Dose-Finding For Combination Trials Based On Toxicity Probability Intervals

• Activity Number: 350
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Fang Liu*+ and Zijiang Yang
• Companies: Octagon Research Solutions, Inc. and TechData Service Company, LLC
• Keywords: Adaptive design ; Dose-finding ; Bayesian design ; Combination trials ; Dose escalation

Abstract

When two drugs are combined in the drug development, the toxicity rate for the combo is expected to be higher than the individual drug and there could be multiple combination levels yielding the same toxicity rate, i.e. the MTD is not unique. It is crucial to identify the contour lines with the same toxicity rate and carry the most efficacious combo to later phase trials. This paper proposed a new dose-finding design for combination trials, which extended the design proposed by Ji's simple Bayesian design for mono-therapy. The new design is based on a beta/binomial Bayesian model and uses a zone-based escalation rule which can extend to all pre-specified dose combinations. The new design can be easily implemented by physicians without any input from statistician during the escalation process. Simulation studies indicate that the new design outperforms the traditional combo escalation design and competitive with the complicated CRM-based designs.