Inconsistent Reporting and Definitions of Time-to-Event Endpoints in CAR T Clinical Trials: A Review

Clinical trials evaluating chimeric antigen receptor T-cell therapy (CAR T) commonly report time-to-event (TTE) endpoints. However, definitions are not necessarily comparable across studies and variability can lead to misinterpretation of results or inappropriate comparisons across products and studies. Amid the rapidly increasing number of published CAR T trials-many of which were used for regulatory approval-this study aims to summarize the variation in the use and reporting of TTE endpoints in CAR T trials. We include CAR T trials published January 2008 to January 2023 on PubMed that reported at least one of these TTE endpoints: overall survival (OS), progression-free survival (PFS), duration of response/remission (DOR), disease-free survival, event-free survival (EFS), relapse-free survival (RFS), time to relapse, time to progression, or time to treatment failure. We abstracted and summarized endpoint definitions, including the time origin, events, competing events, and censoring. We assessed the completeness of endpoint reporting, overall and by subgroups such as study phase, publication year, and the journal's impact factor. We included 116 publications in the analysis. The most frequently reported TTEs were OS (83%,), PFS (56%), DOR (55%), and EFS (23%). Complete reporting of endpoints was poor overall: 32%, 24%, 25%, and 56% for OS, PFS, DOR, and EFS respectively. Complete reporting was lower in articles published before 2018, in lower impact factor journals, and in phase I trials. There was also a large variability in TTE definitions among those reported. For example, among 64 studies reporting DOR, 48% used the date of response as the time origin while 20% used the date of infusion, and 31% did not report a time origin. There is substantial heterogeneity and incompleteness of TTE endpoint definitions in CAR T trials that could impact the interpretation of the study results. Improving TTE reporting, by stating the time origin, event(s) of interest, competing event(s) if any, and censoring, is required to ensure valid assessment of clinical benefit and cross-trial comparison.