Biostatistics Seminar - Dr Peter Thall - Utility Based Dose-Finding with Ordinal Toxicity and Efficacy


Room 25


Phase I-II clinical trials combine the first two phases of conventional treatment evaluation by using both efficacy and toxicity to decide whether a new treatment is acceptably safe and promising and, if so, determine an optimal dose.  After preliminaries on commonly used phase I designs,  I will describe a new sequentially adaptive Bayesian phase I-II dose-finding design based on elicited utilities of a bivariate ordinal (toxicity, efficacy) outcome.  The methodology assumes parametric models for the marginal dose-toxicity and dose-efficacy distributions, and a copula is used to obtain a joint distribution. Using computer intensive methods, prior means are derived from elicited outcome probabilities, and prior variances are calibrated to control prior effective sample size and obtain a design with good frequentist operating characteristics. Subject to posterior dose acceptability criteria to control the risk of severe toxicity and exclude unpromising doses, patients are randomized adaptively among the doses having posterior mean utilities near the maximum. The method is illustrated by two phase I-II trials: (i) optimizing biologically equivalent dose of radiation therapy for children with brain stem gliomas and (ii) optimizing dose of a BRAF inhibitor, given with a fixed dose of tumor infiltrating lymphocytes, for advanced melanoma.


Peter Thall received his PhD in statistics and probability from The Florida State University Department of Statistics in 1975. He has worked at M.D. Anderson Cancer Center since 1990, where he holds the Anise J. Sorrell Professorship. Dr. Thall has pioneered application of Bayesian methods in medical research. He has designed over 300 clinical trials, published over 180 papers and book chapters, and presented 25 short courses and over 160 invited talks. He has served as an Associate Editor for Statistics in Medicine, Journal of the National Cancer Institute, and Biometrics. Currently, he is an Associate Editor for Clinical Trials and Statistics in Biosciences, an American Statistical Association Media Expert, a member of the International Biometric Society ENAR Regional Advisory Board, and is the JSM 2013 Program Chair for the ASA Section on Bayesian Statistical Science.  He is the Principal Investigator of the NCI RO1 grant “Statistical Methods for Complex Cancer Trials.”