Estimation of Within Subject Coefficient of Variability (Wscv) in Clincial Positron Emission Tomography (Pet) Test-Retest Studies

• Activity Number: 157
• Type: Topic Contributed
• Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Richard Baumgartner*+ and Dai Feng and Aniket Joshi
• Companies: Merck and Merck and Merck
• Keywords: Test Retest ; Within Subject Coefficient of Variation ; Clinical Trial ; Positron Emission Tomography

Abstract

Positron Emission Tomography (PET) is an imaging modality that enables imaging of molecular targets such as neuroreceptors. PET tracers are radiolabeled molecules designed to bind to specific targets. Determining the baseline test-retest variability of the PET tracers is an essential part of the tracer validation process.

In these baseline test-retest studies, typically a small number of subjects (~5-10) is included. In our presentation, we will review metrics of test-retest, developed and commonly used in PET literature including the "percent variability" measure obtained from mean normalized differences. We will elucidate its relationship with the within-subject coefficient of variation (WSCV) based on random effects ANOVA.

We will present small sample size properties of the confidence intervals for WSCV as obtained in a simulation study for a range of typical WSCV values (< 10%) using delta method with and without variance stabilizing transformation (VST). We will also discuss possible ways to address slightly liberal coverage of the WSCV CI with VST.

Finally, we will introduce examples of PET test-retest data from clinical studies and demonstrate the utility of the WSCV.