There Is More to Length of Survival Than Survival: Endpoints in Oncology Trials.

Progression-free survival as a primary endpoint for comparative trials does not fully capture the therapeutic risk/benefit ratio. Additionally, summarization of the treatment effect via a hazard ratio is problematic when the proportional hazards assumption is violated. Restricted mean survival time metrics may address these challenges but have other limitations. See related article by Chang et al., p. 3282.