Generalized Pairwise Comparisons in Dose Optimization Oncology Trials: Beyond Safety to Multi-outcome Dose Selection

Area of Science:

• Clinical Pharmacology
• Biostatistics
• Oncology Drug Development

Background:

• Traditional dose-finding oncology trials (DFOTs) primarily rely on safety outcomes for dose selection.
• Modern DFOTs increasingly integrate safety, activity, and tolerability for optimal dose determination.
• Existing decision frameworks struggle with quantifying trade-offs due to variable clinical judgment.

Purpose of the Study:

• To introduce WIN-DOSE, a novel hierarchical, multi-outcome win ratio-based approach for dose optimization in DFOTs.
• To provide a statistical framework that mirrors clinical decision-making by evaluating treatment benefits across prioritized outcomes.
• To enhance transparency and clinical relevance in dose selection for early-phase oncology trials.

Main Methods:

• Utilized generalized pairwise comparisons (GPCs), specifically the win ratio (WR), to compare doses based on prioritized outcomes.
• Developed WIN-DOSE, a method where the optimal dose maximizes favorable patient-pair comparisons.
• Demonstrated WIN-DOSE in a two-arm randomized trial assessing dose-limiting toxicities (safety), preliminary response (activity), and dose intensity/patient-reported outcomes (tolerability).

Main Results:

• WIN-DOSE consistently identified the optimal dose when one dose was clearly superior.
• The win ratio framework effectively accommodated various trade-offs between safety, activity, and tolerability.
• The approach supports transparent and clinically relevant dose selection.

Conclusions:

• The win ratio offers a robust statistical method for transparent and clinically relevant dose selection in oncology.
• WIN-DOSE aligns with patient-centric decision-making goals in early-phase DFOTs.
• This method facilitates better integration of multiple outcome assessments for optimizing drug dosage.