Optimal designs for estimating the most successful dose.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacometrics

Background:

• Traditional Phase I clinical trials focus on identifying the maximum tolerated dose (MTD) based on toxicity.
• Emerging Phase I/II dose-finding designs integrate both toxicity and efficacy data.
• The goal is to identify the most successful dose (MSD), balancing safety and therapeutic response.

Purpose of the Study:

• To present an abstract, theoretical 'virtual' dose-finding design for Phase I/II studies.
• To establish a benchmark for optimal dose-finding by achieving the Cramer-Rao bound for toxicity and efficacy estimates.

Main Methods:

• Developed a 'virtual' design based on a bivariate response (toxicity and efficacy).
• The design theoretically achieves optimal properties, reaching the Cramer-Rao bound for dose-specific probability estimates.
• Utilized theoretical investigations and examples from existing literature for performance analysis.

Main Results:

• The proposed virtual design offers theoretical optimality for Phase I/II dose finding.
• While not practically implementable, it serves as a benchmark for evaluating real-world designs.
• Demonstrates how this theoretical framework provides insights into the relative performance of competing dose-finding strategies.

Conclusions:

• The virtual design provides a theoretical optimum for Phase I/II dose finding, integrating toxicity and efficacy.
• It serves as a valuable theoretical tool for assessing the potential for improvement in practical dose-finding designs.
• Facilitates deeper understanding and comparison of different clinical trial designs for dose optimization.