Bayesian Dose Finding for Combined Drugs with Discrete and Continuous Doses

Area of Science:

• Oncology
• Biostatistics
• Clinical Trial Design

Background:

• Cancer clinical trials increasingly use drug combinations to enhance patient response.
• Evaluating drug synergism is a key goal in combination therapy studies.
• Standard of care (SOC) agents are often combined with new therapeutics, requiring careful dose management.

Purpose of the Study:

• To propose a novel two-stage Bayesian adaptive dose-finding design for cancer trials.
• To optimize the dosing of a discrete-dose agent and a continuous-dose SOC agent.
• To effectively manage toxicity and improve patient outcomes in combination drug trials.

Main Methods:

• A two-stage Bayesian adaptive design is proposed for trials with continuous and discrete dose agents.
• Stage 1 uses a continual reassessment method to find the optimal discrete dose, fixing the continuous SOC dose.
• Stage 2 fine-tunes the continuous SOC dose to meet target toxicity rates, using adaptive dose escalation/de-escalation based on toxicity probabilities.

Main Results:

• The proposed design adaptively assigns patients to optimal dose combinations as toxicity data accumulate.
• Extensive simulation studies demonstrate the design's good performance in practical scenarios.
• The design effectively balances dose finding and toxicity management for combined cancer therapies.

Conclusions:

• The two-stage Bayesian adaptive design offers an effective strategy for optimizing drug combinations in cancer clinical trials.
• This adaptive approach allows for flexible and efficient dose adjustments throughout the trial.
• The design shows promise for improving therapeutic efficacy and patient safety in oncology drug development.