Multi-stage enrichment and basket trial designs with population selection

Area of Science:

• Clinical trial methodology
• Biostatistics
• Pharmaceutical research

Background:

• Early-phase biomarker data in clinical trials often exhibit high variability, impacting reliability.
• Two-stage trial designs are logical for late-phase confirmatory trials, refining populations before hypothesis testing.
• Existing literature on multi-stage designs, especially for biomarker enrichment and basket designs with interim analyses, is limited.

Purpose of the Study:

• To expand on two-stage confirmatory trial designs by incorporating multiple stages for ongoing efficacy monitoring.
• To address the limited literature on multi-stage designs, specifically focusing on biomarker enrichment and basket designs.
• To provide methods for implementing interim efficacy analyses in these designs while maintaining statistical integrity.

Main Methods:

• Focuses on two popular designs: biomarker enrichment and basket designs.
• Discusses the implementation of interim efficacy analyses following population refinement.
• Explores statistical aspects including type I error control, power, and stopping probabilities.

Main Results:

• Provides a framework for conducting interim efficacy analyses in multi-stage biomarker enrichment and basket trials.
• Demonstrates methods to control Type I error rates throughout the trial.
• Evaluates the impact of these designs on statistical power and the probability of early stopping.

Conclusions:

• Multi-stage designs offer a structured approach to refining patient populations and monitoring efficacy in late-phase trials.
• The proposed methods enhance trial efficiency and statistical validity for biomarker enrichment and basket designs.
• Further research into optimizing these multi-stage designs is warranted for improved clinical trial efficiency.