Phase I design for partially ordered groups with late-onset toxicity.

Area of Science:

• Clinical Trial Design
• Pharmacometrics
• Biostatistics

Background:

• Phase I clinical trials often face challenges with dose finding, especially with late-onset toxicities.
• Partially ordered groups, where frailty order is only partially known, add complexity to dose determination.
• Simultaneous consideration of partially ordered groups and late-onset toxicity in dose finding remains an unmet need.

Purpose of the Study:

• To develop and evaluate a novel dose-finding method for Phase I clinical trials.
• To address dose finding in partially ordered groups with late-onset toxicity.
• To improve the accuracy and safety of dose selection in heterogeneous patient groups.

Main Methods:

• Extension of the continual reassessment method (CRM) shift model framework.
• Incorporation of late-onset toxicity outcomes into dose-finding algorithms.
• Simulation studies comparing the proposed method with existing approaches.

Main Results:

• The proposed method demonstrates comparable performance to existing methods in correct dose selection.
• Extensive simulations confirm favorable operating characteristics for the new method.
• The approach effectively accounts for late-onset toxicities in dose recommendations.

Conclusions:

• The developed method successfully integrates partially ordered groups and late-onset toxicity for dose finding.
• The approach avoids problematic dose reversals, ensuring more sensitive groups receive appropriate doses.
• Favorable operating characteristics support the method's utility in clinical trial design.