Abstract W MP21: Diagnostic Accuracy of CT-Angiography for Internal Carotid Artery Stenosis - Comparison with a Revisited Multiparametric Ultrasound Approach

Abstract

Background: A novel multiparametric approach from the Neurosonology Research Group (NSRG) of the World Federation of Neurology has been proposed for duplex ultrasound (DUS) grading of internal carotid artery (ICA) stenosis. We aimed to determine the accuracy of CT-angiography (CTA) compared with the widely used NSRG ultrasound criteria for diagnosis of extracranial ICA stenosis. Methods: We studied consecutive acute ischemic stroke patients who were admitted to our tertiary stroke center from 01/2012 to 12/2012 and underwent CTA and DUS. On DUS, ICA stenosis and complete occlusion were graded according to the published NSRG multiparametric criteria. Axial CTA source images were independently assessed and NASCET-type measurements were manually performed for each extracranial ICA. In addition, each ICA was reviewed for complete occlusion. We calculated accuracy parameters of CTA for identification of 50-99% and 70-99% ICA stenosis, as well as complete ICA occlusion using DUS as reference. Results: A total of 339 patients (age 72±12yrs; 58% men; median NIHSS 4, interquartile range 8) provided 651 DUS and CTA measurement pairs for comparison. On DUS, 50-99% stenosis was found in 28, 70-99% stenosis in 16, and complete occlusion in 29 internal carotid arteries. The sensitivity of CTA against DUS was low for both detection of 50-99% and 70-99% ICA stenosis (64% and 31%, respectively). Further accuracy parameters of CTA were as follows: specificity 97%, PPV 49%, NPV 98% and overall accuracy 96% (18 true positive, 604 true negative, 19 false positive and 10 false negatives), and 99%, 56%, 98% and 98% (5 true positive, 631 true negative, 4 false positive and 11 false negatives), respectively. The highest accuracy of CTA was obtained for discriminating between complete ICA occlusion and non-occlusion: 100%, 96%, 91%, 100% and 99%. Conclusions: In comparison with the new and widely used NSRG criteria, CTA NASCET-type measurements seem to underestimate the degree of extracranial ICA stenosis. The validity of the NSRG criteria is currently evaluated in a prospective multicenter validation study utilizing invasive angiography as the gold standard.