1688O - Informative censoring of surrogate end-point data in FDA-approved cancer drugs

Background

The use of surrogate endpoints in oncology is widespread and increasing, but their association with survival is often weak. Since informative censoring can affect survival analysis, we sought to investigate the patterns of censoring in oncology randomised controlled trials (RCTs) FDA-approved based on surrogate endpoints and examine their association with overall survival.

Methods

In this cross-sectional study we reconstructed K-M curves of RCTs leading to FDA drug approval based solely on surrogate endpoints between 2010 and 2020. The reverse K-M method (i.e., events and censoring are flipped) was used to examine differential censoring with the analogous reverse hazard ratio and restricted mean survival time. Sensitivity analysis was performed by restoring the balance in censoring between study arms.

Results

Of the 65 eligible studies, 22 (34%) reported significant overall survival benefit (concordant trials), and 43 (66%) did not (discordant trials). The proportion of studies with significant differential censoring in surrogates was 51% (22/43) and 45% (10/22) in discordant and concordant trials, respectively. Although censoring imbalance occurred in both groups, after sensitivity analysis, 74% (32/43) of the discordant trials lost their statistical significance, compared to 9% (2/22) of concordant trials (OR = 29.09 95% CI 5.83 – 145.08, P = 0.0018).

Conclusions

Drugs approved based on surrogate endpoints frequently exhibit uneven dropout rates between study arms. Our analysis indicates that studies without survival benefit often fail to demonstrate improvement in surrogates after accounting for excessive censoring.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

T. Meirson: Financial Interests, Personal, Full or part-time Employment, I report receiving personal fees from Purple Biotech: Purple Biotech. G. Markel: Financial Interests, Personal, Financially compensated role: MSD, Roche, BMS, Novartis, 4C Biomed, Nucleai, Biond Biologics, Ella Therapeutics. All other authors have declared no conflicts of interest.