Design, Implementation, and Results for a Bayesian Adaptive Randomization Trial for Targeted Therapy in Lung Cancer

• Activity Number: 360
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Suyu Liu*+ and J. Jack Lee
• Companies: MD Anderson Cancer Center and MD Anderson Cancer Center
• Keywords: Bayesian adaptive design ; clinical trial ; hierachical Bayes model

Abstract

The BATTLE trial applies response adaptive randomization (AR) to identify promising treatment agents for lung cancer patients. A hierarchical Bayes model is used to characterize disease control rates in four treatment arms by incorporating biomarkers. After an equal randomization phase, patients were adaptively randomized to treatments based on the posterior disease control rates. A web-based application has been developed to conduct this trial. We have successfully randomized 255 patients, characterized the treatment efficacy and identified the associated predictive markers. The adaptive randomization, however, did not work as well as we planned. We will report the lessons learned from running this innovative trials for personalizing medicine. Practical considerations for running a successful Bayesian adaptive trial will be given.