Optimal Designs with Futility Stopping for Diagnostic Imaging Studies

• Activity Number: 517
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Aida Aydemir*+ and Igor Zurbenko and Daniel Haverstock and Christoph Gerlinger
• Companies: Bayer HealthCare Pharmaceuticals Inc. and The State University of New York at Albany and Bayer HealthCare Pharmaceuticals Inc. and Bayer Schering Pharma AG
• Keywords: Futility ; Conditional power ; Futility treshold ; Interim analysis ; Current trend ; Confidence intervals

Abstract

For early phase drug development studies early stopping for futility is essential to minimize efforts on non performing products. For diagnostic imaging studies where specificity and sensitivity of the testing procedure have to be analyzed simultaneously this is even more crucial. We restrict our study to futility design because in the therapeutic area of DI, there is some uncertainty with respect to performance of the new diagnostic test which makes it difficult to determine when a study will be adequately powered. Our aim is to find the optimal futility threshold for which the trial will be declared futile based on conditional power (CP). We investigate key factors affecting CP of interim analysis with a single futility look. These factors include, assumed future trend of the data, timing of interim analysis, the threshold to be used in CP evaluation and overall type II error. These operating characteristics are evaluated in a large simulation study and described with graphical presentations. A practical application of a planned DI study is given. We select the most optimal plan for various values of timing and overall type II error, for various CP thresholds.