Bayesian Designs for Phase I-II Clinical Trials

Reliably optimizing a new treatment in humans is a critical first step in clinical evaluation since choosing a suboptimal dose or schedule may lead to failure in later trials. At the same time, if promising preclinical results do not translate into a real treatment advance, it is important to determine this quickly and terminate the clinical evaluation process to avoid wasting resources. Bayesian Designs for Phase I–II Clinical Trials describes how phase I–II designs can serve as a bridge or protective barrier between preclinical studies and large confirmatory clinical trials. It illustrates many of the severe drawbacks with conventional methods used for early-phase clinical trials and presents numerous Bayesian designs for human clinical trials of new experimental treatment regimes.Written by research leaders from the University of Texas MD Anderson Cancer Center, this book shows how Bayesian designs for early-phase clinical trials can explore, refine, and optimize new experimental treatments. It emphasizes the importance of basing decisions on both efficacy and toxicity.

Why Conduct Phase I-II Trials?The Conventional Paradigm The Continual Reassessment Method Problems with Conventional Dose-Finding Methods The Phase I-II Paradigm Efficacy and Toxicity Elements of Phase I-II Designs Treatment Regimes and Clinical Outcomes Sequentially Adaptive Decision Making Risk-Benefit Trade-Offs Stickiness and Adaptive Randomization Simulation as a Design Tool Establishing Priors Pathological Priors Prior Effective Sample Size Computing Priors from Elicited Values Efficacy-Toxicity Trade-Off–Based Designs General Structure Probability Model Admissibility Criteria Trade-off Contours Establishing a Prior Steps for Constructing a Design Illustration Sensitivity to Target ContoursSensitivity to Prior ESS Trinary Outcomes Time-to-Event Outcomes Designs with Late-Onset Outcomes A Common Logistical Problem Late-Onset Events as Missing Data Probability Model Imputation of Delayed Outcomes Illustration Utility-Based Designs Assigning Utilities to Outcomes Subjectivity of Utilities Utility-Based Sequential Decision Making Optimizing Radiation Dose for Brain Tumors Personalized Dose Finding The EffTox Design with Covariates Biomarker-Based Dose Finding Combination Trials Bivariate Binary Outcomes Bivariate Ordinal Outcomes Optimizing Molecularly Targeted Agents Features of Targeted Agents One Targeted Agent Combining Targeted and Cytotoxic Agents Combining Two Molecularly Targeted Agents Optimizing Doses in Two Cycles The Two-Cycle Problem A Two-Cycle Model Decision Criteria Illustration Simulation Study Optimizing Dose and Schedule Schedule Dependent Effects Trinary Outcomes Event Times Outcomes Dealing with Dropouts Dropouts and Missing Efficacy Probability Model Dose-Finding Algorithm Simulations Optimizing Intra-Arterial tPA Rapid Treatment of Stroke Probability Model Decision Criteria and Trial Conduct Priors Simulations Optimizing Sedative Dose in Preterm Infants Respiratory Distress Syndrome in Neonates Clinical Outcomes and Probability Model Prior and Likelihood Decision Criteria Simulations Bibliography

Ying Yuan is a professor and co-chief of the Section of Adaptive Clinical Trials in the Department of Biostatistics at the University of Texas MD Anderson Cancer Center. He is also an adjunct associate professor in the Department of Statistics at Rice University. Dr. Yuan has published over 100 peer-reviewed research papers in top statistical and medical journals. He is an associate editor of Biometrics and a board member of the International Chinese Statistical Association. He received his PhD in biostatistics from the University of Michigan. His research interests include Bayesian adaptive clinical trial design, statistical analysis of missing data, and Bayesian statistics.Hoang Q. Nguyen is a senior computational scientist in the Department of Biostatistics at the University of Texas MD Anderson Cancer Center. He received his PhD in computational and applied mathematics from Rice University. His research interests include Bayesian clinical trial design, computational algorithms, regression modeling, and Bayesian data analysis. Peter F. Thall is the Anise J. Sorrell Professor in the Department of Biostatistics at the University of Texas MD Anderson Cancer Center. He is also an adjunct professor in the Department of Statistics at Rice University. Dr. Thall is a fellow of the American Statistical Association (ASA) and the Society for Clinical Trials, an associate editor for Clinical Trials and Statistics in Biosciences, and an ASA Media Expert. He has published over 200 papers and book chapters in the statistical and medical literature. He received his PhD in statistics and probability from the Florida State University. His research interests include clinical trial design, dynamic treatment regimes, prior elicitation, Bayesian nonparametric statistics, and personalized medicine.