Combined criteria for dose optimisation in early phase clinical trials

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacometrics

Background:

• Traditional dose-finding methods may not efficiently balance efficacy and toxicity.
• Adaptive designs offer flexibility in clinical trials but require robust optimization criteria.

Purpose of the Study:

• To develop and evaluate combined optimality criteria for dose optimization in seamless phase I/II adaptive clinical trials.
• To investigate the connection between "best intention" and D-optimum designs in this context.

Main Methods:

• Introduction of combined criteria integrating probability of success and Fisher information matrix determinant.
• Incorporation of penalties for toxic or inefficacious doses.
• Adaptive dose selection for subsequent cohorts based on maximizing defined criteria, using a trinomial response model.

Main Results:

• The proposed adaptive design method accurately identifies optimal doses.
• Simulation studies demonstrate effective dose identification with minimal exposure to toxic doses.
• The approach provides a bridge between "best intention" and D-optimum design principles.

Conclusions:

• The developed combined criteria offer a robust approach for dose optimization in adaptive phase I/II trials.
• This methodology enhances patient safety by minimizing exposure to suboptimal doses.
• The findings support the use of adaptive designs with combined optimality criteria in early-phase drug development.