Subgroup-Based Adaptive (SUBA) Designs for Multi-Arm Biomarker Trials

Area of Science:

• Oncology
• Biostatistics
• Clinical Trial Design

Background:

• Biomarker-driven targeted therapies are crucial in cancer research, but identifying responsive patient subgroups remains challenging.
• Current treatments often apply to broad patient populations, leading to unpredictable responses and limited efficacy for many.
• Effective targeted therapies for specific subgroups, like HER2+ breast cancer, are rare.

Purpose of the Study:

• To introduce Subgroup-Based Adaptive designs (SUBA) for simultaneously identifying prognostic subgroups and adaptively allocating patients to optimal treatments.
• To address the limitations of current targeted therapy approaches in clinical trials.

Main Methods:

• SUBA employs a random partition model for continuous patient subgroup reclassification.
• Adaptive patient allocation is based on posterior predictive probabilities to the best subgroup-specific treatment.
• The SUBA design was compared against equal randomization, outcome-adaptive randomization, and a probit regression design via simulation studies.

Main Results:

• Simulation studies indicated that SUBA performs favorably compared to alternative trial designs.
• SUBA demonstrated effectiveness in simultaneously identifying subgroups and optimizing treatment allocation.

Conclusions:

• Subgroup-Based Adaptive designs offer a promising approach to enhance the efficiency and effectiveness of targeted cancer therapy clinical trials.
• SUBA has the potential to improve patient outcomes by ensuring more precise treatment assignments based on identified prognostic subgroups.