Duplex ultrasound may lead to misclassification of carotid artery stenosis.

OBJECTIVE: Carotid duplex ultrasonography (CDUS) is often the first-line imaging modality for clinical evaluation of carotid disease. Despite improved equipment and standardization of vascular imaging protocols, CDUS can be technician-dependent and unreliable in calcified blood vessels. The aim of this study was to assess variability in carotid stenosis severity measured using CDUS compared with reference high-resolution magnetic resonance imaging (MRI).

METHODS: In this prospective cohort study designed to study long-term brain health outcomes, we enrolled 123 participants with a diagnosis of asymptomatic carotid disease and/or ≥2 cardiovascular risk factors between the ages of 50 to 85 years from vascular and cardiology clinics in Tucson, Arizona, for the ongoing Carotids and Minds (CAM) study. Using a specialized radiofrequency coil for carotid imaging, high-resolution time-of-flight (0.63 × 0.63 × 1.0 mm3) and T1-weighted (in plane resolution = 0.63 × 0.63 mm2; slice thickness = 2.0 mm) MRIs were acquired to evaluate carotid stenosis severity. A neuroradiologist, vascular surgeon, and MRI specialist utilized cross-sectional measurements of carotid images to adjudicate stenosis percentage based on NASCET criteria. Agreement between percent carotid stenosis categories (<50%, 50%-69%, 70%-99%, 100%) derived from MRI and clinical CDUS was assessed using Cohen's Kappa with linear weights (κw). Sensitivity and specificity were calculated to evaluate how well CDUS identifies clinically significant stenosis defined by MRI.

RESULTS: A total of 246 carotid arteries were evaluated for percent stenosis using both MRI and clinical CDUS imaging. Twenty-five percent of the carotid arteries with severe (70%-99%) stenosis based on MRI were misclassified as having mild (<50%) or moderate (50%-69%) stenosis based on CDUS. Conversely, 33% of arteries with mild disease were overestimated by CDUS into higher stenosis categories. For detecting ≥70% carotid stenosis, clinical CDUS had a sensitivity of 0.77 (95% confidence interval, 0.67-0.87) and specificity of 0.92 (95% confidence interval, 0.88- 0.96), indicating a false negative rate of 23%.

CONCLUSIONS: CDUS provides a rapid, noninvasive assessment of carotid disease in clinical settings, yet in this cohort, it misclassified carotid stenosis severity (both over and under) 32% of the time. Thus, reliance on CDUS alone could lead to inappropriate treatment decisions. High-resolution MR angiography or computed tomography angiography should be considered for more accurate stratification of disease severity in patients with asymptomatic carotid disease.