Adaptive seamless designs with interim treatment selection: a case study in oncology

Area of Science:

• Oncology
• Clinical Trials
• Biostatistics

Background:

• Seamless phase II/III adaptive designs offer efficiency in oncology clinical trials.
• Early selection of experimental treatment regimens is crucial but challenging due to immature primary endpoints.

Purpose of the Study:

• To discuss the planning of oncology clinical trials using a seamless phase II/III adaptive design.
• To evaluate methods for selecting the most promising experimental treatment regimen at an interim analysis.
• To assess statistical approaches for hypothesis testing and error rate control in such designs.

Main Methods:

• Comparison of two experimental regimens against a control at an interim analysis.
• Incorporation of primary and surrogate endpoints for regimen selection.
• Assessment of statistical methods for power, type I error rate, and bias.
• Use of benchmark scenarios (perfect surrogate, no surrogate) to define operating characteristics.

Main Results:

• The study evaluates various statistical approaches for hypothesis testing, considering power and type I error rates.
• Benchmark scenarios provide a framework for understanding the expected performance of the adaptive design.
• The final analysis utilizes all collected data, both pre- and post-regimen selection.

Conclusions:

• Seamless adaptive designs require careful planning regarding interim analysis and endpoint utilization.
• Statistical methods must account for potential bias and control family-wise error rates for primary and secondary endpoints.
• The proposed framework aids in optimizing oncology trial design and interpretation.