An Alternative Treatment Effect Measure for Time-to-Event Oncology Randomized Trials

Area of Science:

• Biostatistics
• Clinical Trials
• Survival Analysis

Background:

• Time-to-event endpoints like Overall Survival (OS) are crucial in phase III oncology trials.
• Current methods (Cox models, log-rank tests) rely on proportional hazards and large-sample assumptions.
• These standard methods can be unreliable with heavy censoring or non-proportional hazards.

Purpose of the Study:

• Introduce the univariate martingale residual (UMR) as a novel endpoint and summary measure.
• Develop an exact inference framework using randomization testing for UMR.
• Provide a robust alternative to traditional survival analysis in oncology trials.

Main Methods:

• The UMR quantifies the difference between observed and expected events per subject.
• Average UMRs per arm provide an absolute measure of excess events.
• A randomization-based testing framework computes exact p-values, bypassing proportional hazards and asymptotic assumptions.

Main Results:

• UMRs provided stable and interpretable summaries under heavy censoring and non-proportional hazards.
• The UMR-based randomization test maintained Type I error control.
• The UMR test was competitive or more powerful than the log-rank test when proportional hazards were violated.

Conclusions:

• The UMR offers an intuitive, assumption-free summary of treatment effects.
• UMR supports exact inference, crucial for reliable clinical trial results.
• UMR is a practical and robust alternative for phase III oncology trials, especially in complex survival situations.