A Bayesian adaptive design with biomarkers for targeted therapies

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacogenomics

Background:

• Targeted therapies rely on biomarkers to identify patient subgroups.
• Traditional clinical trials face challenges in evaluating biomarker implications.
• Existing methods provide limited insights into biomarker predictive values.

Purpose of the Study:

• Develop novel clinical trial designs for targeted therapies.
• Effectively evaluate both diagnostic and therapeutic roles of biomarkers.
• Optimize patient stratification for personalized medicine.

Main Methods:

• Propose a Bayesian covariate-adjusted response-adaptive randomization design.
• Integrate biomarker profiles and clinical outcomes adaptively.
• Utilize partial least squares regression for predictive subgroup identification.

Main Results:

• Simulations demonstrate efficient identification of patients benefiting most from targeted therapy.
• The proposed design offers substantial sample size savings compared to alternative methods.
• Adaptive randomization optimizes treatment assignment based on emerging data.

Conclusions:

• The Bayesian adaptive design shows promise for early-phase targeted therapy development.
• Positive findings require confirmation via independent Phase III studies.
• This approach enhances the efficiency of biomarker-driven clinical trials.