Bayesian adaptive dose-escalation procedures for binary and continuous responses utilizing a gain function.

Area of Science:

• Clinical Pharmacology
• Biostatistics
• Drug Development

Background:

• Early phase clinical trials aim to determine safe and effective doses for further studies.
• Dose-limiting events (DLEs) and efficacy responses are crucial for dose-escalation procedures.
• Integrating both safety and efficacy data in dose-finding is challenging but essential.

Purpose of the Study:

• To describe and evaluate a Bayesian adaptive approach for dose escalation in early phase trials.
• To incorporate both binary DLEs and continuous efficacy responses into the dose-finding process.
• To assess the robustness and performance of the proposed method against existing approaches.

Main Methods:

• A Bayesian adaptive design utilizing logistic regression for DLEs and a linear log-log model for efficacy.
• A gain function guides dose selection for subsequent cohorts based on both safety and efficacy.
• Proposed stopping rules facilitate efficient decision-making during the trial.
• Simulations were conducted to compare the approach with DLE-only methods and assess robustness.

Main Results:

• The proposed Bayesian adaptive approach outperformed methods considering only DLE responses.
• The linear log-log model for efficacy provided robust dose recommendations, even when efficacy data followed an Emax model.
• The approach demonstrated comparable performance to alternative methods that integrate safety and efficacy data.
• The method proved to be a useful approximation for estimating Emax models.

Conclusions:

• The developed Bayesian adaptive approach effectively integrates safety and efficacy for dose escalation in early phase trials.
• The approach offers a robust and efficient method for dose-finding, improving upon DLE-only strategies.
• This method provides a valuable tool for optimizing clinical trial design, balancing therapeutic benefit with patient safety.