ABSTRACT
This chapter presents some recent development in the application of cure modeling in the statistical design and planning and hypothesis testing in cancer clinical trials. Because of the breakthrough in cancer treatment and dissemination of cancer screening, many types of cancer are curable, especially in early stages. The usual unweighted log-rank tests may not be efficient in the presence of cured patients. The optimal weighted log-rank tests and the associated sample size calculation are presented for the clinical trial with delayed onset of treatment effect and cure. Importantly, a delay in the separation of survival curves violates the fundamental assumption of proportional hazards for typical clinical trial design and analysis methods, which might reduce the statistical power of a study to differentiate between two treatment arms. The power and sample size calculation for immuno-oncology clinical trials need to deal with the delayed onset of treatment effects.
