Design and sample size determination for multiple-dose randomized phase II trials for dose optimization.

Area of Science:

• Clinical Trial Design
• Pharmacology
• Biostatistics

Background:

• The U.S. Food and Drug Administration (FDA) initiated Project Optimus to transition cancer drug dose selection from the maximum tolerated dose (MTD) to a dose optimizing the risk-benefit tradeoff.
• Current FDA guidance suggests randomized phase II trials comparing multiple doses to identify the optimal dose for maximizing patient benefit and minimizing risk.

Purpose of the Study:

• To present a novel methodology for determining sample size in multiple-dose randomized trials for dose optimization.
• To introduce the MERIT (Multiple-dosE RandomIzed Trial) design, a framework for selecting optimal doses in early-phase oncology trials.

Main Methods:

• Generalized standard definitions of type I error and power to suit dose optimization objectives.
• Developed a decision rule and algorithm for calculating optimal sample size in multiple-dose randomized trials.
• Introduced the MERIT design, incorporating toxicity and efficacy data for dose selection.

Main Results:

• Simulation studies confirmed MERIT design's desirable operating characteristics for dose optimization.
• A sample size of 20-40 participants per dosage arm typically provides adequate power and acceptable type I error rates.
• MERIT design facilitates patient safety and maximizes therapeutic benefit through optimized dosing strategies.

Conclusions:

• The MERIT design offers a statistically sound approach to sample size determination for dose optimization trials.
• This methodology supports Project Optimus by enabling the identification of optimal doses with favorable risk-benefit profiles.
• Software for implementing the MERIT design is available to aid clinical trial researchers.