Risk of bias and clinical benefit of phase III trials supporting FDA-approved anticancer medicines, 2018-2025: A...

Phase III randomised controlled trials often support FDA approval of new anticancer drugs, but trial quality reporting and clinical relevance vary. We assessed the risk of bias and clinical benefit of phase III randomised trials supporting FDA approvals of new anticancer drugs.

Cross-sectional study of pivotal phase III randomised trials supporting initial FDA approvals of anticancer drugs (January 1, 2018-May 15, 2025). Risk of bias was assessed using Cochrane Risk-of-Bias Tool 2.0 (low, some concerns, high), and clinical benefit with the European Society for Medical Oncology Magnitude of Clinical Benefit Scale (substantial, intermediate, low).

Among 43 indications supported by randomised trials, 21 (51%) were open-label. Overall survival benefit was demonstrated in 10 (23%) studies and quality of life improved in 6 (14%). Using a combined assessment incorporating publications and FDA regulatory documents, 13 trials (30%) were judged to be at low risk of bias, 24 (56%) raised some concerns, and 6 (14%) were at high risk of bias. Source-specific assessments differed: 19/42 trials (45%) were classified as high-risk based on published reports alone, versus 7/43 (17%) when assessed using FDA review documents alone. Substantial clinical benefit was observed in 14/43 (33%), but only 4 (9%) combined low risk of bias with substantial benefit. Among the 16 trials (37%) with low clinical benefit, 12 (75%) had methodological concerns or high risk of bias. Risk of bias was inconsistently associated with clinical benefit, overall survival, or quality-of-life outcomes.

Recent FDA approvals of anticancer drugs frequently rely on trials with methodological limitations and limited clinical benefit. Integrating risk-of-bias and clinical benefit assessments may strengthen pre-approval evaluation and post-approval oversight, support patient communication, and inform value-based pricing decisions.