Department of Biostatistics Seminar/Workshop Series
Phase I clinical trials in VICC: statistical issues in dose-finding studies
Tatsuki Koyama, PhD
Associate Professor Department of Biostatistics and Center for Quantitative Sciences, Vanderbilt University School of MedicineThe Scientific Review Committee in the Vanderbilt-Ingram Cancer Center (VICC SRC) meets biweekly to review clinical trial protocols to ensure that cancer clinical trials conducted within the institution meet peer-review standards of scientific design. Biostatisticians in the committee are responsible for validity of design and the statistical analysis plan. Most of the VICC phase I clinical trials contain a dose-finding component, and reviewing these trials has been a frustrating experience. The most common dose-finding design is the 3+3 design, and its operational characteristics are known to be suboptimal; however, the NCI seems to advocate its usage, and the VICC SRC has always approved protocols proposing a 3+3 design. In this talk, I will review the 3+3 design, along with two other dose-finding designs, the continual reassessment method, and modified toxicity probability interval design (mTPI). I will present a case study where the proposed mTPI design was met with resistance from the NCI and subsequently changed to a 3+3 design.The talk will conclude with a brief discussion of large sample expansion cohorts, an increasingly common design feature that may further jeopardize the validity of dose-finding phase I clinical trials.