Clinical trial designs for testing biomarker-based personalized therapies

Area of Science:

• Clinical Trial Design
• Personalized Medicine
• Biomarker Discovery

Background:

• Molecular therapeutics have advanced personalized cancer treatment using biomarkers.
• Challenges persist in developing and validating biomarker-guided personalized therapies.
• Existing frequentist and Bayesian approaches have limitations.

Purpose of the Study:

• To develop a novel clinical trial design addressing limitations in biomarker-guided personalized therapy development.
• To integrate strengths of frequentist and Bayesian methods for improved trial design.
• To enhance the validation of promising biomarker-guided strategies.

Main Methods:

• Utilized generalized likelihood ratio tests for intersection null and enriched strategy null hypotheses.
• Derived a novel clinical trial design for advancing biomarker-guided strategies.
• Investigated adaptive randomization and futility stopping for trial efficiency.

Main Results:

• Simulation studies confirmed advantages of testing both enriched strategy and intersection null hypotheses.
• Adaptive randomization and early termination increased patient probability of receiving preferred treatments and improved response rates.
• Inference complexity and power reduction were noted for small-to-moderate sample sizes.

Conclusions:

• Innovative clinical trial designs are crucial for biomarker-based personalized therapies.
• The proposed design leverages likelihood inference and interim analysis to manage sample size and evolving technologies.
• The design aims to validate biomarker-guided strategies, acknowledging limitations in direct comparison to standard of care.