Optimal three-stage designs for phase II cancer clinical trials

Area of Science:

• Clinical Trials
• Biostatistics
• Oncology

Background:

• Phase II cancer clinical trials aim to identify effective treatments by comparing response rates to current standards.
• Existing two-stage designs, like Simon's, allow early termination for ineffective treatments.
• Prior three-stage extensions had limitations on early rejection criteria and minimum sample size.

Purpose of the Study:

• To extend two-stage clinical trial designs to a more flexible three-stage framework.
• To develop optimal and minimax three-stage designs without prior restrictions on rejection regions or sample sizes.
• To provide a method for reducing expected sample size in early-phase cancer trials.

Main Methods:

• Development of a generalized three-stage clinical trial design.
• Tabulation of optimal and minimax designs for practical application.
• Comparative analysis of sample size reductions compared to two-stage designs.

Main Results:

• The proposed three-stage design offers flexibility by removing restrictions on the first stage rejection region and sample size.
• Tables for optimal and minimax designs are provided, facilitating implementation.
• An average reduction of 10% in expected sample size is observed compared to two-stage designs.

Conclusions:

• The three-stage design is advantageous for reducing sample size, particularly when treatment efficacy is uncertain or patient accrual is slow.
• This approach enhances the efficiency of phase II cancer clinical trials.
• The design provides a valuable alternative for optimizing resource allocation in clinical research.