American Statistical Association New York City Metropolitan Area Chapter Memorial Sloan-Kettering Cancer Center Biostatistics Seminar

Traditional designs for phase I clinical trials require each patient (or small group of patients) to be completely followed before the next patient or group is assigned. In situations where late toxicities are important, this may result in trials of impractically long duration. In the first half of the talk, I will introduce the time to event continual reassessment method (TITE-CRM), which allows patients to be entered in a staggered fashion. It is an extension of the continual reassessment method (CRM), a model-based method for phase I trials. We prove that the recommended dose given by the TITE-CRM converges to the correct level under certain conditions. In the second half, I will describe a technique to evaluate the sensitivity of the dose-toxicity model in the CRM. Although the technique is based on asymptotic arguments, we find that it is indicative of when the method will fail in small sample settings.