BIOS-08. A COMPREHENSIVE ANALYSIS OF METRICS AFFECTING GLIOBLASTOMA RANDOMIZED CONTROLLED TRIALS ROBUSTNESS

Abstract

Abstract INTRODUCTION Randomized controlled trials (RCTs) are the gold standard for evaluating the effectiveness of oncological interventions. RCTs are influenced by various factors, including trial design, outcome measures, sample size, recruitment, and retention. These factors can impact the robustness of a trial. OBJECTIVE We comprehensively evaluate the characteristics and robustness of RCTs supporting current GBM management by evaluating trial design, treatment effects, reporting strategies, primary outcome description, patients’ loss to follow-up as well as author and patient demographics. METHODS We conducted database searches and all GBM RCTs with significant outcomes published between 2010-2022 were included. Trial characteristics, survival times, the fragility index (FI-minimum number of patients needed to switch to nullify trial outcome) were assessed. Multivariable regression was performed to assess association of trial characteristics with robustness. RESULTS Of the 310 RCTs identified, 95 (30.6%) trials reported overall survival as primary end point, 92 (29.7%) progression free survival and 123 (39.7%) examined safety profiles of drugs. Of these, 178 (57.4%) were published from 2015-2022. The trials had a median power of 80.0% (interquartile range [IQR] 80.0-90.0). The median number of enrolled patients was 46 (IQR 24-108) with a median of 2 (IQR 1-5) loss to follow up. Of these trials, 68 were eligible for FI calculation. The median FI was 4.0 (IQR 2.0-11.50), with 38 trials (55.9%) having an FI < 4 and a third of the trials had FI less than number of patients loss to follow up. After adjusting for co-variates, recently published trials, multiple centers trials and commercial sponsorship was associated with significantly higher FI (p< 0.05). CONCLUSION Trials supporting current GBM management guidelines have alarmingly low robustness. While the robustness of trials has improved over time, future trials must consider additional metrics in their planning.