Designs for Phase-II Window Studies

• Activity Number: 361
• Type: Contributed
• Date/Time: Wednesday, August 10, 2005 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Myron Chang*+ and Meenakshi Devidas and James Anderson
• Companies: University of Florida and University of Florida and University of Nebraska
• Keywords: likelihood ratio test ; two-stage design ; power function ; composite hypothesis ; binomial test

Abstract

Phase-II window cancer studies involve the administration of an investigational drug or combination to patients for a short period (4 to 12 weeks) prior to the administration of standard chemotherapy. This design has been used mostly in patient populations who have a high rate of failure on standard therapy. The primary goal is testing new agents or combinations in newly diagnosed patients who are less likely to have multiple-drug-resistant tumors and are potentially in better physiologic condition. This approach may reduce the chance of abandoning a clinical useful agent that might show negligible activity in heavily pretreated patients. As outlined in Smith (2001), the statistical considerations for Phase-II window studies should have early stopping rules for both insufficient activity and excessive early progression rates. In this presentation, we will propose the likelihood ratio test, which provides a natural tradeoff between response and early progression. The interesting features of the likelihood ratio test will be given. One- and two-stage tests will be considered and practical examples will be presented.