ALBERT

Description: BACKGROUND The Hedgehog (Hh) signaling pathway, which plays an important role during embryogenesis, can be reactivated in a wide range of cancers. Vismodegib, a selective hedgehog pathway inhibitor, demonstrated antitumor activity in medulloblastoma and basal-cell carcinoma (BCC) patients and has been approved by the FDA and EMA for the treatment of BCC. Preclinical data indicate that Hh pathway can also be activated in lymphoid malignancies and that its inhibition has antitumor activity (Dierks et al, Nature Med 2007; Singh et al, Leukemia 2012; Decker et al, Blood 2012). METHODS We conducted a phase II trial testing vismodegib in patients with relapsed/refractory lymphoma and CLL. Vismodegib was given orally at the dose of 150mg/day until disease progression or unacceptable toxicities for a maximum of one year. The primary objective was to evaluate the efficacy of vismodegib as measured by the best overall response rate (ORR) during the treatment period. RESULTS Between February 2013 and January 2014, 31 patients were recruited including diffuse large B-cell lymphoma (DLBCL, N=12), indolent lymphoma (iNHL, N=6), primary CNS lymphoma (PCNSL, N=10) and chronic lymphocytic leukemia (CLL, N=3). Patient characteristics are summarized in Table 1. We found that Hh signaling pathway (measured by IHC and/or PCR based on the expression of SHH, PTCH 1 and 2, SMO, GLI 1, 2 and 3, and ABCG2) was frequently activated at baseline in lymphoma patients. PK analysis demonstrated bioavailability of vismodegib in blood and CSF (median concentration at day 28 = 13398 [6970-20700] and 323 [99-717]ng/L, respectively). Nevertheless, none of the patients responded to vismodegib therapy, except for one. This patient had an iNHL (grade 3a follicular lymphoma). He experienced a partial remission after 2 cycles that lasted for 4.8 months. Interestingly, this patient had the strongest expression of GLI (1, 2 and 3) by PCR compared to all the other patients tested. All patients discontinued treatment prematurely (Table 1), mostly due to disease progression (90.3%). Adverse events (AE) and serious AEs (SAE) related to vismodegib were observed in 51.6 and 12.9% of the patients, respectively. SAEs related to vismodegib were diarrhea, vomiting, lung infection, hypoglycemia, pulmonary embolism, and skin rash (1 patient each). CONCLUSIONS Despite frequent Hh pathway activation, treatment with vismodegib did not show significant clinical efficacy in patients with lymphoma or CLL. This study was supported by Roche. Table 1. DLBCL iNHL PCNSL CLL Patients (N) 12 6 10 3 Median age (years) 74 74.5 65 74 Median number of prior therapies (min-max) 3 [1-6] 2.5 [1-4] 2 [1-3] 3 [1-4] Median duration of treatment (months) 1.7 1.8 1.2 1.9 Best overall response (CR/uCR/PR) 0% 16.7% (N=1*) 0% 0% Median PFS (months) 1.7 2.2 1.2 3 Median OS (months) 5.4 21.3 16.4 19.4 (*) Duration of response = 4.8 months Disclosures Off Label Use: Vismodegib in lymphoid malignancies. Haioun:Roche: Honoraria. Thieblemont:St. Louis Hospital, Paris, France: Employment. Casasnovas:Roche: Consultancy, Research Funding; Takeda: Consultancy; Gilead: Consultancy. Morschhauser:Genentech Inc./Roche: Other: Advisory boards. Salles:Roche: Honoraria, Research Funding.

Description: Abstract 1693 Poster Board I-719 A central goal of cancer immunotherapy is an adaptive immune response against tumors. The memory T cell is a critical mediator of this response as it can give rise to effector cells as well as self-renew. Regulatory T cells (Tregs) present a barrier to effective cancer immunotherapy. Indeed, cancer patients have increased numbers of CD4+CD25+ Tregs and cancer vaccination strategies in some cases expand this cell population. Here, we demonstrate that (1) CD4+CD44hi memory T cells are effective in mediating anti-tumor immunity and (2) that expression of CD137 can be used to exclude tumor-reactive Tregs from the CD4+CD44hi population. We have established a model for adoptive therapy of mantle cell lymphoma in which CD4 T cells mediate anti-tumor immunity. Specifically, we use a whole tumor-cell vaccine to induce anti-tumor immune cells in vivo. These cells are isolated and adoptively transferred into lethally irradiated, syngeneic, recipient mice. We show that CD4 but not CD8 T cells from vaccinated donor mice can prevent tumor growth when adoptively transferred into irradiated recipient mice. We observed that a majority of anti-tumor T cells display a memory phenotype. 10 days after transfer, T cells from recipient mice were co-cultured with irradiated lymphoma cells for 24 hours. Only CD4 T cells produced IFN-γ in response to co-culture and greater than 95% of IFN-γ+ CD4 T cells expressed the memory marker CD44. Finally, we observed that transferred CD4+CD44hi T cells persisted for over 100 days suggesting that this CD4 subset is important for lasting anti-tumor immunity. Contaminating Tregs may limit the effectiveness of our CD4 T cell adoptive therapy. In order to functionally deplete these Tregs from the CD4 population, we sought to identify a surface marker that could uniquely distinguish tumor-reactive Tregs from other CD4 T cells. T cells from vaccinated donor mice were co-cultured with irradiated lymphoma cells for 24 hours. We evaluated a panel of activation markers and observed that CD137 expression defined a distinct population of Tregs. Based on these results, we used flow cytometry to isolate a sub-population of CD4+CD44hiCD137- T cells from vaccinated donor mice. Adoptive transfer of less than 100,000 CD44hiCD137- but not other sub-populations of CD4 T cells provided significant protection from tumor challenge. These results suggest that CD137 defines a novel population of tumor-reactive Tregs. This marker can facilitate the enrichment of anti-tumor CD4 memory T cells within the CD44hiCD137- population. In conclusion, these findings highlight the potential use of CD4 T cells in adoptive therapy for mantle cell lymphoma. Disclosures No relevant conflicts of interest to declare.

Description: Abstract 422 Background: Antibody-dependent cell-mediated cytotoxicity (ADCC), largely mediated by natural killer (NK) cells, is thought to play an important role in the efficacy of rituximab, an anti-CD20 monoclonal antibody (mAb) used to treat patients with B cell lymphomas. CD137 is a costimulatory molecule expressed on a variety of immune cells following activation, including NK cells. We hypothesized that the anti-lymphoma activity of rituximab could be enhanced by stimulation of NK cells with an anti-CD137 agonistic mAb. Methods: Rituximab induced upregulation of CD137 on NK cells was assessed using lymphoma cell lines and primary lymphoma patient samples. In-vitro NK cell degranulation and cytotoxicity were assessed by CD107a mobilization and chromium release. A murine lymphoma tumor model targeted by mouse anti-mouse CD20 mAb was used to assess in-vivo synergy of anti-CD20 and anti-CD137 mAbs. Mechanism of synergy was explored by T cell, NK cell, and macrophage depletion in the immune competent mouse model. A xenotransplant model in SCID mice with disseminated, luciferase-labeled lymphoma was used to demonstrate efficacy of anti-CD137 mAb and rituximab, and sufficiency of an innate immune response. Results: NK cells in human primary lymphoma samples do not express CD137 at baseline, however these cells highly upregulate CD137 when encountering rituximab-coated tumor B cells. Rituximab-induced NK cell degranulation and cytotoxicity as measured by CD107a mobilization (p=.006) and chromium release (p=.01) are enhanced by anti-CD137 agonistic mAb. In a murine lymphoma model, anti-CD137 mAb significantly enhances anti-tumor activity of anti-CD20 mAb leading to complete tumor resolution (p

Description: Introduction: in large unselected series, the median age of Primary CNS Lymphoma (PCNSL) patients (pts) is about 70 years. In one USA cancer registry study for PCNSL patients (pts) older than 65 years, 14% of them are older than 80 years. Data on clinical characteristics, therapeutical management, toxicity of treatment and outcome of these very elderly pts are limited. Methods: We reviewed PCNSL pts aged of 80 years or older included in the database of the French Oculo-Cerebral lymphoma (LOC) network. From January 2011, this network prospectively recorded all newly diagnosed PCNSL from 22 regional expert centers in France. For this study, 110 PCNSL pts with a histological diagnosis of diffuse large B-cell lymphoma (DLBCL) aged of 80 years or older were analyzed. All medical records were reviewed for clinical and biological characteristics, modality of treatment and supportive care, toxicities and outcome. Results: 110 pts with a DLBCL PCNSL aged of 80 years or older were diagnosed between January 2011 and January 2018 representing 8% of pts available in the LOC database. The clinical characteristics were as follows: 63% of females; median age: 83y (80-92); performance status (PS) according to EORTC scale: 1, 24% of pts, 2, 21% of pts, 3-4, 55% of pts. 23.6% of pts had a CIRS-G grade 3 or 4 in at least one category and 40.9% had a cumulative CIRS-G score more than 6. Diagnosis procedure was biopsy (87%), tumor resection (5%), vitrectomy (4%), CSF cytology (4%). At diagnosis, 9/68 (13%) of evaluable pts had ocular involvement, 13/61 (21%) cerebrospinal (CSF) involvement, 79/110 (72%) involvement of the deep structures of the brain and 35/86 (41%) had elevated LDH level. Median creatinine clearance (CKD.EPI) was 70ml/min (min: 24, max: 102). Treatment was initiated either by a neuro-oncology or a hematology team in 35% and 65% of cases, respectively. Median delay between first symptoms and treatment was 60 days. First line treatment was high-dose (HD) methotrexate (MTX) based chemotherapy (CT) in 85 pts (77%), other chemotherapy regimen in 13 pts (12%) and palliative care in 12 pts (11%). Median number of CT cycles was 3 (1-11) with a median dose of MTX of 3g/m2 (0.5-5.0). Interestingly, no difference of distribution for the main clinical and biological characteristics (median age, PS, symptoms, tumor localization, albumin level, creatinine clearance) was observed between these three groups. Rituximab was used in combination with CT in 53/98 treated pts (54%). After first-line induction chemotherapy, response rate for evaluable patients (n=85) were as follows: 37% of complete response, 9% of partial response, 54% of stable or progressive disease. Finally, 27 pts (32%) received consolidation treatment with high-dose cytarabine after MTX-based CT. For toxicity, among the 351 infusions performed for the 85 pts who received MTX-based CT, grade 3-4 toxicities were: 46% of any events, 15% of infection, 13% of cytopenia, 10.5% of acute renal failure and 8% of elevated liver enzymes. 13% of pts presented toxic death. Median progression free survival (PFS) and overall survival (OS) were 3.9 months and 7 months, respectively. Pts treated with MTX-based CT had a significantly prolonged PFS and OS as compared to patients treated without MTX or with palliative care (Figure 1A, 1B). In the univariate analysis performed for the 85 pts treated with MTX-based CT, no initial clinical and biological characteristics (age, PS, type of symptoms, CIRS-G, tumor localization, LDH level, albumin level, hemoglobin level, lymphocyte count, creatinine clearance) influenced PFS or OS. The initial dose of MTX did not influence outcome but intravenous rituximab used in first line therapy significantly improved PFS and OS (Figure 1C, 1D). Conclusions: to the best of our knowledge, this is the largest series of consecutive PCNSL pts aged of 80y or over prospectively recorded in a national database. This study showed that the prognosis remains poor with major toxicity under conventional treatment. No clinical predictor of survival was highlighted in our series but patients initially treated with MTX-based CT in combination with rituximab had an improved outcome. The development of target and innovative therapies is needed for this category of patients representing 8% of all PCNSL in the database of the LOC network. Disclosures Houot: Celgene: Honoraria; BMS: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees; Janssen: Honoraria; Novartis: Honoraria.

Description: Several gene-expression signatures predict survival in diffuse large B-cell lymphoma (DLBCL), but the lack of practical methods for genome-scale analysis has limited translation to clinical practice. We built and validated a simple model using one gene expressed by tumor cells and another expressed by host immune cells, assessing added prognostic value to the clinical International Prognostic Index (IPI). LIM domain only 2 (LMO2) was validated as an independent predictor of survival and the “germinal center B cell–like” subtype. Expression of tumor necrosis factor receptor superfamily member 9 (TNFRSF9) from the DLBCL microenvironment was the best gene in bivariate combination with LMO2. Study of TNFRSF9 tissue expression in 95 patients with DLBCL showed expression limited to infiltrating T cells. A model integrating these 2 genes was independent of “cell-of-origin” classification, “stromal signatures,” IPI, and added to the predictive power of the IPI. A composite score integrating these genes with IPI performed well in 3 independent cohorts of 545 DLBCL patients, as well as in a simple assay of routine formalin-fixed specimens from a new validation cohort of 147 patients with DLBCL. We conclude that the measurement of a single gene expressed by tumor cells (LMO2) and a single gene expressed by the immune microenvironment (TNFRSF9) powerfully predicts overall survival in patients with DLBCL.

Description: Signal transduction plays a key role in cell survival, and changes to signaling are frequently implicated in tumor initiation and progression. We sought to identify abnormal variation in signaling network activity within primary tumor samples obtained prior to treatment from patients with follicular lymphoma (FL). We previously showed that altered B cell receptor (BCR) signaling distinguishes tumor B cells from the non-malignant host B cells in FL tumors. Here we extend this approach and use flow cytometry to measure 648 signaling events in live lymphoid cells from more than 25 lymphoma specimens and healthy controls. We combined 9 previously identified BCR stimulation conditions with inputs from CD40, interleukin 4, interferons (IFNs), and more than 10 other environmental cues that govern the development and activity of lymphocytes. Fluorescent cell barcoding allowed simultaneous staining and analysis of phospho-protein activation under all 27 stimulation conditions within a single tube. The activation of key phospho-protein nodes throughout lymphocyte signaling networks, including Syk, Erk1/2, Btk, Src family kinases, cCbl, p38, NFkB, Akt, Stat1, Stat3, Stat6, and Stat5, was measured under each of the 27 stimulation conditions. Measurements of phospho-protein responses to stimulation were combined with detection of the Bcl-2 oncogene, B and T cell lineage markers in each cell. This panel allowed us to characterize signaling in the heterogeneous cell subsets found within each patient’s tumor sample. Tumor B cells, host tumor infiltrating T cells, non-malignant B cells were all distinguished by contrasting signaling profiles. In some cases, subsets of tumor B cells with differences in signaling network topology were observed within the tumor B cell population. This result suggests that signaling can distinguish between tumor sub-clones and could be used to measure tumor heterogeneity. As previously reported, little variation in signaling was observed among healthy peripheral blood B and T cell samples from different individuals. Abnormally low host T cell signaling was commonly observed within the tumor infiltrating T cells infiltrating FL tumors. Further analysis of tumor T cell subsets indicated that a high proportion of infiltrating T cells expressed CD4 and FoxP3. Taken together, these results support the hypothesis that FL tumor B cells promote suppressed signaling in the T cells of the patient and may modulate the immune response against the tumor. In FL tumor B cells, BCR and IFN signaling frequently triggered Stat5 phosphorylation, but not Stat1 phosphorylation. These results are consistent with the hypothesis that Stat5 initiates genetic programs that support cancer cell survival and proliferation, whereas Stat1 promotes immunogenicity and cooperates with the p53 tumor suppressor protein. In contrast with healthy B cells, loss of the response to CD40L, altered PKC signaling, and variable responses to BCR crosslinking were all seen in FL tumor B cells. The patterns of abnormal signaling we observed in tumor B cells and tumor infiltrating T cells suggest that measuring the activity of key signaling network nodes can identify targets for therapeutic attention in FL.

Description: RATIONAL Primary CNS lymphoma (PCNSL) is a diffuse large B-cell lymphoma (DLBCL), predominantly of non-germinal center (non-GC) subtype, carrying a pejorative prognosis. Constitutive activation of the NF-kB pathway via mutations in B cell receptor (BCR) pathway (CD79B) and mutation of MYD 88 and TBL1XR1 plays an important role in PCNSL. Ibrutinib, an inhibitor of BCR signaling, has been found to have significant therapeutic activity in relapsed or refractory non-CNS non-GC DLBCL. METHODS In this prospective, multicenter, open-label phase II, we enrolled immuno-competent patients over 18 with a refractory or relapse of PCNSL or primary vitreo-retinal lymphoma (PVRL) of DLBCL type. The treatment consisted in ibrutinib monotherapy given orally at 560 mg daily until disease progression or unacceptable toxicity. Additional corticosteroids treatment was allowed during the first 4 weeks of treatment in case of a threatening or symptomatic edema. Therapeutic responses were assessed according to the international primary CNS lymphoma collaborative group (IPCG) criteria. The primary objective of the study was the disease control (DC) rate (CR + CRu + PR + SD) after two months of treatment. This study is a two-stage Simon's design. Patients were evaluable for response if they received 〉 90 % of the expected dose during the first month of treatment. An interim analysis for futility was planned when 18 patients were evaluable for response. P0 and P1 hypotheses were 〈 10 % and 〉 30 % respectively. A total of 35 evaluable patients are required for the final analysis. Exploratory ancillary studies are planned and consist in dosage of ibrutinib in the cerebrospinal fluid after one cycle of treatment, and correlation of therapeutic response with mutational status of the disease. This study is registered with ClinicalTrials.gov, number NCT02542514. RESULTS BetweenSeptember 25, 2015 and June 30, 2016, 52 patients were recruited in 10 French centers of the French LOC network for PCNSL. The interim analysis was done on the first 18 patients evaluable for response (median age: 70 y, range 49-80). At initial diagnosis, diagnoses were PCNSL (n = 12) and PVRL (n = 6). Patients were included in the study for a relapse (n = 13) or a progressive disease (n = 5). At time of inclusion in the study, disease status was PCNSL (n = 11) and PVRL or isolated intra-ocular relapse of a PCNSL (n = 7). ECOG performance status was 0, 1 and 2 in 4, 10 and 4 patients respectively. All the patients had previously received high-dose methotrexate-based chemotherapy. Four patients had previously received high-dose chemotherapy followed by autologous stem cell transplantation. Patients had received 1, 2 or 3 prior treatments in 12, 5 and 1 cases respectively. Three patients had a concomitant meningeal involvement. Five patients received concomitant corticosteroids during the first month of treatment. At the time of analysis (median follow-up = 6.6 months), nine patients discontinued ibrutinib after a median duration of 3 months (range, 0.9 -6.4) because of a disease progressive(n = 8) or a concurrent illness (n=1). Median number of treatment cycles was 5 (range, 1-9). One patient experienced a pulmonary aspergillosis with a favorable outcome. No hemorrhagic complication was reported. Five patients died due to disease progression (n = 4), and concurrent illness (n = 1). After two months of treatment, a DC was achieved in 15/18 patients (83 %, IC 95 %, [59-96%]) (complete and unconfirmed complete response: n =3; partial response: n = 7; stable disease: n =5). CONCLUSION In this interim analysis, Ibrutinib monotherapy demonstrated a high DC rate of 83%, including 56% objective responses in patients with relapse/refractory PCNSL or PVRL. Regarding safety, Ibrutinib might be a risk factor for aspergillosis in this population of PCNSL patients, otherwise not exposed to fungal infection. A security warning was sent to all the investigators for a close monitoring of infections. The second cohort of patients has been recruited. Thirty-three patients are currently on study treatment. The final analysis of the iLOC study is awaited to confirm these encouraging results and better define the positioning of ibrutinib in the therapeutic strategy of PCNSL and PVR patients. Disclosures Choquet: Janssen: Consultancy; Celgene: Consultancy. Ghesquieres:Mundipharma: Consultancy; Roche France: Research Funding; Celgene: Consultancy, Membership on an entity's Board of Directors or advisory committees. Soussain:Celgene: Research Funding; Roche: Research Funding; Pharmacyclics: Research Funding.

Description: Introduction: In diffuse large B-cell lymphoma (DLBCL), the number of circulating monocytes and neutrophils represents an independent prognostic factor. These cells include monocytic and granulocytic myeloid derived suppressor cells (M- and G-MDSCs) defined by their ability to suppress T-cell responses. MDSCs are a heterogeneous population described in inflammatory or infectious diseases as well as in numerous tumors including multiple myeloma, chronic lymphocytic leukemia and DLBCL. However, their mechanisms of action in DLBCL remain unclear. The aim of the study was to investigate the presence and mechanisms of suppression of MDSC subsets in DLBCL. Methods: Whole blood from 56 DLBCL at diagnosis and 45 healthy donors (HD) were analyzed by flow cytometry. Quantitative Real Time PCR were analyzed on Taqman@ array microfluidic cards. Cytokine levels in plasma and culture supernatants were measured by Luminex and Elisa. T-cell proliferation after monocyte depletion or treatment with inhibitors, was analyzed by CFSE assay. Results: By flow cytometry, we identified an expansion of G-MDSC (Linneg HLA-DRneg CD33pos CD11bpos) and M-MDSC (CD14pos HLA-DRlow) in DLBCL compared to HD (p

Description: Abstract 2093 Background: Adoptive immunotherapy is a promising novel approach to the treatment of cancer. However, clinical translation of adoptively transferred CD4 T cells is limited by cotransfer of an inhibitory population of regulatory CD4 T cells (Tregs). We identified a method of isolating viable antitumor CD4 T cells while excluding Tregs based on two surface markers—CD44 and CD137. Methods: We have developed a model for adoptive cell therapy of lymphoma whereby anti-tumor T cells are generated in vivo through vaccination with a CpG-loaded whole cell vaccine (CpG/H11). These vaccine-induced cells can protect from lethal tumor challenge when isolated and transferred into lethally irradiated, syngeneic recipient mice. We investigated the subsets of T cells involved in the anti-tumor response through a combination of in vitro and in vivo assays. Results: Adoptive transfer of CD137negCD44hi CD4 T cells, but not other sub-populations, provided protection from B cell lymphoma. We demonstrate that the population of CD137posCD44hi CD4 T cells consists primarily of activated Tregs. In vitro, these CD137pos cells suppressed the proliferation of effector cells in a contact-dependent manner. We observed that this CD137pos Treg population persisted following adoptive transfer and maintained expression of FoxP3 as well as CD137. Moreover, the addition of CD137posCD44hi CD4 cells to CD137negCD44hi CD4 cells suppressed the antitumor immune response. In the presence of CD137posCD44hi CD4 T cells, homing of other T cell populations to tumor sites was disrupted. These results suggest that CD137 expression on CD4 T cells defines a population of activated Tregs that prevent antitumor immune responses. Conclusions: Our findings identify two surface markers that can be used to facilitate the enrichment of anti-tumor CD4 T cells while depleting an inhibitory Treg population. This approach has direct applicability to clinical trials for patients with lymphoma. Disclosures: No relevant conflicts of interest to declare.

Description: BACKGROUND: the combination of rituximab (RTX) and lenalidomide (LEN) has been shown to be synergistic in pre-clinical models and in patients (pts) with previously untreated and relapsed follicular lymphoma (FL). Furthermore, LEN demonstrated single agent efficacy in pts with mantle cell lymphoma (MCL) and diffuse large B cell lymphoma (DLBCL). We hypothesized that the combination of obinutuzumab (GA), a glycoengineered type II anti-CD20 antibody, with LEN (GALEN) might be even more efficient while retaining a similarly manageable safety profile. In 2012, we started a phase Ib/II study to assess the safety and efficacy of this combination (ClinicalTrials.gov: NCT01582776) in pts with relapsed/refractory lymphoma. Phase IB defined the recommended dose (RD) of LEN as being 20mg in combination with GA. Here, we report the results of the Phase II part assessing efficacy and safety of GALEN therapy during a 6 months induction followed, for responders, by a reduced LEN maintenance GALEN regimen (in a cohort of relapsed/refractory aggressive NHL (DLBCL or MCL) pts. METHODS: pts with a histologically confirmed DLBCL (de novo or transformed low-grade NHL) or MCL were eligible if they had a ECOG PS of ²2, a life expectancy 〉3 months (mo) and had previously received ³1 prior therapy including at least one RTX-containing regimen. Induction treatment consisted of LEN 20 mg given orally once daily on days 1-21 of a 28-day cycle for the first cycle and on days 2-22 of a 28-day cycle from cycles 2 to 6. Intravenous infusions of GA were given at a flat dose of 1000mg on days 8, 15, and 22 of cycle 1 and at D1 of cycles 2 to 6 (total of 8 infusions). Responding pts then received 12 cycles of LEN at 10 mg daily on days 2-22 every 28 days for a total of 18 cycles and GA 1000mg every 8 weeks for 12 additional cycles until progression or unacceptable toxicity. All pts were required to take daily aspirin (100 mg) for deep vein thrombosis (DVT) prophylaxis during study period. Pts who were unable to tolerate aspirin and had a prior history of DVT or were at high risk for a DVT received low molecular weight heparin therapy or warfarin (coumadin) treatment. The primary study endpoint was overall response rate (ORR) by investigator assessment at the end of induction according to 1999 IWG criteria. Secondary endpoints included ORR and complete response (CR) according to IWG 2007, best ORR, duration of response, progression-free survival, overall survival (OS) and safety. RESULTS: ninety-one pts (DLBCL [n=77 including 18 transformed], MCL [n=13], and other [n=1]) were enrolled between June 2014 and March 2015 at 22 LYSA centers in France and Belgium. At baseline, median age was 70 (range, 48-84) years with 64.7% men, ECOG PS of 1 (47.1%) or 2 (18.8%), 83.5% Ann Arbor stage III-IV and 66.3% with elevated LDH. Median number of prior systemic therapies was 2 (range, 1-9) including haematopoietic stem cell transplantation in 17pts (20%); 61% were refractory to a RTX-containing regimen and 53% refractory to last prior therapy. Three pts were withdrawn before receiving any treatment due to major protocol violation, concurrent illness or ulceration of the tumor in the stomach and 3 pts received one of both study drugs, leaving 85 pts assessable for efficacy. Cut-off date was March 25th 2016. With amedianfollow-up of 14.5 mo,39pts (45.8%) completed induction (32 DLBCL and 7 MCL) and 20 (23.5%) are still ongoing in maintenance (16 DLBCL, 4 MCL). At the time of the cut-off, 44pts (51.8%) had died mainly due to lymphoma (88.6%). Response at the end of induction, best ORR during induction and OS are summarized in the table. Table Analysis of other secondary endpoints and correlation with cell of origin is ongoing. The most common treatment emergent AEs during induction (all grades 〉10%pts/ % grade 3/4) were neutropenia (51.1/31.8), infections (45.5/12.5), constipation (30.7/0.0), cough (21.6/1.1), diarrhea (17.0/3.4), nausea (17.0/0) anemia (15.9/10.2), thrombocytopenia (15.9/10.2), muscle spasms (13.6/0.0), fatigue (12.5/1.1) and dyspnea (11.4/4.5). Febrile neutropenia occurred in 4.5% pts. There was onlyone secondprimary malignancy (basal carcinoma). CONCLUSION: oral LEN plus GA infusion is effective in relapsed or refractory DLBCL and MCL pts with no unexpected toxicity. Longer follow-up and additional translational studies are needed to identify the subset of pts most benefiting of this regimen. Correlative studies will be presented at the meeting. Table Table. Disclosures Morschhauser: celgene: Honoraria, Membership on an entity's Board of Directors or advisory committees; roche: Honoraria, Membership on an entity's Board of Directors or advisory committees; servier: Consultancy, Honoraria; gilead: Consultancy, Honoraria; janssen: Honoraria. Cartron:Roche: Consultancy, Honoraria; Celgene: Honoraria; Gilead: Honoraria; Jansen: Honoraria. Salles:Janssen: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Celgene: Consultancy, Honoraria; Gilead: Honoraria, Research Funding; Mundipharma: Honoraria; Novartis: Consultancy, Honoraria; Roche/Genentech: Consultancy, Honoraria, Research Funding. Maerevoet:ARGN-X: Membership on an entity's Board of Directors or advisory committees; roche: Membership on an entity's Board of Directors or advisory committees; BMS: Membership on an entity's Board of Directors or advisory committees. Bonnet:ROCHE: Membership on an entity's Board of Directors or advisory committees; SERVIER: Membership on an entity's Board of Directors or advisory committees; BMS: Membership on an entity's Board of Directors or advisory committees; JANSSEN: Membership on an entity's Board of Directors or advisory committees.