Dose Optimization Design for Randomized Phase II Trials With Toxicity and Efficacy Endpoints

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Pharmacology

Background:

• Existing optimal biological dose (OBD) identification methods for early-phase anticancer drug trials, such as the multiple-dose randomized phase II trial (MERIT) design, face limitations in controlling statistical error rates.
• These limitations stem from how the null hypothesis is defined, potentially compromising the reliability of OBD selection for molecular-targeted therapies and immune checkpoint inhibitors.

Purpose of the Study:

• To develop and validate a novel clinical trial design that ensures strict type I error control for identifying the optimal biological dose (OBD).
• To address the limitations of existing methods in consistently controlling statistical error rates across the entire null space for toxicity and efficacy.

Main Methods:

• A new design was developed, treating toxicity and efficacy as co-primary endpoints to ensure robust dose-level type I error control.
• Bonferroni correction was applied to manage multiplicity in dose selection, achieving overall type I error control.
• The rejection region is determined analytically, reducing computational burden compared to simulation-based methods.

Main Results:

• The proposed design maintains the type I error rate below the significance level across the entire null space, unlike existing methods.
• The analytical determination of the rejection region offers computational efficiency.
• While conservative due to the co-primary endpoint framework and Bonferroni adjustment, the method ensures statistical rigor.

Conclusions:

• The developed co-primary endpoint design provides a statistically sound framework for OBD selection in phase I/II clinical trials.
• This approach enhances the reliability of identifying optimal biological doses for novel anticancer agents.
• Future research may explore optimizations to balance statistical power and sample size requirements.