ABSTRACT
A phase III cancer clinical trial typically compares an experimental therapy to a standard therapy. It is widely acknowledged that the most appropriate way to compare treatments is through a randomized clinical trial. This chapter aims to extend the sample size calculation to the Weibull model. It illustrates the motivation to use the Weibull model for survival trial design with an example given by Cox and Oakes. When designing an actual trial with accrual time, calculating the sample size is usually impractical because not able to enroll the planned number of patients within the given accrual duration. The asymptotic distribution of the test statistic derived, allows one to calculate the required sample size for the trial design. The chapter proceeds with a parametric maximum likelihood estimates (MLE) test under the Weibull model because it is simple and its asymptotic distribution is easy to derive and hence the sample size can be calculated.
