A Bayesian decision-theoretic design for simultaneous biomarker-based subgroup selection and efficacy evaluation.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacogenomics

Background:

• Targeted therapy success relies on selecting patient populations based on biomarkers.
• Phase II trial planning often lacks a defined biomarker threshold for patient selection.
• Existing biomarker-guided designs struggle when thresholds are initially unavailable.

Purpose of the Study:

• To propose a two-stage clinical trial design for simultaneous biomarker threshold selection and efficacy evaluation.
• To accommodate targeted therapies effective in the entire patient population.
• To provide a flexible design for early-phase drug development.

Main Methods:

• A Bayesian decision-theoretic approach is utilized.
• A utility function incorporates study benefit and cost considerations.
• Simulations are employed to investigate operating characteristics under various scenarios.

Main Results:

• The proposed two-stage design effectively selects biomarker thresholds and evaluates efficacy.
• The design accommodates scenarios where the drug benefits all patients.
• Simulation results demonstrate the design's performance across different conditions.

Conclusions:

• The novel two-stage design offers a robust solution for biomarker-guided targeted therapy trials.
• It addresses the critical issue of unknown biomarker thresholds in early development.
• The Bayesian approach provides a framework for optimizing study design based on utility.