DOD-Combo: Bayesian dose finding design in combination trials with meta-analytic-predictive prior

Area of Science:

• Oncology
• Biostatistics
• Clinical Trial Design

Background:

• Combination therapy is crucial in oncology for enhanced efficacy and overcoming resistance.
• Determining optimal doses for combination therapies is challenging due to complex drug interactions and toxicity.

Purpose of the Study:

• To introduce a novel Bayesian dose-finding design for combination therapies called DOD-Combo.
• To improve the efficiency and reduce sample sizes in identifying the maximum tolerated dose (MTD) for combination treatments.

Main Methods:

• The DOD-Combo design utilizes a Bayesian framework with information borrowing from historical single-agent trials.
• It employs a meta-analytic-predictive (MAP) power prior and a copula-type model to integrate historical data and model joint toxicity.

Main Results:

• Simulations demonstrate that DOD-Combo outperforms designs without information borrowing.
• The adaptive incorporation of historical data leads to reduced sample sizes and enhanced efficiency in MTD selection.

Conclusions:

• The DOD-Combo design offers a superior and more efficient approach to dose-finding for combination cancer therapies.
• This method effectively addresses the complexities of drug interactions and toxicity in clinical trials.