e13707 Background: Excessively stringent clinical trial eligibility criteria may hinder trial accrual, differentially exclude racial-ethnic minority groups, and limit generalizability. This study reviewed the association between trial accrual and concordance of eligibility criteria with known drug safety data in DLBCL clinical trials. Methods: Eligibility criteria and accrual of completed phase II/III clinical trials for DLBCL that opened in the US between 2010-2020 were abstracted from ClinicalTrials.gov. Trials missing posted results or studying cellular therapy / transplant were excluded. Concordance with known safety data was assessed by comparing study eligibility criteria for QTc, renal, hepatic, and cardiac function to FDA labels and phase I safety signals of study drugs. Excessively strict discordant criteria (ExSDC) were defined as eligibility criteria without supporting safety signals. Logistic regression +/- dispersion was used for analysis. Results: 58 trials met criteria for inclusion, accruing 6,232 participants. 88% of studies involved a non-FDA-approved agent, 74% were multicenter, 55% were industry sponsored, 66% only accrued in the US, and 86% were phase II. Eligibility criteria were highly variable (Table). Discordance of renal, hepatic, cardiac function, and QTc criteria from known safety data occurred in 31%, 26%, 41%, and 29% of studies, respectively. Race and ethnicity data were not reported in 22% and 45% of studies, respectively. Trials after 2015 were more likely to report race (p=0.04) and ethnicity (p=0.01). In a multivariable model adjusted for international study sites, greater number of ExSDC was associated with a lower proportion of black patients in the trial (OR 0.45, CI 0.29-0.71, p=0.001). ExSDC in cardiac function were most associated with reduced accrual of this minority group (OR 0.17, CI 0.05-0.58, p=0.007). 11 studies (19%) had early termination for lack of accrual (ETLA). In multivariable analysis, greater number of ExSDC was independently associated with ETLA (OR 3.53, CI 1.45-8.64, p=0.006), while multicenter status protected against ETLA (OR 0.05, CI 0.01-0.31, p=0.002). Industry sponsor and concordance of cardiac criteria protected against ETLA in univariable analysis only. Other eligibility criteria were not associated. Conclusions: Eligibility criteria in DLBCL trials varied widely and were often discordant with known safety data. Higher number of ExSDC was associated with lower accrual of black patients and higher risk of ETLA. Optimizing concordance of eligibility criteria with safety data may improve trial accrual and diversity. [Table: see text]