Accrual speed and characteristics of phase III U.S. cancer clinical trials with frequently used biomarkers in eligibility criteria.

Abstract

475 Background: Patient accrual remains a key rate-limiting step in cancer clinical trials, often influencing statistical power, trial budgets, and delayed access to therapies. In particular, trials involving biomarker measurement may have accrual concerns related to the additional screening procedures in supporting important precision oncology decisions. However, few studies have measured accrual speeds and characterized trials that incorporate biomarkers in eligibility criteria versus those that do not. Methods: We identified U.S.-based phase III, adult, interventional therapeutic oncology trials from ClinicalTrials.gov, initiated after 1/1/2000 and completed before 1/1/2025. We defined accrual speed as patients per site per month (PPSM). Using the FDA Table of Pharmacogenomic Biomarkers, we identified biomarker trials via keyword searches of eligibility criteria and included the most common biomarker targets (occurrence > 8). We compared trial characteristics and accrual speeds between trials with versus without frequently used eligibility biomarkers. Results: Of the 517 trials, 132 (25.5%) incorporated the most frequently used biomarkers (ERBB2, EGFR, PD-L1, PD-1, ALK, CD20, and BRAF) in their eligibility criteria (frequent-biomarker trials). Trials with frequently used eligibility biomarkers had lower accrual speed than trials without these biomarkers overall (PPSM: 0.2 ± 0.7 vs. 1.3 ± 5.3; p < .001), with consistently lower PPSM across individual biomarker categories (all p < .01). Compared to non-frequent-biomarker trials, frequent-biomarker trials were more likely to be multi-site (99.2% vs. 84.4%, p<.001), have larger enrollment size (median: 575 vs. 400, p<.001), be randomized (97.7% vs. 90.1%, p=.009), and be industry-sponsored (88.6% vs. 64.2%, p<.001). Conclusions: Trials with frequently used eligibility biomarkers, representing over one-fourth of phase III U.S. cancer trials, have significantly prolonged accrual and are associated with multi-site recruitment, large enrollment size, and industry sponsorship. Findings underscore a need for structural and workflow optimizations to improve accrual efficiency for large-scale biomarker trials. Trials with frequently used eligibility biomarkers (n, %)　 No (385, 74%) Yes (132, 26%) P PPSM, mean (SD) 1.3 (5.3) 0.2 (0.7) <.001 Multi-site 325 (84.4) 131 (99.2) <.001 Enrollment size, median [Q1,Q3] 400 [165,672] 575 [417,766] <.001 Randomized 347 (90.1) 129 (97.7) .009 Double/triple blind 136 (35.3) 55 (41.7) .231 Industry sponsored 247 (64.2) 117 (88.6) <.001 Top-common biomarker (n, %) PPSM (mean, SD) ERBB2 (49, 9.5%) 1.1 (4.8) 0.3 (1.0) .008 EGFR (44, 8.5%) 1.1 (4.8) 0.2 (0.6) <.001 PD-L1 (35, 6.8%) 1.1 (4.8) 0.1 (0.1) <.001 PD-1 (26, 5.0%) 1.0 (4.7) 0.1 (0.1) <.001 ALK (20, 3.9%) 1.0 (4.7) 0.1 (0.1) <.001 CD20 (13, 2.5%) 1.0 (4.7) 0.1 (0.1) <.001 BRAF (9, 1.7%) 1.0 (4.7) 0.2 (0.1) <.001