Sample-Size Determination to Assess Diagnostic Accuracy Based on Area Under Receiver Operating Characteristic Curve Analysis

• Activity Number: 80
• Type: Contributed
• Date/Time: Sunday, August 4, 2013 : 4:00 PM to 5:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Ya-Hui Hsu*+
• Companies: AbbVie
• Keywords: sample size determination ; diagnostic tool ; sensitivity ; specificity ; area under curve

Abstract

Targeted therapies which optimize the treatment on certain patient population, for example, patient populations with some biomarker positive or some oncogenic overexpression/mutation, are important in oncology research and development. The identification of an appropriate patient population will be critical for the success of the targeted therapy. Sample size determination to assess diagnostic accuracy for a candidate diagnostic tool for patient selection is needed before conducting a clinical trial. Area under receiver operating characteristic curve (AUC) analysis is proposed and we compare this method to joint sensitivity and specificity analysis. These two methods depend on different parameter sets to obtain sample size estimates. We will provide an example using simulated preclinical data to demonstrate how to implement these two methods for a clinical trial in the relevant patient population.