Abstract
1731
Objectives Procedures for the measurement of atherosclerosis with 18FDG vary in published literature. Measurement variability and high contrast between target and background are key considerations for experimental and clinical use of a quantitative imaging biomarker. We aim to describe the characteristics of several methods to measure activity in the abdominal aorta.
Methods 16 clinical PET-CT exams (18-70yrs) scanned with a Biograph128 (256 matrix, iterative reconstruction with 2 iterations/24 subsets, 2.0 FWHM Gaussian filter, 3mm slice thickness) at 60(±8)min uptake time were included. Maximum SUV of the vessel wall on 5 consecutive slices centered at the aortic bifurcation were used to form target to background (TBR) ratios using the mean activity within regions of interest drawn within the inferior vena cava (IVC), aortic lumen (AL), and liver. With the exception of the liver, reference regions were from the same slice. Comparisons of TBR (mean±SD) used Wilcoxon matched pairs testing. A p-value of less than 0.05 was used to determine significance.
Results TBR for Liver, AL, and IVC regions were 1.11(0.15), 1.50(0.13), and 1.69(0.21). Wilcoxon testing demonstrated significant differences among all mean TBRs (maximal p<0.02).
Conclusions Using both high ratio contrast (high TBR) and low SD as criteria for a clinically and meaningful quantitative variable for 18FDG measurement, the use of the AL as a reference had the least variability. However, when the IVC is used as reference, IVC-TBR had the highest ratio but also the highest SD. The liver has the lowest TBR with a comparable SD compared to AL. A combination of these metrics may be useful to balance TBR and SD for experimental power calculations or clinical application.
Research Support Supported by the NIH intramural research program and the Imaging Sciences Training Program (ISTP).