Better Learning About the MTD in Phase I Oncology Trial: A Case Study Implementing Bayesian Design Based on Continuous Reassessment Method

• Activity Number: 18
• Type: Topic Contributed
• Date/Time: Sunday, August 1, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Inna Perevozskaya*+
• Companies: Pfizer Inc.
• Keywords: maximum tolerated dose ; phase 1 trials ; oncology ; continual reassessment method ; toxicity ; Bayesian methods

Abstract

We review a continual Reassessment Method (CRM) design implemented for Phase I trial in oncology, which is currently on-going. A modification of CRM method targeting pre-specified toxicity rate is used. Similar to the traditional 3+3 designs used in oncology Phase 1 trials, learning about drug's toxicity profile occurs in real time. However, this design is expected to greatly improve our ability to learn about the dose-toxicity profile and to identify the correct maximum tolerated dose because it utilizes all available toxicity information via Bayesian approach, including those toxicities that occur after the initial follow-up period; with appropriate choices of design parameters, the overall proportion of toxicities experienced can stay within acceptable limit. We will present operating characteristics of the design with some examples and discuss alternative design options considered.