Date & Time: 5th January, 2011; 15.00-16.00

Venue: Ramanujan Hall

Title: Bivariate Semiparametric Bayesian Approach for Finding Minimum Effective Dose and Maximum Safe Dose

Speaker:Satrajit Roychoudhury, BDM Oncology, Novartis Pharmaceutical Company, Florham Park, NJ

Abstract: In early stage clinical trials the dosing range or therapeutic window is jointly determined by dose-efficacy curve and dose-safety curve. Minimum effec- tive dose(MED) refers to the minimum dose that has clinically significant effect over the placebo/control and maximum tolerated dose (MTD) refers to the maximum dose which is comparably safe with the placebo/active con- trol dose. In this paper a hierarchical semi-parametric bayesian approach is explored to find the MED and MTD simultaneously. Both monotonic and non-monotonic dose response relations are explored. The model described detection of MED and MTD event when the dose effect between different dose groups are heterogeneous. The model is illustrated with simulations and real dose-response data.

Key words: Minimum effective dose, Maximum tolerated dose, Hierarchical bayesian model, Dirichlet process.