Einstein Circle Lecture | Miquel Serra-Burriel (University of Zurich)

Talk

Control Groups, Censoring, and Progression Outcomes in Cancer Trials

A key aim of oncology trials is to estimate whether treatments delay cancer progression. Progression-free survival, however, depends on determining when progression has occurred, which is often uncertain in practice. When such uncertainty leads to censoring that depends on the patient’s true progression status, the standard assumption of non-informative censoring breaks down and effect estimates may become biased.

In this lecture, Miquel Serra-Burriel will present recent work in which publicly available trial curves and censoring patterns were examined to assess how reported treatment effects varied by endpoint, control-group design, and censoring imbalance across arms. He will show how estimated effects on progression-free survival differed from those on overall survival and how these differences relate to imbalances in censoring patterns, taking a particular look at placebo-blinded and open-label trials.

He will discuss implications of his work for reported progression-free survival benefits in published oncology trials and how individual patient data re-analyses may shed light on whether apparent treatment effects on progression-free survival reflect true benefit or bias introduced by trial conduct and analysis.

The lecture will be followed by a lively discussion, moderated by Jessica Rohmann, Head of the Responsible Health Data Analysis Team at the BIH QUEST Center for Responsible Research.