Practical and robust test for comparing binomial proportions in the randomized phase II setting.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Oncology Research

Background:

• Phase II oncology trials commonly use one-arm, non-randomized designs to evaluate response rates.
• The FDA's Fast Track Designation has increased the use of randomized phase II trials for stronger evidence.
• Existing randomized phase II designs for proportions have limitations in sample size, cost, duration, or operating characteristics.

Purpose of the Study:

• To propose a novel statistical test for comparing two binomial proportions in randomized phase II oncology trials.
• To address limitations of existing methods regarding sample size and operating characteristics.
• To provide an efficient and robust alternative for phase II trial designs.

Main Methods:

• Development of a modified test for comparing two binomial proportions, building upon the standard z-test and Jung's test.
• Numerical evaluations to contrast the proposed method with existing approaches.
• Application and validation of the new method using a real-world oncology clinical trial.

Main Results:

• The proposed test demonstrated improved efficiency and robustness compared to existing methods.
• Numerical evaluations confirmed the advantages of the new approach.
• Application to a real-world trial showed potential for significant cost and time savings, particularly with a high control arm rate.

Conclusions:

• The new test for comparing binomial proportions is an efficient and robust alternative for randomized phase II oncology trials.
• This method is especially beneficial when the control arm exhibits a high response rate.
• The proposed approach can optimize resource allocation and accelerate drug development timelines.