Optimal two-stage designs for exploratory basket trials.

Area of Science:

• Oncology
• Clinical Trial Design
• Biostatistics

Background:

• Exploratory oncology trials aim to identify effective drugs for further development.
• Basket trials test a single drug across multiple tumor indications to mitigate selection errors.
• Existing designs may not optimally balance efficiency and statistical rigor in multi-indication settings.

Purpose of the Study:

• To propose optimal and minimax two-stage basket trial designs for exploratory oncology studies.
• To provide methods that explicitly control type I and type II error rates.
• To offer a framework extending single-arm trial designs to multi-arm basket trials.

Main Methods:

• Development of optimal and minimax two-stage basket trial designs.
• Incorporation of a stage 1 to prune inactive tumor indications.
• Stage 2 involves pooling active indications for overall drug effectiveness assessment.
• Derivation of closed-form sample size formulas.
• Simulation studies to compare performance against existing methods.

Main Results:

• The proposed designs effectively prune inactive tumor indications early in the trial.
• Active indications are pooled in stage 2 for robust effectiveness evaluation.
• The designs offer explicit control over type I and type II error rates.
• Simulation results demonstrate favorable operating characteristics compared to alternative methods.

Conclusions:

• The proposed two-stage basket trial designs provide an efficient and statistically sound approach for exploratory oncology.
• These designs represent a valuable extension of established single-arm trial methodologies to multi-arm settings.
• The methods facilitate better drug candidate selection in early-phase oncology research.