Sample Size Recalculation in Adaptive Group Sequential Study Designs for Comparing Restricted Mean Survival Times

Area of Science:

• Biostatistics
• Clinical Trial Design
• Survival Analysis

Background:

• Non-proportional hazards are common in time-to-event clinical trials, making hazard ratios potentially misleading.
• Restricted Mean Survival Time (RMST) is an increasingly promoted alternative endpoint for treatment effect comparison.
• Sample size calculation for RMST trials requires precise nuisance parameter estimates, posing planning challenges.

Purpose of the Study:

• To evaluate adaptive group sequential designs for clinical trials comparing RMST, particularly with delayed treatment effects.
• To compare the performance of adaptive RMST designs against other common endpoints in scenarios with delayed treatment effects.
• To provide a thorough explanation and methodology for adaptive RMST designs.

Main Methods:

• Extensive simulation study comparing adaptive designs with RMST as the primary endpoint.
• Focus on sample size adaptations within adaptive group sequential designs.
• Development and evaluation of a combination test applicable to various adaptations.

Main Results:

• Adaptive designs with RMST demonstrate robust performance in handling non-proportional hazards and delayed treatment effects.
• The proposed adaptive RMST design offers a viable solution to sample size uncertainties in trial planning.
• The combination test shows utility beyond sample size adaptations.

Conclusions:

• Adaptive group sequential designs utilizing RMST are effective for clinical trials with non-proportional hazards and delayed treatment effects.
• The methodology presented enhances the reliability and power of clinical trials by allowing sample size adaptations.
• The developed combination test offers flexibility for various adaptive trial strategies.