug induced liver failure. Disclosure of Interest: None declared.

Validation of two new prognostic scores to predict nonrelapse mortality in patients undergoing reduced-intensity conditioning allogeneic hematopoietic cell transplantation P. Barba Introduction: In 2014, 2 new pretransplant predictive models of non-relapse mortality (NRM) for patients undergoing allogeneic hematopoietic cell transplantation (all-HCT) have been created based on modifications of the HCT Comorbidity Index (HCT-CI) and the EBMT score. The first model (HCT-CI/ age) consisted of the addition of an extrapoint for patients440 years to the HCT-CI (Sorror et al. JCO.2014 ). The other model developed by the ALWP of the EBMT combined 16 categories of the HCT-CI and the EBMT score into an integrated score (Versluis et al. Leukemia. 2014 ). None of these models have been validated in independent cohorts. Materials (or patients) and methods: We analyzed the predictive capacity of these new models and compared it with the HCT-CI and the EBMT score in a population of reducedintensity conditioning allo-HCT (allo-RIC) consecutively transplanted patients in 2 Spanish centers during an 11 year period (2000) (2001) (2002) (2003) (2004) (2005) (2006) (2007) (2008) (2009) (2010) (2011) . The scores of all models were calculated by a single investigator as originally defined. Risk-groups stratification was also performed as originally defined, except for the EBMT score in which patients were divided into low-(score 0-3), intermediate-(score 4-5) and high-risk (scores 6) according to percentiles 33, 66 and 100. For the HCT-CI/Age patients scoring 0 points (n ¼ 3) were grouped with those scoring 1-2 points. The predictive capacity of all models was calculated by means of the Harrell's c-statistic and were compared by calculating a z-score and p values from the estimated standard error.

Results: A total of 232 patients were included. Median age at HCT was 55 years (range 18-71). Most patients received allo-HCT from HLA identical sibling donors (n ¼ 170, 73%) mainly for acute myeloid leukemia and myelodisplastic syndromes (n ¼ 73, 32%). Median follow-up for survivors was 5.5 years (range 0.3-11) . The median HCT-CI/age and the ALWP model scores were 4 (range 0-14) and 5 (range 1-15), respectively. The median HCT-CI, EBMT scores were 3 (range 0-13) and 5 (range 1-7), respectively. Risk group distribution of patients according to each model is summarized in Introduction: Systemic inflammatory response syndrome (SIRS) is defined as an inflammatory state induced by infections or toxic damages. SIRS is diagnosed when two or more of the following criteria are met: body temperatureo36 1 C or438 1C, heart rate490 beats/minute, tachypnea420 breaths/minute or PaCo2o32 mmHg, leukocyteso4000 cells/ mm 3 or412000 cells/mm 3 or presence of410% immature neutrophils. The goal of this study was to assess the incidence of SIRS early after an allogeneic stem cell transplantation (alloSCT) (from day 0 to hematopoietic recovery) and evaluate whether SIRS may influence the occurrence of acute S36 graft-versus-host disease (aGVHD) and non-relapse mortality (NRM Introduction: Pure red cell aplasia (PRCA) after allogeneic hematopoietic stem cell transplantation (HSCT) is a relatively rare complication after major ABO incompatible allogeneic transplant. Although reduced intensity transplant was a potential risk factor for PRCA, the impact of stem cell source has not been fully evaluated. We conducted a retrospective risk factor analysis for developing PRCA in 163 major ABO incompatible transplant including 106 cord blood transplantation (CBT) and 74 reduced-intensity conditioning.

Materials (or patients) and methods: We reviewed the medical records of 668 adult patients who underwent allogeneic HSCT for the first time at the Toranomon hospital from 2006 to 2013. PRCA after HSCT was defined as anemia with low reticulocyte counts (o1%) in peripheral blood for more than 60 days after transplantation in association with neutrophil engraftment and a lack of erythroid precursors in bone marrow. Results: One-hundred and sixty-three patients with major or bi-directional ABO incompatibility who achieved neutrophil engraftment and survived more than 60 days after HSCT were included in this study. Seventy four patients received reducedintensity conditioning, 106 patients underwent CBT, 39 did bone marrow transplantation (BMT) and 18 did peripheral blood stem cell transplantation (PBSCT). Reticulocyte engrafted (reticulocyte 4 ¼ 1%) in 160 patients with a median time of 29 days after HSCT during this study, which was significantly longer after CBT compared to BMT/PBSCT (31 days vs. 26 days, P ¼ 0.008). In 9 patients, reticulocyte count remained o1% beyond 60 days post-transplant, 5 of whom were diagnosed as PRCA with a cumulative incidence of 3.1%. PRCA was not observed in CBT patients, and the cumulative incidence of PRCA was significantly lower after CBT compared to BMT/PBSCT (0% vs. 8 is the only curative treatment in FA. Immune reconstitution after HSCT is increasingly recognized as a critical determinant of morbidity and mortality in HSCT. The aim of the study was to better understand the kinetics of immune reconstitution in children with FA who underwent allogeneic HSCT after a Fludarabine based reduced intensity conditioning regimen.

Materials (or patients) and methods: In this study, lymphocyte subgroups of children who underwent HSCT were evaluated before HSCT and 1, 3, 6, 12, and 24 months after HSCT. Children with FA (n:21) comprised the study group and children with non-malignant diseases (n:36) comprised the control group. In addition to classical lymphocyte subgroups; activated T lymphocyte subgroups including CD8/57( þ ), CD8/56( þ ), CD3/HLA-DR( þ ), CD4/25( þ ), CD4/ 28( þ ) T lymphocytes were evaluated in study and control groups.

Results: When absolute levels of lymphocyte subgroups were evaluated in children with FA, CD3( þ ) lymphocyte count returned to pre-HSCT levels at 12 months. CD4( þ ) T lymphocyte count reached to pre-HSCT levels at 24 months. CD8( þ ) T lymphocyte count returned to pre-HSCT levels at 3 months. CD19( þ ) B lymphocyte count turned to pre-HSCT levels at 3 months. CD4/8 ratio returned to pre-HSCT levels at 12 months. CD16/56( þ )CD3( þ ) NK-T and CD16/56( þ )CD3(-) NK lymphocytes returned to pre-HSCT levels within 1 month after HSCT. Among HSCT related complications; acute GvHD developed in 2/21 (9.5%) children in study group and in 10/36 (27.7%) children in control group. On the other hand, chronic GvHD developed in 1/21 (4.7%) children in study group and in 5/36 (13.8%) children in control group. When specific subgroups reflecting lymphocyte activation were evaluated in study and control groups; activated CD8/ 56( þ ) NK lymphocyte and CD8/57( þ ) T lymphocyte count returned to pre-HSCT levels at 3 months in both groups. While activated CD3/HLA-DR( þ ) T lymphocyte count returned to pre-HSCT levels at 1 months in both groups; activated CD3/ HLA-DR( þ ) T lymphocyte count was higher at 1, 6, and 12 months in control group. Activated CD4/25( þ ) T lymphocytes returned to pre-HSCT levels at 12 months in study group and those returned to pre-HSCT levels after 24 months in control groups. CD4/25( þ ) activated T lymphocyte count was higher at 24 months in study group (pr0.05). CD4/28( þ ) activated T lymphocytes reached pre-HSCT levels at 12 months in control group and at 24 months in study group. Besides, CD4/28( þ ) activated T lymphocyte count was higher at 6 and 12 months in control group (pr0.05).

In this study, we show that the kinetics of recovery of the lymphocytes subgroups in children with FA after HSCT follows those patterns also described for children with other diseases: early recovery of NK cells (1 month), followed by effector cytotoxic T cells (3 months) and B cells (3 months) , and finally, CD4( þ ) T-helper cells (24 months). High levels of CD3/DR( þ ) activated T lymphocyte count at 1, 6, and 12 months and high levels of CD4/28( þ ) T lymphocyte count at 6 and 12 months in control group are attributable to the high frequencies of acute and chronic GvHD in control group than those of study group. Disclosure of Interest: None declared. Introduction: Although HLA haploidentical HSCT has been largely employed in children with life-threatening nonmalignant disorders, the survival of patients given this type of allograft has been reported to be inferior to that of patients transplanted from a compatible unrelated volunteer (UV). We implemented a novel method of ex vivo T-and B-cell depletion based on the selective elimination of ab þ T cells and B cells.

We herein report an update of 31 children with non-malignant disorders who were given this type of allograft. Materials (or patients) and methods: Twenty-two patients were males and 9 females, median age at HSCT being 3.5 years (range 0.3-13.2) . Nine patients had severe combined immunedeficiency (SCID), 8 Fanconi Anemia (FA), 4 Severe Aplastic Anemia (SAA), 2 Thalassemia Major, 2 Hemophagocytic Lymphohistiocytosis (HLH) and 1 each Immunedeficiency with Polyendocrinopathy Enteropaty X-linked (IPEX), Kostmann Syndrome, Hyper IgE Syndrome, Osteopetrosis, Swachmann-Diamond Syndrome and Congenital Amegakaryocytic Thrombocytopenia (CAMT). All patients were transplanted from 1 of the 2 parents (21 from the mother and 10 from the father), the median number of CD34 þ and ab þ T cells infused being 22.57 x 10 6 /kg and 4 x 10 4 /kg. The original conditioning regimen consisted of treosulphan and fludarabine (FLU) þ thiotepa in 13 (9 SCID, 1 IPEX, 1 CAMT, 1 Kostmann Syndrome and 1 Swachmann-Diamond Syndrome), FLU and cyclophosphamide þ single dose TBI in 12 (8 FA and 4 SAA) and busulphan, FLU and thiotepa in 6 (2 Thalassemia, 2 HLH, 1 Osteopetrosis and 1 Hyper IgE Syndrome). No patient received immunosuppression after HSCT. All patients received Fresenius rabbit ATG (4 mg/kg/day) on days -5 through -3 before allografting and rituximab (200 mg/m 2 ) to prevent EBV-related PTLD on day -1.

Results: All patients but 6 engrafted, the median time to reach neutrophil and platelet recovery being 13 days (range 9-23) and 9 days (range 7-40), respectively. The 6 patients (2 with SAA and 1 each with thalassemia, FA, HLH and Osteopetrosis) who had primary graft failure were successfully re-transplanted (2 from the same parent, 3 from the other relative and 1 from an 1-HLA locus disparate UV Figure 1a ). In the latter group, aGVHD started median 21 days after HSCT and was Zgrade III and steroid refractory in 3/6 cases. A higher aGvHD rate was also observed in male patients who received a graft from a female donor (F-4M mismatch) compared to all other sex-matches (5/9 vs. 2/16, P ¼ 0.02). In multivariate cox regression analysis, both omission of MTX and F-4M mismatch were associated with aGvHD occurrence. In line with this, the protective effect of MTX was most pronounced in the subcohort of patients with a F-4M mismatch: grade II-IV aGvHD occurred in 5/5 F-4M mismatched non-MTX recipients whereas 0/4 F-4M mismatched MTX recipients developed aGvHD (P ¼ 0.005). ALL relapse occurred in 9/25 patients (36%). We did not observe differences in relapse rate between non-MTX recipients and MTX recipients (4/11 vs. 5/14, P ¼ 0.97, Figure 1b) . Although non-MTX recipients were more often transplanted for ALL in 2 nd remission, pre-transplantation minimal residual disease (MRD) status did not differ between the groups. In line with earlier reports, MRD positivity was the most important risk factor for ALL recurrence. In MRD positive patients, the omission of MTX and aGvHD occurrence did not prevent relapse after HSCT.

Conclusion: MTX prophylaxis reduced the occurrence of severe aGvHD, without compromising relapse free survival in pediatric ALL patients after T cell replete HLA-identical bone marrow transplantation. Prevention of aGvHD reduces morbidity and the need for high-dose immunosuppressive agents, allowing for alternative immunotherapy-based therapeutic interventions in individuals at high risk for disease recurrence. Introduction: Hematopoietic stem cell transplantation (HSCT) has contributed to improved outcome in childhood acute leukemia (AL). However, post-HSCT relapse is associated with a dismal prognosis and its optimal treatment remains unclear. We aimed to compare patients' related factors and treatment strategy, in case of relapse or progression post-allogeneic HSCT in children with AL in a recent ten-year period. Materials (or patients) and methods: A total of 334 children who received a first allogeneic HSCT for ALL or AML from January 2000 to December 2009 experienced a relapse or progression thereafter. They were treated in the 33 centers of the SFGM-TC, among them 279 cases were analysable. Primary endpoint was overall survival (OS) after diagnosis of relapse or progression post first HSCT whatever the treatment post relapse was. . Lymphocyte (sub)populations were analysed frequently post transplantation by flow cytometry. CD3 þ T-cell recovery was defined as appearance ofZ100 cells/mL, B-cell and NK cells recovery asZ50 cells/mL.

The median active ATG serum concentration at time of T-cell or B-cell reappearance was 0.03 AU/mL, the maximum level was 1 AU/mL. In 25% of the patients, NK cells re-appeared when the median active ATG level was higher (0.16 AU/mL) up to a maximum level of 7 AU/mL (see Figure) . For alemtuzumab the median concentration at T-cell recovery was 0.008 mg/mL, max. 0.2 mg/mL, but in 80% of the patients NK cells reappeared at a higher concentration up to 2.2 mg/mL. For both drugs, T-cell recovery was significantly correlated with the serum level of the drug (Po0.001 for ATG, and Po0.01 for C1H), whereas a significant correlation was absent for NK cells.

Conclusion: ATG and alemtuzumab are both able to deplete T-, B-, and NK cells. For both drugs, the exposure is highly variable even in patients with an equal weight receiving the same dose. Here, we report that T-cell recovery is closely related with serotherapy exposure. NK-cells are the first cells that re-appear post HSCT . All of the patients received the same RIC regimen based on the use of fludarabine in combination with melphalan and antithymocyte globulin (ATG). Prophylaxis against GvHD was achieved via cyclosporine and methylprednisolone.

Results: All patients were engrafted. The median times to neutrophil and platelet engraftments were 11 days (range: 8-33), and 22 days (range: 10-67), respectively. Patients underwent HSCT from HLA matched sibling donors (n ¼ 9), full matched other related donors (n ¼ 6), unrelated matched donor (n ¼ 1) and unrelated mismatched donor (n ¼ 3). The source of graft were peripheral blood (n ¼ 11), bone marrow (n ¼ 6) and cord blood (n ¼ 2 Materials (or patients) and methods: Twenty-nine patients with MIOP were treated by HSCT at the University childreń s hospitals in Paris (n ¼ 9, since 2008) and Ulm (n ¼ 20, since 1998). MIOP was caused by mutations in TCIRG1 (n ¼ 20), CLCN7 (n ¼ 5), SNX10 (n ¼ 1), RANK (n ¼ 2), and FERMT3 (n ¼ 1); age at transplant was between 2 and 72 months (median 6 months); donors were haploidentical family donors (n ¼ 16), MUD (n ¼ 7), phenoidentical relatives (n ¼ 4), and MSD (n ¼ 2). Thiotepa and serotherapy was added to the Busulfan and Fludarabine based regimen in all patients with donors other than MSD. Results: All but 6 patients showed a primary and sustained engraftment; 4 of 6 patient, who rejected their haploidentical graft, could be rescued by a second graft from the second parent. Severe VOD, which had to be treated by ascites puncture, was seen in 2 patients only. Only one case of GvHD4 12 and no case of chronic GvHD was observed. Cause of death in 4 patients were liver toxicity in conjunction with CMV and fungal infection after prolonged aplasia (MUD transplant at age of 1 month) and complications in conjunction with engraftment failure ( Introduction: BMT is the only proven curative treatment available for haemoglobinopathies. However, the number of patients who can benefit is seriously restricted by the lack of HLA-matched related donors not suffering from the condition and the limited number of unrelated donors available for the ethnic groups in which these conditions are prevalent. In order to expand the donor pool, haploidentical transplantation with a post-infusion of stem cells cyclophosphamide approach has been developed for young adults, but whilst well tolerated it has resulted in relatively high rates of rejection and the need for a prolonged period of immunosuppression 1 . Materials (or patients) and methods: 12 consecutive parental haploidentical transplants (11 for sickle cell disease and one for b halassaemia major) were performed at St. Mary's Hospital, London, from June 2013 to November 2014. The median age was 9.5 years of age (range 3 to 14). All patients lacked a suitable HLA-matched related donor and an unrelated search had not identified a 10/10 or 9/10 donor. Endogenous haemopoieis was suppressed with hypertransfusions, hydroxycarbamide 30 mg/kg and azathioprine 3 mg/kg for at least two months pre-transplantation. The conditioning included fludarabine 150 mg/m 2 , thiotepa 10 mg/kg was added, cyclophosphamide 29 mg/kg, TBI 2 Gy and ATG (Thymoglobulin) 4.5 mg/kg. GvHD prophylaxis was provided with cyclophosphamide 50 mg/kg on days þ 3 and þ 4, MMF and sirolimus. The minimum follow-up was 40 days post-transplantation and half of the patients are 4150 days post-transplantation and have completed all treatment. The source of stem cells was G-CSF primed bone marrow in all cases, aiming 48 x 10 8 TNC/kg. Results: All patients engrafted, though one patient subsequently suffered secondary graft failure following macrophage activation syndrome and died. The median neutrophil engraftment was 17 days (range 16 to 19). All 11 surviving patients are cured from the manifestations of the original disease. None of the patients suffered VOD, though infectious complications occurred at a higher rate than seen for related transplants for the same conditions at our institution. Five patients had no evidence of acute or chronic GvHD. Five patients developed stage 1 acute GvHD (median presentation day þ 36, range 21 to 66) responding to topical steroids; one patient suffered skin GvHD stage 3 on day þ 35 and one patient gut GvHD stage 4 on day þ 24, both treated with MSC. All patients responded to first line treatment with no recurrence of disease. All patients but one achieved 490% donor fraction both in whole blood and T cells. One patient requires immunosuppression beyond day þ 180 with stable 59% donor fraction in whole blood and 10% in T cells. Introduction: Bone marrow transplantation (BMT) offers a definitive cure for thalassemia in over 90% of low-risk children with a matched related donor. Many centers currently incorporate thiotepa in busulfan-or treosulfan-based BMT regimens for thalassemia. This combination, however, may permanently impair fertilty in most patients. In the era of increasingly effective supportive care in which many thalassemia patients may have children, this is concerning. Very longterm follow up studies have shown how the standard Bu-Cy regimen may be associated with birth rates comparable to the control population (La Nasa et al. Blood 2013). This study retrospectively compares BMT outcomes in two groups of low risk patients (defined as livero2cm and ageo12y) with severe thalassemia (ST) (defined as a thalassemia syndrome with spontaneous hemoglobino7 g/dL), receiving oral busulfan (14 mg/kg), cyclophosphamide (200 mg/kg) and either thiotepa (10 mg/kg) (TT-Bu-Cy) or rabbit ATG (Fresenius 16 mg/kg or Thymoglobulin 4 mg/kg total doses from day -12 to -10) (ATG-Bu-Cy) as preparative regimen. Standard cyclosporine and short-term methotrexate plus low dose methylprednisolone were used for GvHD prophylaxis. Materials (or patients) and methods: This is a retrospective multicentre comparative study of the safety and efficacy of substituting TT with ATG in low-risk ST patients undergoing matched-related BMT. Between January 2009 and July 2013, a group of 35 patients were transplanted after conditioning with TT-Bu-Cy, while between August 2013 and July 2014, 30 patients were conditioned with ATG-Bu-Cy. Results: The actuarial overall survival in the TT-Bu-Cy and ATG-Bu-Cy groups is 91.3% and 93.4%; thalassemia-free survival is 88.4% and 93.4% at a median follow up of 23.5 and 3.1 months respectively, with no statistically significant difference by logrank test. Conclusion: Substituting thiotepa with ATG in the standard Bu-Cy context seems safe and effective. Higher fertility rates are expected for patients on the ATG-Bu-Cy regime. Disclosure of Interest: None declared. Materials (or patients) and methods: Data has been collected retrospectively from 234 infants with SCID who received transplants at 8 different centers over a 20-year period . The differences between groups were compared by using Chi-Square or Fisher's Exact test, where appropriate. A p value of less than 0.05 was considered statistically significant. Results: 145 boys (62%) and 89 (%38) girls with SCID whose ages ranged between 0.25-176 months (median 5 months) at the time of diagnosis were transplanted. Parental consanguinity was identified in 171 (76%) of 224 infants. 72% of the patients had received BCG vaccination before diagnosis. B þ and B-phenotypes were detected in 41.8% (n ¼ 97) and 51.3% (n ¼ 119) respectively, while ADA deficiency was recognized in 4.7%, RD (reticular dysgenesis) in 2.2% of the cases. RAG1, JAK3, RAG2 and Artemis defects were the leading genotypes among the patients with molecular diagnosis (42.3%; n ¼ 101). Out of 234, 114 patients (48.7%) had either a matched sibling or a family donor, while 81 (34.6%) and 39 (16.7%) children received haploidentical (MMFD) and MUD transplants respectively. The HSCT source was bone marrow in 113 (48.3%), peripheral blood in 89 (38%) and cord blood in 32 (13.7%) of the patients. Among a total of 24 (10.25%) retransplants, 18 received a second transplant while 6 children received a boost only. 153 children survived, 80 died and 1 was lost to followup. The overall survival rate was 65.7% over a 20 years period. It increased from 54% (1994) (1995) (1996) (1997) (1998) (1999) (2000) (2001) (2002) (2003) (2004) to 69% (P ¼ 0.052) during the latter 10 years (2005-2014) and even to 72,9% during the last 4 years (2010-2014). The survival rates with relation to donor types were as follows: MSD ¼ 85.7% (n ¼ 77), MRD ¼ 70.3% (n ¼ 37), haploidentical 47.5% (n ¼ 80) and MUD 59% (n ¼ 39). Age at diagnosis significantly (r5 months or 45months) influenced the survival rate of the patients (P ¼ 0.002). Immunophenotype did not seem to have an effect on survival rate and immunoglobulin (Ig) requirement following HSCT did not differ between B þ vs B-phenotypes (P4 0.05). Conclusion: This is the first multicenter study with the largest data obtained from SCID patients transplanted in Turkey. The median age at diagnosis was 5 months, B-phenotype and RAG were the most common among other defects. Age at diagnosis (45 months), and donor type (haploidentical) (Po0.01) were two major factors significantly related to poor outcome. Expanded donor availability, advances in intensive care facilities, diversity of transplantation centers and specialized teams are among major factors contributing to the longterm outcomes of HSCT. However, Newborn Screening is of paramount importance in ensuring early diagnosis and timely transplantation thus improving the survival of SCID patients in Turkey. Disclosure of Interest: None declared.

Oral session: Stem cell mobilization & regenerative medicine Haematology, BMT Unit, Hospital Clínico Universitario Virgen de la Arrixaca, IMIB, University of Murcia, 2 Surgery, Hospital Clínico Universitario Virgen de la Arrixaca, 3 Haematology, BMT Unit, Hospital Clínico Universitario Virgen de la Arrixaca, IMIB, University of Murcia, Murcia, Spain Introduction: Amniotic membrane (AM) is a non-tumorigenic tissue attributed with various biological properties (low immunogenicity, anti-inflammatory, anti-fibrotic and antimicrobial effects) related to its ability to synthesize and release cytokines and growth factors. AM, that is usually discarded after birth, is in our experience an easily obtained tissue which processing, storage and management can be included in the daily routine of the Cryobiology Laboratory. MA can be used as a ''biologic bandage'' for healing management of chronic wounds in diabetic and non-diabetic patients. In our Hospital there is an ongoing clinical trial to study the use of AM to improve epithelialization (NCT01824381). Here, we describe the results obtained prospectively after the compassionate use of cryopreserved human amniotic membrane allografts in 6 patients with chronic diabetic foot ulcers. Materials (or patients) and methods: AM was obtained from healthy mothers who had programmed an elective caesarean operation for obstetric reasons after they signed the informed consent. Donors were screened by reviewing their medical records and by performing laboratory test to discard transmissible disease agents. AM processing was done under sterile conditions in the GMP facility; the process involves: washing the AM to eliminate blood traces, cut AM into several fragments, sew each fragment on an impregnated dressing sheet and introduce them on cryopreservation bags adding the cryoprotectant solution based on human albumin, TC-199 medium and 9% DMSO. The AM fragments were storaged at -1961C and delivered once we the negative viral serology of the donor was confirmed 3 months later. Cryopreserved AM was applied to six consecutive patients with diabetic foot ulcers under a compassionate use program of the Diabetic Foot Unit from May to November 2014. Wound size reduction and rates of complete healing were evaluated. Results: Patients were aged between 45 and 65 years. They were affected by grade II diabetic foot ulcers in the Wagner Ulcer Classification Scale that had failed previous treatments for periods between 2 months and 4 years. AM was applied weekly or every ten days until complete healing or partial reepithelialization of the ulcers. A median of 5 (3-8) cryopreserved AM fragments were applied for an average treatment period of 45 days. The mean size of the ulcer was reduced by 76%.l wreduced in size by 76%. At last follow-up, 4 out 6 patients have total epithelialization of the ulcer. No adverse events related to its application were observed. Conclusion: Our results show that the application of cryopreserved amniotic membrane is a feasible and safe treatment in complex diabetic foot ulcers. More rapid healing may decrease clinical operational costs and prevent long-term medical complications. Furthermore, the treatment achieves re-epithelialization of long evolution wounds that were not reached with conventional therapies. Disclosure of Interest: None declared.

Bone as a regulator of human hematopoietic stem cell (HSC) trafficking: Study of biochemical markers of bone remodeling and angiogenic cytokines during HSC mobilization, in patients with lymphoma and myeloma P. Tsirkinidis 1,* , E. TERPOS 2 , G. BOUTSIKAS 1 , A. PAPATHEO-DOROU 3 Introduction: Bone is not considered just a structural, supportive tissue of bone marrow, but an HSC-niche regulator. Data regarding the role of bone turnover in HSC mobilization in humans are scarce. The aim of the present study was to study bone remodeling and vessel equilibrium during HSC mobilization in lymphoma and multiple myeloma patients. Materials (or patients) and methods: Forty-six patients (32 lymphoma and 14 multiple myeloma) were studied. Serum samples were collected at two time points: before mobilization (pre-mobilization sample) and on the day of HSC collection, which coincided with the peak circulating CD34 counts (collection sample). In 19/46 patients, 3 additional serum samples were collected, between mobilization and collection. The following molecules were measured by ELISA in patients' sera: 1) bone resorption markers: carboxyterminal telopeptide of collagen type 1 (CTX), aminoterminal telopeptide of collagen type 1 (NTX), tartrate resistant acid phosphatase isoform 5b (TRACP-5b), 2) bone formation markers: bone alkaline phosphatase (BALP), osteocalcin (OSC), osteopontin (OPN), 3) the osteoblastogenesis inhibitor dickkopf-1 (DKK-1) 4) the osteoclastic regulators: receptor activator NF-kB ligand (RANKL), osteoprotegerin (OPG), 5) angiogenic cytokines: angiopoietin-1 (ANGP1), angiopoietin-2 (ANGP2), angiogenin (ANG). Values were compared with non-parametric methods. Patients who had either a collection of CD34 þ cells o2.0x10 6 /kg, or a circulating CD34 count peak o20/mL were considered poor mobilizers. Results: The comparison of the aforementioned molecules between the pre-mobilization and collection samples revealed: BALP (P ¼ 0.000) and OPN (P ¼ 0.049) increased significantly, while OSC, a marker of bone turnover, and DKK-1 decreased significantly (P ¼ 0.000 and P ¼ 0.041, respectively). These findings reveal a significant increase of bone formation during mobilization. RANKL (P ¼ 0.000) and OPG (P ¼ 0.000) increased significantly, leading to an increase of RANKL/OPG ratio (P ¼ 0.000), consistent with osteoclastic activation. However, there was no evidence of increased osteoclastic activity, as CTX decreased significantly (P ¼ 0.026), while both NTX and TRACP-5b did not change. ANGP-1 showed a dramatic reduction (P ¼ 0.000), while ANGP-2 increased (P ¼ 0.000), resulting in a significant decrease of the ANGP-1/ANGP-2 ratio, a finding indicating vessel destabilization during mobilization. These results were further supported by the intermediate measurements, which showed a straightforward alteration of bone metabolism early in HSC mobilization. Poor mobilizers had significantly higher CTX levels both at premobilization (P ¼ 0.004) and collection samples (P ¼ 0.001), higher NTX levels at collection (P ¼ 0.02), lower ANGP-1 premobilization (P ¼ 0.004) and higher OSC at collection (P ¼ 0.000) compared to good mobilizers. Thus, CTX, NTX S48 and ANGP-1 pre-mobilization levels may serve as reliable predictors of poor mobilization. Conclusion: Our study showed for the first time in humans, that bone plays a dynamic role during HSC mobilization: Bone formation and vessel destabilization are the two major events and osteoblasts seem to be the orchestrating cells during this process. Osteoclasts are stimulated, but not fully active. Moreover, some of these markers may identify poor mobilizers. Disclosure of Interest: None declared. Introduction: Allogeneic bone marrow transplantation is a curative treatment for leukemia and genetic disorders. Although some patients are cured of their underlying illness, they are at risk of developing potentially fatal graft-versus-host disease (GvHD). A recent phase 3 randomized trial conducted by the Canadian BMT Group (N ¼ 230) comparing the impact of G-CSF mobilized peripheral blood (PB) to bone marrow (BM). A representative group of 85 donor samples were evaluated and identified that high concentrations of CD56 bright NK cells in the donor product is strongly associated with a lack of development of both acute and chronic GvHD (odds ratio 0.11 ; P ¼ 0.0002) in G-CSF stimulated BM. We found that the CD56 bright NK population was CD335 (NKp46) positive with comparable expression of CD337 in both sub-populations consistent with regulatory NK cells (NK reg ). We hypothesized that alternate strategies utilizing the mobilizing agent, Plerixafor, may enrich for CD56 bright NKp46 cells in PB thus reducing the risk of GvHD and circumventing the need of harvesting donor cells from the BM. Materials (or patients) and methods: We performed a pilot study to examine the impact of Plerixafor, that blocks the SDF-1 CXCR4 interaction, in mobilizing CD56 bright NK reg cells in the PB and BM, to determine the whether we can optimize a donor source with the highest potential for post transplant tolerance, resulting in a lack of both aGvHD and cGvHD. We recruited 9 healthy adult human volunteers. Five subjects received one dose of Plerixafor at 240 micrograms/kg/day and PB and BM samples were harvested prior to, at 4, and 24 hours after Plerixafor administration. The subsequent 4 participants each received 4 daily doses of granulocyte colony stimulating factor (G-CSF) at 5 micrograms/kg/day for 4 days, followed by a dose of Plerixafor on day 5. PB and BM samples were collected prior to G-CSF, after G-CSF, and at 4 and 24 hours after Plerixafor administration. We used multi-parametric flow cytometry to identify CD56 bright NK cells. The phenotype of CD56 bright NK cells was CD56 bright CD335 þ (NK46 þ ) CD3 -CD16perforingranzyme B -. Results: We observed a significant rise in PB nucleated cells at both 4 and 24 hours after Plerixafor administration with a peak increase at 4 hours. Plerixafor alone induced a significantly higher proportion of CD56 bright NK reg cells in PB when compared to G-CSF for 4 days (2.5 vs. 0.94 fold, P ¼ 0.041) and G-CSF for 4 days followed by Plerixafor treatment (2.5 vs. 0.43 fold, P ¼ 0.011). Moreover, cells collected 4 hours after Plerixafor alone had a significant increase in the proportion of CD56 bright NK reg cells (2.5 vs. 0.89 fold, P ¼ 0.031) in PB compared to BM.

This study suggests that Plerixafor alone is able to mobilize a high number of CD56 bright NK reg cells in PB harvested after 4 hours. We were not able to increase that population in a similar manner by using G-CSF alone or G-CSF plus Plerixafor. These findings suggest that allogeneic donor mobilization of peripheral blood may give a product with a low rate of GvHD potentially superior to G-CSF stimulated marrow and warrants further testing in a randomized clinical trial. Disclosure of Interest: None declared. Introduction: Only small studies have compared performance of the novel Spectra Optia apheresis system (Terumo BCT) with the widely used COM.TEC (Fresenius Healthcare) and the late Cobe Spectra (Terumo BCT) device in allogeneic stem cell collections. Our collaborative working group compared performance data of all 3 devices from two collection centers to analyze device-as well as center-specific performance parameters. Materials (or patients) and methods: We analyzed 5288 firstday apheresis collections in G-CSF-stimulated healthy donors that were performed in Cologne (CGN) using Spectra Optia MNC (n ¼ 1818), Cobe Spectra MNC (n ¼ 877), or COM.TEC (n ¼ 1657) and in Dü sseldorf (DUS) using Spectra Optia MNC (n ¼ 63) or Cobe Spectra MNC (873). Peripheral blood and product samples were analyzed in center-specific laboratories. Results: In both centers, irrespective of the apheresis device, similar yields of CD34 þ cells per kg donor bodyweight were reached. In CGN, collection rates (collected CD34 þ cells per kg donor bodyweight per CD34 þ cell count in peripheral blood before apheresis) were similar between the three apheresis systems. The continuously collecting Cobe Spectra yielded more CD34 þ cells over time (CR per h: 0,034 ± 0,009) than the discontinuously collecting apheresis systems 030±0, 009, Po0, 001) and Spectra Optia MNC (0,025 ± 0,008, Po0,001). The COM.TEC collected CD34 þ cells with highest efficiency (CE2: 63 ± 15% vs 61 ± 16%, Po0,001 and 55 ± 15%, Po0,001 for Spectra Optia MNC and Cobe Spectra, respectively). In comparison to Cobe Spectra (178±49 min), procedure times were longer when using COM.TEC (199±50 min, Po0, 001) or Spectra Optia MNC (237±58 min, Po0,001) systems. The product purity, measured as percentage of MNC was highest in products collected with the Spectra Optia MNC (79 ± 15% vs 76 ± 10%, Po0,001 vs 75 ± 14%, Po0,001 for Cobe Spectra and COM.TEC, respectively). The relative differences between Cobe Spectra and Spectra Optia collection performance parameters were similar in the DUS apheresis center: shorter procedures with higher CR per h for Cobe Spectra, higher MNC purity in Spectra Optia MNC products. Absolute results differed between the two centers. Conclusion: The highest collection efficiency (CE2) was seen when the COM.TEC device was used. However, this was accompanied by low product purity. Compared to any other apheresis device that has been analyzed in this study the novel Spectra Optia allows collection of higher MNC purity apheresis products from allogeneic donors. However, this was associated with a significant prolongation of procedure timethe major disadvantage of this device. The Cobe Spectra collects more cells over time by allowing higher mean inlet flow rates despite inferior collection efficiency. Therefore, a continuous collection procedure on the Spectra Optia device that allows high inlet flow rates would be an ideal collection setting in terms of collection efficiency and collection rate per unit of time. Disclosure of Interest: None declared. Introduction: Our group has previously established the safety and effectivity in terms of CD34 þ cell recovery and viability of the automated washing of cryopreserved hematopoietic progenitors (HP) with the Sepax â device (Biosafe) using normal saline supplemented with 2.5% albumin (NSA), as well as the absence of infusion-related events and of a negative impact on engraftment dynamics (Sánchez-Salinas et al. 2012 ). In our present study we compare this solution with a ready-to-use free of human derived components solution: 6% hydroxyethyl starch 130/0.4 in 0.9% sodium chloride injection (Voluven â , Fresenius Kabi). Materials (or patients) and methods:: 411 peripheral blood HP units apheresis cryopreserved using autologous plasma plus 9% DMSO corresponding to 158 autologous peripheral blood HP transplants were studied. After rapid thawing in a water bath at 371C, an automatic wash with the Sepax (2 washes cycle) was performed using either NSA (229 units) or Voluven (182 units). Nucleated cell levels determined by an hematology analyzer, flow cytometry CD34 þ cell counts, Trypan Blue cell viability and granulocyte macrophage (GM-CFU) and erythroid (E-BFU) colony forming cell cultures were performed on aliquots collected prior to and after the washing technique. Statistical analysis was performed using descriptive statistics and a simple-measures ANOVA. Results: The mean total nucleated cell (TNC) and CD34 þ cell recovery was 75,12% ±14,66, and 113,18% ±57,76 respectively for NSA and 79,08% ± 14,75 and 110,02% ± 44,02 for Voluven. The mean GM-CFU and E-BFU cell recovery was 163,80% ± 152,64 and 144,62% ± 176,61 respectively for NSA and 187,59% ± 232,48 and 141,82% ± 148,75 for Voluven. The mean viability recovery was 102.02 ±17,79 for NSA and 101,59 ±15,04 for Voluven without differences between both solutions (p 0,795). There were no significant differences between both solutions in none of this parameters In spite of the TNC significant loss (p o0,001), there were no significant differences between the pre and post-washing CD34 þ cell numbers (P ¼ 0,146), GM-CFU (p 0,051), E-BFU (p 0,952) or viability (p 0,537). In contrast with the 40% of untoward reactions recorded in our historical data of 226 DMSO containing cell infusions, we observed just three adverse effect with the washed cells with Voluven (1,6%) and none with NSA. One patient experienced a epileptic fit related to the infusion speed, another two suffered grade 1 nausea and transient hipotension respectively. Median time to neutrophil engraftment (4500 cells/mL) and platelet engraftment (420.000 cells/mL) for NSA were 11,29 ± 3,32 and 13,19 ± 6,8 days respectively and 12,02 ± 3,01 and 14,77 ± 14,87 for Voluven. When comparing both solutions, there were no significant differences in neutrophil engraftment (p 0,169) or in platelet engraftment (p 0,376). Conclusion: Both NSA and Voluven are equally effective for washing cryopreserved HP, ensuring a good CD34 þ cell recovery and preserving their viability and engraftment potential. Both solutions avoid the DMSO infusion related adverse events. As so, Voluven constitutes an excellent alternative free of human-derived products and ready-to-use solution to our previous NSA standard washing solution. References: Sánchez-Salinas A. et al. Transfusion. 2012 Nov; 52(11) :2382-2386 Disclosure of Interest: None declared. Introduction: Ruxolitinib is a JAK inhibitor that was recently approved for treatment of primary and secondary MF and shows impressive symptom control by suppression of inflammation. Ruxolitinib is also a promising drug for treatment of acute and chronic GvHD. The immune-modulatory effects of ruxolitinib are at least in part due to an inhibitory effect on dendritic cell biology (Heine et al., Blood 2014) . Dendritic cells (DCs) are important antigen-presenting cells. Upon antigen contact they migrate into the draining lymph nodes to prime T cells. The aim of this study was to define the impact of the JAK inhibitor ruxolitinib on DC migration. Materials (or patients) and methods: CD14 þ cells isolated from human buffy coats were differentiated for seven days in the presence of GM-CSF and IL-4 to moDCs and finally matured with LPS. Murine bone marrow-derived DCs (bmDCs) were generated by flushing bone marrow from femur and tibia of mice, plating the cell in medium containing GM-CSF and maturing the cells with LPS. Migration of DCs was analyzed in Transwell assays or dynamically by time-lapse microscopy within three dimensional collagen gels towards CCL-19 gradients. Signaling events were analyzed by Western Blot to evaluate changes in phosphorylation levels.

Results: 2D migration of ruxolitinib-exposed DC is dosedependently reduced in vitro. By analyzing the migratory phenotype of human moDCs within three dimensional collagen gels, ruxolitinib-exposed DCs are still able to sense chemokine gradients and form lamellipodia at the leading edge of the cell, whereas the retraction of the uropod is inhibited. Additional in vivo studies could show that the JAK inhibitor potently reduces homing of bmDCs into draining lymph nodes in mice. siRNA knockdown experiments revealed that this inhibitory effect is JAK1-and JAK2-independent. On a molecular level we could show a reduced phosphorylation of Rho-associated protein kinase (ROCK) in ruxolitinib-treated moDC upon CCL-19 stimulation. Finally, the migration phenotype of moDC exposed to ruxolitinib could be mimicked by the ROCK inhibitor Y-27632. Conclusion: RhoA family members are key proteins controlling important steps of cell migration, such as protrusion formation at the front of the cell and consequently cell polarization. ROCK is a downstream effector of RhoA and leads to stabilization of the actin cytoskeleton via cofilin and actomyosin II contraction by the myosin light chain. The observed reduction of ROCK phosphorylation may reveal an important mechanism of ruxolitinib-induced inhibition of DC migration. Our current efforts focus on the validation of ROCK as offtarget JAK1/2-independent target kinase of ruxolitinib as potential mediator of the observed effects, which may at least in part also explain the potential therapeutic impact of ruxolitinib for therapy of GvHD. Introduction: Alemtuzumab, a monoclonal CD52-antibody, is used for T-cell depletion (TCD) in the context of allogeneic hematopoietic stem cell transplantation (HSCT) to prevent graft versus host disease (GvHD). When we applied this approach in a clinical trial, nearly half of our patients developed acute GvHD (aGVHD) overall Grade I-II1. Since regulatory T cells (Treg) play a major role in controlling GvHD, we hypothesized that they might be functionally impaired in our patients after Alemtuzumab based treatment and further investigated on these cells. Materials (or patients) and methods: We analyzed peripheral blood samples of 20 patients with aGvHD, 10 patients with chronic GvHD, and 12 patients who never developed GvHD after Alemtuzumab-mediated TCD. Treg were identified as CD3 þ CD4 þ CD25 þ CD127-or CD3 þ CD4 þ CD25 þ FoxP3 þ and subsets described by expression of CD52. Since CD52 is linked to the membrane by a glycosylphosphatidylinositol (GPI)-anchor, we used FLAER to stain for GPI-anchors themselves. To further investigate Treg activation, we stained additional markers: CD39, CD44, GITR, CXCR3, CCR5, CTLA-4, GARP and Granzyme. To observe how Treg reconstitute after HSCT, we analysed samples from patients with aGvHD at different time points after transplantation and correlated our findings with the clinical course of GvHD. Treg function was evaluated in CFSE-suppression-assays: patient derived ex vivo Treg were FACS-sorted according to their CD52 expression and incubated with proliferating CFSEstained CD4-effector T cells from healthy donors. Reduced CD4-proliferation was used as indirect marker for the suppressive function of the CD52 positive and negative Treg. Results: Patients with aGvHD showed significantly elevated percentages of CD52 negative Treg: mean 55.3% (range 34.4-79.7% ) in comparison to only 10.1% (range 1-21.3%) in patients with chronic our without GvHD. By FLAER-staining we confirmed that loss of CD52 correlates with the absence of GPI-anchors on the cell surface, these cells do also lack other GPI-anchored proteins (e.g. CD56, CD59). Patients who overcame aGvHD over time reconstituted GPIanchor positive Treg, whereas GPI-anchor negative Treg remained the dominant population in patients with ongoing aGvHD. The fraction of activated GARP positive Treg was mainly detected among the GPI-anchor positive Treg population. All other markers showed a heterogeneous expression profile with a tendency towards lower expression on GPI-anchor negative Treg. Suppression-assays showed a higher suppressive capacity of GPI-anchor positive ex vivo Treg in contrast to GPI-anchor negative Treg of the same patient. Conclusion: CD52 negative / GPI-anchor negative Treg reconstituted in patients after Alemtuzumab mediated TCD and mainly persisted in patients with acute GvHD. These Treg were less likely to express the activation marker GARP. Functional assays demonstrated that GPI-negative Treg were functionally impaired -in patients developing acute GVHD, these impaired Treg represented the major Treg-population. Our preliminary data suggest that GPI-anchor negative Treg, like other GPI-anchor negative T-cell subpopulations (previously shown), are functionally altered. We hypothesize that this might promote acute GvHD. These GPI-anchor negative Treg could be useful to diagnose and guide immunosuppressive therapy in patients with acute GVHD. Disclosure of Interest: None declared. Introduction: We previously showed that enhanced human invariant NKT (hiNKT) lymphocytes recovery after allogeneic stem cell transplantation was associated with a reduced risk of acute graft versus host disease (aGVHD). Using a humanized mouse model of GVHD, we aimed to determine whether hiNKT cell subsets could be involved in the regulation of allogeneic immune response and to elucidate their mechanisms of action. Materials (or patients) and methods: Human iNKT were obtained by in vitro expansion of total peripheral blood mononuclear cells (PBMC) from healthy donors followed by electronic sorting of CD4 þ and CD4 -iNKT subsets among the CD3 þ CD1d-tetramer positive total iNKT population. Xenogeneic GVHD was assessed in 2 Gy irradiated NOD/SCID/ gamma-c -/-(NSG) mice previously injected with hPBMC alone or hPBMC enriched with CD4 -(PBMC þ iNKT4 -) or CD4 þ (PBMC þ iNKT4 þ ) hiNKT cells. The effects of hiNKT lymphocytes on the survival and phenotype of human monocyte derived dendritic cells (hMo-DC) were assessed by flow cytometry (on the expression of Annexine V, propidium iodide, and that of CD86, CD80, CD40, respectively). The effects of CD4or CD4 þ hiNKT cell subsets on the expression of the T cell activation marker CD25 and on the cytokine expression profile of CD4 þ T lymphocytes was assessed in vivo in the NSG model and in vitro, during the allogeneic mixed lymphocyte reaction (MLR). Results: In vivo, NSG mice transplanted with PBMC þ iNKT4cells showed a significantly prolonged overall survival in comparison to NSG mice transplanted with PBMC alone (P ¼ 0.001) or with PBMC þ iNKT4 þ iNKT cells (P ¼ 0.01). Improved survival observed with hPBMC þ iNKT4cells was associated with lower clinical and histological GVHD scores. In vitro, at 2 iNKT/ 1 DC (2 N) ratio, hCD4 -iNKT cells significantly increased the apoptosis of mature hMo-DC compared to the CD4 þ iNKT subset (P ¼ 0.001). hMo-DC apoptosis in the presence of CD4 -iNKT cells was contact dependent (21% in contact vs. 3% in transwell, Po0.0001), occurred as early as 4 hours after cell-contact and was associated to the degranulation of CD4 -iNKT cells as shown by CD107 expression. While h iNKT CD4lymphocytes could inhibit the maturation of hMo-DC in vitro, their CD4 þ counterpart strongly induced high levels of expression of CD80, CD86 and CD40 on Mo-DC in a contact dependent way. When allogeneic MLR was performed, in the presence of CD4 -iNKT lymphocytes, proliferating alloreactive T CD4 þ lymphocytes showed a lower median fluorescence intensity of CD25 (P ¼ 0.02), and intracellular INF-g (P ¼ 0.005) IL-17 (P ¼ 0.002) and IL-21 (P ¼ 0.021) in comparison to MLR without addition of iNKT cells. Compared to NSG mice injected with PBMC alone, mice injected with PBMC þ iNKT4cells showed lower CD25 expression on circulating T CD8 lymphocytes (P ¼ 0.001), and reduced intracellular expression of IL-17 on circulating T CD4 lymphocytes (P ¼ 0.01). Conclusion: These results are in line with our clinical observations and suggest that h CD4 -iNKT cell subset could directly modulate the allogeneic immune response by downregulating antigenic presentation and subsequently T cell activation and Th1/Th17 cytokine production. The NSG preclinical mouse model suggest that developing a cellular therapy based on iNKT CD4lymphocytes might be interesting for the prevention of human aGVHD. Disclosure of Interest: None declared.

In vivo expansion of host type regulatory T cells via a selective TNFR2 agonist protects from acute GvHD Medicine II, Wuerzburg University Hospital, 2 Pathology, Wuerzburg University, Wuerzburg, Germany Introduction: CD4 þ Foxp3 þ regulatory T cells (Tregs) suppress graft-versus-host disease (GvHD) after allogeneic hematopoietic stem cell transplantation (allo-HCT). By controlling the magnitude of the alloresponse Tregs still allow for efficient anti-leukemia (GvL) effects of transplanted conventional T cells in preclinical mouse models. Current clinical study protocols for donor Tregs in the treatment or prophylaxis of GvHD rely on their ex vivo expansion and infusion in high numbers. Here we present a fundamentally novel strategy for inhibiting GvHD based on the in vivo expansion of recipient Tregs prior to allo-HCT, exploiting the crucial role of tumor necrosis factor receptor 2 (TNFR2) in Treg biology. Materials (or patients) and methods: To expand Tregs in vivo we developed a highly selective TNFR2 agonist (that would not bind to TNFR1) and treated allo-HCT recipients two weeks before allo-HCT. Expansion of radioresistant host type Tregs, alloresponses of conventional T cells, and tumor progression were monitored with bioluminescence imaging, fluorescence microscopy, and flow cytometry in different MHC major mismatch mouse models (C57BL/6, H-2 b 4Balb/c, H-2 d and FVB/N, H-2 q 4C57BL/6, H-2 b ) of GvHD and GvL. Results: In vitro, this new TNFR2-agonist expanded natural Tregs from wild type but not from TNFR2-deficient mice. Accordingly, a human variant of this TNFR2-specific agonist expanded human Tregs in vitro. In vivo treatment of healthy mice with the murine TNFR2-agonist significantly increased Treg numbers in secondary lymphoid organs and peripheral tissues, particularly in the gastrointestinal tract, a prime target of acute GvHD. Consequently, pre-treatment of recipient mice with this novel TNFR2-agonist expanded host-type radiation resistant Tregs prior to allo-HCT in two models across MHC barriers. TNFR2agonist pre-treatment resulted in significantly prolonged survival and reduced GvHD severity when compared to TNFR2-deficient recipients or untreated allo-HCT recipients. This was accompanied by reduced donor T cell proliferation and infiltration into GvHD target organs. In vivo depletion of host derived Tregs completely abrogated the protective effect of TNFR2-agonist pre-treatment. While in vivo TNFR2-agonist pre-treatment protected allo-HCT recipients from GvHD, antitumor effects of transplanted T cells remained unaffected in two different murine B cell lymphoma models. Conclusion: Our novel strategy demonstrates that the expansion of host Tregs by selective in vivo TNFR2-activation significantly improves the outcome after allo-HCT and results in prolonged tumor-free survival. Disclosure of Interest: None declared. Introduction: We reported that miR-155 expression is upregulated in donor T cells during aGVHD and mice receiving miR-155 knock-out (KO) donor splenocytes do not exhibit lethal GVHD and have improved survival as compared to mice receiving wild type (WT) splenocytes. While we showed that miR-155 does not affect the alloreactive or homeostatic proliferative potential of T cells, a significant decrease in the expression of the homing receptors CCR5, CXCR4, and S1P1 were found on miR-155-KO T cells, suggesting that the loss of miR-155 could impair the migration of these cells to the peripheral target organs resulting in less aGVHD. Here, we further investigate the impact of miR-155 expression in T cell migration and elucidate the T cell population responsible for aGVHD modulation. Materials (or patients) and methods: Lethally irradiated BALB/c or B6D2F1 recipients were infused with T cell depleted bone marrow (BM) cells (5x10 6 ) and GFP expressing miR-155 KO or GFP-B6 WT T cells (1x10 6 ). Transplants with CD4 þ (2x10 6 ) only T cell subsets were performed to identify the lymphocyte population that contributes to the miR-155 mediated effects on aGVHD. Results: On days 7, 14 and 21 post transplant, recipient mice were sacrificed, and tissues harvested in order to study the kinetics of miR-155 KO T cell migration following allogeneic HSCT. There was a dramatic decrease in T cell infiltration of peripheral organs (Peyer's patches, liver, lung and skin) in recipients of miR-155-KO T cells as compared to WT T cells as evidenced by confocal microscopy of GFP labeled donor cells. We reasoned that these effects could be due to the modulation of CCR5 and other chemokine receptors by miR-155. We found that a recently described long noncoding RNA (LincR-Ccr2-5 0 AS) located in the vicinity of several chemokine receptor encoding genes including CCR1, CCR2 and CCR5 and that is involved in chemokine regulation and T cell migration has 3 conserved potential miR-155 binding sites. We then set out to determine if miR-155 regulates the expression of this lncRNA in relationship with CCR5 expression. There was a significant decrease in CCR5 mRNA expression in miR-155-KO versus WT donor T cells obtained from recipient mice at the time of aGVHD, but no significant difference in the levels of LincR-Ccr2-5 0 AS. A luciferase reporter assay, however indicate that there was an interaction between miR-155 and LincR-Ccr2-5 0 AS. Our result does not exclude the possibility that miR-155 might influence the activity rather than LincR-Ccr2-5 0 AS levels, a hypothesis that is currently being tested. To identify the lymphocyte population involved in miR-155 mediated modulation of aGVHD we performed a B6 into F1 transplant using CD4 þ T cells alone as the source of donor T cells. Median survival of recipients of T cell depleted BM þ WT CD4 þ T cells (n ¼ 12) was 48 days as compared to 100% survival of all recipient mice of miR-155 KO CD4 þ T cells (n ¼ 12) on day 100, (P ¼ 0.02). Recipients of miR-155 KO CD4 þ T cells exhibited also significant lower aGVHD clinical (Po0.01) and pathological scores (Po0.01) than WT recipients. Conclusion: Our data suggest that miR-155 may exert its modulating effects in aGVHD by affecting T cell migration. Our results also point to the CD4 þ T cell subset seems to play an important role in the miR-155 regulation of aGVHD. Experiments are currently underway to determine the role of miR-155 in regulatory T cells and CD8 þ T cell subsets in the modulation of aGVHD. Disclosure of Interest: None declared.

Oral session: Solid tumors Introduction: HDCT has a recognized indication in the salvage setting of advanced GCT and is steadily utilized worldwide. While the prognostic impact of response to prior lines of CT (i.e. definition of chemoresistance) is ascertained, that of response to induction/mobilization CT preceding single or multiple HDCT cycles is unknown. We present the results of a retrospective study of the EBMT- STWP. Materials (or patients) In the multivariable model for PFS, tumor primary (overall p value ¼ 0.039), IGCCCG category (overall p value ¼ 0.033), and progression to induction (HR: 2.02, 95%CI: 1.31-3.12 , P ¼ 0.001) were significantly prognostic. The latter was also significantly prognostic for OS (HR: 2.36, 95%CI, Po0.001) together with the transplant setting (overall p value ¼ 0.014). Results for response to induction CT were confirmed in the models that included missing data. The c-index of the model was 0.62 for PFS and 0.63 for OS. Conclusion: Progression to induction CT prior to HDCT was independently and significantly associated with shorter PFS and OS, while response or progression to prior CT lines was not. This information could have important implications to refine patient eligibility to transplantation and enhance the prognostic risk grouping. These data need external validation. Disclosure of Interest: None declared.

KIR-ligand incompatibility in the graft-versus-host direction improves progression-free survival in patients with primary high risk neuroblastoma after umbilical cord blood transplantation with nonmyeloablative conditioning Y. Takahashi Introduction: Donor NK cells expressing inhibitory killer cell immunoglobulin-like receptors (KIRs), which do not recognize their cognate ligands ('KIR ligands') on recipient targets, are free from HLA inhibition, resulting in a decreased incidence of relapse and improved the outcome after HLA haploidentical stem cell transplantation (HSCT) (Ruggeri, Blood 2007) or umbilical cord blood transplantation (UCBT) (Willemze, Leukemia 2009) in leukemia patients. Venstrom reported that significant association between KIR/HLA genotypes predictive of missing KIR ligands have a better outcome without allogeneic HSCT following anti GD2 monoclonal antibody therapy in patients with high risk neuroblastoma (Venstrom, Clin Cancer Res 2009). These observations led us to start the clinical trial of allogeneic UCBT from KIR ligand incompatible donor for patients with high risk neuroblastoma in 2008. Materials (or patients) and methods: Eligibility criteria of this study were newly diagnosed stage IV neuroblastoma patients with one of the following 1) Chemo-resistant disease defined as MIBG positive score after 4 courses of induction chemotherapy, 2) Older age defined as 10 years old and older at diagnosis or 3) MYCN amplification defined as greater than 10 copies per tumor cell. KIR ligand mismatched UCBT donor was prospectively selected from the Japan Cord Blood Bank Network according to the genotyping of HLA-C or B as previously reported (Willemze, R. Leukemia 2009). We scheduled UCBT with reduced intensity conditioning regimen about three months after conventional high dose chemotherapy followed by autologous peripheral blood stem cell transplantation. Reduced intensity conditioning regimen for CBT was Flu, L-PAM 140 mg/m2 and TBI 2 Gy. Tacrolimus and methotrexate were used for GVHD prophylaxis. Single inhibitory KIR expressed NK cells with no corresponding recipient's HLA were monitored by flowcytometry before and after UCBT to access the expansion of alloreactive NK cells in vivo. We retrospectively analyzed the outcome of 82 patients with high risk neuroblastoma treated in Nagoya University hospital between April 1982 and June 2014. Results: Fifty four patients were treated before Dec. 2007 . After this study started in Jan. 2008 , 15 patients underwent KIR ligand incompatible CBT who match eligibility criteria (7 chemo-resistant, 6 MYCN amplification and 2 older age) and 13 patients received standard treatment. All patients achieved engraftment in UCBT group and no patients developed grade III or more acute GVHD and chronic GVHD. Only two patients died in this group because of BU related lung toxicity. Surprisingly, no patient in this group relapsed with the median follow-up period of 51 (6-71) months. On the other hand, cumulative incidence of relapse in others was 51.0%. 3year progression free survival in CBT group was significantly better than in others (83.6% vs 40.7%, P ¼ 0.048). Single inhibitory KIR expressed NK cells significantly expanded after CBT (P ¼ 0.0009). Finally, multivariate analysis revealed only KIR ligand incompatible CBT and stage III were significant better covariate factors for relapse. Conclusion: This is the first report that KIR ligand incompatible allogeneic UCBT significantly reduced the incidence of relapse in high risk neuroblastoma. Disclosure of Interest: None declared.

Better prognosis for BRCA-mutated breast cancers treated with high-dose chemotherapy and autologous hematopoietic progenitor cell transplantation. A singleinstitution retrospective study L. BOUDIN 1, 2, 3, * , A. GONCALVES 1, 2, 4 , J. M. EXTRA 1, 2 , R. SABATIER 1, 2 , H. SOBOL 2, 5 , F. EISINGER 2, 5 , J. MORETTA 2, 5 , C. TARPIN 1, 2 , J. CAMERLO 1, 2 , B. CALMELS 2, 6 , C. LEMARIE 2, 6 , A. GRANATA 2, 6 , 4, 7 , F. BERTUCCI 1, 2, 4 , A. MADROSZYK 1, 2 , P. VIENS 1, 2, 4 , C. CHABANNON 2, 4, Introduction: Hereditary BRCA genetic mutation is responsible for approximately 2-3% of breast cancers (BC).The objective of this study was to compare the outcome of BRCA-mutated (BRCA mut ) versus BRCA wild type or unknown (BRCA wt/uk ) BC following treatment with high dose chemotherapy (HDC) and autologous hematopoietic progenitor cell transplantation (ACT). Materials (or patients) and methods: All female patients treated for breast cancer (BC) with HDC and ACT at Institut Paoli-Calmettes between 2003 and 2012 were included. Patients were divided into 2 groups depending on the indication of HDC with ACT: Metastatic breast cancer (MBC) or high risk breast cancer (HRBC) including inflammatory breast cancer (IBC) and/or massive lymph node involvement (LNI). All patients were examined for the presence of known BRCA 1 or 2 mutations. Information regarding patient, tumor characteristics and treatment was also collected. The main objectives were the analysis of overall survival (OS) and progression-free survival (PFS) according to the BRCA mutation status in the two groups. Survival curves were generated using Kaplan-Meier method and compared using Log-rank test.

Results: A total of 377 patients were included, 235 MBC and 142 HRBC (73 LNI and 69 IBC). Among MBC and HRBC, 10 (6 BRCA1, 4 BRCA2) and 5 (3 BRCA1, 2 BRCA2) patients were BRCA mut , respectively. In MBC, median age was respectively 36 (range 29-53) and 42 years (range 24-61) for BRCA mut and BRCA wt/uk ; 70% of BRCA mut patients had triple negative (TN) BC subtype (HER2-negative and Hormonal receptor-negative) versus only 21% in BRCA wt/uk individuals. In BRCA mut and BRCA wt/uk median number of metastatic sites was 2 and 1, with 90% and 69% visceral metastases, respectively. In HRBC, median age for BRCA mut and BRCA wt/uk was 36 (range 28-58) and 49 (range 20-62), respectively; 60% of BRCA mut had TN BC subtypes versus 24,6% of BRCA wt/uk . IBC represented 49% of HRBC in BRCA wt/uk and 40% in BRCA mut . In MBC, 90% (all except one) of BRCA mut remained alive at 5 years; the median overall survival (OS) was not reached, compared with a median OS of 3.62 years for BRCA wt (CI95% ¼ 3.07-4.68), (P ¼ 0.048 Log-rang test); median progression free survivals (PFS) were 26,4 months (Cl95% ¼ in BRCA mut patients versus 15,6 months (Cl 95% ¼ 11. 64-18.48) , and in BRCA wt/uk patients (P ¼ 0.073 Log-rank test). In HRBC, 8 years probabilities of OS were respectively 1 (all patients alive) vs for BRCA mut and BRCA wt/uk individuals, while 8 years probabilities of disease free survival (DFS) were ) and , respectively (P ¼ 0.374 Log-rank test). Conclusion: The prognostic impact of BRCA mutation in BC remains controversial. In this series of MBC and HRBC, we have reported an outstanding survival outcome in a small subset of patients with documented BRCA mutation. In spite of a higher proportion of TN and more aggressive features, BRCA mut BC had a better outcome than their BRCA wt/uk counterpart, the difference reaching statistical significance in MBC population. Hypersensitivity to DNA-damaging agent, possibly due to defect in homologous recombination associated with BRCA mutation, could explain these results. Despite the low numbers of cases in our series, HDC with ACT, may be considered as an option for BRCA mut , women with advanced BC. Disclosure of Interest: None declared.

Introduction: Paclitaxel-based regimens are now commonly employed for second or third-line salvage therapy for GCT. This might have an impact on the results with subsequent salvage HDCT in these patients (pts). The EBMT-STWP is sponsoring a retrospective study on the outcomes of HDCT administered in the last 10 years. Hence, we aimed to study outcomes with HDCT after relapse to paclitaxel-CT to identify the level of chemoresistance in these pts. Materials (or patients) and methods: Data have been collected from EBMT registry, including 24 European centers. Eligibility included adult male patients (pts) with GCT, and treatment with second or further-line HDCT between the years 2002 and 2012. Both paclitaxel used in prior CT lines of therapy and in induction-mobilization regimens pre-HDCT were considered. Multivariable Cox regression analyses (MVA) were undertaken to evaluate the association of prespecified factors (site of tumor primary, IGCCCG category, response to induction CT, response to prior lines [chemosensitive vs chemorefractory], line of HDCT, year of HDCT) including prior-paclitaxel therapy with progression-free (PFS) and overall survival (OS). Results: Since 10/2013, 442 pts have been registered, 324 of them were suitable for present analysis. 165 pts (51%) received HDCT in second-line, 102 (31%) in third and 57 (18%) beyond the third-line. HDCT regimens were as follows: 177 (55%) HD-CBDCA-VP16 (CE), 41 (12%) adding ifosfamide (CEI), 106 (33%) other mixed regimens. 76 (23%) were taxane-containing. 120 pts received a single HDCT course, 99 pts double and 104 multiple courses (1 missing). Median follow up was 36 months (IQR: 19-70). Prior paclitaxel was significantly associated with shorter OS in the univariable model (P ¼ 0.032). However, on MVA prior paclitaxel-therapy was not significantly prognostic for both PFS and OS, as shown in the Table. A separate model evaluated the interaction between prior paclitaxel-therapy and taxane-containing HDCT: no significant interaction was found (P ¼ 0.221 1.18-5.76 .017 Conclusion: The administration of paclitaxel-based regimens before HDCT did not affect PFS/OS. Results were confirmed when excluding pts who were administered taxane-containing HDCT. Line of HDCT was not significantly prognostic too. There is no evidence to disallow patients who have been treated with taxanes in second or third line to receive HDCT as futher salvage therapy. These data need external validation. Disclosure of Interest: None declared.

Favorable outcome of a cohort of metastatic breast cancer patients treated with high dose chemotherapy and autologous transplantation at a single institution does not change the prognostic significance of histopathological subtypes L. BOUDIN Introduction: Studies of high dose chemotherapy (HDC) in breast cancer (BC) often lack of biomarker information, notably the human epidermal growth factor receptor 2 (HER2) status. The objective of this study was to evaluate the outcome of patients affected with different subtypes of BC following treatment with HDC and autologous hematopoietic progenitor cell transplantation (ACT). Materials (or patients) and methods: All female patients treated for metastatic breast cancer (MBC) with HDC and ACT at Institut Paoli-Calmettes between 2003 and 2012 were included. Patient, tumor and treatment characteristics were collected. Patients were categorized in three subtypes based on hormonal receptor (HR) and HER2 status of the primary tumor: Luminal (L), (HR þ /HER2-), HER2 (HER2 þ , any HR), and triple negative (TN) (HER2-and HR-). Main study endpoints were overall survival (OS) and progression-free survival (PFS) categorized by BC subtypes. Treatment related mortality (TRM) was also evaluated. Survival curves were generated using Kaplan-Meier method and compared using Log-rank test.

Results: A total of 231 MBC patients are included; Median age was 47 (range 24-61); Metachronous and synchronous MBC were 64% and 36%, respectively. 96% patients received HDC with ACT as part of their first-line treatment following diagnosis of metastases. Median number of metastatic sites was 1 (range 1-6), including 69% of visceral metastases. L, HER2 and TN subtypes were 61, 18 and 21% of patients, respectively. All but one HER 2 patients (98%) received trastuzumab during the metastatic phase of the disease before (88%) and/or after (91%) Oral session: Acute leukemia 2 Introduction: The outcome of patients 455 years old with ALL is poor with no clear recommendations regarding the indications for HCT. These patients are usually considered ineligible for myeloablative alloHCT and offered transplantation with reduced intensity conditioning (RIC). Autologous HCT is an alternative. However, the role of both treatment options has not been established so far. The aim of this study was to retrospectively compare results of RIC-alloHCT and autoHCT in ALL 455 years old and to identify factors affecting outcome. Materials (or patients) and methods: 267 patients treated with RIC-alloHCT from either HLA-identical sibling (n ¼ 154) or matched unrelated donor (n ¼ 113) and 179 treated with autoHCT in first complete remission between 2000 and 2011 have been included in this analysis. Median age in both groups was 60 (55-74)y and 60 (55-76)y, while median interval from diagnosis to HCT was 5.9 months and 6.6 months, respectively. The proportion of Ph( þ ) ALL among those with reported cytogenetics was 71% and 66%, respectively. Results: With a median follow-up of 33 months, the probability of OS at two years was 44% for RIC-alloHCT and 57% for autoHCT (P ¼ 0.02), while LFS rates were 34% and 41%, respectively (P ¼ 0.06). Relapse incidence at two years was comparable for RIC-alloHCT and autoHCT (42% vs. 48%, P ¼ 0.39) while non-relapse mortality was significantly reduced for autoHCT (23% vs. 11%, respectively, P ¼ 0.002).

The advantage in favor of autoHCT was significant for Ph(-) ALL (OS: 61% vs. 38%, P ¼ 0.02; LFS: 54% vs. 21%, P ¼ 0.005) while no significant differences could be observed for Ph( þ ) ALL (OS: 55% vs. 47%, P ¼ 0.6; LFS: 42% vs. 35%, P ¼ 0.4).

In a multivariate analysis adjusted for recipient age and gender as well as interval from diagnosis to transplantation the use of autoHSCT was independently associated with reduced risk of mortality (HR ¼ 0.69, P ¼ 0.01), treatment failure (HR ¼ 0.76, P ¼ 0.03) and non-relapse mortality (HR ¼ 0.39; P ¼ 0.0004) with no effect on relapse incidence (HR ¼ 0.98, P ¼ 0.88).

In the RIC-alloHSCT subgroup LFS was negatively affected by female donor/male recipient combination (HR ¼ 1.64, P ¼ 0.01). LFS rates for both sibling and MUD transplants were comparable (32% vs. 35%, P ¼ 0.18). The use of peripheral blood cells compared to bone marrow was associated with reduced risk of relapse (HR ¼ 0.5, P ¼ 0.03). In the autoHSCT setting there was a tendency to higher risk of treatment failure with increasing recipient age (HR ¼ 1.05, P ¼ 0.06). Other variables including type of conditioning (TBI-based vs. chemotherapy-based) did not affect survival in any of the study cohorts. Conclusion: Considering poor overall prognosis of ALL patients 455 years old, results of both RIC-alloHCT and autoHCT appear enhancing and both types of transplantation may be considered valuable treatment options. Potential advantage of autoHCT as suggested by results of our analysis should be further explored including data on disease-related prognostic factors and the status of minimal residual disease. Disclosure of Interest: None declared. Introduction: The karyotype of the leukemic cells at diagnosis is one of the strongest prognostic factors in acute myeloid leukemia (AML). But the major cytogenetic risk categorizations were based on the large clinical studies of conventional chemotherapy for AML. In this retrospective study, we analyzed the influence of the cytogenetics of AML at diagnosis on the outcomes of allogeneic hematopoietic cell transplantation (HCT). Materials (or patients) and methods: From the database of JSHCT, we extracted the data of adult patients with AML who receive first HCT between 2006 and 2010. A total of 4,241 recipients were included. Additional survey for the recipients who had been reported to have cytogenetic abnormalities at diagnosis (n ¼ 2,384) confirmed the data of karyotype of 1,360 cases.

Results: Cytogenetics at diagnosis were categorized into (A) normal karyotype (41.5%), (B) inv(16) or t(16;16) (2.9%), (C) t(8;21) (9.1%), (D) 11q23 abnormality (4.9%), (E) complex karyotype (4 ¼ 3 abnormalities) (12.3%), (F) monosomal karyotype (MK) (6.3%), and others. Overall survival (OS) and cumulative incidence of relapse at 2 years of all cases were 48.2% and 37.5%, respectively. These recipients were classified into 4 groups: favorable (Fav) included (B) (n ¼ 121); intermediate (Int), (A), (C), t(9;11) and others (n ¼ 3,169); unfavorable-1 (Unf-1), (D) except t(9;11), and (E) except (F) (n ¼ 404); unfavorable-2 (Unf-2), (F) (n ¼ 265). Adjusted OS at 2 years were 65.2% in Fav, 53.2% in Int, 37.8% in Unf-1 and 24.2% in Unf-2 ( Figure) . Risk of overall mortality (ROM) and risk of relapse (RR) were compared among these four groups, adjusted by age, sex, disease status at transplantation and stem cell source. In Fav group, ROM (HR 0.71, , P ¼ 0.0038) and RR (0.44, 0.27-0.71 , P ¼ 0.0007) were significantly lower than in Int group. On the other hand, compared with Int group, recipients in Unf-1 and Unf-2 group had significantly (Po0.001) higher ROM (Unf-1 1.48, 1.29-1.70 ; Unf-2 2.22 1.93-2.57 ) and RR (Unf-1 1.65, 1.41-1.92 ; Unf-2 2.35, 1.94-2.73 ).

These results showed that the karyotype of leukemic cells at diagnosis was also one of the powerful prognostic factors in HCT for AML. But the impact of each cytogenetic presentation on the outcomes of transplantation may have some differences from that of chemotherapy. To improve this risk stratification system, molecular abnormalities of AML at diagnosis are also considered for the influence on outcomes of HCT. Disclosure of Interest: None declared.

Late intensification with autologous HSCT after nonmyeloablative BEAM conditioning has shown to be effective approach in adults with T-cell acute lymphoblastic leukemia treated by non-intensive protocol: results of the RALL study group E. Parovichnikova 1,* , L. group is conducting the ALL-2009 trial where the main principle is non-intensive but non-interruption treatment and autologous HSCT after BEAM followed by maintenance in T-cell ALL pts without HLA-identical donors (EBMT 2014, OP-009). The autologous HSCT was planned for all patients as late high dose consolidation (on the 20-22 weeks of the protocol).

In patients with early and late T-cell immunophenotype having HLA-identical siblings allogeneic BMT was an option. From Nov, 2008 , till Nov, 2014 , 30 centers enrolled 264 ALL patients. In 6phenotype was unknown (2,2%). B-cell precursor phenotype was diagnosed in 62,6% (n ¼ 166), T-cell precursor phenotypein 34,8% (n ¼ 91); biphenotypic AL -in 0,4% (n ¼ 1). T-cell ALL patients were young -28 y (16-56); male gender prevailed -33 f/58 m. T-ALL subtypes distribution in our study was as follows: 44 (47,8%) patients had early T-ALL (T-I/II), 36 (40,1%)thymic (T-III), 11 (11,1%) -mature (T-IV). For the whole group medians were: Hb-112 g/l (42-180), L -22,3*10/ 9 l (0,5-313), plt -90*10/ 9 (5-943), b/m blasts -74,9% (0-99), LDH -995 IU (131-12 000). No born marrow involvement was detected in 4 pts, b/m blasts 5-25% -in 10 (so T-LBL -14,5%). Cytogenetics was available in 63,2% of pts (n ¼ 55/87), 45,5% of them (n ¼ 25/55) had normal (NK), 25 pts -abnormal karyotype, 5 pts (9%) had no mitosis. Mediastinum involvement was registered in 48/87 (54,5%) and CNS disease -in 11/87 (12,5%).

Results: Induction and follow-up data were available in 87 pts. CR was achieved in 78/87 (89,8%): after prephase ¼ 12, after 1 st ind.phase ¼ 45, after 2 nd ind.phase ¼ 21. Induction death occurred in n ¼ 5 (5,7%). Primary resistance and even progression during induction was registered in n ¼ 4 (4,5%). 28 pts proceeded to autologous HSCT, almost all (n ¼ 26) were successfully harvested at a median time of 20 weeks. 2 pts were poor mobilizers and bone marrow exfusion was carried out. No data on MRD level at time of harvesting is available so far. Auto-HSCT was applied at a median 6 months from CR. Allogeneic HSCT from sibling donors was performed in 6 CR pts (4 with TI/II and 2 with T-IV). No deaths in allo-HSCT group have occurred. The land mark analysis at 6 months for chemotherapy group (n ¼ 27) and at day of HSCT for transplanted groups was done. Disease free survival constituted for chemotherapy -55% at 5 years, and 100% -in both transplanted groups. None of the transplanted patients relapsed so far. The overall and diseasefree survival in the whole cohort of T-cell ALL patients was 58% and 68%. In a multivariate analysis only autologous HSCT influenced DFS. Conclusion: Our study demonstrates that T-ALL may be treated by non-intensive, but non-interruption approach. Even without high dose consolidation 5-years OS and DFS (land mark analysis) constituted 55%. Autologous HSCT with BEAM conditioning after mild induction/consolidation followed by prolonged maintenance seems to add benefit to the overall optimistic results decreasing the relapse rate from 34% to 0% within 5 years. Disclosure of Interest: None declared. With a median follow up from time of HLA typing of 14 months (range, 0.3-97) for all patients and 30 months (range, 3-97) for survivors, the 1-and 8-year probability of survival for all 225 eligible AML patients was 68±3% and 40±4%, respectively. At the same time points, for the 172 transplanted patients the probability of survival from time of graft was 60 ± 4% and 43 ± 4%, respectively. The probability of survival was particularly dismal for the 37 patients transplanted in advanced disease phase (4±4% at 4 yrs), whereas for the 135 patients transplanted in early (CR1 þ CR2) phase the 8-yrs probability of survival was 54±5% and, by excluding the low number of CB recipients, it was 55 ± 7% for 56 HLA ID Sib, 58 ± 8% for 40 HRD and 68 ± 9% for 28 MUD recipients (P ¼ ns). Conclusion: RTN policy allows a donor identification in 94% of all evaluable AML patients and provides an allogeneic transplant to 86% of them with no substantial differences in S58 terms of long-term survival between initially eligible and definitively transplanted patients or by comparing the different donor stem cell sources. Transplant results should be given following information on the specific transplant policy and only the ITT analysis allows to know the real impact of each transplant program. Disclosure of Interest: None declared. Materials (or patients) and methods: Oral PAN was administered in two sequential schedules, either thrice weekly (TIW) every week (A) or every other week (B). Arm A, in which PAN was started at a dose of 10 mg TIW and escalated to 30 mg TIW using a 3 þ 3 design was followed by arm B, in which PAN doses increasing from 20 mg TIW to 40 mg TIW were investigated. PAN was initiated between day þ 60 and þ 150 after HSCT and given for up to 1 year. Eligibility criteria included: ANCZ1,000/mL, plateletsZ75,000/mL, adequate organ function and no severe GvHD. DLT was defined as prolonged G4 hematologic toxicity or any non-hematologic toxicityZG3 unrelated to disease progression within 28 days of the first PAN dose.

Results: 24 pts (21 AML, 3 MDS), median age 52 yrs (range, 21-71), are evaluable for MTD. Cytogenetics were classified as low (n ¼ 2), intermediate-1/2 (n ¼ 11) or adverse risk (n ¼ 11) according to ELN criteria. PAN was started a median of 98 days (range, 60-147) after HSCT from a MRD (n ¼ 8), MUD (n ¼ 10) or mismatched donor (n ¼ 6) which was performed in active disease (n ¼ 17, median bone marrow blasts 23%, range, 8-77), CR1 (n ¼ 5) or CR2 (n ¼ 2). The RP2D was 20 mg TIW in arm A and 30 mg TIW every other week in arm B based on 5 DLTs: fatigue G3 at 20 mg, colitis and nausea/emesis G3 each at 30 mg (arm A), diarrhea and headache G3 at 40 mg each (arm B). PAN-related or unrelated G3/4 AEs were reported in 20 of 24 pts (83%) and included hematologic toxicity (50%), laboratory alterations (42%), infections (29%), constitutional (29%) and gastrointestinal symptoms (25%) and neurologic, pulmonary, pancreatic/hepatobiliary or vasculary toxicity (4-8% each). AEs were rapidly reversible after interruption and/or dose reduction (n ¼ 6) and no study-related deaths occurred. Fifteen pts (63%) have completed one year of PAN, 9 pts discontinued treatment prematurely after 7-217 days due to AEs (n ¼ 8) or relapse (n ¼ 1). Prophylactic or preemptive DLIs (1-6) were administered to 15 pts (63%). Eleven pts (45%) developed mild (n ¼ 6), moderate (n ¼ 3) or severe (n ¼ 2) chronic GvHD. Patients died of AML (n ¼ 3) or severe chronic GvHD (n ¼ 1 Introduction: Our previous research has provided the evidence that an autoreactive immune system can be ?reset? into a healthy, tolerant state by immunoablative treatment to eradicate pathogenic effector cells, followed by transplantation of hematopoietic progenitor cells (HSCT). Here, we present the clinical and serologic responses and long-term immune reconstitution in 20 patients with severe ADs for up to 15 years after receiving immunoablation and ASCT. Materials (or patients) and methods: Since 1998, 20 patients with refractory autoimmune diseases (SLE, n=10; SSc, n=4, MS, n=2; polychondritis, n=1; panniculitis, n=1, granulomatosis with polyangiitis, n=1, and chronic inflammatory demyelinating polyneuropathy, n=1) received a CD34 + -selected autologous stem cell transplantation after immunoablation with ATG (Neovii, formerly Fresenius) and cyclophosphamide as part of a prospective monocentric phase I/II clinical trial. Autoantibody titers were evaluated with ELISA, peripheral Tand B lymphocyte subsets immunophenotyped using multicolor flow cytometry.

Results: With a median follow-up of 12 years after immune reset (range, 24 months to 16 years), 15 of 20 patients (75%) achieved a progression-free survival (PFS), defined as survival without major organ failure. 50% of these patients showed durable clinical remissions for up to 15 years despite discontinuation of immunosuppressive treatment, while 25% of patients had a stabilization of their underlying disease under maintenance therapy. Disease relapse occurred in three SLE patients (at 18, 36, and 80 months, respectively), two of whom died later from uncontrolled disease and infection, respectively. 3 of 20 patients died from infection (n=2) or cardiac failure (n=1). Anti-dsDNA antibodies completely normalized in all SLE patients and ANA significantly declined from a median titer of 1:5120 at baseline to 1:160 six months after transplantation in patients with connective tissue diseases. Recovery of recent thymic emigrants (RTEs) was completed between 12 and 24 months after immune reset, reaching on average 3.6 to 5.1 times the baseline levels.

Recurring Foxp3 + Tregs had significantly higher expression levels for CD31 and CD45RA compared to age-matched healthy controls, indicating their recent thymic origin. Numeric recovery of the naÿve B cell compartment was completed within 1 year after immune reset in responding patients. Conclusion: These data confirm our assumption that the reprogramming of an autoreactive immune system into a juvenile and self-tolerant immune system is feasible and associated with long-term remissions. Our findings propose that chronic autoimmunity is not an end point depending on continuous treatment with specific anti-inflammatory agents, but may be cured by combining specific targeting of autoreactive memory and effector cells with a reactivation of thymic activity. A future challenge is to make this therapeutic approach attractive for a larger number of patients by reducing the rate of severe infections. This may be achieved by either using a more selective graft purging, e.g., depletion of T cell receptor alpha/beta and CD19 + cells from apheresis products with the CliniMACS Device, or an adoptive transfer of microbe-or virus-specific memory T and/or B cells. Disclosure of Interest: None declared. In the present report we summarize the long term effects of transplantation. Materials (or patients) and methods: The mobilization protocol included cyclophosphamide and G-CSF. The conditioning consisted of cyclophosphamide (50 mg/kg/day on days -5,-4,-3, -2 prior to transplantation) and antithymocyte globulin (Thymoglobulin -of 0.5 mg/kg/day given on day -5, and 1.0 mg/kg/day given on days -4,-3, -2 and -1) except for three patients, who received ATG Fresenius at adjusted doses.

Results: One of the transplanted patients died during neutropenia due to the sepsis and its complications. The mean time of observation of remaining 23 patients as of December 2014 was 57 months (range 33 -80 months). The independence of exogenous insulin after the transplantation was achieved in 20 of them. Three patients were lost to follow up. The median time without exogenous insulin for 20 patients in follow up was 30 months (range 0-80 months). Four out of 20 (20%) remain in remission of diabetes (exogenous insulin free) with median follow up of 54 months (range 34-80 months). Notably, three patients in the series were treated with ATG Fresenius (due to transient problems with Thymoglobulin availability) -the median follow-up of these patients is 65 months with 58 months of remission of diabetes. The treatment led to significant reduction of HbA1c and good glycemic control in patients. No attempt was made to repeat procedure in relapsing patients.

The HSCT leads to remission of new onset T1DM in majority of patients but the response in most of them is limited to average of 30 months. Patients receiving ATG Fresenius tended to have longer remission but small number prevented drawing conclusions. There is a need to develop procedure to either prolong response or to effectively treat the relapse of diabetes. Disclosure of Interest: None declared.

Hematopoietic stem cell transplantation for refractory crohn's disease: feasibility and toxicity J. Introduction: Autologous hematopoietic stem cell transplantation (HSCT) is a salvage treatment for severe refractory Crohń s disease (CD) patients. We evaluated feasibility and toxicity of autologous HSCT for refractory Crohn's disease (CD). Materials (or patients) and methods: In this prospective study, refractory CD patients with an aggressive course despite medical treatment, impaired quality of life and no surgical options were included. Hematopoietic stem cells were mobilized with cyclophosphamide and G-CSF and collected by leukapheresis from pheripheral blood. In a second step, conditioning regimen with cyclophosphamide and rabbit antithymocite globulin (rATG) was used and previously collected stem cells were infused. Toxicity and complications during the procedure and within first-year after transplantation were evaluated, along with the impact on safety after the introduction of supportive measures over the study. Results: Twenty-six patients were enrolled. During mobilization, 16 patients presented febrile neutropenia, including 1 bacteremia and 2 septic shocks. Neutropenia median time after mobilization was 5 days. Five patients withdrew from the study after mobilization and 21 patients entered into the conditioning phase. Hematopoietic recovery median time for neutrophils (40.5x10 9 /L) was 11 days and for platelets (420x10 9 /L) was 4 days. Ninety-five percent of patients suffered febrile neutropenia and 3 patients presented worsening of perianal CD activity during conditioning. Among non-infectious complications, 6 patients presented antithymocyte globulin reaction, 12 patients developed mucositis and 2 patients had hemorrhagic complications. Changes in supportive measures over the study, particularly antibiotic prophylaxis regimes during mobilization and conditioning, markedly diminished the incidence of severe complications. During the first 12-month follow-up, viral infections were the most common observed complications, and one patient died due to systemic cytomegalovirus infection. Conclusion: Autologous HSCT for refractory CD patients is feasible but extraordinary supportive measures need to be implemented. We consider that this procedure should only be performed in highly-experienced centers. Disclosure of Interest: None declared.

Autologous haematopoietic stem cell transplantation in aggressive multiple sclerosis: a UK cohort from two centres J. Clay 1,* , P. 0.5 (range 0.5-4.5) , whilst the remaining 4 had a deterioration of 0.5-1.0 Materials (or patients) and methods: 25 CB Units (CBU) were frozen and thawed across the participating Banks, according to local SOPs; pre-freezing and post-thaw samples were assessed for clonogenic potential (CFU) and CD34 þ recovery and viability through standard ISHAGE protocol (SI) (Brocklebank AM 2001). A modified ISHAGE gating strategy (MI) was developed by reduction of the debris gate and by extension of Lympho-monocytes gate in order to detect CD34 þ events with abnormal physical parameters (CD34 þ gate). FACS sorting was performed to collect CD45 þ /CD34 þ events included inside (P8) and outside(P9) the SI CD34 gate. Sorted cells were then characterized by FC, confocal microscopy analysis and CFU assay in order to confirm that CD34 þ cells detected out of the SI CD34 gate are not an artifact. Confocal microscopy analysis was performed by labeling with a CD34 Ab recognizing a different epitope and marked with a nuclear counterstain.

Results: SI and MI strategies showed no significant discrepancies when determining CD34 þ absolute number and viability in pre-freezing CBUs. The recovery of viable CD34 þ cells assessed by SI or MI in the thawed products did not show any differences statistically significant, whilst the recovery of total CD34 þ cells was significantly lower in SI (58,3 ± 18%) than in MI (80,1±19.7%)-analyzed samples (mean±SD, Po0,05 with pair T test). This difference resulted in a lower CD34 þ viability with MI than SI (79,5% ± 15.9 vs 96,7% ± 3%, respectively, Po0.02). FC post-sorting analysis showed that the sorted P8 (SI CD34 gate) population was CD45 dim CD34 þ CD133 þ CD5 -7-AAD-, whereas P9 (MI extended CD34 gate) showed the same markers expression but a 7-AAD þ staining, therefore determining such events as CD34 þ dead cells. This data was confirmed by confocal microscopy analysis. Finally, P8 events showed a 67% clonogenic efficiency, whereas P9 events were not able to generate any colony.

Conclusion: A CD34 þ count after CBU thawing lower than the fresh sample is often reported, usually with a high rate of viability. We showed that the standard methodology for CD34 þ counting leads to the exclusion of a number of dead CD34 þ cells, thus resulting in a viability overestimation. We propose here a modification of the ISHAGE counting standard for thawed samples, which provides a more reliable assessment of CD34 þ viability. The content of pDC in BM grafts had a significant beneficial effect on post-transplant survival, an association not seen in GPB grafts 1 . The 20% improved survival seen in recipients of more donor BM pDC was due to decreased deaths from graft rejection and acute and chronic GvHD. Donor BM pDCs in murine BMT favorably modulated T-cell activation, decreasing GvHD through gamma-Interferon-dependent upregulation of IDO, while preserving GvL effects 2 . The present study tested whether the biological activity of pDC from different graft sources are explained by the differential expression of homing receptor molecules or immuneregulatory pathway genes. Materials (or patients) and methods: FACS isolated pDC from allogeneic stem cell grafts obtained from 10 BM, 8 GPB, 4 umbilical cord blood (UCB) and 2 following mobilization with plerixafor (PLXPB) samples. Expression of chemokine receptor, integrin and selectin were measured by flow cytometry. Gene expression on flow cytometry-sorted pDC was analyzed with Illumina chips. Differentially expressed genes using a 2-sided t-test comparing groups of replicate samples with Po0.05 were selected for exploratory analyses of regulatory pathways. MHC mismatched mouse H2b-H2k transplants compared outcomes with pDC from CCR7 KO vs WT donor. Indirect presentation of allogeneic peptides by pDC and classical DC was studied using H2b-H2 b/d transplants with H2Kd peptide recognition by TEA transgenic donor T-cells.

Results: pDC from BM and PLXPB grafts had higher CD62L expression but lower CCR7 and CXCR3 expression than pDC from GPB and UCB grafts. High CXCR4 and low CCR9 expression were seen in pDC from all graft sources. While differences in CCR7 expression between pDC from BM vs GPB grafts suggested homing differences might explain observed differences in TRM associated with the pDC content of the allograft, donor pDC from CCR7 KO mice had equal ability to modulate GvHD as pDC from WT mice. pDC in grafts from both human and mice were phenotypically immature and were relatively ineffective in activating T-cells through indirect presentation of alloantigen peptide, indicating that immune regulatory effects might be more important than antigen presentation. Supporting this hypothesis, gene expression patterns of human pDC from BM showed lower expression levels of genes related to adaptive immunity, and higher

[O108]

expression of genes associated with innate immunity, than pDC from GPB, UCB, or PLXPB grafts. Conclusion: Functional differences in immuneregulatory genes explain some of the observed differences in transplant outcome when stratifying recipients based upon between the content and source of donor pDC. BM pDC appear to be polarized towards activation of innate immunity while pDC from other graft sources are more polarized towards antigen presentation. Novel approaches of stem cell mobilization, such as plerixafor, may increase the content of immature pDC enriched for expression of that might be effective in recapitulating the beneficial effects of BM pDC with respect to preventing early posttransplant transplant related mortality.

References : Results: With a median follow-up of 43, 43 and 49 months in the 10/10 URD, 9/10 URD and UCB, hematopoietic recovery was slower in UCB compared to URD recipient: the cumulative incidence of neutrophils recovery at day 28 and platelet recovery 450 Â 10 9 /L at day 180 were respectively 73% and 56% in UCB against 96% and 85% in 10/10 URD and 95% and 75% in 9/10 URD (Po0.0001). While there was no significant difference in the day 100 cumulative incidence of grade 2-4 aGVHD: 31% in 10/10 URD, 39% in 9/10 URD and 32% in UCB, the 2 years cumulative incidence of cGVHD was significantly lower in UCB recipients: 16% against 37% in 10/10 URD (Po0.0001) and 29% in 9/10 URD (P ¼ 0.004). In multivariate analysis, there was no statistically significant difference in NRM between UCB recipients and 9/10 recipients (HR 1.58, P ¼ 0.13) or 10/10 recipients (HR 1.35, P ¼ 0.25) . In multivariate analysis, the cumulative incidence of relapse/progression was significantly lower in 10/10 URD recipients compared to UCB recipients (HR 0.60, P ¼ 0.02) and there was also a trend towards a decreased incidence of relapse in the 9/10 URD recipients (HR 0.62, P ¼ 0.07). Compared to UCB transplants there was no significant difference in PFS after 9/10 URD (HR 1.17, P ¼ 0.29 ) and 10/10 URD (HR 1.10, P ¼ 0.49 (Table 1) .

Results: There were no significant differences in pre-transplant characteristics between both groups ( Table 1) . The 30day cumulative incidence of neutrophil engraftment was similar in both groups (93% vs 100%), with a median time of 16 and 17 days, respectively (P ¼ 0.28). Similarly, the 60-day cumulative incidence of platelet engraftment was 70% vs 76%, in a median time of 39 and 35 days, respectively (P ¼ 0.11). Four cases among the haplo-cord group showed primary CB graft failure (3 of them showing third party donor cells engraftment) who were rescued by a second CB transplant in two cases and haplo-SCT in two cases. There were no graft failures in the haplo-SCT group. Grade II-IV acute GvHD was more frequent in the haplo-SCT group (20% vs 41%, P ¼ 0.06) and chronic GvHD showed a higher tendency in the haplo group (18% vs 35%, P ¼ 0.24). With a median follow-up of 51 months (10-122) in the haplo-cord group and 19 months in the haplo-SCT group, OS was 52% and 78% (P ¼ 0.15), respectively. EFS was 41% vs 72% (P ¼ 0.06), relapse rate was 22% vs 20% (P ¼ 0.71) , and TRM was 22% vs 8% (P ¼ 0.16), respectively. Conclusion: In our experience, albeit differences in follow-up and limited number of patients, myeloablative single cord supported by HLA-mismatched cells and haploidentical transplantation with post-transplant cyclophosphamide both offer valid alternatives for patients with acute leukemia. Engraftment rates are similar in both groups as well as relapse rate. Early toxic mortality is higher in the haplo-cord group while GvHD rates seem to be higher in the haplo-SCT patients. Further analysis including larger series and longer follow-up are needed to confirm these observations. Disclosure of Interest: None declared. Introduction: Studies on conditioning regimens prior to autologous stem cell transplant (ASCT) in lymphomas generally present a major pitfall, the miscellaneous mix of histologies composing the series, resulting in a reduced statistical power when focusing on a specific subset. We previously demonstrated (Visani et al, Blood 2011) the safety of a new conditioning regimen with bendamustine, etoposide, cytarabine, and melphalan (BeEAM) prior to ASCT in resistant/relapsed lymphoma patients (EUDRACTnumber2008-002736-15). The regimen showed long-lasting significant antilymphoma activity, with a 3-year PFS of 75%. Materials (or patients) and methods: We thus designed a single histology, phase II study to evaluate the efficacy of the BeEAM conditioning in resistant/relapsed diffuse large B-cell non-Hodgkin lymphoma (DLBCL) patients. The study was registered at at European Union Drug Regulating Authorities Clinical Trials (EudraCT) N. 2011-001246-14. The primary endpoint of the study is to evaluate the 1-year complete remission rate. Fixing the lowest acceptable rate as 55% and the successful rate as 70%, with a significance level a ¼ 0.05 and a power 1-b ¼ 0.90, the sample size was estimated in 88 patients. Results: Until now, 53 patients (median age 54 years, range 19-69) were enrolled. At the time of writing, we have data available for 44 patients with resistant/relapsed diffuse large B cell non-Hodgkin lymphoma were consecutively enrolled in the study. Briefly, 33/44 patients had advanced stage disease (III-IV), 14 were primary refractory and 30 had relapsed after a median number of 2 lines of therapy (range: 2-3). Eleven patients had 1 or more relevant comorbidities (range: [1] [2] [3] [4] [5] . 22 patients were in II or subsequent CR after salvage therapy, whereas 19 were in PR and 3 had progressive disease. A median number of 5.90x10 6 CD34 þ /kg cells (range 2,8-9,42) collected from peripheral blood was reinfused to patients. All patients engrafted, with a median time to ANC40.5x10 9 /L of 10 days. Median times to achieve a platelet count 420x10 9 /L and 450x10 9 /L were 12 and 16 days respectively. Ten out of 44 patients presented a fever of unknown origin (23%), whereas 21 patients (47%) presented a clinically documented infection. All patients received G-CSF after transplant for a median time of 8 days (range: 8-13). One patient died due to an incomplete hematological recovery after transplant, producing an overall transplant related mortality of 2.7%. Thirtyeight out of 44 patients are evaluable up to now for response to treatment. 31/38 (81.5%) obtained a CR, 3/38 a PR, whereas 4/38 did not respond to therapy. After a median follow-up of 12 months after transplant (range 2-30), 4/38 patients were refractory, 7/38 relapsed, and 27/38 (71%) are still alive, in continuous CR. Conclusion: In conclusion, this data provide evidence that the BeEAM regimen is safe and has promising efficacy in resistantrelapsed aggressive diffuse large B cell lymphoma.

pts with HL at risk of relapse post-ASCT (ClinicalTrials. gov #NCT01100502). The primary results showed that BV significantly improved progression-free survival (PFS) per independent review vs. placebo. Efficacy analyses by investigator review were performed in subgroups as these data may provide useful information for making treatment decisions. Materials (or patients) and methods: The AETHERA trial is a phase 3, randomized, double-blind, placebo-controlled study. After ASCT, pts received BV 1.8 mg/kg q3wk or placebo for up to 16 cycles. Pts were enrolled in 1 of 3 high-risk categories: refractory to frontline (FL) therapy, relapse o12 mos after FL therapy, and relapse Z12 mos after FL therapy with extranodal disease. The primary endpoint was PFS per independent review. PFS by investigator was also evaluated. Results: 329 pts were randomized at 78 sites in the US and Europe. Median age was 32 yrs (range, 18-76) and 53% were male. ASCT conditioning regimens were BEAM: 61%, CBV: 11%, or other regimens: 28%. 6% of pts received radiation as part of their transplant regimen. All pts had completed the treatment phase as of Aug 2013. Subgroup analyses by demographics, stratification factors, pt characteristics, and number of risk factors showed that PFS by investigator was improved across all groups; the hazard ratio comparing BV to placebo was o1 for all analyses (Table 1) . In an ad hoc analysis, pts with increasing numbers of risk factors for progression had progressively greater PFS benefit with BV vs. placebo. The most common adverse events (AEs) in the BV arm (vs. placebo) were peripheral sensory neuropathy (56% vs. 16%) and neutropenia (35% vs. 12%). A higher incidence of herpes simplex and zoster infections were observed with BV (11%) vs. placebo (3%); otherwise, opportunistic infections were rare and balanced between the arms. The incidence of pulmonary toxicity (by MedDRA SMQ) was low, but slightly higher with BV vs. placebo (5% vs. 3%). Two pts died r40 days of last dose of BV; cause of death was acute respiratory distress syndrome. Response to salvage therapy pre-ASCT CR 123 0.78 (0.42, 1.46 ) PR 113 0.44 (0.26, 0.76 ) SD 93 0.40 (0.23, 0.71) HL status after frontline therapy Refractory 196 0.55 (0.37, 0.83 ) Relapse o12 months 107 0.45 (0.27, 0.93 ) RelapseZ12 months with extranodal disease 26 0.30 (0.08, 1.16) 42 treatments pre-ASCT 149 0.32 (0.19, 0.54 ) B symptoms after failure of frontline therapy 87 0.29 (0.15, 0.55) Extranodal disease at time of pre-ASCT relapse 107 0.37 (0.20, 0.68) Risk Factors a Z1 329 0.50 (0.36, 0.70 ) Z2 280 0.40 (0.28, 0.57 Introduction: although high-dose chemotherapy (HDC) is the gold standard for the treatment of many relapsed or refractory lymphomas, the outcome is still unsatisfactory in some subsets of patients with adverse prognostic features. We treated 111 high-risk patients with a tandem strategy associating debulking with HDC followed by autologous stem cell transplantation (ASCT) and subsequent allogeneic SCT (tandem auto-allo) Materials (or patients) and methods: adult patients consecutively treated at two centers were included. Criteria for receiving tandem auto-allo were: HL and NHL refractory to first-line therapy, less than CR after first salvage treatment, relapse after prior ASCT, multiple relapses, histology of transformed follicular, mantle-, T-and NK-cell lymphoma and documented infections during hospital stay in 30 (27%), without significant differences between BEAM and HD-Mel groups. Among the 65 patients with active disease before ASCT, the overall response rate was 86% (n ¼ 56 responders) and those in CR were 34 (52%). No difference was observed in terms of response in the BEAM and HD-Mel groups (P ¼ 0.28). Allogeneic SCT donor was either HLA-identical (n ¼ 86), MUD 9/10 (n ¼ 2), haploidentical (n ¼ 20) or cord blood (n ¼ 3). 3-y OS of entire cohort was 68% (95% CI: 59-77), 3-y PFS was 61% (52-70), rates of acute GvHD grade 2-4 and chronic GvHD were 28% and 38% respectively. TRM rate was 18% (n ¼ 20). No difference between BEAM and HD-Mel group was observed for OS (73% and 64% respectively, P ¼ 0.40) or TRM (19% and 13%, P ¼ 0.44). Of note, OS of patients in CR before and after ASCT and OS of those obtaining CR after ASCT was superimposable ( Figure 1) Figure) . A multivariate analysis revealed that late allo-HSCT was a unfavorable prognostic factor for OS with a statistical significance (hazard ratio, 1.46 ; 95% CI, 1.01-2.11; P ¼ 0.04 (3) or complete response after autoSCT after a third line of chemotherapy (9 pts). Donors of alloSCT were HLA identical siblings in 38% of patients, matched unrelated in 35% and haploidentical donors in 27% of patients. Median age at transplant was 33 (range, 17-60). At alloSCT, 43% of patients were in CR, 30% were in PR, 27% in SD or PD. 83% patients performed alloSCT having relapsed o12 months from autoSCT or with primary refractory disease, 17% having relapsed 412 months after autoSCT.

Results: Median follow-up of surviving patients was 4.8 years (range, 0.5-10.5) . Overall survival (OS) was 76% at 1 year, 59% at 3 years and 55% at 5 years of follow-up. In multivariate analysis, only disease status before alloSCT significantly impacted the OS (P ¼ 0.003, HR 1.6, CI95% 1.2-2.2) , whereas donor and timing or relapse/refractoriness did not change OS (P ¼ 0.149, HR ¼ 1.3, and P ¼ 0.501, HR ¼ 1.3, respectively better control of the disease: this has occurred together with changes in donor type (from unrelated to family haploidentical), in GvHD prophylaxis (from MTX CyA ATG to post transplant cyclophosphamide), and in the conditioning regimen, combining thiotepa and intravenous busulfan; in addition DIPSS was somewhat lower in more recent patients, and also spleens were smaller. We believe these data suggest that alternative donor grafts are currently a therapeutic option for patients with MF. Disclosure of Interest: None declared.

Long follow-up data show that survival rate after allogeneic stem cell transplantation in patients with CLL is higher for younger patients because of significantly lower NRM during the whole follow-up period: a retrospective study of the CMWP M. Van Gelder 1,* on behalf of CMWP . Among pediatric pts, 28% were aged 0-23 months, 52% were aged 2-11y, and 20% were aged 12-16y. Day þ 100 survival data by age group are available for 526 pts post-HSCT and 62 pts post-CT (Table) , with rates of sVOD post HSCT of 55% in pediatric pts and 50% in adults. HPSE has been shown to be involved in inflammation and may therefore play an important role in the pathogenesis of VOD of the liver. The purpose of this study was to identify an association between HPSE gene single nucleotide polymorphisms (SNPs) and VOD of the liver after allogeneic HSCT in childhood.

Materials (or patients) and methods: 160 children (median age, 14 years) who underwent allogeneic bone marrow (n ¼ 91) or peripheral blood stem cell transplantation (n ¼ 69) in a single center and their respective donors were genotyped of HPSE for rs4693608 and rs4364254 using TaqMan real-time polymerase chain reaction. The donor was HLA-matched unrelated in 63% of transplants and HLAidentical related in 25% of transplants. Conditioning regimen was myeloablative in all cases and based on busulfan in 46% of transplants or total body irradiation in 33% of transplants. Two forms of post-transplant immunosuppression predominated, cyclosporine A and methotrexate in 50% of transplants and cyclosporine A alone in 30% of transplants.

Results: 160 donor/patient pairs were analyzed. Cell samples from the patient were available in 155 cases and from the donor in 153 cases. Genotype AG of HPSE rs4693608 SNP was found in 82 patients (53%), AA in 49 patients (32%), and 24 patients were homozygous for GG (15%). Analysis of HPSE rs4364254 SNP revealed a similar distribution for TC (n ¼ 69, 44%) and TT (n ¼ 68, 44%) and a frequency of 18 patients (12%) for CC. VOD of the liver was observed in 12/160 patients (8%). If VOD of the liver was diagnosed all of our patients were treated with defibrotide as early as possible. Altogether, 4/12 patients with VOD of the liver (33%) died of VOD, whereas 8/ 12 patients with VOD of the liver (67%) survived VOD. Early medical intervention is most probably the reason for this high cure rate. Patients with HPSE genotypes GG or AG of rs4693608 (G4A) had a significantly reduced incidence of VOD of the liver on day 100 after HSCT compared to patients with genotype AA (5% vs. 14%, P ¼ 0.038). In addition, incidence of VOD in patients with genotype CC or CT of rs4364254 (C4T) was significantly decreased in comparison to patients with genotype TT (2% vs. 15%, P ¼ 0.004). Interestingly, no patient with genotype CC developed VOD. Because both SNPs co-occur in vivo, we generated subsets: AA-TT, GG-CC and a group with remaining SNP combinations. We found significant differences between all three patient groups (P ¼ 0.035). Patients with AA-TT showed the highest incidence of VOD of the liver (17%), while VOD was not observed in patients with GG-CC (0%) and residual combinations were numerically in-between (5% allogeneic, n ¼ 42) were considered at risk of HBV reactivation for the following criteria of positivity: 1) donor and/or recipient anti-HBc and/or 2) anti-HBe and/or 3) anti HBs þ /-. Lamivudine prophylaxis was given to 51 out of 58 patients (88%). Overall, 4 patients, 1 autologous and 3 allogeneic recipients, developed HBV reactivation at a median time of 40 months (range, 28-53) after HSCT. Of these 4 patients, 3 were at risk of reactivation and 2 of them were not receiving prophylaxis. One was HBV-negative at the time of pretransplant screening. The cumulative incidence of HBV reactivation was 4,9% for patients at risk and 1%.for the entire HSCT population studied. One patient reactivated HBV infection during lamivudine prophylaxis. HBV isolate genotype was studied in all reactivated patients. Two isolates showed HbsAg escape mutations (123 N, 124Y, 126I, 145 K, 145 R, 144E, R122K, T140S) and in 1 lamivudine drug resistance mutations (181 S and L801L), 3 of them resulted HBV genotype D and one resulted genotype F. Conclusion: The low rate of reactivation in our cohort is probably due to the accuracy of pre-transplant donor/ recipients screening and to the extensive prophylaxis program. However, a case of HBV reactivation occurred in a patient negative at serological screening (sero-negative HBV occult infection). The detection of immune-escape HBsAg mutations, associated with lack of recognition by neutralizing immune activity and by diagnostic assay in most HBVreactivated patients, supports the role of these mutations in the process of immune-suppression driven HBV reactivation. Disclosure of Interest: None declared. Introduction: Polyomavirus (PV) hemorrhagic cystitis (HC) is a severe complication after haploidentical hematopoietic stem cell transplantation (haplo HSCT). In the setting of solid organ transplantation, the use of Tacrolimus (FK) has been associated with higher risk of PV-HC compared with cyclosporine A (CsA). The aim of our study was to investigate risk factors of PV-HC in haplo-HSCT recipients, with particular focus on immunosuppressive agent used as graft-versus-host disease (GVHD) prophylaxis. Materials (or patients) and methods: We retrospectively analyzed the incidence, risk factors and outcome of PV-HC in 149 consecutive adult patients (pts) undergoing haplo HSCT due to hematological malignancies between 2009 and 2014 at our two institutions. All HSCTs were T-cell replete and included post-transplant high-dose cyclophosphamide as part of GVHD prophylaxis. PV-HC was defined as PV detection in urine by PCR testing in association with clinical symptoms of HC without other concurrent genitourinary conditions. Variables tested as potential risk factors to develop a PV-HC were: conditioning regimen, age, diagnosis, stem cell source, levofloxacin prophylaxis, acute GVHD, interval between diagnosis and HSCT, pre-HSCT therapy lines, previous subdiaphragmatic radiotherapy, n. of CD34 þ cells infused, neutrophil and platelet engraftment, FK vs CsA-based immunosuppressive regimen.

Results: Main pts' characteristics are shown in Table 1 . After a median follow-up time of 13 months from transplantation, ten (7%) pts developed a PV-HC. PV-HC occurred in 6/52 (12%) pts receiving FK and 4/97 (4%) pts receiving CsA (P ¼ 0.10). Etiology was BK virus in 9 (90%) cases and JC virus in one (10%) case. Median time of PV-HC diagnosis was 30 (7-68) days post-HSCT. In the multivariate analysis, myeloablative (MYA) or reduced-intensity conditioning (RIC) regimens (HR ¼ 4.25, 95% CI: 1.18-15.33 ; P ¼ 0.03) and FK-based immunosuppression (HR ¼ 3.78, 95% CI: 1.05-13.64 ; P ¼ 0.04) were the only two independent factors associated with PV-HC. Regarding treatment, two pts received Cidofovir therapy, one pt benefited from immunosuppression tapering, whereas only supportive therapy was required for the remaining 7 pts. Notably, all pts achieved complete remission of symptoms. 3.9 (0.8-14.0) relapse who would benefit from further treatment intensification early after transplant. Introduction: The prognosis for patients with hematologic malignancies (HM) who relapse after alloHCT is dismal, and immune checkpoint modulation with CTLA4 blockade is a novel strategy to enhance the graft versus tumor (GVT) effect. Materials (or patients) and methods: This is a phase I/Ib study of the CTLA4 blocking antibody ipilimumab to treat patients with any HM histology who relapse after alloHCT. The primary objectives are to determine the MTD and evaluate safety. Secondary objectives include a preliminary evaluation of efficacy and changes in immune cell phenotype. Ipilimumab was given at 3 mg/kg or 10 mg/kg IV q3 weeks for 4 cycles of induction, followed by maintenance dosing q12 weeks up to 1 year. Disease-specific response criteria were assessed at the mid-point (7 weeks), end of induction (13 weeks), and throughout maintenance. Immunophenotyping was performed by 8-color flow cytometry and analyzed by FACSDiva. Results: Twenty-three patients are enrolled to date. In the phase I portion, 6 patients were treated at 3 mg/kg and 7 patients were treated at 10 mg/kg. An MTD was not reached, and 10 patients subsequently enrolled in the phase Ib expansion cohort at 10 mg/kg. The median number of prior therapies excluding transplant was 3 (range 2-11), and 14 patients had received prior therapy for their post transplant relapse. Histologies included cHL (n ¼ 7), AML (n ¼ 7), NHL (n ¼ 4), and MDS (n ¼ 2), and 1 patient each had MM, MPN, and ALL. The median age at enrollment was 55 (range 22-75). Immune-related adverse events (irAEs) were observed in 4 patients, and included ITP (n ¼ 1), diarrhea (n ¼ 2), pneumonitis (n ¼ 3), and colitis (n ¼ 1), all of which were generally reversible with steroids with most patients resuming ipilimumab dosing. Three DLTs have been observed, including 2 cases of chronic GVHD (both liver, mild) and one patient death due to presumed sepsis in the context of severe irAEs, including grade 3 colitis and grade 4 pneumonitis. Nine patients remain on treatment, and 11 patients discontinued due to progressive disease, 2 patients due to cGVHD, and 1 patient with TRM. In an interim efficacy analysis, 7/19 (36.8%) patients evaluable for response had anti-tumor activity with clinical benefit. Four patients achieved formal responses by disease-specific criteria, including a cHL patient with PR with dramatic reduction in nodal and extranodal disease with a complete marrow response at 7 weeks (baseline 90% involvement), a MM patient with a PR with near resolution of a lung plasmacytoma, and two patients with AML who achieved CR by 7 weeks. The median follow-up time among survivors is 5.4 months, and 6 month overall survival is currently 57%. Immunophenotyping studies revealed that the ratio of regulatory T cells to conventional T cells decreased between 24% and 41% after ipilimumab treatment.

Conclusion: Multiple doses of ipilimumab given to patients with relapsed HM after alloHCT were generally well-tolerated, with cGVHD and irAEs observed in a small number of patients. Anti-tumor activity was observed, both in lymphoid malignanices and for the first time in myeloid malignancies, including 2 patients with AML who achieved CR. The ratio of regulatory to conventional T cells decreased with CTLA4 blockade. The phase Ib expansion cohort at 10 mg/kg continues to accrue and will provide additional efficacy, safety, and correlative data. Disclosure of Interest: None declared. adverse risk cytogenetic (P ¼ 0.04). In patients who relapsed between 6 and 12 months post-transplant more than one course of chemotherapy to achieve CR1 (P ¼ 0.02), adverse risk cytogenetics (P ¼ 0.05) and FLT3 ITD positivity (P ¼ 0.00002) all predicted for relapse. Finally only CMV positivity predicted for relapse risk for relapse more than 12 months post-transplant (P ¼ 0.05).

Conclusion: This study demonstrates for the first time that a complex interaction of disease specific and transplant factors determine the kinetics of relapse post-transplant. In addition to identifying that heterogenous factors related to transplant characteristics and disease biology determine the timing of disease relapse these data confirm the importance of early intervention post-transplant in patients allografted for AML and and will assist in the design of novel therapeutic strategies.

Disclosure of Interest: None declared. Introduction: Bone marrow (BM) is the recommended stem cell source in hematopoietic stem cell transplantation (HSCT) for bone marrow failure (BMF). Recent large studies in aplastic anemia showed that the use of peripheral blood stem cells (PBSC) increased the risk for chronic graft-versus-host disease without reducing graft failures and lead to an inferior survival after HSCT. However, a substantial proportion of HSCT is still performed with PBSC. Ease of collection and a more rapid engraftment might favor PBSC in the short term and more background information to clarify the situation is warranted. Materials (or patients) and methods: The current global practice of HSCT for BMF was studied for potential macroeconomic factors associated with the selection of stem cell source. Introduction: The outcome of alternative donor hematopoietic stem cell transplantation (HSCT) for the patients with severe aplastic anemia (SAA) who lack an HLA-matched familial donor has improved recently. This study was aimed to compare the treatment outcome of immunosuppressive therapy (IST) with alternative donor HSCT in children with SAA. Materials (or patients) and methods: Medical records of SAA patients who received IST and/or alternative donor HSCT from umbilical cord blood (UCB), haploidentical peripheral blood stem cells (PBSC), unrelated bone marrow (UBM), or unrelated PBSC (UPBSC) between June 1996 and December 2012 were reviewed, and data were analyzed retrospectively. The Kaplan-Meier method was used to calculate the estimated survival rates which were compared with log-rank test.

Results: Of 59 patients with SAA, 42 patients received IST and/ or alternative donor HSCT. Nineteen patients received IST as an initial treatment modality, of whom 13 patients failed to respond at 6 months from the initiation of treatment. Thirtyfour patients received alternative donor HSCT either following IST (n ¼ 11) or as frontline therapy (n ¼ 23; 11 UBMT, 11 UPBSCT, 8 UCBT, 4 haploidentical PBSCT). The failure-free survival rate (FFS) of IST group was 31.6%, while that of frontline HSCT group was 91.3% (Po0.001). Patients who received HSCT following IST showed an inferior event-free survival rate as compared with those who received frontline HSCT (50.9% vs 91.3%; P ¼ 0.015).

Conclusion: The outcome of alternative donor HSCT in our cohort was higher than usually expected, especially in those who received HSCT without prior IST. These results suggest that frontline alternative donor SCT might be a better treatment option than IST for children with SAA who lack an HLA-matched familial donor. Disclosure of Interest: None declared.

Increased risk of development of malignancies in adult patients treated with antithymocyte globulin as first line treatment for aplastic anaemia during a follow-up period of thirty years. J. V. D Introduction: Adult aplastic anemia (AA) is considered to be an immune-mediated disease with bone marrow aplasia and pancytopenia. Patients can be treated with either hematopoietic stem cell transplantation (HSCT) or immunosuppressive therapy (IST) with antithymocyte globulin (ATG) and achieve long-term survival. Treatment related long term toxicity and outcomes should guide the decisions about first line therapy. A long-term toxicity of both treatments is the development of malignancies. Follow-up studies until 15 years after IST in AA patients showed cancer development in up to 25% of patients. Analysis of cohorts with a longer duration of follow up are lacking and the incidence of malignancies in AA patients was never compared with a control population and the contribution of second line HSCT to this risk is not clear. We assessed the long-term cumulative incidence of malignancies in a cohort of adult patients with AA who were treated with ATG as first line treatment at Leiden University Hospital between 1980 and 2008 with HSCT and death as competing risks and compared this incidence to cancer development in the general Dutch population. Materials (or patients) and methods: 93 Patients treated with first line ATG between 1980 and 2008 were entered in this single-centre retrospective cohort study. Primary endpoint was the cumulative incidence of malignancies with death and second line treatment with HSCT as competing risks. Secondly, the cumulative incidence in our cohort was compared to that of the sex-and age matched general Dutch population derived from the Dutch cancer registration, using standardized incidence ratio (SIR). Furthermore, several patient-dependent (age), disease-dependent (severity of disease) and treatmentdependent (addition of ciclosporin to ATG) variables were assessed as possible prognostic factors for cancer development, using cumulative incidence curves, Gray's test and the Cox proportional hazard model for the cause-specific hazard. SPSS Statistics 20.0 and R were used for all data analyses. Ethical approval for data collection and data analysis was obtained.

Results: Median length of follow-up of surviving patients without HSCT was 17.4 years (range 0.2-32.2 years) , with a cumulative malignancy rate of 29% after 30 years. The incidence of malignancies was increased compared with the general Dutch population (SIR: 3.35 (95% CI: 1.98-5.08 )). The cumulative incidence of developing MDS/AML was 6% after 30 years (occurring at 5-17 years after start of treatment). In the multivariate analysis, age at time of IST was significantly associated with a higher malignancy rate (HR: 1.52 per 10 years; P ¼ 0.01). HR for addition of ciclosporin was1.87 (P ¼ 0.25) and HR for NSAA vs (V)SAA HR was 1.65 (P ¼ 0.32)).

The risk of the development of malignancies in AA patients treated with ATG is increased compared to the risk in the age and sex matched Dutch population for both solid and hematologic malignancies. We are not able to discern whether the increased risk is due to the treatment with IST or due to the biology of the disease. Long follow up studies in age matched AA patients receiving HSCT are needed to compare this long term risk between both treatments. These results underscore the recommendation in international Guidelines to follow all patients with AA after initial IST lifelong.

Disclosure of Interest: None declared. 1 Hematopoietic stem cell transplantation, 2 Molecular Biology, DMITRIY ROGACHEV CENTER FOR PEDIATRIC HEMATOLOGY, ONCOLOGY AND IMMUNOLOGY, Moscow, Russian Federation

Introduction: Results of matched unrelated and, to a lesser extent, haploidentical transplantation in severe aplastic anemia have improved over the last decade. The major factors behind this improvement are high-resolution HLA-typing, intensified immune suppression with fludarabine and alemtuzumab and modern supportive care. We implemented a new graft manipulation method, TCR-alpha/beta depletion, to further improve GVHD control and overall results of transplantation in a group of SAA patients, refractory to two courses of ATG/CsA immune suppression. Materials (or patients) and methods: Twenty six patients with SAA, median age 11,4 (2,5 -22) years, 16 male/ 10 female, received MUD (20) or haploidentical (6) transplantation with TCR alpha/beta and CD19 depletion between 01. 09.2012 and 01.09.2014 . Patients received a median of 2 ATG/CsA courses. Core preparative regimen included cyclophosphamide 25 mg/ kg x4, fludarabine 30 mg/kg x 5, ATG (horse) 25 mg/kg x 4 and total lymphoid irradiation 6 Gy total dose. One patient received alemtuzumab instead of ATG due to anaphylaxis. In all cases G-CSF-mobilized PBSC were used as graft source. TCR-alpha/beta and CD19 depletion were performed with CliniMACS Plus according to manufacturer's recommendations. Final graft contained 10x106 of CD34 þ /kg and 17x103 of TCRa/b þ /kg. Post-transplant immune suppression included short methotrexate and tacrolimus till day þ 45-60.

Results: Primary engraftment was registered in all patients, median 15 days for neutrophils and 14 for platelets. Secondary graft failure/rejection developed in 2 patients, both transplanted from MUD. Both patients were salvaged with a second MUD transplantation. Acute GVHD was registered in 4 patients, grade 2 skin in 2 and grade 3 skin þ gut in 2 patients. Cumulative incidence of aGVHD grade 2-3 was 15%(95%CI:6-38). Chronic GVHD was registered in 1 patient, cumulative incidence 4%(95% CI:0,7-30). Two patients (1 from each cohort) died of CMV pneumonia at 4 and 8 months respectively. In both cases CMV sero status was D-/R þ . At 2year follow-up cumulative incidence or TRM is 12%(95%CI:0,3-46), KM estimte of overall survival is 88%(95%CI:72-100), 92%(95%CI:76-100) in the MUD group, 67%(95%CI:13-100) in the haplo group. Conclusion: TCR alpha/beta depletion is a highly effective method of GVHD prophylaxis in SAA patients. Use of this platform makes alternative donor transplantation in SAA a safe therapeutic option. Further improvement will depend on new strategies to control viral infections. Disclosure of Interest: None declared.

TP:129 (34) vs 173 (36) cm/s, respectively; (Po0.001), and were decreased more significantly following HSCT than on TP: mean(SD) D: -42 (31) vs þ 6 (34), respectively. The percentage of patients with normal velocities was significantly higher post-HSCT (29/32) than in the TP arm (16/32) (P ¼ 0.001) Conclusion: This prospective national trial comparing TP vs. HSCT in SCA-patients with a history of abnormal velocities shows for the first time that HSCT significantly results in a greater decrease in velocities than TP, and has very little toxicity. These results suggest HSCT is the treatment of choice for SCA-children with a history of abnormal-TCD and genoidentical donor. Disclosure of Interest: None declared. Introduction: Busulfan (Bu) is the backbone of the conditioning regimen for patients with sickle cell anemia (SCA) undergoing BMT. Patients with SCA might predispose to transplant-related neurological and pulmonary toxicities due to pre-existing disease-related cerebrovascular and lung injury. Bu therapy appears to be an important contributing factor in this context. To date, no studies have evaluated the pharmacokinetic (PK) parameters of intravenous Bu (IVBu) for subsequent doses in patients with SCA. Materials (or patients) and methods: We studied IVBu PK parameters and clinical outcomes of 36 children with SCA undergoing BMT from HLA matched siblings. The median age of patients was 10.4 (range, 1.7-17.1 ) years. Conditioning regimen consisted of Bu/Cy200/ATG (n ¼ 12) or Flu/Bu/CY200 (n ¼ 24). Six patients (17%) had stroke, 11 (31%) exhibited gliosis due to previous brain injury, 13 (36%) had repeated acute chest syndrome, and 8 (22%) patients were on chronic transfusions. IVBu was administered every 6 h for 4 days with PK-guided dose adjustment to target a conservative area under the concentration versus time curve (AUC) range of 900-1350 mMol*min. The role of glutathione S-transferase (GST) polymorphisms has also been investigated. Results: All patients had sustained engraftment. A repeated measures ANOVA showed that the first-dose Bu clearance (4.24 ml/min/kg) was significantly higher than after dose 5 (3.70 ml/min/kg; Po0.0005), dose 9 (3.58 ml/min/kg; Po0.0005), and dose 13 (3.42 ml/min/kg; Po0.0005). Such differences in clearance have never been described in patients with SCA. After the first-dose, 69% of patients achieved the target range.None of the patients developed hepatic VOD. No patient developed grade Z3 toxicity. There was no association between GST polymorphism and the PK of Bu. We adapted a new dose-adjustment strategy targeting exposures to the lower end (900 mMol*min) of the AUC range after the first dose of Bu to avoid unnecessary dose increases on subsequent days due to differences in clearance. This strategy enabled most patients (90%) to maintain the AUC within therapeutic range following dose adjustments. A 3-year probability of OS and sickle-cell free survival were 91% (95% CI, 75-97%). There was no correlation between any Bu PK parameters and survival, toxicity, acute GVHD, chronic GVHD, and transplant-related mortality.

Conclusion: This study found that the PK behavior of IVBu in children with SCA is characterized by significantly higher Bu clearance after the first-dose compared with subsequent daily doses. This finding allowed us to adapt a new dose adjustment strategy after the first dose of Bu, which effectively prevented subsequent dose readjustments. Conservative AUC range and targeting exposures to the lower end of the range after the first dose was associated with negligible toxicity, and high engraftment and sickle cell-free survival rates. Disclosure of Interest: None declared. However, the probability to find a suitable door is only 25-50%.

Since most of these patients (pts) could not find the potential donor, we would like to investigate the haploidentical donor HSCT (Hapo-SCT) in thalassemia. Materials (or patients) and methods: Between Jan 2013 and December 2014, a total of 116 severe thalassemia patients (pts) underwent HSCT in our center. Sixty five pts underwent MRD-HSCT, 33 pts underwent MUD-HSCT and 18 pts underwent haplo-HSCT. For haplo-HSCT, 10 subjects were male and 8 were female. The median age was 16 yrs (range; 2-22). Thirteen of 18 received stem cells from mother and 5 from father. Ten of 18 were high risk class 3. These high risk class 3 pts received hydroxyurea 20 mg/kg/d for at least 3 months prior to HSCT. All pts received 2 courses of pre-transplant immunosuppression (PTIS) consisting of fludarabine (Flu) 40 mg/m 2 /d together with dexamethasone (dex) 25 mg/m 2 /d for 5 days. After 2 courses of PTIS, all pts received a reducedtoxicity conditioning (RTC) regimen consisting of thymoglobulin 1.5 mg/kg/d (d-11 to d-9) , Flu 35 mg/m 2 /d i.v. (-7 to -2) each dose immediately followed by busulfan (Bu) 130 mg/ m 2 once daily i.v. (d-7 to d -4) GVHD prophylaxis consisted of cyclophosphamide (Cy) 50 mg/kg/d (d þ 3 to d þ 4). Tacrolimus or sirolimus was given for 6 months to 1 yr started together with mycophenolate mofetil on d þ 5, the latter was quickly tapered after 2 months. T-cell repleted peripheral blood stem cells (PBSC) were given to all pts, targeting a CD34 þ dose of 7-16 x 10 6 cells/kg. Results: Sixteen of 18 were engrafted with full donor chimerism (100%) while 2 pts suffered graft failure. However, these 2 pts received second transplant on day þ 30 with minimal conditioning regimen and additional PBSC after which they achieved full donor chimerism. The median time to neutrophil engraftment was 18 days (range; 14 -22) . Six pts had acute GVHD gr I, 3 grade II and 1 gr IV. Only one had extensive chronic GVHD. At this time, 17 of 18 pts survive thalassemia-free and have sustained full donor chimerism (100%). Event free survival (EFS) and overall survival (OS) rates are 95%. The median follow up time is 12 months (range 4-20). The EFS rates among MRD, MUD and Haplo-HSCT in our center are 88%, 82% and 95% respectively (P ¼ 0.46). Conclusion: Haploidentical HSCT for high risk thalassemia pts with our novel approach is safe and should be considered as modality to secure thalassemia-free survival with a low risk for graft rejection and treatment-related mortality. In view of our results, we suggest that all thalassemia pts even those with high risk class 3 features should be offered the chance for cure with HSCT. Disclosure of Interest: None declared. Introduction: Allogeneic hematopoietic stem cell transplantation (HSCT) is a curative option for many patients with hematological malignancies. Graft versus host disease (GVHD) mediated by donor alloreactive T cells remains a major obstacle and limits its wider application. It is now well established that T regulatory cells (Tregs) are critical for the maintenance of self-tolerance, and a deficiency or dysfunction in these cells is thought to be a key factor in the development and progression of GVHD. Numerous murine studies have shown the benefit of adoptive transfer of Tregs in the setting of HSCT. Notably, the success of this approach is improved if the transferred Tregs are specific for antigens expressed by the host. A robust method to generate homogeneous populations of alloantigen specific Tregs in humans would be a major step towards realizing the goal of using these cells to suppress GVHD. Traditional approaches to generate antigen specific T cells include repetitive cycles of re-stimulation with antigen in vitro, sorting of tetramer positive cells, or over-expression of a T cell receptor for a specific antigen. These methods produce limited cell numbers and require modification for each individual patient depending on their own tissue type and that of the transplanted patient. We hypothesized that a more efficient approach to generating antigen specific Tregs would be to genetically engineer them to express alloantigen-specific chimeric antigen receptors (CARs). Materials (or patients) and methods: We generated an HLA-A2 specific CAR by cloning the immunoglobulin heavy and light chain variable regions from an anti-HLA-A2 antibodysecreting hybridoma. These regions were fused into a single chain antibody, which was then linked to intracellular signalling domains from human CD28 and CD3zeta. Surface expression and antigen-specificity of the A2-CAR was confirmed by flow cytometry to detect expression of a myc epitope tag on the extracellular portion of the CAR and binding to an HLA-A2 tetramer, respectively. CD4 þ CD25 hi CD45RA þ naïve Tregs (nTregs) and CD4 þ CD25 lo CD45RA þ naïve Tconv (nTconv) cells were sorted from the blood of HLA-A2donors, stimulated with artificia APCs and transduced with lentivirus encoding the A2-CAR or a control CAR specific for HER2. Results: Transduced Tregs maintained their expected phenotype, including high levels of FOXP3, CTLA-4, CD25 and Helios, and low levels of CD127 and IL-2 compared to Tconv cells. To test antigen specificity, the transduced T cell lines were stimulated with K562 cells expressing HLA-A2 or HER2: only K562 cells expressing HLA-A2 stimulated proliferation of A2-CAR expressing Tregs and Tconv cells. CAR-stimulated Tregs also suppressed the in vitro proliferation of CAR-stimulated Tconv cells. To demonstrate the in-vivo suppressive capacity of A2-CAR Tregs, NSG mice were reconstituted with HLA-A2 þ PBMCs in the absence or presence of different ratios of control-transduced or A2-CAR transduced Tregs. Preliminary results suggest that A2-CAR Tregs have a superior ability to delay the onset of GVHD compared to polyclonal Tregs. Conclusion: Human Tregs can be efficiently transduced to express functional, alloantigen-specific CARs that confer antigen-specific specific suppression of effector T cells both in-vitro and in-vivo. These data supports the feasibility of this strategy to re-direct Tregs for transplant-related therapies. Disclosure of Interest: None declared.

CAR spacers including NGFR domains allow efficient T-cell tracking and mediate superior antitumor effects M. Casucci 1,* , L. Falcone 1 , B. Camisa 2 , F. Ciceri 3 , C. Bonini 2 , A. Bondanza 1 1 Innovative Immunotherapies Unit, 2 Experimental Hematology Unit, 3 Clinical Hematology and Bone Marrow Transplantation Unit, San Raffaele Hospital Scientific Institute, Milano, Italy Introduction: Chimeric antigen receptors (CARs) frequently include an IgG1-CH2CH3 spacer conferring optimal flexibility for antigen engagement and allowing the selection and tracking of CAR-expressing T cells. A serious drawback of CH2CH3-spaced CARs is however their interaction with Fcg receptors (FcgRs). Indeed, this antigen-independent binding may lead to the unintended elimination of cells expressing these receptors (mainly phagocytes), foster the development of non-specific immune reactions and drastically decrease the efficacy of the strategy due to the premature clearance of transduced T cells in vivo. Materials (or patients) and methods: We designed and constructed novel CAR backbones by substituting the IgG1-CH2CH3 spacer with regions from the extracellular portion of the low-affinity nerve-growth-factor receptor (LNGFR), differing for the length and potential binding to NGF. In particular, we used our recently developed CD44v6-specific CAR as a model for comparing the antitumor activity of the different LNGFR-based designs both in vitro and in vivo. Results: After transduction, all constructs could be identified on the T-cell surface using anti-LNGFR antibodies, indicating that they were correctly processed, mounted on the cell membrane and still recognized by anti-NGFR antibodies. As a consequence, all the LNGFR-based spacers allowed selecting CAR-T cells with immune-magnetic beads coupled to anti-NGFR antibodies, without interfering with their expansion and functional differentiation after activation with CD3/CD28 beads plus IL-7 and IL-15. Most importantly, LNGFR-spaced CAR-T cells maintained potent cytotoxic, proliferative and cytokine-release activity in response to CD44v6-expressing leukemia and myeloma cells, while lacking antigen-independent recognition through the FcgRs. Noticeably, even at supraphysiological NGF concentrations, the LNGFR-spaced CD44v6-CAR.28z CAR T cells were not induced to proliferate, indicating the absence of signaling via soluble NGF. Strikingly, LNGFRspaced CAR-T cells better expanded and persisted in vivo compared to CH2CH3-spaced CAR-T cells and mediated superior antitumor effects in a well-established tumor disease model. Interestingly, we demonstrated that the premature disappearance of CH2CH3-spaced CAR-T cells was due to engulfment by mouse phagocytes, a phenomenon not occurring with LNGFR-spaced T cells.

In conclusion, we have demonstrated that the incorporation of the LNGFR marker gene directly in the CAR sequence allows for a single molecule to work as a therapeutic and as a selection/tracking gene and shows an increased efficacy/safety profile compared to the IgG1-CH2CH3 spacer. Disclosure of Interest: None declared.

University Children's Hospital, Würzburg, 10 Clinic of Pediatric Oncology, Heinrich-Heine-University Düsseldorf, 11 University Children's Hospital, Essen, 12 University Children's Hospital, Münster, 13 Department of Pediatrics, Jena University Hospital, Germany

Introduction: Viral infections represent an important cause of morbidity and mortality in immunosuppressed patients post hematopoietic stem cell transplantation (HSCT). As viral infections often remain refractory to pharmacologic treatment, alternative treatment strategies such as immunotherapy are required. Adenovirus (AdV) is the predominant diseasecausing pathogen in pediatric HSCT. Materials (or patients) and methods: In a clinical trial we analyzed safety and efficacy of ex vivo adoptive T-cell transfer (ACT) with hexon-specific T cells, predominantly of T HELPER -1 phenotype. Thirty patients suffering from chemo-refractory AdV disease or viremia after HSCT were treated with AdVspecific T cells generated by IFN-g capture technique.

Results: AdV-specific T cells were successfully isolated in 100% of cases using the adenoviral hexon antigen and were directly infused into the patients without further ex vivo expansion steps. AdV-specific T-cell grafts were composed of a mixture of naïve, central memory, effector memory and effector T-cell populations with a predominance of late effector stages, indicating proliferation and effector potential of the transferred T cells. In all thirty patients, ACT was feasible without acute toxicities or significant onset of graft-versus-host disease. ACT led to antiviral immunity in vivo up to 6 months with viral control, resulting in complete clearance of viremia in 86% of patients with antigen-specific T-cell responses. Efficacy of adoptive T-cell transfer was independent of the initial T-cell dose. After a follow up of 6 months post ACT, overall survival was markedly increased in responders (mean: 122 days, 15 survivors) as compared to non-responders who all died shortly after ACT (mean: 24 days, no survivors). AdV-related mortality was 100% in non-responders compared to 9.5% in responders (Z1 log reduction of DNA copies/ml post ACT), indicating a strong correlation between virus-specific immunity and virus control. Conclusion: In conclusion, ex vivo ACT of AdV-specific T HELPER -1 cells was well tolerated and led to successful and sustained restoration of T-cell immunity, correlated with virological response and protection from virus-related mortality. This cellular immunotherapy is a short-term available and broadly applicable treatment. Disclosure of Interest: None declared.

[O164]

Putkonen 6 , T. Kuittinen 1 , J. Pelkonen 2,7 , P. Mäntymaa 7 , K. Remes 6 , V. Varmavuo 1 , E. Jantunen 1 O094 O096 Comparison of RIC-alloHCT and autoHCT for 455 years old patients with acute lymphoblastic leukemia: an analysis from Acute Leukemia Working Party of the EBMT S

United Kingdom, 7 Erasmus MC-Daniel den Hoed Cancer Centre

Double-Blind, Placebo-Controlled Phase 3 Study of Brentuximab Vedotin in Patients at Risk of Progression Following Autologous Stem Cell Transplant for

United States, 4 Szent Istvan & Szent Laszlo Corporate Hospital Hematology & Stem Cell Dept

United States, 6 Department of Bone Marrow Transplantation & Oncohematology, Maria Sklodowska-Curie Institute of Oncology

United States, 10 Istituto Nazionale dei Tumori, Milano, 11 IRCCS Azienda Ospedaliera Universitaria

United States Introduction: The AETHERA trial was initiated to evaluate whether brentuximab vedotin (BV) can prevent progression in Conclusion: BV improved post-ASCT PFS across all subgroups of pts. Pts with more risk factors had the most benefit from BV. AEs were consistent with the known safety profile of BV

References: 1. Lion T. Adenovirus Infections in Immunocompetent and Immunocompromised Patients

European guidelines for diagnosis and treatment of adenovirus infection in leukemia and stem cell transplantation: summary of ECIL-4

Adoptive transfer and selective reconstitution of streptamer-selected cytomegalovirus-specific CD8 þ T cells leads to virus clearance in patients after allogeneic peripheral blood stem cell transplantation

Lowest numbers of primary CD8 þ T cells can reconstitute protective immunity upon adoptive immunotherapy

Donor EBV status has

Introduction: Epstein-Barr virus (EBV) has been a major cause of post-transplant lymphoproliferative disorder after allogeneic stem cell transplantation (HSCT). The impact of the donor (D) and recipient (R) serologic status on survival, relapse-free survival, relapse incidence, non-relapse mortality and incidence of graft-versus-host disease was unknown so far. Objective: We analyzed the influence of the donor's and recipient's EBV status on allo-HSCT transplant outcomes. Materials (or patients) and methods: 11,364 allo-HSCTs performed due to acute leukemia

EBV-seropositive donors also had no influence on relapse-free survival

57), and non-relapse mortality

HSCT recipients receiving grafts from EBV-seropositive donors had higher risk of acute GVHD

After adjusting for confounders (donor type, conditioning, stem cell cource, patient age, gender match, T-cell depletion, year of transplant), D þ serostatus had an impact on development of acute GvHD

P ¼ 0.09, and for cGVHD

All patients engrafted, with a median time to neutrophil and platelet recovery of 18 days (13-45) and 16 days (9-100), respectively. Post-HSCT recovery of lymphocyte subsets was broad and fast: median time to CD34100/ml was 28 days, to CD44200/ml 41 days and to CD1940/ml 41 days. Circulating T cells comprised naïve and memory subsets, with a recovery of CD31 recent thymic emigrants (RTEs) from day 30. All patients had a significantly higher proportion of RTEs at day 30 and 180 compared to their pre-HSCT levels, suggesting an improvement in their thymic function after HSCT. With a median follow-up for living patients of 15 months (5-24), the 1-year cumulative CI of NRM and relapse were 17% and 35%. Three of the 11 acute leukemia relapses were HLA-loss variants. Notably, one was observed for the first time in ALL. CI of aGvHD grade II-IV and III-IV at 6 months were 17% and 7%, while 1-year CI of cGvHD was 20%

Conclusion: Myeloablative haploHSCT with PBSC, PT-Cy and sirolimus is a valid option for patients with aggressive/ advanced disease. The acceptable rates of GvHD and NRM as well as the favorable immune reconstitution profile open the way for combining it with novel immunomodulatory/ cellular therapies to improve DFS in patients at high risk for O142 Factors determining the Kinetics of Disease Relapse after Allogeneic Stem Cell Transplantation (allo-SCT) for Acute Myeloid Leukaemia (AML): a survey from the Acute Leukaemia Working Party of the EBMT C

O143 A novel quantitative PCR approach targeting insertion/ deletion polymorphisms (indel-PCR) for chimerism quantification: finally high sensitivity and quantification capacity together A

Post-hematopoietic stem cell transplantation (SCT) chimerism monitoring is important to assess engraftment, anticipate relapse and provide information on the development of graft versus host disease, facilitating therapeutic intervention. The aim of this study was to test the technical efficacy and clinical utility of a novel quantitative PCR approach targeting insertion/deletion polymorphisms

Of note, analysis of artificial mixtures provides evidence of significantly (Z2 logs) higher sensitivity by indel-PCR (0.01%) than by STR-PCR (1%, Table 1). Moreover, indel-PCR shows unprecedented quantification capacity (Table 1). Out of the 113 samples analyzed, 29 were positive and 6 negative by both methods, while 78 were positive only by indel-PCR (95% with o1% recipient). Hematological relapse occurred in 5 patients, molecular relapse/persistence in 2 patients. All of them presented a positive indel-PCR (with increasing %R in 4/ 5) and a negative STR-PCR result in the sample before relapse (Table 2). The 12 patients in complete remission, although presented positive indel-PCR, showed stable or decreasing %R chimerism dynamics in 10/12 (data not shown). Conclusion: This novel indel-PCR is a simple and accurate technique that, in comparison with the current gold standard STR-PCR, shows very good concordance and provides higher rates of informative loci per patient, as well as unprecedented combined sensitivity and quantification capacity

Introduction: The immune recovery after CD34 þ cell selection is slow and patients tend to remain susceptible to opportunistic infections for several months after HSCT. To hasten and improve post-transplant immune reconstitution broad repertoire various strategies

Under this approach, a rapid immunological reconstitution and very promising outcome have been reported in pediatric patients. With the aims of confirming these results even in adults, we have recently launched this programme and here we report our preliminar clinical data in 22 leukemia patients we have so far treated over the past 26 months. Materials (or patients) and methods: Twenty-two patients, median age 44 years (range 19-67), with AML (n ¼ 16), ALL (n ¼ 5), MDS (n ¼ 1) entered the study. Eight patients were in CR1, 3 in CR2, and 11 in advanced-stage disease at transplant. Conditioning consisted of ATG 1,5 mg/kg from day -13 to day -10, Treosulfan 12gr/sqm from -9 to -7, Fludarabine 30 mg/sqm from -6 to -2 and Thiotepa 5 mg/Kg on days -5 and -4. No patient received any post transplantation pharmacologic prophylaxis for GVHD. Ten mg/kg G-CSF was used to mobilize PBPCs from one-haplotype mismatched donors (3 mothers

Median CD4 þ cell/mL counts at 30, 60 and 90 days since the transplant were 36, 80 and 110, respectively. CMV antigenemia reactivation occurred in 6 cases (in 2, CMV serology was unfavourable). No patients has so far developed CMV disease. Invasive fungal disease was prevented in all cases using L-AmB-based prophylaxis over the neutropenic phase. Overall,10 patients have so far died (7 relapse,3 non-hematologic causes). 12 survive (10 diseasefree,2 in early relapse) at a median follow-up of 13 months (range 2-24) (Fig.1). Conclusion: The infusion of ab/CD19-depleted grafts was safe and effective also in adult setting, resulting into rapid engraftment and fast immunological reconstitution

Haploidentical Stem Cell Transplantation In Very Poor Risk Cytogenetics Acute Myeloid Leukemia: results in 40 consecutive patients

Cytogenetic prognostic risk was defined according to the revised Medical Research Counsil (MRC) classification, from diagnostic bone marrow samples with standard methods and in accordance with International System of Human cytogenetics guidelines. OS and Disease Free Survival (DFS) were calculated using the Kaplan-Meier methods. Results: Median age of the patients at time of transplant was 50 years (range 28 to 67).Cytogenetics:chromosome 7 abnormalities 16 pts, monosomal karyotype 3 pts and complex karyotyping 21 pts

Non relapse mortality (NRM) was 17% at 1 year after transplant.Estimated LFS from day þ 30 after transplant was 34.5% and 28% at 3 years. Conclusion: Haploidentical stem cell transplant (haplo-SCT) is a valid treatment option for the patients with very poor risk AML. References: 1

Haploidentical, unmaipulated, primed bone marrow for pts with high risk hematologic malignancies

Haploidentical transplantation using T cell replete PBSC and myeloablative conditioning in pts with high risk hematologic malignancies who lack conventional donors

Kodera 7 on behalf of the Worldwide Network of Blood and Marrow Transplantation O149 Graft depletion of TCRalpha/beta and CD19 in matched unrelated and haploidentical transplantation for severe aplastic anemia: high survival with low incidence of graftversus-host disease and transplant-related mortality M

At enrollment all patients were on TP and paired analysis showed that mean (SD) maximum velocities had significantly decreased (Po0.001) under TP:169 (46) cm/s vs 219 (26) cm/s at TP initiation. Following HLA-typing, 35 without genoidentical donor were included in the transfusion arm and 32 with genoidentical donor were transplanted in 6 HSCT-centers. During the 12 months follow-up, no stroke was observed but one patient in the TP arm experienced a hyperammonemic reversible coma, without MRI/MRA alteration. In the HSCT arm, all patients successfully engrafted, one grade II and two grade III acute GVHD, and no chronic GVHD were observed. Complications were seizures (n ¼ 2), CMV (n ¼ 9) or EBV replications (n ¼ 5), hemorrhagic cystitis (n ¼ 4), aspergillosis (n ¼ 1), prolonged but reversible thrombopenia (n ¼ 2), transitory hemolytic anemia (n ¼ 1)

Granda Ospedale Maggiore Policlinico, 5 Hematology and Bone Marrow Transplantation Unit

Metabolic syndrome (MS) is defined as a clustering of five factors including (1) hyperglycaemia (2) hypertriglyceridaemia; (3) low HDL cholesterol; (4) hypertension; (5) obesity (high waist circumference or body mass index (BMI)). It is associated with raised risk of cardiovascular disease (CVD) and is increasingly recognised in patients after HCT and revised guidelines for long-term HCT survivors recommend screening for MS. Previous studies have been small and the definition of MS variable, although harmonised criteria are now agreed

Materials (or patients) and methods: This was an EBMT approved cross-sectional, non-interventional study of consecutive HCT patients aged 18 þ years and a minimum of 2 year post-transplant attending routine follow-up HCT and/or late effects clinics in 9 centres. Centres completed MED C forms incorporating routine recording of the MS parameters (given above) as well as performance status (ECOG); evidence of cardiovascular events

Using the harmonised definition of MS (at least 3/5 factors), the prevalence of MS was 47.2%. There was no difference in time since HCT but there was a significant difference in prevalence by age at diagnosis, HCT, follow-up (all Po0.001 with increasing age). As expected, statistical differences (Po0.001) between patients with and without MS were observed for BMI

Routine screening and early intervention may reduce the risk of cardiovascular events in HCT survivors, and should ideally be tested in a randomised controlled trial setting. Meanwhile, screening and management of reversible features of the metabolic syndrome should be robustly integrated within routine HCT long-term follow-up care. References: Disclosure of Interest: None declared. Introduction: The introduction of less toxic conditioning regimens for hematopoietic cell transplantation (HCT) has led to an increase in eligible patients, although their benefit on patient's perceived wellbeing remains unclear. We aimed to prospectively study patients' quality of life (QoL) and emotional wellbeing (EW) in consecutive HCT recipients

depression and anxiety (HADS), and sleep quality (PSQI) at pre-HCT, at hospital discharge (HD) and at 3 months post-HCT. Results: Out of 223 transplanted patients, 191 (86%) consented to participate. Those who refused (n ¼ 32, 14%) more frequently had active disease at HCT (31% vs. 21% for PR/CR, P ¼ 0.032) and/or had a prior HCT

Among included patients (57% men; median age 53

83 received an Auto-HCT (n ¼ 79 at HD; n ¼ 52 at þ 3 m)

55 received an allo RIC-HCT (n ¼ 41 at HD; n ¼ 33 at þ 3 m)

At baseline, clinically significant depressive symptoms were reported by 5% of the patients, with a slight increase (7%) at HD and at þ 3 m (P ¼ 0.058). Again, there was a strong interaction between depressive symptoms and HCT groups in the early post-HCT phase (P ¼ 0.002); depression decreased in Auto-HCT after HD and, on the contrary, it increased in the same time-point in MAC-HCT (P ¼ 0.041). Borderline differences were seen between Auto-and RIC-HCT (P ¼ 0.062) but not between RIC-and MAC-HCT (P ¼ 0.827). Clinically significant anxiety was observed at baseline in 14% of the patients and significantly decreased at the time of HD (6%) and afterwards (5%)

Conclusion: MAC-HCT recipients reported the greatest impairment in the parameters studied; other variables such as gender, age and baseline EW/QoL should be considered for specific psychological/clinical follow-up and eventual intervention

Unit of molecular and functional immunogenetics, 4 Unit of Hematology and Bone Marrow Transplantation

Conflict with: Scientific consultant of MolMed S.p.A. O161 ErbB2-chimeric antigen receptor (CAR) modified cytokineinduced killer (CIK) cell intervention for refractory solid tumors after allogeneic stem cell transplantation M

Already at day 10 of culture, up to 90% of transduced cells showed surface expression of the ErbB2-CAR (n ¼ 4). There were no significant phenotypic differences (CD3 þ /CD56 þ /CD4 þ /CD8 þ /TCR-a/b and TCR-g/d) between unmodified, empty control vector and ErbB2-CAR transduced CIK cells. ErbB2-CAR CIK cells efficiently lysed ErbB2-overexpressing breast carcinoma cells (MDA-MB 453) in a 3 hour short-term cytotoxicity (europium release) assay in vitro. Compared with unmodified and empty-vector transduced CIK cells

E:T; 46.0% vs. 18.1% specific lysis, n ¼ 4; Po0.019). Long-term cytotoxicity analysis (16 h, brightfield imaging cytometry) demonstrated comparable results even at low effector to target ratios of 1:1 (36.3% vs. 11.8% specific lysis, n ¼ 3). Comparable results for shortand long time cytotoxicity could be obtained for all other tested rhabdomyosaroma cell lines in vitro. Conclusion: ErbB2-CAR engineered CIK cells are highly specific and efficient against ErbB2-antigen expressing tumor cell lines in vitro. Our experiments may help to develop an approach for improved treatment of patients with high-risk

Adoptive T-cell immunotherapy with hexon-specific THELPER-1 cells as a treatment for refractory adenovirus infection after allo-SCT -safety and efficacy results from a

Anna Children's Hospital

University Children's Hospital, Frankfurt, 5 Dr. von Haunersches Kinderspital

Introduction: Allogeneic hematopoietic cell transplantation (HCT) offers the chance of cure for patients with nontransformed follicular lymphoma (FL) but is associated with the risk of non-relapse mortality (NRM). The aim of this study was to identify subgroups of FL patients who benefit from HCT. Materials (or patients) and methods: The Minimum Essential A Data of 146 consecutive patients who received HCT for FL between 1998-2008 were extracted from the database of the German Registry ''DRST''. Diagnosis of FL was verified by contact with reference pathologists. Results: The median patient age at time of transplantation was 48 years (range 29-71). Prior to allogeneic HCT 90/146 patients (62%) had undergone autologous HCT. At time of HCT, 110 patients (77%) had sensitive disease while 33 patients (23%) had chemorefractory disease (RD). Engraftment was achieved in 99% of evaluable patients. Day 100 NRM was 16%. The median follow-up of surviving patients was 9.1 years (range 3.6-15.7) . Estimated 1, 2, 5, 10-year overall survival (OS) was 67%, 60%,53%, and 48%, respectively. The corresponding estimates for EFS were 63%, 53%, 47%, and 40%, respectively. 40% of the 33 patients with RD at time of transplantation survived long-term. Of the n ¼ 116 patients with documented CR after HCT only n ¼ 17 (15%) relapsed. Only two late relapses (beyond year 3) were diagnosed among the 77 patients with a follow-uP45 years. Patients with chronic GvHD (irrespective of stage) had a lower risk of relapse, if transplanted in CR, and a higher chance to achieve CR, if transplanted in PR or with PD (no chronic GvHD: 19/48, chronic GvHD: 11/75, P ¼ 0.0019). therefore a reduced intensity conditioning approach might be considered in future prospective trials. HSCT is most successful prior to leukemic transformation. Given implications for treatment decisions and donor selection GATA2 mutation screening should be performed on all patients with molecularly undefined MDS and BMF disorders and potential related donors. Disclosure of Interest: None declared.

Pre-transplant weight loss predicts non-relapse mortality and relapse rates in patients with myelodysplastic syndrome after allogeneic stem cell transplantation A. Radujkovic 1, * Introduction: We have recently provided evidence that weight loss and minor metabolic changes prior to alloSCT were able to predict relapse and death of acute myeloid leukemia patients using data from two independent patient cohorts. This retrospective study investigated the influence of pre-transplant weight loss on the outcome of MDS patients after alloSCT in three independent cohorts. Materials (or patients) and methods: A total of 111 patients (59% male) with a median age of 52 years were included into the analysis. Patients have been diagnosed with MDS according to WHO criteria and received an alloSCT between 2000 and 2012 in three different German referral centers (Heidelberg, Dresden and Berlin). Weight data were raised from medical records by three independent researchers in three independent institutions. Weight loss (expressed in percent) was calculated on the basis of recorded weight data at the time of alloSCT and the maximum weight in the time period of 3-6 months prior to alloSCT. The MDS WHO subtype was RA(RS)/RCMD in 31 patients (28%), RAEB1 in 30 patients (27%) and RAEB2 in 49 patients (45%). According to IPSS 34%, 45% and 21% of the patients were in the risk groups intermediate-1, intermediate-2 and high, respectively. The majority of the patients (n ¼ 72, 65%) was previously untreated. Nineteen patients (17%) and 14 patients (13%) received hypomethylating agents and induction type chemotherapy prior to alloSCT, respectively. Thirty-one patients (28%) received transplants from related donors, 59 patients (53%) from matched unrelated donors and 21 (19%) from mismatched unrelated donors. Ninety-three patients (84%) received reduced intensity conditioning and 18 patients (16%) received standard myeloablative conditioning. Survival times were measured from date of alloSCT. Overall survival (OS), relapse-free survival (RFS), relapse incidence and non-relapse mortality (NRM) were calculated from date of alloSCT to the appropriate endpoint. Cox regression analysis was applied for OS, RFS, relapse and NRM. Relapse and NRM were considered as competing risks. Results: Estimated median follow-up at the time of analysis of surviving patients was 36 months. A total of 34 (31%) patients experienced weight loss 42% with 17 (15%) patients losing more than 5% weight in the period of 3-6 months prior to alloSCT. Patient, disease and transplant characteristics did not differ between patients with weight loss (42%, n ¼ 34) and those without (n ¼ 77). In multivariate analysis, weight loss and donor type were independently associated with shorter OS and RFS (Po0.001 and Po0.05, respectively). NRM was predicted by donor type (P ¼ 0.006), IPSS (P ¼ 0.015) and pre-transplant weight loss (P ¼ 0.014) in multivariate analysis. Furthermore, weight loss was also an independent predictor of relapse (cause-specific HR 11.52 95%CI Po0.001) . In a mixed effect model with weight loss as outcome only IPSS prior to alloSCT had significant impact on weight loss (P ¼ 0.046 Introduction: HLA-C-encoded KIR ligands (C1/C2) have been identified as important factors for the outcome of unrelated allogeneic HSCT: In a previous retrospective study CML recipients bearing at least one C2 ligand showed worse survival when compared to C1C1 recipients (HR 5.9 , Po0.01), especially when peripheral blood progenitor cells (PBPC) were used or in advanced disease stages. These initial findings were confirmed in a second cohort in advanced AML/CML, (but not in MDS or ALL/NHL) receiving PBPC. Notably, HLA-C allele matching contributed differentially to the transplantation outcome: it was beneficial in C1 patients, but was detrimental in C2 recipients (increased TRM, HR 3.5, Po0.012 ; increased relapse, HR 2.7, P ¼ 0.06). We hypothesized that C1 patients have a high frequency of immuno-competent NK cells (icNK) enabling eradication of residual disease due to the genetically hard-wired sequential acquisition of KIR receptors during early reconstitution phase post HSCT, with C1-specific NK cells emerging first. Alloreactive T cells -resulting from HLA-C mismatch-might thus not have an additional beneficial effect in C1 patients and might even be detrimental (e.g. increased GvHD), but may serve an important function in relapse control in C2 patients. The lack of disease control in HLA-C-matched C2 patients would thus be explained by a combination of insufficient numbers of icNK cells in the early phase and the lack of alloreactive T cells later post HSCT. Consequently, this group exhibited poorest clinical outcome of all four groups defined by recipients KIR ligands and HLA-C allele matching in the investigated cohorts. Materials (or patients) and methods: The aim of the present retrospective study was to determine the influence of HLA-C allele matching on the background of HLA-C encoded KIR ligand status in a large patient cohort (n ¼ 7327, provided by CIBMTR). Statistical analysis was performed by CIBMTR. Patients received unrelated allografts between 1988 and 2009 , with the majority of patients after 2000 (70%). 30% of the recipients were younger than 30y, 67% younger than 40y. 70% of patients had early, 5% intermediate, and 25% advanced disease (AML 40%, ALL 21%, CML 22%, MDS 17%). 56% received bone marrow, 44% PBPC. 80% received a myeloablative conditioning.

Endpoints were OS, aGVHD II-IV and III-IV, extensive chronic GVHD, relapse, DFS, and NRM. Due to multiple comparisons and multiple endpoints, Po0.01 was considered as significant. All models were adjusted for significant clinical covariates. Stratification was used in cases of non-proportional hazards. Patient-donor pairs were classified according the recipient HLA-C KIR ligand expression (C1C1, or C2 ( ¼ C1C2 or C2C2)) and the degree of the HLA-C allele match: Results: Introduction: Hepatic veno-occlusive disease (VOD), also called sinusoidal obstruction syndrome, is a potentially fatal complication of hematopoietic stem cell transplantation (HSCT). Severe VOD (sVOD), clinically characterized by multiorgan failure (MOF), has been associated with a 480% mortality rate; it may develop in a substantial number of highrisk patients (pts). Defibrotide (DF), a sodium salt of complex single-stranded oligodeoxyribonucleotides, is thought to protect injured endothelium and to restore thrombo-fibrinolytic balance. In a phase 3 trial in sVOD, DF improved complete response rate and survival at day þ 100 post HSCT vs historical controls, with a favorable safety profile. In the European Union, DF is approved for treatment of severe hepatic VOD in HSCT therapy in adults and children. In the US, DF is available through an expanded access, protocol-directed treatment IND (T-IND) collecting data on safety/efficacy in children and adults with sVOD and non-severe VOD post HSCT or post chemotherapy (CT).

The original T-IND protocol required VOD diagnosed by Baltimore criteria (total bilirubin Z2.0 mg/dL with Z2 of hepatomegaly, ascites, or 5% weight gain) plus MOF (renal and/or pulmonary) following HSCT; the study was amended to include non-severe VOD (ie, Materials (or patients) and methods: In a single center retrospective study 382 patients who underwent allogeneic hematopoietic stem cell transplantation (HSCT) for various diseases (51% acute leukemia) were genotyped for CYP 1B1 (C432G) expression and their influence on outcome was analyzed. Genotyping of CYP 1B1 (C432G) was performed by real-time PCR.Results: 169 patients (44%) were genotyped as homozygous wild-type (wt) gene C/C, 157 (41%) as heterozygous genotype C/G and 56 (15%) as homozygous gene mutated G/G. Calculated genotype frequencies did not differ from that reported earlier by other studies for Caucasians. Patients' demographic and treatment characteristics showed no difference between the three groups except that CYP 1B1 CC was more common in females 52% than in males 38% (Po0.02).Five-year estimate for overall survival (OS) was 58 þ 4% for the CC group and 48 þ 3% for the C/G-and G/G groups (Po0.036). Surprisingly, this difference was primarily evident in males (Po0.009), where the group with CYP gene mutations did significantly worse (3-year estimate for OS: 65 þ 5% vs. 47 þ 4%), whereas it was virtually absent in females (P ¼ 0.99). TRM and RR were higher for the group with mutated genes in regard to the group with wt gene (although not significant). 2006 . In phylogenetic analysis of the E3 gene, a two-step PCR amplification of almost the entire adenovirus E3 gene was performed using primers designed from the known sequence of the HAdV A31 reference strain (AM749299.1). Results: All 9 patients had been admitted to the ward, but the two last patients (patient 8 and 9) had no timely connection to other known HAdV A31 cases ( Figure) . In addition, four of the patients (1, 5, 6 and 7) made visits to the out-patient clinic on the same day as one or several other HAdV positive patients.Sequencing the hexon gene resulted in 100% homology between the patient samples but also to the reference strains of HAdV A31 (Accession number AB330112. 1 (CMV) is an important cause of morbidity after allogeneic hematopoietic stem cell transplantation (HSCT). The latent virus reactivates in immune-compromised patients, due to both post-HSCT immunosuppressive therapy and impaired T-cell reconstitution/function. Here we report the impact of mismatches in HLA-molecules between donor and recipient on CMV-reactivation and CMV-specific immune reconstitution. Materials (or patients) and methods: This retrospective study included 752 patients, who received a transplant from a matched related donor (MRD n ¼ 234), matched unrelated donor (MUD n ¼ 384), mismatched unrelated donor (MMUD n ¼ 115) or mismatched related donor (MMRD n ¼ 19). HLAtyping (10/10) of patients and donors was conducted via highresolution multiplexed PCR. Blood samples were routinely monitored for CMV pp65 antigen expressing cells per 400,000 leukocytes. CMV-cytotoxic lymphocytes (CMV-CTL) reconstitution was analyzed in the blood from 246 patients at days þ 50, þ 100, þ 200, þ 300 after HSCT, using 6 HLA-CMV tetramers. Results: The Fisher's exact test was used to analyze the data. The outcome was that HLA mismatch (class I or II) showed significant influence on recurrent (multiple) CMV reactivations (mCMV-R) (Po0.001). We analyzed the relative risk (RR) in the subgroups with different levels of HLA-matching for 1CMV-R and mCMV-R. The group transplanted from MRD served as reference (ref. ) . Shortly, we found significantly higher risk for mCMV-R in the MMUD group (RR 2.6, 95% CI 1.63-4.15 , P ¼ 0.0001). In the MRD 24 (10%) patients had mCMV-R, while in the MUD 59 (15%), in the MMRD 3 (16%) and in the MMUD group 31 (27%) had mCMV-R. Furthermore, we investigated the mean numbers of CMV-CTL/ ml of blood in the groups with different levels of HLA-match. We divided CMV-CTL levels into 3 ranges: o1, 1-10 and 410 CMV-CTL/ml. In the MMUD group we observed a trend for an increased risk for the lack of CMV-CTL (25 (48%) patients; RR 1.5 , 95% CI 0.96 to 2.38, P ¼ 0.07) compared to 21 (32%) patients from the MRD group. Significantly less patients (17; 33%) had more than 10 CMV-CTL from the MMUD group (RR 0.6, 95% CI 0.39 to 0.95 Thus the focus of the present study was the selection of HAdVstreptamer þ T-cells and EBV-streptamer þ T-cells.Materials (or patients) and methods: Cells from leukapheresis healthy donors were prepared in large (1 À 6 Â 10 9 ) and small (25 x 106) cell batches. Whereas the larger batch was directly labelled with streptamers to select HAdV and/or EBVspecific T-cells (large-scale), the smaller batch was used to generate in vitro virus-specific T-cell lines before streptamerlabelling for streptamer selection (small-scale). Isolation of HAdV-and/or EBV-specific T-cells was performed using the CliniMACS device.Results: The purity of HAdV-streptamer þ T-cells among CD3 þ cells, obtained from large-scale selection was only 7.6%, but reached up to 56% when HAdV-and EBVstreptamers were applied simultaneously. A further increase in purity of HAdV-specific T-cells reaching up to 98% was achieved by small-scale selection. All final products fulfilled the microbiological and chemical release criteria. IFN-g-response indicating functional activity was seen in 6/9 HAdV and 2/3 EBV large-scale selections and in 2/3 HAdV small-scale selections.Conclusion: The use of HAdV-streptamers for clinical applications is feasible particularly when combined with other streptamers or when performed after a previous in vitro expansion period.

In this cohort of 149 T-cell replete haplo HSCTs using post-transplant high-dose cyclophosphamide, we found that a higher intensity of conditioning (MYA or RIC vs non-MYA) as well as the use of more immunosuppressive calcineurin inhibitor (FK vs CsA) were both significantly associated with a higher incidence of PV-HC. Results: In CD34 þ Lin -CD10cells, 1609 probes were deregulated between patients without aGVHD and patients with aGVHD (1560 of this probe were up-regulated and 49 were down-regulated, Po0.05, Fold Change41.5) . In CD34 þ Lin -CD10 þ CD24progenitors, 987 probes were deregulated between patients without aGVHD and patients with aGVHD (941 of this probe were up-regulated and 46 were downregulated, Po0.05, Fold Change41.5). 273 probes were deregulated in both CD34 þ Lin -CD10 þ CD24and CD34 þ Lin -CD10populations. Genes from ribosome protein biogenesis, translation machinery (EEF1D, EEF1G, EIF3K) and cell cycle (CCND1, CDK6) were over-expressed in CD34 þ Lin -CD10 þ CD24and in CD34 þ Lin -CD10populations from patients without aGVHD compared with those from patients affected by aGVHD and from healthy donors. Expressions of genes from the oxidative phosphorylation metabolic pathway (NDUFS2, SDHA, ATP5A1) and genes involved in stress resistance (BTG2, MGST3, HPX) were specifically increased in CD34 þ Lin -CD10 þ CD24lymphoid progenitors and not in CD34 þ Lin -CD10non-lymphoid progenitors from patients without aGVHD compared with patients suffering from aGVHD and from healthy donors.

We show for the first time that circulating lymphoid T-cell progenitors undergo profound changes in metabolism favoring energy production and response to stress after allo-HSCT in Humans. These mechanisms are abolished in case of aGVHD, indicating a persistent cell-intrinsic defect in addition to the impact of aGVHD on the bone marrow environment. Disclosure of Interest: None declared. Introduction: Post transplant interventions such as donor lymphocyte infusion (DLI) or administration of pharmacological agents, represent important novel strategies with the potential to reduce the risk of disease relapse after allo-SCT in Acute Myeloid Leukaemia (AML). Such approaches are critically dependent on timely intervention post-transplant but despite this the factors determining the kinetics of disease relapse in patients allografted for AML have not been defined. Materials (or patients) and methods: 1052 adults who received an allo-SCT for AML in first complete remission (CR1) between 2000 and 2012 were studied. 544 patients were transplanted using a sibling donor and 508 from an adult matched-unrelated donor. 538 patients received a myeloablative conditioning (MAC) regimen and 514 a reduced intensity (RIC) regimen. A series of landmark analyses were performed at 3, 6 and 12 months in order to identify prognostic factors of relapse for patients alive and well at the beginning of each time interval. The probabilities of relapse were calculated by using the cumulative incidence estimator to accommodate for death as a competing risk. Factors predicting relapse were studied using Cox regression model including time dependent variables. A backward stepwise procedure was used for variable selection. Results: With a median follow-up of 26 months, 244 patients relapsed. The 3 year cumulative incidence of relapse was 26% [95%CI: 23-28]. Overall 84% of patients destined to relapse did so within the first year post-transplant. The overall factors predicting disease relapse for the whole population were more than one course of chemotherapy to achieve CR1, FLT3 ITD positivity, adverse risk cytogenetics, shorter interval from CR1 to transplant. The occurrence of acute GVHD grade II or greater (P ¼ 0.05) and chronic GVHD (P ¼ 0.03) were both associated with a lower risk of relapse. Using landmark analyses the factors determining relapse at different stages post transplant were observed to differ. In the first 3 months post-transplant the significant factors determining relapse risk were: patient age (P ¼ 0.03), prolonged interval from diagnosis to CR1 (P ¼ 0.05), flt3 ITD positivity (P ¼ 0.002), adverse risk cytogenetics (P ¼ 0.02) and use of in vivo T cell depletion (P ¼ 0.003). The only factors observed to determine relapse risk between 3 and 6 months post-transplant were Introduction: The number of haematopoietic stem cell transplants being performed worldwide is increasing as is interest in side effects. Despite the increase in published data on late effects in the last decade, data on very long term survivors (425 years) is lacking. In this study we describe the outcome of all the patients transplanted at our centre more than 25 years ago. Materials (or patients) and methods: Between June 1979-January 1990, 216 patients had received allogeneic SCT for haematological malignancies at the Hammersmith hospital. Most patients (180/216) were transplanted for CML with CY/ TBI conditioning and the majority were in chronic phase at SCT (n ¼ 140). At the time of analysis in December 2014, 151/216 (70%) patients had died. Of 65 presumed survivors, 14 patients had moved abroad and an additional 13 patients were considered lost to follow up as they had had no contact with our centre within 5 years of the study. Of the remaining 41 patients, detailed follow up information was available for 34. Results: The majority of deaths (94/151) occurred within 2 years of SCT. A further 27 patients died between 2-10 years after SCT the most frequent cause of death being relapse (17/ 27) and infection (5/27). Between 10-20 years after SCT there were 18 deaths; the most frequent causes were relapse (n ¼ 4), second malignancy (n ¼ 4) and GVHD (n ¼ 4). Between 20-35 years there have been 9 deaths and the most frequent causes were second malignancy (n ¼ 3) and respiratory (n ¼ 2). The latest recorded relapse was at 14.8 years. For 34 survivors for whom we had detailed follow up information the median follow up time was 29y 10 months (range 25y 8 months -35 years 7 months). The median age at follow up was 61y (range 45-80). 11/34 patients had had a diagnosis of cancer at the following sites: skin (BCC or melanoma) n ¼ 5, oral or tongue n ¼ 3, oesophagus n ¼ 1, breast n ¼ 1 and a further patient had had testicular and bladder cancer. 16/34 patients had dyslipidaemia and 14/34 were being treated for hypertension. 5/34 were diabetic and 12/34 were hypothyroid. Of 15 male patients, 3 had low testosterone levels requiring treatment. 8/34 had vascular complications including three with ischaemic heart disease one of whom also had PVD, two with venous thrombosis, one with a TIA and two patients with renovascular disease. DXA data was available for 25/34 of these patients and BMD was recorded as low (osteopenia or osteoporosis) in 12/25. Conclusion: We conclude that late deaths more than 20 y after SCT are more likely to be due to second malignancy than relapse. Appropriate screening identifies a large number of abnormalities in the surviving patients most of which are amenable to treatment. Disclosure of Interest: None declared. The aim was to evaluate the survival and late toxicities, defined as any disease condition other than lymphoma occurring after at least 6 months after ASCT. The median patient age at ASCT was 52 years (range, 20-70). All patients relapsed after at least one chemotherapy line (previous treatments range 1-8), 26% of them received radiotherapy. At ASCT, 76% of patients were in CR, 15% in PR, 1% stable, 8% in progression. Full dose BEAM was given to 87% of patients while 13% received dose reductions for comorbidities. Results: The median follow-up was 5.4 years (range 0.5-12.2) . The 5-year OS and PFS were 81 and 69% (median not reached for both). The non-lymphoma-associated mortality was 5%, 8% and 9% at 3, 5 and 7 years of follow-up. The OS was impacted in multivariate analysis by disease status before ASCT (P ¼ 0.001, HR 2.2, CI95% 1.4-3.6 ), and radiotherapy (P ¼ 0.017, HR 6.5, CI95% 1.4-30.7) . PFS was impacted by female gender (P ¼ 0.029, HR 0.4, ), pre-transplant disease status (P ¼ 0.001, HR 1.9, ), and radiotherapy (trend, P ¼ 0.064, HR 2.7, . None of the factors analyzed impacted late non-lymphoma-associated mortality, except for a trend given by age (P ¼ 0.075). Late toxicities after BEAM ASCT occurred in 61% of patients, and included infection (32% -most frequently pulmonary), hypogammaglobulinemia (30%), pulmonary complications (21% -mostly reduction of pulmonary function tests scores), metabolic syndrome (17%), cardiovascular complications (12%), second tumors (10%), hypothyroidism (8%), diabetes (5%), chronic kidney failure (4%), hepatitis B reverse seroconversion (2%) and ocular complications (1%). The cumulative incidence of second tumors was 1, 6, and 10% at 3, 5, and 7 years of follow-up, and reached a plateau of 16% at 10 years of follow-up. 17 patients had a second cancer, of whom 12 had a solid tumor (skin [4] , colorectal, prostate, lung [2 each], breast and oropharingeal [1 each]), 5 a hematologic tumor (secondary MDS or AML [4] or NHL [1] ). Age (P ¼ 0.013, HR 1.5 per year, CI95% 1.0-2.1) , and male gender (P ¼ 0.043, HR 0.05 favorable for female sex, CI95% 0.0-0.9) , increased risk of a second tumor. Of 13 patients who died without lymphoma, 6 died of second tumors, 2 died of cardiovascular complications, 2 of late infections, and 3 for other causes. In multivariate logistic regression, the incidence of second tumors was associated with age (P ¼ 0.04), and there was a trend for patients receiving radiotherapy for late cardiovascular complications (P ¼ 0.07). Conclusion: BEAM conditioning is associated with a 61% crude incidence of late effects, mostly infections, hypogammaglobulinemia, and pulmonary complications. The most important preventive measures for late mortality could be the screening for cancer, especially for older patients, screening for heart disease particularly for patients receiving radiotherapy, and prompt and aggressive treatment of late infections. Disclosure of Interest: None declared. Introduction: The advent of highly active antiretroviral therapy (HAART) in 1996 had led to a suppression of HIV viral load, to an improved immune function resulting in a significant reduction of opportunistic infections and HIV related morbidity and mortality. Consequently, more intensive treatments, including autologous stem cell transplantation (ASCT), have been extended also to the HIV-positive population. However, in the literature data are scarce concerning the long-term events (incidence of lymphoma relapse, of second cancers and AIDSdefining conditions) in HIV-positive patients (pts) affected by relapsed lymphoma who underwent ASCT. Materials (or patients) and methods: We treated 36 HIVpositive pts affected by relapsed/refractory lymphomas with ASCT consecutively in our cancer center. Ten pts died during or early after ASCT due to progressive disease (4 pts), chemotherapy toxicity (1 pt) and infection (5 pts). We analyzed the post-transplantation long-term data of 26 HIV-positive lymphoma pts, reaching a complete response after ASCT. Eighteen pts were male (69%) and 8 pts were female. Our cohort of pts included 17 non-Hodgkin's lymphomas (NHL) and 9 Hodgkin's lymphomas (HL), respectively. Twenty-two pts (85%) received one, and 4 pts received two second-line chemotherapy regimen before ASCT, respectively. The majority of the pts were submitted to a single (BEAM conditioning regimen) and only two pts to a tandem ASCT procedure (high dose Melphalan followed by BEAM conditioning regimen). All pts received HAART concomitantly to cancer treatment. Results: Two pts experienced a lymphoma relapse, after 4.27 and 3.08 years from ASCT, respectively. Three pts presented with a secondary malignancy (1 pt an anal squamous cell cancer, 1 pt a squamous cell carcinoma of the larynx and 1 pt a CIN2, respectively), with a median time of 3.01 years from ASCT. Eight pts had opportunistic infections (OI): 2 pts developed a Pneumocystis Carinii pneumonia, 1 pt a Cytomegalovirus pneumonia, 1 pt a Mycobacterium Avium Complex pneumonia, 1 pt a Herpes Simplex Chronic Ulcer, 3 pts cutaneous relapsing Herpes Zoster, respectively. The median time of OI appearance was 0.25 years (IQR: 0.11-2.33 ). Two pts died: one of lymphoma relapse, the other of car accident. With a median of 6-years follow up (IQR: 4.55-9.87 ) the OS and PFS of the entire sample of pts were 91% and 36% at 10 years, respectively.

Our results may be summarized as follows: 1) 24 out of 26 pts are still alive and in long-term complete remission after ASCT. These data confirm the long-term efficacy of ASCT in HIV-positive pts affected by relapse/ refractory lymphoma. 2) The appearance of OI is earlier than that of second malignancies after ASCT. 3) The secondary malignancies developed by our pts are non-AIDS-defining cancers, in agreement with the increased incidence reported by the literature in the HAART-era and at least two cases are linked to a viral pathogenesis (HPV for both anal cancer and cervical cancer precursor lesion). 4) Both OI and second malignancies in our pts series were successfully managed and cured and the only long-term death occurred due to lymphoma relapse. Disclosure of Interest: None declared. Introduction: Allogeneic hematopoietic cell transplantation (HCT) is an effective therapeutic option for high risk hematological malignancies; 80% of those who survive the first 2 years are expected to become long-term survivors. The prevalence of chronic health conditions approaches 75% among HCT survivors and that for severe or life-threatening conditions exceeds 20%. 1 Materials (or patients) and methods: A standardized followup of HCT survivors is applied at our center, according to Jacie Standards. Here we report the analysis of data collected between Nov 2013 and Nov 2014 in 249 adult patients (ptsmedian age at follow-up 54y -r19-81) who underwent an HCT between 1992 and 2013 at our Institution. Data on 7 items were prospectively collected in an Institutional database. A written consent was given by pts allowing the use of medical records for research in accordance with the Declaration of Helsinki. Results: Overall 40% of pts received an haplo, 30% a MUD and 30% a match related HCT; 13 pts deceased in the last year (7 because of disease relapse, 4 of late major infectious complication, 2 of second cancer). At a median follow-up of 4y (r1-22; cumulative follow-up 1277y) we observed: -chronic Graft-versus-Host-Disease (c-GvHD): at a median follow-up of 3y (r1-21) 61 (25%) pts are presenting c-GvHD features. According to NIH 2004 consensus criteria 15 cases were classified as mild, 25 moderate, 21 severe. Median number of involved organs 3 (r1-4), 32 pts were experiencing skin lesions, 38 eyes impairment, 23 mouth alterations.-Late infectious manifestation: 54 (22%) pts present late infections, 4 pts deceased. Pneumonia was reported in 22 pts, Varicella Zoster virus reactivation in 12, encephalitis in 5 (3 virus related, 2 Toxoplasma related), hepatitis in 4, EBV reactivation in 2.-Second cancers: second malignancies were diagnosed in 32 (13%) pts, 5 pts are actually under work-up for diagnosis. Nonmelanoma skin cancer was the most frequent diagnosis (17 cases); 3 pts were diagnosed with cervix cancer, 2 with prostate cancer, 2 with lung cancer and 2 with bladder cancer. Single cases of thyroid, parathyroid, colon, gastric, kidney, larynx, endometrial and breast malignancies were also reported. All pts were treated according to standard policy for general population, 30/32 are alive.-Thyroid dysfunction: 38 (15%) pts presented overt hypothyroidism.-Cardiovascular diseases: arterial diseases were reported in 17 pts, atrial fibrillation in 5 and cardiomyopathy in 2 pts -overall 10%.-Metabolic syndrome (MS): 84 (34%) pts were presenting features of MS (3/5 features among hypertension, dyslipidemia -raised triglycerides and lowered high-density lipoprotein cholesterol-, raised fasting glucose and central obesity). -Secondary hemosiderosis: iron overload was documented (with MRI and blood parameters) and treated in 38 pts (15%). According to donor source no difference were observed (Chisquare test -p ns) except for higher incidence of moderate/ severe GvHD incidence in match related HCT (p 0.0097) as compared to alternative HCT. Conclusion: HCT survivors are at a defined relevant risk of developing long-term complications that have a direct impact on quality of life, morbidity and mortality. 1 Introduction: Long-term survivors of allogeneic HSCT now form an expanding and unique patient population with often complex physical and psychological late effects (LE) and associated unmet needs. Despite international guidelines 1 , optimal delivery models of LE services are unclear from clinical, organisational and economic viewpoints. Materials (or patients) and methods: In order to scope current models of care for LE service delivery within the UK, we undertook a survey of the 27 NHS adult allogeneic HSCT centres during 2014. Centres were invited to participate in an online survey composed of 30 questions examining service organisation, multi-disciplinary team (MDT) provision, access to other specialist services and patient engagement. Results: A 100% response rate was achieved from programme directors or delegated specialist staff. Around half of centres also treated patients r18 years and all centres had achieved or were working towards JACIE accreditation. In 480% of centres, the LE service was led, coordinated and delivered by consultant medical staff, with the remainder being nurse-led. Most centres (490%) provided follow-up in a dedicated allograft or LE clinic for the first year, but thereafter attrition resulted in B50% of patients being followed after 5 years, and B30% after 10 years. Most centres had easy access to medical specialities necessary for LE management, but specialist interest in long-term HSCT complications was uncommon. Only 18% of centres had access to a LE MDT, often limited to patients o25 years. Despite specific JACIE competency S91 standards, a third of centres had held no LE educational event in the previous three years. Most centres (70%) had an SOP for long-term monitoring and LE management, with the focus predominantly on physical LE with only 28% including a formal psychological screening assessment. Only 39% of centres had audited the performance of the SOP. Screening for endocrinopathies, iron overload and cardiovascular complications was near universal, but access to mammography and cervical smear testing was more limited. Revaccination rates were high, but only 23% of centres routinely tested antibody responses. Despite recommendations, most (59%) centres never used standard templates 2 to communicate LE risk to GPs or referring consultants. Only 41% of centres had a patient support group accessible to HSCT survivors with equivalent numbers having undertaken patient satisfaction surveys related to LE service provision. The most commonly perceived barriers to implementation of LE services were funding of psychological and other clinical staff and extra investigation costs.Conclusion: This survey has demonstrated variation and limitations in the provision of long-term follow up of allogeneic HSCT survivors within the UK NHS. Although patients are seen in specialist clinics and have access to other specialities, there are limitations in sustaining long-term screening, MDT working, education, audit and patient engagement, as well as perceived barriers to resourcing staff and investigations. Further work is warranted to optimise effective, sustainable and affordable models of care for delivery of LE services in this expanding specialised patient population. References: 1. Majhail et al. Biol 2 Hematology Department, Hospital Sant Pau, 3 Hematology Department, Hospital Vall d'Hebrón, 4 Epidemiology Department, Hospital de Sant Pau, 5 Hematology Department, Hospital del Mar, Barcelona, Spain in vivo and to unravel the requirements for their long-term persistence directly in humans. Materials (or patients) and methods: We studied the immune system of 10 patients who underwent haploidentical HSCT and infusion of donor lymphocytes transduced to express TK suicide gene (median dose: 1.9x10 7 cells/kg) for high-risk hematologic malignancies. In case GvHD, proliferating TK-cells were promptly eliminated upon ganciclor (GCV) administration with complete resolution of the adverse reaction without immunosuppressive treatments. Results: At a median follow-up of 7 years after HSCT (range 2-12.3), a complete recovery of NK cells, B lymphocytes and ab or gd T cells was observed. The CD8 þ and CD4 þ T cell compartment of TK patients were characterized by level of naïve and memory cell comparable to age and sex matched healthy controls. TK-cells were detected in all patients, at low levels (median ¼ 4 cells/uL), even in patients treated with GCV. Ex vivo selection of pure TK-cells confirmed the presence of functional transduced cells, thus directly demonstrating the ability of memory T cells to persist for years. Importantly, GCV sensitivity was preserved in long-term persisting TK-cells, independently from their differentiation phenotype. Longitudinal follow up revealed that TK-cells circulated in patients at stable levels and displayed a conserved phenotype comprising effector memory (T EM ), central memory (T CM ) and stem memory (T SCM ) T cells. The low level of Ki-67 positivity suggested the maintenance of a pool of gene-modified memory cells through homeostatic proliferation. Polyclonality was demonstrated by sequencing among TK-cells of thousands of diverse TCRs with a broad usage of V and J alpha and beta genes. The number of TK-cells persisting at the longest follow-up did not correlate with the amount of infused cells, but instead with the peak of TK-cells measured within the first months after infusion, suggesting that antigen recognition is dominant in driving in vivo expansion and persistence of memory T cells. Accordingly, we documented the persistence of CMV and Fluspecific TK-cells only after post-transplant CMV reactivation or after Flu infection. We observed that the number of infused T SCM cells positively correlated with early TK-cell expansion and with their long-term persistence, suggesting that T SCM might play a privileged role in the generation of a long-lasting immunological memory. Conclusion: After infusion, gene-modified memory T cells persist for up to 12 years within a physiological immune system. Antigen exposure and a T SCM phenotype were associated with long-term persistence of infused TK-cells. Further studies on TK-cell TCR repertoire and vector integrations are currently being performed to elucidate the in vivo dynamics of infused memory T cells.

Use of zoledronic acid after TcRab/CD19-depleted haploidentical transplantation to enhance gd T cells anti-leukemia effect P. Merli Introduction: HSCT is a potentially curative option for a number of malignant disorders; however, up to 30% of patients lack a suitable HLA-matched either related or unrelated donor. In order to optimize haploidentical transplantation, we recently developed a new method of graft manipulation (i.e. TCRab/CD19 negative selection), which retains in the final product large numbers of effector cells, namely NK and TCRgd lymphocytes. Relapse remains the main cause of treatment failure. Based on preclinical data showing bisphosphonates-mediated improvement of TCRgd cells-blast killing through accumulation of phosphoantigens, we started a prospective trial based on post-transplant infusion of zoledronic acid, with the aim of enhancing TCRgd cells anti-tumor effect. Materials (or patients) and methods: Enrolled in the study were 35 pediatric patients (median age at transplantation 10.3 years, range 1-18) affected by either ALL and AML (26 and 9 patients, respectively) at very-high risk for relapse/TRM due to disease status (cytogenetic/molecular characteristics, lack of remission or previously failed HSCT). All of them underwent a TCRab/CD19-depleted HSCT from an HLA-haploidentical donor (one of the two parents). According to the model of KIR/KIR ligand mismatch, 13 patients were transplanted from an NK-alloreactive donor. The median number of infused gd þ T cells was 7.9 x 10 6 /kg (range 0.9-42.7) . Zoledronic acid was administered monthly at the dose of 0.05 mg/kg per dose (maximum dose 4 mg), starting from day þ 30 after transplantation. Patients underwent zoledronic acid infusions, together with oral calcitriol and calcium supplementation, in the outpatient unit. Results: A total of 102 infusions were administered with a mean of 2.9 infusions per patient (range [1] [2] [3] [4] [5] ; only one episode of symptomatic hypocalcemia (at first administration) occurred and was rapidly corrected with parenteral calcium supplementation. None of the patients experienced de novo onset or worsening of previously developed acute GVHD, this finding supporting the observation that gd T-lymphocytes do not cause GVHD. In the study period, six patients relapsed and 2 died due to infectious complications. With a median follow up of 9 months (range 4-22) the 2-year Kaplan-Meyer estimate of OS and LFS were 88.1% (SE 6.6) and 62.2 (SE 10.1), respectively ( Figure 1a) . The cumulative incidence of relapse and TRM were 31.3% and 6.6%, respectively (Figure 1b) . Repeated infusions of zoledronic acid (i.e. more than 3) seem to offer an advantage in terms of DFS (87.5% vs 48.6%, P ¼ 0.13), although the difference was not statistically significant (Figure 1c and d) .Conclusion: These data suggest that the infusion of zoledronic acid after TCRab/CD19-depleted haplo HSCT is safe. Repeated infusions of zoledronic acid after haploidentical HSCT seems to be more effective in preventing leukemia recurrence. More patients and a longer follow-up are needed to establish the efficacy of this approach. Disclosure of Interest: None declared.

Inducible T-cell receptor expression in precursor T-cells for leukemia control Introduction: The co-transplantation of hematopoietic stem cells (HS) with those that have been engineered to express tumor-reactive T cell receptors (TCRs) and differentiated into precursor T cells (preTs) may optimize tumor reduction. Since expression of potentially self-(tumor-) reactive TCRs will lead to negative selection upon thymic maturation, we investigated whether preTs forced to express a leukemia-reactive TCR under the control of a tetracycline-inducible promoter would allow timely controlled TCR expression thereby avoiding thymic negative selection. Materials (or patients) and methods: Using lentiviral vectors, murine LSK cells were engineered to express a Tetracyclineinducible TCR directed against a surrogate leukemia antigen. TCR-transduced LSK cells were co-cultured on T cell development-supporting OP9-DL1 cells to produce preTs. Lethallyirradiated B6/NCrl recipients received syngeneic T celldepleted bone marrow and 8 Â 10 6 syngeneic or allogeneic (B10.A) TCR-engineered preTs. An otherwise lethal leukemia cell (C1498) challenge was given 28 days later. Results: After in vivo maturation and gene induction up to 70% leukemia free survival was achieved in recipients of syngeneic TCR-transduced preTs (Po0.001) as shown in figure 1 A. Importantly, transfer of allogeneic gene-manipulated preTs increased the survival of recipients (Po0.05) without inducing graft versus host disease (GVHD). Nontransduced preTs provided significantly lower leukemia protection being not significantly superior to the PBS controls. The progenies of engineered preTs gave rise to effector and central memory cells providing protection even after repeated leukemia challenge. In vitro transduction and consecutive expansion of mature T cells required at least 40 Â 10 6 cells/ recipient to mediate similar anti-leukemia efficacy, risking the development of severe GVHD if of mismatched origin, and providing no long-term protection. Importantly, while transgene induction starting immediately after transplant forced CD8 þ T cell development and was required to obtain a mature T cell subset of targeted specificity, late induction favored CD4 differentiation and failed to produce a leukemiareactive population due to missing thymic positive selection. Conclusion: Co-transplanting TCR gene-engineered preTs is of high clinical relevance since small numbers of even mismatched HS can be transduced at a reasonable cost, expanded in vitro, stored if needed, and provide potent and long lasting leukemia protection. Disclosure of Interest: None declared.