Impact of modernizing eligibility criteria on enrollment and representation in acute myeloid leukemia clinical trials

Abstract

Abstract Clinical trial eligibility criteria select a target population and reduce anticipated risks for participants but may unnecessarily limit participation both overall and differently across demographic groups. We previously abstracted eligibility criteria for 190 phase 2/3 acute myeloid leukemia (AML) trials and used US Food and Drug Administration and professional society guidance on modernizing criteria to develop alternative, safety-based eligibility criteria for each trial. In this analysis, these trial- and safety-based eligibility criteria sets were applied to a retrospective cohort of 2226 newly diagnosed patients across 8 hospitals to assess the impact on eligibility. Eligibility proportions increased from a median of 47.9% with trial-based criteria to 84.2% with safety-based criteria (median difference, 30.0%; P< .001); excluding age criteria, the increase was 11.5% (P< .001). Non-Hispanic (NH) Asian, NH Black, NH White, and Hispanic patients were eligible for median proportions of 41.1%, 44.0%, 47.9%, and 50.0%, respectively, with trial-based criteria, increasing by 27.9% to 31.6% when using safety-based criteria (within-group changes, all P< .001; between-group changes, all P> .05). Excluding age criteria, increases were between 10.0% and 11.9%. Moving from trial- to safety-based criteria decreased the proportion of trials with significant eligibility differences between NH White and NH Asian (−11.1%), NH Black (−4.2%), and Hispanic (−12.1%) patients. Criteria significantly associated with increased eligibility and decreased between-group differences in eligibility were coronary artery disease, congestive heart failure, aspartate transaminase level, upper age limits, and previous malignancy. These data suggest that modernization of eligibility for AML trials to focus on safety-based criteria can improve both overall enrollment and population representation.