Bayesian Adaptively Randomized Clinical Trial of Late-Stage, Non-Small Cell Lung Cancer

• Activity Number: 350
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Chunyan Cai*+ and Valen E. Johnson and Ying Yuan
• Companies: The University of Texas MD Anderson Cancer Center and The University of Texas MD Anderson Cancer Center and The University of Texas MD Anderson Cancer Center
• Keywords: Bayesian adaptive randomization ; Bayes factor ; Symptom interventions ; Factorial design

Abstract

Bayesian adaptively randomized clinical trials assign patients to treatments with probabilities that are calculated using the outcomes of previous patients. Recently, the use of Bayesian adaptive trials has increased due to their potential for increasing the number of patients assigned to efficacious treatments. In this proposal, we develop a Bayesian adaptive randomization design to test the efficacy of 16 combinations of 4 trial agents for reducing symptoms of late stage non-small cell lung cancer (NSCLC) patients. To obtain initial estimates of treatments effects, we assign the first 32 patients to treatments following a randomized factorial design; subsequent patients are assigned to treatments according to the posterior probability that each treatment is most efficacious. Bayes factors are used in the computation of posterior probabilities in the test-based design. Compared with equal randomization design, we show that our Bayesian adaptively randomized design assigns more patients to better treatments and efficiently identify the best treatment from multiple treatments.