Biostatistics Weekly Seminar

The novel mechanism of action of immunotherapy agents, in treatment of various types of cancer, poses unique challenges during the design and analysis of clinical trials. The presence of non-proportional hazards (NPH) treatment effect has been well documented in the context of immuno-oncology (IO) studies, where a delayed separation of the Kaplan-Meier (KM) curves has often been observed. Other more complex forms of NPH, such as crossing KM curves are also seen in practice, which makes analytical approaches and interpretation of study results, even more challenging. A cooperative effort with the Pharmaceutical Industry was initiated by the FDA in late 2016 to take a holistic approach in understanding the impact of NPH in the design, analysis and interpretation of clinical trials. In this talk, we provide an overview of some of the efforts and then proceed to discuss the recommendations of a combination test based on weighted logrank tests. Power and Type I error will be discussed along with realistic examples to illustrate the method in practice. We will also discuss clinical trial design considerations in the context of a new treatment with a potentially delayed treatment effect.

MRBIII, Room 1220
7 November 2018
1:30pm