A Bayesian three-tier quantitative decision-making framework for single arm studies in early phase oncology.

Area of Science:

• Oncology Drug Development
• Clinical Trial Design
• Biostatistics

Background:

• Single-arm proof-of-concept (POC) studies are crucial for early-phase oncology drug development decisions.
• These studies often involve small sample sizes and surrogate endpoints, making decision-making challenging, especially with competitive immunotherapy development.
• Expediting promising oncology drug candidates is a key objective.

Purpose of the Study:

• To propose a novel Bayesian three-tier approach to facilitate decision-making in early oncology drug development.
• To formally incorporate the option of accelerating promising drug programs.
• To introduce a Bayesian adaptive three-tier (BAT) design for continuous study monitoring and adaptive decision thresholds.

Main Methods:

• Development of a Bayesian three-tier decision framework with pre-specified criteria.
• Extension to a Bayesian adaptive three-tier (BAT) design allowing adaptive thresholds and continuous monitoring.
• Performance comparison of proposed methods against existing approaches via simulation studies.

Main Results:

• The proposed Bayesian approach provides a structured framework for making critical early-phase drug development decisions.
• The Bayesian adaptive three-tier (BAT) design enhances flexibility through continuous monitoring and adaptive thresholds.
• Simulations demonstrate the performance of the proposed methods in guiding decisions on program termination, further investigation, or acceleration.

Conclusions:

• The proposed Bayesian three-tier approach and BAT design offer a robust methodology for optimizing early-phase oncology drug development.
• These methods support data-driven decisions, balancing the need for rigorous evaluation with the imperative to expedite promising immunotherapies.
• The framework facilitates informed go/no-go decisions, contributing to more efficient drug development pipelines.