Bayesian Ordered Lattice Design for Phase I Clinical Trials.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacology

Background:

• Phase I clinical trials often have small sample sizes, limiting the effective use of prior toxicity data.
• Existing methods may not fully leverage the inherent order of toxicity probabilities across dose levels.
• Accurate identification of the maximum tolerable dose (MTD) is crucial for patient safety and drug development.

Purpose of the Study:

• To develop a novel Bayesian framework, Bayesian Ordered Lattice Design (BOLD), for early Phase I clinical trials.
• To enhance dose selection, toxicity monitoring, early stopping, and MTD identification by incorporating prior information and dose-level ordering.
• To improve the accuracy of MTD determination compared to existing popular methods.

Main Methods:

• Development of a Bayesian methodology that incorporates prior information and posterior updating.
• Utilization of the natural ordering of toxicity probabilities across different dose levels.
• Application of straightforward dose-level Bayesian specifications and clinically interpretable posterior probabilities for decision-making.

Main Results:

• The proposed BOLD framework effectively guides dose selection and toxicity monitoring.
• BOLD leverages data from other dose levels by utilizing their order relationship for enhanced analysis.
• BOLD demonstrates superior accuracy in determining the MTD compared to popular existing methods.
• The approach is computationally simple and avoids the need for simulation.

Conclusions:

• The Bayesian Ordered Lattice Design (BOLD) offers a robust and accurate framework for Phase I clinical trials.
• BOLD enhances decision-making in dose-finding studies by integrating prior knowledge and dose-response relationships.
• This computationally efficient Bayesian approach provides a valuable alternative for identifying the MTD in early-phase drug development.