To evaluate the diagnostic accuracy of photon-counting detector CT angiography (PCD-CTA) for grading internal carotid artery (ICA) stenosis compared to color Doppler ultrasonography (CDUS). In this study, patients underwent both contrast-enhanced PCD-CTA and CDUS within six months. Four board-certified radiologists graded ICA stenosis using the North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria (grades 1-5) through visual and measurement-based methods from PCD-CTA. CDUS measurements served as the reference standard. Measurements were analyzed via Spearman's correlation and Fleiss's Kappa for inter-reader reliability. Image quality was assessed visually and through coefficient of variation measurements. A simulated intention-to-treat analysis compared that potentially require treatment (stenoses ≥ 70 % - grade 4 or higher) with stenosis < 70 %. A total of 31 patients (mean age 70.5 ± 10.2, 48.4 % female) and 60 ICAs were analyzed. CDUS identified 12/60 (20 %) cases of clinically significant stenosis. PCD-CTA stenosis measurements showed strong agreement with CDUS, with a mean difference of 5.85 % ± 9.8 % and a significant correlation (Spearman's rho = 0.951, p < 0.001). Inter-reader reliability was moderate (Fleiss's kappa values 0.56 for visual, 0.53 for measurement-based grading). In the intention-to-treat analysis, PCD-CTA visual assessments had 93.8 % sensitivity and 93.8 % specificity for detecting clinically significant stenosis, while quantitative CT measurement-based grading showed 75 % sensitivity and 96.9 % specificity. Misclassifications occurred in cases with severe calcifications. Image quality was excellent or good in 96 % of cases. PCD-CTA demonstrated high diagnostic accuracy with moderate inter-reader reliability for ICA stenosis grading, comparable to CDUS. However, misclassification due to calcified stenosis suggests the need for further improvements in imaging and analysis.
Rau et al. (Thu,) studied this question.