Model-based phase I designs incorporating toxicity and efficacy for single and dual agent drug combinations: methods...

Area of Science:

• Clinical pharmacology
• Biostatistics
• Drug development

Background:

• Novel therapeutic agents present unique challenges in drug development due to specific targets, unknown dose-efficacy relationships, and limited toxicity data.
• Traditional dose-finding methods may be insufficient for these novel agents, necessitating advanced statistical designs.
• Phase I clinical trials are critical for establishing safe and effective doses.

Purpose of the Study:

• To review model-based designs for identifying optimal doses of single or combination agents in Phase I trials.
• To explore the use of proportional odds and continuation ratio models for dose optimization.
• To address practical challenges in implementing these advanced designs in clinical practice.

Main Methods:

• Review of two model-based designs: proportional odds and continuation ratio models.
• Utilizing a continual reassessment method (CRM) with accumulated patient data for dose selection.
• Incorporating separate toxicity and efficacy curves for each drug in two-agent combination studies.

Main Results:

• Simulation studies indicate theoretical promise for model-based designs in identifying optimal doses.
• The proposed methods allow for flexible modeling of toxicity and efficacy for individual agents.
• Practical challenges to implementation were identified and potential solutions discussed.

Conclusions:

• Model-based designs offer a promising approach to optimize dosing for novel therapies in early-phase trials.
• Addressing implementation challenges is crucial for the routine adoption of these advanced statistical methods.
• Closer alignment between design development and clinical paradigms can facilitate quicker integration.