Determining doses for backfill cohorts based on patient-reported outcome

Area of Science:

• Clinical Trials Methodology
• Oncology Drug Development
• Biostatistics

Background:

• Phase I oncology trials utilize backfill cohorts for dose optimization.
• Current methods face delays in identifying ineffective doses, leading to patient exposure.
• There's a need to integrate patient-reported outcomes (PRO) for better dose determination.

Purpose of the Study:

• To propose a novel Bayesian design framework, 'Backfill-QoL', for phase I oncology trials.
• To incorporate patient-reported quality of life (QoL) data into dose-finding and selection processes.
• To evaluate the efficiency and patient-centeredness of the proposed design.

Main Methods:

• Developed a unified Bayesian design framework ('Backfill-QoL').
• Integrated methods for trial monitoring, dose-finding algorithms, and dose selection criteria.
• Conducted simulation studies and sensitivity analyses to validate the design.

Main Results:

• The Backfill-QoL design demonstrated greater efficiency compared to traditional strategies.
• Fewer patients were assigned to doses with unacceptable QoL profiles.
• A user-friendly application for implementing the design is available.

Conclusions:

• The Backfill-QoL design facilitates continuous monitoring of safety, efficacy, and QoL.
• It enables a more patient-centered approach to identifying the recommended phase II dose (RP2D).
• This framework improves dose optimization in early-phase oncology trials.