Accuracy of the Society of Radiologists in Ultrasound (SRU) Carotid Doppler Velocity Criteria for Grading North American Symptomatic Carotid Endarterectomy Trial (NASCET) Stenosis: A Meta‐Analysis

PURPOSE: The Society of Radiologists in Ultrasound (SRU) consensus panel proposed six Doppler velocity cut points for classifying internal carotid artery (ICA) stenosis of 50% and 70% according to the North American Symptomatic Carotid Endarterectomy Trial (NASCET) method. Their relative accuracies have not been compared. MATERIALS AND METHODS: Meta-analysis performed following comprehensive literature review and identification of manuscripts with graphs of individual patient NASCET ICA stenosis measured by arteriography versus ICA peak-systolic velocity (PSV), end-diastolic velocity (EDV) and ICA PSV to common carotid artery (CCA) PSV. True positives, true negatives, false positives, and false negatives were calculated and used in two-level mixed effects models. Hierarchical summary receiver operating characteristic (ROC) curves were generated. Areas under the ROC curves were estimated. RESULTS: Nine studies performed between 1993 and 2016 were identified after review of 337 manuscripts. There were 1738 bifurcation data points extracted for PSV, 1026 for EDV, and 775 for ICA/CCA ratio. The highest sensitivity was 96% (95% CI: 93%, 98%) for PSV of 125 cm/s (50% stenosis) and highest specificity 86% (95% CI: 71%, 93%) for PSV of 230 cm/s (70% stenosis). Areas under the ROC curves ranged from a high of 0.93 (95% CI: 0.92, 0.95) for PSV (50% stenosis) to a low of 0.86 (95% CI: 0.84, 0.88) for EDV (70% stenosis). CONCLUSIONS: The SRU consensus Doppler cut points vary in their accuracies for predicting ICA stenosis. The PSV cut points have tradeoffs: high sensitivity/low specificity for 50% stenosis and high specificity/moderate sensitivity for 70% stenosis.