Randomized phase III clinical trial designs for targeted agents

Area of Science:

• Clinical trial design
• Biostatistics
• Oncology

Background:

• Development of novel cancer therapies with distinct mechanisms of action from conventional chemotherapy.
• Need for robust clinical trial designs to evaluate targeted agents, especially in specific patient subgroups.
• Importance of tumor markers in identifying patient populations likely to benefit from targeted treatments.

Purpose of the Study:

• To investigate the performance of various phase III clinical trial designs for targeted cancer therapies.
• To compare designs for testing overall efficacy versus efficacy in marker-positive subgroups.
• To assess the trade-off between study size and treatment effectiveness in marker-positive patients under different scenarios.

Main Methods:

• Simulation studies were employed to evaluate sample size and statistical power for different trial designs.
• Binary outcomes were analyzed, considering continuous markers and various underlying scenarios.
• Marker prevalence and misclassification effects on power and sample size were simulated.

Main Results:

• Targeted designs, randomizing patients with the specific marker, demonstrated superior performance when a true predictive marker exists.
• Randomizing all patients, irrespective of marker status, often performed comparably or better than marker-based randomization strategies.
• The performance of different designs was evaluated against the standard of care in simulated clinical trials.

Conclusions:

• Randomizing all patients is recommended if the new treatment may benefit marker-negative patients or if marker status is uncertain.
• Designs allowing testing of both overall and targeted subgroup hypotheses are advisable in such cases.
• Careful consideration of marker characteristics and potential benefits across patient groups is crucial for optimal trial design.