This study evaluated the quality and clinical relevance of phase III randomized controlled trials that served as pivotal evidence for initial FDA approvals of anticancer drugs from January 1, 2018, through May 15, 2025. The authors sought to quantify trial risk of bias and to measure the magnitude of clinical benefit offered by these approvals.
The primary objective was to assess both the risk of bias and the clinical benefit of pivotal phase III randomized trials underpinning FDA approvals of new anticancer indications during the specified period. The analysis was cross-sectional and focused on trials cited as supporting initial approvals.
The analysis included pivotal phase III randomized trials that supported initial FDA approvals for anticancer indications between 2018-01-01 and 2025-05-15. The authors used information from published trial reports and FDA regulatory review documents to perform assessments. The exact list of included trials, trial names, individual drugs, and indication-level identifiers were reported in the original article (not reproduced here).
Risk of bias was evaluated using the Cochrane Risk-of-Bias Tool 2.0, classifying trials as low risk, some concerns, or high risk. Clinical benefit was measured with the European Society for Medical Oncology Magnitude of Clinical Benefit Scale (ESMO-MCBS), categorizing benefit as substantial, intermediate, or low. The authors performed assessments using published reports alone, FDA review documents alone, and a combined assessment integrating both sources.
Across 43 indications supported by randomized trials, 21 (51%) trials were reported as open-label, indicating that about half lacked blinding. The study noted variability in trial reporting and methodological transparency between public publications and FDA review materials.
Overall survival (OS) benefit was demonstrated in 10 of the included studies, representing 23% of the sampled indications. Improvements in quality of life (QoL) were reported for 6 studies, representing 14% of trials. These endpoints were used alongside ESMO-MCBS scoring to contextualize the magnitude of clinical benefit.
Using a combined assessment that integrated data from publications and FDA regulatory documents, the distribution of risk of bias was: 13 trials (30%) at low risk, 24 trials (56%) raising some concerns, and 6 trials (14%) at high risk. Assessments differed notably when performed using single sources: based on published reports alone, 19 of 42 trials (45%) were classified as high risk of bias, whereas FDA review documents alone yielded 7 of 43 trials (17%) classified as high risk. These discrepancies highlight how source selection influences bias judgments.
Per ESMO-MCBS, 14 of 43 indications (33%) achieved a rating of substantial clinical benefit. However, only 4 trials (9% of the sample) combined a designation of low risk of bias with substantial clinical benefit. Conversely, among 16 trials (37%) judged to have low clinical benefit, 12 (75%) had either methodological concerns or were at high risk of bias. Thus, a meaningful proportion of approvals provided limited clinical benefit and also exhibited methodological limitations.
The study found that risk of bias was inconsistently associated with measured clinical benefit, overall survival gains, or quality-of-life improvements. That is, methodological quality as judged by RoB 2.0 did not uniformly predict whether a trial showed OS benefit, QoL improvement, or achieved high ESMO-MCBS scores. The inconsistency underscores the complexity of equating trial internal validity with downstream clinical impact in oncology approvals.
The authors conclude that recent FDA approvals of anticancer drugs frequently rely on phase III randomized trials that have methodological limitations and often provide limited clinical benefit. They recommend that integrating risk-of-bias assessments with clinical benefit evaluations could strengthen pre-approval review and post-approval oversight, improve patient communication about expected benefits, and inform value-based pricing decisions. The findings suggest a need for regulators and stakeholders to consider both trial quality and magnitude of benefit when judging new oncology therapies.
The abstract and article note that assessments depended on available information in publications and FDA review documents; the observed differences between source-specific assessments indicate that reporting completeness affects judgments. The full article includes a conflict-of-interest statement: one author (Dr. Kesselheim) has received personal fees and served as an expert witness in unrelated litigation; other authors declared no known competing interests that could have influenced the work. Further methodological details, trial-level lists, and supplementary analyses are reported in the full text of the original publication.
Overall, this cross-sectional analysis raises concerns about the methodological rigor and clinical impact of many phase III trials that underpin recent FDA anticancer approvals and argues for combined assessments to better inform regulatory decisions, clinical practice, and health-policy discussions.