Abstract:
|
Positron Emission Tomography (PET) is a functional imaging technique to study metabolic activity or receptor density. Assessment of test-retest variability of the PET tracers is an indispensable part of tracer validation and it is usually performed in a small number of subjects. In our presentation, we will review test-retest metrics that naturally ensue from random effects Analysis of Variance (ANOVA) model including intraclass correlation coefficient (ICC), repeatability coefficient (RC) and within subject coefficient of variation (WSCV). We will discuss their relative merits and small sample size performance. Although the ICC is frequently reported in PET studies, reporting of the other ANOVA-based metrics is usually lacking. On the other hand, percentage test-retest difference (PTRT), a measure of test-retest variability unique to PET imaging, is typically evaluated. We will relate the PTRT to ICC, RC and WSCV and provide recommendations on their use in test-retest PET imaging experiments. We will also illustrate the utility of the test-retest metrics in applications of PET tracers in receptor occupancy studies as well as in beta amyloid tracers in Alzheimer's disease.
|