A Phase I Trial Design for Incorporating Efficacy Outcomes That are Conditional Upon Absence of Dose-Limiting Toxicities

• Activity Number: 404
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Shan Kang*+ and Thomas M Braun and Jeremy Michael George Taylor
• Companies: University of Michigan and University of Michigan and University of Michigan
• Keywords: Adaptive Design ; Phase I/II trial ; Dose-Finding Study

Abstract

We propose a Phase I trial design in which there are three possible outcomes for each patient: dose-limiting toxicity (DLT), absence of therapeutic response without DLT, and presence of therapeutic response without DLT. We define the latter outcome as a 'success.' The goal of the trial is to find the most successful dose (MSD), the dose with the largest probability of success. We propose a design that accumulates information on patients with regard to both DLT and response conditional on no DLT. Bayesian methods are used to update the estimates of DLT and response probabilities when each patient is enrolled, and use these methods to determine the dose level assigned to each patient. Due to the need to explore doses more fully, each patient is not necessarily assigned the current estimate of the MSD; our algorithm instead will assign a dose that is in a neighborhood of the current MSD. We examine the ability of our design to correctly identify the MSD in a variety of settings via simulation and compare the performance of our design to that of a competing approach.