INTRODUCTIONThe phase 1 selection for clinical trials of patients with relapsed or refractory Diffuse Large-B cell lymphoma (R/R DLBCL) with a potentially rapid tumor growth is a challenge. Prognostics' factors identification could help to better orient patients in appropriate clinical trials. This study is focused on DLBCL to evaluate their outcomes in phase I clinical trials. Our main objectives were to identify prognostic factors, attest the investigational drugs' safety and perform a preliminary assessment of drugs' efficacy.METHODAll consecutive patients with R/R DLBCL included in phase I clinical trial at a single cancer center in France between 2008 and 2017. If a given patient had participated in several phase I clinical trial, only the data from his first trial were examined. The patients' and DLBCL characteristics, the safety data and the efficacy outcomes were recorded. Reponses were assessed according to the International Harmonization Project in Lymphoma Cheson 2007 criteria. Overall responses rates (ORR) included partial responses (PR) and complete response (CR) and tumor control rates included overall responses and stable disease (SD).RESULTSA total of 101 patients (males: 63,4%) with R/R DLBCL were included in a panel of 21 clinical trials. The median age was 64 (range 21-86). Before their inclusion in a phase I trial, patients had received a median of 3 (1-7) lines of treatment and 25,7% of patients had undergone an autologous stem cell transplantation. At the cute-off date, 5 of the 101 patients (4,95%) were still taking the investigational drug. The median progression-free survival and overall survival (OS) were 1,8 and 9,7 months. High-grade toxicity (grade 3 or higher) occurred in 47 of the 101 patients (46,5%), and was related to the investigational drug in 29 of these cases (61,7%). The most important drugs-related toxicity was a hematological toxicity (neutropenia and thrombocytopenia) for 79,3% of patients, other toxicities were digestive toxicity (5,9%), cutaneous toxicity (2,9%), hepatotoxicity (2,9%) and pneumopathy (2,9%). Dose-limiting toxicity (DLT) was experienced by 7 (6,9%) of the 101 patients. The time from D1C1 (day 1, cycle 1) to the occurrence of high-grade toxicity was 1,2 months (range 0,2-9,5). High-grade toxicity occurred during the DLT period (< 6 weeks) for 27 of the 46 patients (58,7%) and after the DLT period in the remaining 19 (41,3%) patients.Factors associated with poor overall survival (overall survival inferior to the median OS) were Ann Arbor staging (73.6% of poor OS patients (< 9.7 months) with a performance status at 4), serum albumin (30.2% of poor OS patients with a low rate of serum albumin ≤ 35 g/L), serum LDH (79.2% of poor OS patients with a high rate of serum LDH > 250 U/L) and progressive disease (90.6% of poor OS patients had a trial withdrawal because of progression). The overall objective response and disease control rates were 25% and 43%. The median trial duration was 1.9 months and were respectively 12,1 and 5,5 months for responders (PR and CR) and controlled patients (PR and CR and SD). 36.6% of patients were prematurely withdrawing of study (before weeks 6). Factors associated with premature withdrew were performance status at baseline (67.6% of early withdrawal patients at performance status 1); Ann Arbor staging (67.6% of early withdrawal patients at Ann Arbor stage 4); serum LDH (73% of early withdrawal patients with a high rate of serum LDH > 250 U/L) and serum albumin (35,1% of early withdrawal patients with a low rate of serum albumin ≤ 35 g/L).CONCLUSIONHigh-grade toxicity occurred after the DLT period in 41.3% of patients with R/R DLBCL, suggesting that the conventional concept of dose-limiting toxicity should be redefined in the era of modern cancer therapies. Besides, even if the phase 1 selection for clinical trials is very selective, it's necessary to better orient patients in hematology. In fact, 36.6% of patients were prematurely withdrawing of study which could be anticipated thanks to the identification of prognostics' factors. Although the objective response is only a secondary endpoint in phase I clinical trials, the median duration of participation in trials (almost one year for responders and 5.5 months for controlled patients) are relevant for some new possibilities of therapeutics in the field of early drugs clinical trials. [Display omitted] DisclosuresRibrag:NanoString Technologies: Consultancy, Honoraria; epizyme: Consultancy, Honoraria; Incyte Corporation: Consultancy; MSD: Honoraria; Roche: Honoraria, Other: travel; Infinity: Consultancy, Honoraria; BMS: Consultancy, Honoraria, Other: travel; Amgen: Research Funding; Servier: Consultancy, Honoraria; pharmamar: Other: travel; Gilead: Consultancy, Honoraria; argenX: Research Funding.