ObjectiveThis study aimed to validate the efficiency of Doppler ultrasonography for predicting the innominate, subclavian, and common carotid artery stenosis.MethodsThis retrospective multicenter study between 2013 and 2022 enrolled 636 patients who underwent carotid Doppler ultrasonography and subsequent digital subtraction angiography. And 58 innominate artery stenosis, 147 common carotid artery stenosis, and 154 subclavian artery stenosis were included. The peak systolic velocity at innominate, subclavian, and common carotid artery, and velocity ratios of innominate artery to common carotid artery, innominate artery to subclavian artery, and common carotid artery to internal carotid artery were measured or calculated. The threshold values were determined using receiver operating characteristic analysis.ResultsThe threshold values of innominate artery stenosis were peak systolic velocity >206 cm/s (sensitivity: 82.8%; specificity: 91.4%) to predict ≥50% stenosis and >285 cm/s (sensitivity: 89.2%; specificity: 94.9%) to predict ≥70% stenosis. The threshold values of common carotid artery stenosis were peak systolic velocity >175 cm/s (sensitivity: 78.2%; specificity: 91.9%) to predict ≥50% stenosis and >255 cm/s (sensitivity: 87.1%; specificity: 87.2%) to predict ≥70% stenosis. The threshold values of subclavian artery stenosis were peak systolic velocity >200 cm/s (sensitivity: 68.2%; specificity: 84.4%) to predict ≥50% stenosis and >305 cm/s (sensitivity: 57.9%; specificity: 91.4%) to predict ≥70% stenosis.ConclusionsSymptomatic patients with ultrasonic parameters of velocity at innominate artery ≥206 cm/s, velocity at common carotid artery ≥175 cm/s, or velocity at subclavian artery ≥200 cm/s need to be considered for further verification and whether revascularization is necessary.