An optimal two-stage exploratory basket trial design with aggregated futility analysis

Area of Science:

• Clinical Trials
• Biostatistics
• Pharmacology

Background:

• Basket trials evaluate one drug across multiple tumor types, often using individual interim futility analyses per indication.
• Variable enrollment rates can significantly delay overall trial decisions due to slow-enrolling indications.
• Early discontinuation of ineffective indications is crucial for efficient drug development.

Purpose of the Study:

• To propose an optimal two-stage basket trial design with an aggregated futility analysis.
• To accelerate futility decisions and enable earlier resource reallocation in exploratory basket trials.
• To maintain control over global type-I and type-II error rates.

Main Methods:

• An aggregated, two-stage futility analysis is proposed, pooling data across all indications.
• Sample size per indication is adaptive to enrollment rates, with a pre-specified total sample size for the futility analysis.
• Design parameters are optimized to minimize expected total sample size under the null hypothesis.
• The final analysis employs a pruning and pooling approach.

Main Results:

• The proposed aggregated futility analysis design allows for earlier futility decisions compared to individual analyses.
• Simulation studies indicate superior operating characteristics compared to existing designs.
• The design effectively controls global type-I and type-II error rates.

Conclusions:

• The proposed optimal two-stage basket trial design enhances efficiency in early-stage drug development.
• Aggregated futility analysis accelerates decision-making and resource management.
• This approach offers improved performance over traditional individual futility analysis methods.