Adaptive promising zone design for cancer immunotherapy with heterogeneous delayed treatment effect

Area of Science:

• Clinical Trials
• Biostatistics
• Cancer Immunotherapy

Background:

• Cancer immunotherapies often have delayed effects, varying between patients.
• Traditional log-rank tests in trials can lose power and complicate interim analyses for adaptive designs.

Purpose of the Study:

• To propose an adaptive promising zone design for cancer immunotherapy trials accounting for heterogeneous delayed effects.
• To provide methods for conditional power calculation and critical value adjustment for interim analyses.

Main Methods:

• Utilized a piecewise proportional hazard model with random lag time to model patient survival.
• Developed an adaptive promising zone design with three zones (unfavourable, promising, favourable) for sample size adjustment.
• Implemented critical value adjustment for the log-rank test at interim analyses.

Main Results:

• The proposed adaptive design demonstrated greater overall power compared to fixed sample designs.
• The method showed similar statistical power to matched group sequential trials.
• Critical value adjustment effectively controlled type I error rate inflation.

Conclusions:

• The adaptive promising zone design is a powerful and efficient approach for cancer immunotherapy trials with delayed effects.
• The proposed methods offer robust solutions for interim decision-making and sample size adjustments.
• Recommendations are provided for implementing this method in future clinical trials.