Visual analysis of [¹⁸F]FDG PET/CT imaging yields superior diagnostic accuracy compared to semiquantitative analysis in the assessment of giant cell arteritis.

Abstract

[¹⁸F]FDG PET/CT can be used to diagnose suspected new-onset giant cell arteritis (GCA), but it is unclear, whether visual or semiquantitative evaluation yields the highest accuracy. We therefore aimed to compare both methods with regard to diagnostic accuracy within the same patient. This single centre study included all ≥ 50 year-old patients undergoing [¹⁸F]FDG PET/CT between 2012 and 2023 for suspected new-onset GCA. Scans were re-read by two nuclear medicine physicians. Thirty-three arterial segments were scored visually, either with a standard criterion (activity at least equal liver, grade ≥ 2) or with a more specific one (activity above liver, grade 3). Standardized uptake values (SUVs) were measured in each segment, also normalized to liver and blood pool. For SUVs and ratios, ROC-curves were calculated per segment, segment groups and on the patient level to determine optimal cut-offs. The reference diagnosis was determined clinically by chart review after a follow-up of ≥ 6 months. A total of 163 patients, 39 (24%) with GCA, were included. Visual analysis with ≥ 1 segment with uptake grade ≥ 2 showed high sensitivity (80.0%, 95%-confidence interval (CI) 0.64-0.91) and specificity (83.0%, 95%-CI 0.75-0.89) and correct diagnosis in 82.0% (95%-CI 0.76-0.88). Using uptake grade 3, specificity was 96.0% (95%-CI 0.91-0.99) and sensitivity 59.0% (95%-CI 0.42-0.74), correct diagnosis in 87.0% (95%-CI 0.81-0.92). Semiquantitative methods proved less accurate, with SUVmax/SUVmax(liver) ratio at a threshold of 0.917 having the highest accuracy (sensitivity 87.2%, specificity 58.9%, correct diagnosis 65.6%). In suspected, new-onset GCA, expert visual interpretation of [¹⁸F]FDG PET/CT should be recommended, as it outperforms semiquantitative methods.