Abstract:
|
There has been a lot of discussion recently about design and endpoint selection in Phase II cancer clinical trials. Single arm designs have become less popular because of concerns about adequacy of historical controls, bias due to population selection, and recent availability of new classes of agents that lead to clinical benefit through mechanisms other than tumor shrinkage. Consequently, greater use of alternative endpoints and randomization has been encouraged. However, randomized designs often have prohibitively large sample size requirements. We present a randomized trial design which uses continuous tumor size change as the primary endpoint, and incorporates complete responses and deaths. It offers considerable sample size savings relative to the binary response endpoint. We demonstrate how to design such a trial and calculate sample size when preliminary data on tumor size distribution are not available, and show robustness to deviations from normality of tumor size change distribution. Based on the underlying tumor growth model, we show how tumor size changes are related to other common endpoints, and discuss the applicability and limitations of this approach.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.