Quantification of internal carotid artery stenosis with duplex US:: Comparative analysis of different flow velocity criteria

PURPOSE: To compare 13 previously published sets of duplex ultrasonographic (US) criteria with the US criteria used at the authors' institution in terms of agreement with carotid artery angiographic results. MATERIALS AND METHODS: The authors studied 1,006 carotid arteries in 503 patients at duplex US and angiography. The degree of stenosis was determined by using duplex flow US velocities and applying 13 previously published sets of criteria and the criteria used at the authors' institution. Two independent observers evaluated the angiograms according to North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria. kappa statistics, sensitivities, specificities, positive predictive values (PPVs), negative predictive values (NPVs), and generalized linear mixed regression models were used to assess agreement between duplex US and angiographic findings. RESULTS: Stenoses of 0%-29%, 30%-49%, 50%-69%, 70%-99%, and 100% could be differentiated with 73% overall agreement between duplex US and angiographic findings according to flow velocity criteria (kappa = 0.57; 95% confidence interval [CI]: 0.54, 0.60); however, with duplex US, the angiographic degree of stenosis tended to be overestimated. In the differentiation of stenoses of less than 70%, only 45% agreement (kappa = 0.26; 95% CI: 0.23, 0.29) was observed, whereas in the differentiation of high-grade (greater than or equal to70%) stenoses, 96% agreement was observed (kappa = 0.85; 95% CI: 0.83, 0.87). The PPV and NPV for the identification of 70%-99% angiographic stenosis were 69% and 98% respectively with use of the most sensitive duplex US criteria. CONCLUSION: Duplex US is an excellent examination to screen for high-grade carotid artery stenosis; however, it tends to lead to an overestimation of the degree of stenosis. Exclusion of 70%-99% angiographic stenosis can be achieved with a sensitivity of up to 98%. (C) RSNA, 2004.