Pharmaceutical companies frequently explore previously approved cancer drugs for new indications, which may expose trial participants to uncertain benefit and potential harm. Sorafenib has been investigated across numerous cancer types despite limited evidence supporting many off-label uses. This study evaluates sorafenib’s reported risk–benefit profile in clinical trials published during the past decade. To assess the portfolio of sorafenib trials by examining reported efficacy and safety outcomes over the last ten years. A literature search of bibliographic databases was conducted on May 25, 2023, to identify adult clinical trials evaluating sorafenib’s antitumor activity. Eligible studies were published within the prior decade and reported objective response criteria. Screening and data extraction were performed in duplicate. Extracted variables included trial characteristics, objective response rate (ORR), grade 3–5 adverse event (AE) rates, median progression-free survival, and median overall survival. Primary endpoints were categorized as positive, negative, or indeterminate based on author-defined criteria. Since sorafenib’s most recent FDA approval in 2013, investigators have examined its activity in more than 32 off-label indications. Across 154 trials, 12,226 participants were evaluable for AE grading, with 11,003 grade 3–5 AEs reported. Cumulative trends demonstrated persistently low ORR and increasing cumulative rates of severe AEs over time across off-label indications. This descriptive portfolio review highlights that, across published trials of sorafenib monotherapy in non-FDA-approved indications, reported objective responses were infrequent while severe AEs were common. These findings underscore the importance of transparent reporting and careful consideration of the existing evidence base when designing or interpreting future sorafenib trials in off-label settings. Key words: Sorafenib, Off-label indications, Risk-benefit analysis, Clinical trial portfolio, Adverse events
Corwin et al. (Fri,) studied this question.