European Journal of Radiology. 10 (1990) 147-153 Elsevier 147 EURRAD 00019 Carotid artery stenosis - reassessment of the Doppler criteria G.G. Hartnell*, R.N. Gibson **, R. Williams + and A.P. Hemingway+ + Department of Diagnostic Radiology. Royal Postgraduate Medical School, Hammersmith Hospital, London, U.K. (Received 20 May 1989; revised version received 25 September 1989; accepted 12 October 1989) Key words: Carotid arteries, disease; Carotid arteries, ultrasound study; Carotid arteries, flow study; Carotid arteries, digital subtraction angiography Abstract Few previous studies of the accuracy of duplex Doppler in the assessment of carotid artery stenosis have looked at the accuracy in excluding disease in populations with a significant number of normal arteries. In addition, few studies have compared different criteria for stenosis. Duplex Doppler was used to evaluate the carotid arteries in 80 patients. Two well-described methods for analysing Doppler traces were used and the results were compared with those obtained from non-selective digital subtraction angiography. On the basis of these findings, the Doppler criteria used for carotid stenosis were redefined in terms of changes in peak velocity rather than Doppler frequency shift. Using these criteria, the sensitivity of Doppler in identifying carotid stenosis was 97% (98 % sensitivity for > 50% stenosis) and specificity was increased from 84% to 93%. Introduction Doppler ultrasound is an important, safe method for examining the extracranial portions of the carotid arteries. In specialist centres, the reported accuracy rates of duplex Doppler [ 1,2] are better than those reported for real-time ultrasound imaging alone in the assessment of carotid stenosis [3], although ultrasound imaging is superior in the demonstration of non- stenosing ulceration [4,5]. To fulfil its role as a primary diagnostic procedure, it is essential that the non- invasive assessment of carotid artery disease by duplex Present addresses: * ** + ++ Department of Radiology, Bristol Royal Infirmary, Upper Maudlin Street, Bristol, Avon BS2 8HW, U.K. Department of Radiology, Royal Melbourne Hospital, Parkville, Victoria 3050, Australia. Department of Radiology, London Bridge Hospital, London SEl, U.K. Academic Radiology Department, Royal Hallamshire Hospital, Sheffield SlO, U.K. Address for reprints: G.G. Hartnell, M.D., Department of Radi- ology, Bristol Royal Infirmary, Upper Maudlin Street, Bristol, Avon BS2 8HW, U.K. Doppler is as accurate as possible, to ensure that those without serious disease are not exposed to the risks of arteriography, while those with serious disease are not denied the possible benefits of treatment. Although non-selective digital subtraction angio- graphy (DSA) produces adequate results in almost all patients [6], using either arterial or, when appropriate, venous injections, equipment of suitable quality is expensive and not widely available. The risks asso- ciated with conventional selective angiography [ 7-91, and the relatively limited availability of appropriate expertise, make this unsuitable as an initial investigation for many patients with possible carotid disease. Doppler ultrasound should therefore identify patients with severe disease, who may require angio- graphy prior to surgery, patients with moderate disease, who may only require medical treatment with no further investigations, and patients with normal carotid arteries, who may require investigation of other parts of the cardiovascular system to identify causes of neurological symptoms [lo]. To fulfil this role satis- factorily a method of examination and analysis of the Doppler data is required which is highly sensitive to the presence of disease and accurately identifies severe dis- 0720-048X/90/$03.50 0 1990 Elsevier Science Publishers B.V. (Biomedical Division)