# Interobserver Variability in Cardiovascular FDG PET/CT Analysis in Large Vessel Vasculitis

## TL;DR

This study evaluates how consistent different observers are in analyzing PET/CT scans of patients with large vessel vasculitis and mild to moderate COVID-19.

## Contribution

The study provides new evidence on the reproducibility of SUV measurements in vascular structures for diagnosing vasculitis.

## Key findings

- Excellent inter-observer agreement was found for qualitative assessments of aortic FDG uptake.
- Quantitative SUVmean measurements showed very high correlation between observers.
- SUVmax measurements demonstrated strong interobserver reproducibility.

## Abstract

PET/CT has a synergistic value for optimal diagnosis, disease activity monitoring, and evaluation of damage progression in large vessel vasculitis. The use of standardized uptake values (SUV) as a measurement of relative tissue uptake facilitates comparisons between patients, and has been suggested as a basis for diagnosis. The SUVmean and SUVmax reproducibility in vascular structures is not widely studied.

The objective of this study was to evaluate the inter-observer variability of both qualitative visual grading of aortic 18F-FDG uptake and the quantitative aortic mean and maximum SUVs in these patients with mild to moderate covid-19 infection who underwent multimodality cardiac imaging within the COSMIC-19 trial.

This is a sub-study of the COSMIC-19 trial. 30 patients were subjected to a combined Computed Tomography Coronary Angiogram and 18F-FDG PET/CT, followed by cardiac magnetic resonance. Two independent observers measured the Standardized uptake values in five regions of interest at each aortic segment. These were performed sequentially along the length of the aorta every 5 mm on the axial slices. The maximum and mean standard uptake values were measured.

Qualitative assessment showed excellent agreement between observer x and y for the ascending aorta and aortic arch regions with the kappa coefficients for the inter observer agreement of 0.92 (95% CI:0.78–1.0) and 0.91 (95% CI:0.74–1.0) respectively. Quantitative assessment showed a very high positive correlation between the two observers for each of the regions measured for SUVmean as follows; ascending aorta r = 0.96 (p < 0.001), Aortic arch r = 0.90 (p < 0.001) and descending Aorta r = 0.91 (p < 0.001). The correlation coefficients for the SUVmax were substantially strong.

This study shows an excellent inter-observer reproducibility for both qualitative and quantitative SUVmean vascular 18F-FDG measurements in patients with COVID-19 large vessel vasculitis. Quantitative SUVmax demonstrated substantially strong interobserver reproducibility.