A Rank-Based Endpoint Combining Tumor Assessment Information and Overall Survival in Randomized Phase II Trials When Phase III Is Based on Overall Survival

• Activity Number: 454
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Fei Ma*+ and Weichao Bao and Yunro Chung and William Leonard Mietlowski
• Companies: University of Rochester and Novartis and University of North Carolina at Chapel Hill and Novartis
• Keywords: randomized Phase II ; overall survival ; progression-free survival ; tumor burden ; new lesions ; weighted Wei-Lachin test

Abstract

Overall survival (OS) is considered to be the gold standard for evaluating anti-cancer treatments but OS may not be a feasible endpoint for randomized Phase II studies. Intermediate endpoints, such as progression-free survival (PFS) may be used instead. Four 2012 publications independently indicated that the components of the tumor assessment process, especially the occurrence of new lesions, are independent predictors of OS,. PFS also ignores post-progression OS information. For each tumor assessment, we create a score that combines information on the patient's survival status, time of assessment, type of progression and extent of change in tumor burden and apply a weighted version of the Wei-Lachin test to the repeated measures. We simulated randomized Phase II trials from randomized Phase III trials with known OS outcome and compared the proposed rank-based method to PFS with respect to sensitivity, specificity, positive predictive value and negative predictive value.