Adaptive Seamless Phase II/III Randomization Test Considering Treatment Group Selection Based on Short-Term Binary...

Area of Science:

• Clinical Trial Design
• Biostatistics
• Pharmaceutical Development

Background:

• Adaptive seamless phase II/III designs (ASDs) integrate early-phase dose-finding with late-phase confirmatory trials.
• Current ASD methods using combination tests can be overly conservative, especially with low outcome correlation.
• This conservatism can lead to unnecessarily strict Type I error control, impacting trial efficiency.

Purpose of the Study:

• To propose a novel adaptive seamless phase II/III randomization test.
• To address the conservatism of existing combination test methods in ASDs.
• To improve the statistical validity and power of adaptive seamless trials.

Main Methods:

• Developed a randomization test accounting for outcome correlation between phase II (binary) and phase III (overall survival) outcomes.
• Utilized simulation studies to evaluate the proposed method's performance.
• Compared the proposed method against traditional combination test approaches.

Main Results:

• The proposed randomization test effectively mitigated conservatism associated with outcome correlation.
• Type I error rates were controlled appropriately around the nominal level.
• The new method demonstrated higher statistical power compared to the combination test in most simulated scenarios.

Conclusions:

• The proposed adaptive seamless phase II/III randomization test offers a statistically robust alternative.
• This method enhances the probability of successful drug development trials.
• It provides a more efficient approach to adaptive seamless trial design.