Summary
A medRxiv analysis of 354 FDA oncology approvals from 2014 through early 2025 found surrogate endpoints had become the dominant primary measure used for approval. The authors also reported different relationships between surrogate-endpoint benefits and overall survival for immunotherapy and non-immunotherapy approvals.
An analysis of 354 US Food and Drug Administration (FDA) oncology drug approvals found that surrogate endpoints had become more common than overall survival as the primary endpoint supporting approval, according to a medRxiv preprint posted on October 24, 2026. The authors counted 143 immunotherapy and 211 non-immunotherapy approvals from 2014 through early 2025.
The researchers also compared approval timelines, trial sizes and how benefits measured through surrogate endpoints related to overall survival. Their findings highlight how the choice of endpoint can shape what evidence is available when a cancer drug is approved.
Approval timelines and trial sizes
The authors report that immunotherapies initially reached approval more quickly than non-immunotherapies, but the gap narrowed over time. For 2024/2025, they give approval durations of 2,079 days for immunotherapies and 2,357 days for non-immunotherapies. The corresponding median trial sizes were 461 and 698 participants. The abstract links the convergence in timelines largely to these trial-size differences.
An approval timeline is not the same as the time a patient spends in a trial: the abstract does not specify the timeline’s starting point or calculation. The figures therefore describe the authors’ comparison of approval development, rather than a general measure of how long cancer trials take.
Surrogate outcomes and overall survival
Overall survival measures how long patients live. Other endpoints—including progression-free, disease-free, recurrence-free and event-free survival—measure outcomes such as time before cancer worsens or returns, or before a defined event occurs. These measures can provide evidence about a treatment before an overall-survival difference is available, but they do not measure the same outcome.
The authors estimate that, for non-immunotherapy approvals, a median of 60% of the benefit indicated by these surrogate endpoints translated to overall survival. For immunotherapy approvals, they report a corresponding figure of 99%. These are the authors’ comparisons between endpoint benefits, not estimates of an individual patient’s chance of survival.
The study also compared FDA and European Medicines Agency (EMA) approval portfolios. Its abstract describes them as broadly similar overall, while characterising the EMA as more conservative in its approvals.
This analysis is posted as a preprint. The abstract does not explain how the authors calculated the 60% and 99% figures, so those estimates should be read as findings of this particular analysis rather than as a universal measure of how surrogate endpoints predict survival.