Overall survival (OS), defined as the time from treatment initiation until death from any cause, is widely considered the most meaningful endpoint in oncology clinical trials. It directly reflects the primary goal of most cancer therapies: helping patients live longer. However, OS data can be challenging to interpret due to factors such as long follow-up periods, treatment crossover, post-trial therapies, and continued treatment after disease progression.
To address these complexities, the U.S. Food and Drug Administration (FDA) released a draft guidance in August 2025 titled Approaches to Assessment of Overall Survival in Oncology Clinical Trials. This document provides recommendations for trial sponsors on how to appropriately plan, analyze, and interpret OS data in randomized oncology clinical trials to support marketing approval, especially when OS is not the primary endpoint.
Key Recommendations from the FDA Draft Guidance
Prespecification of OS Analyses
When OS is not the primary efficacy endpoint, the FDA recommends that it be pre-specified as a safety endpoint in the clinical protocol and the statistical analysis plan (SAP). This ensures that potential harm such as decrease in survival due to toxicity or treatment-related complications can be adequately assessed, even if efficacy is demonstrated using surrogate endpoints.
Interim Analyses and Safety Monitoring
For large, randomized trials, sponsors are encouraged to include interim analyses to assess futility or emerging safety concerns. These should be reviewed by an independent data monitoring committee to protect participants and minimize exposure to potentially ineffective or harmful treatments.
Intention-to-Treat (ITT) Analysis
The primary OS analysis should use the ITT principle, meaning all randomized participants are included in the analysis according to their assigned groups. ITT analysis helps preserve the integrity of the randomization process and reduces bias.
