Nonparametric adaptive enrichment designs using categorical surrogate data.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Oncology Research

Background:

• Adaptive survival trials are crucial for enrichment designs in life-threatening diseases.
• Current methods for adaptive survival trials have limitations in type I error rate control, especially when using surrogate endpoints for interim decisions.

Purpose of the Study:

• To develop novel statistical approaches for adaptive survival trials that incorporate discrete surrogate endpoints and interim rejection boundaries.
• To address limitations of existing methods regarding data utilization, follow-up restrictions, and early hypothesis rejection.

Main Methods:

• The study proposes new methods based on weighted Kaplan-Meier estimates.
• These methods allow the use of discrete surrogate endpoints (e.g., tumor response) for interim analyses.
• The approaches are integrated into closed combination tests for adaptive enrichment designs.

Main Results:

• The proposed methods enable the use of correlated short-term endpoints alongside the primary survival endpoint.
• They account for nonproportional hazards, which are intrinsic to enrichment designs.
• The methods are robust against informative censoring, often caused by treatment modifications based on surrogate endpoint outcomes.

Conclusions:

• New statistical approaches using weighted Kaplan-Meier estimates provide robust type I error rate control in adaptive enrichment designs.
• These methods enhance flexibility by allowing interim decisions based on surrogate endpoints and addressing nonproportional hazards and informative censoring.
• The findings support more efficient and reliable clinical trial designs in oncology and other critical disease areas.