W J Zwiebel

This review considers the capabilities and limitations of duplex sonography in the diagnosis of abnormalities of the cerebral vasculature. Duplex sonography is an elegant union of B-mode and Doppler sonography that provides valuable information about atherosclerotic obstruction of the carotid arteries. Duplex sonography also can be used to evaluate, in a general way, the composition of carotid atherosclerotic plaque, and in this respect, it is unique among imaging procedures. Duplex sonography is not very effective in providing a "global" perspective of the cerebral vasculature, because only the cervical portion of the carotid arteries can be examined in detail. The best documented and most clearly effective use of duplex sonography is for detecting severe obstructive lesions in the carotid artery that might warrant endarterectomy in patients with cerebral hemispheric symptoms. The role of duplex sonography in the choice between medical and surgical therapy in asymptomatic patients with carotid artery stenosis is controversial, because the indications for endarterectomy are unclear in these patients. The capacity of duplex sonography to assess plaque composition may ultimately prove to be quite valuable for selecting therapeutic options and for evaluating the effectiveness of medical therapy. Meanwhile, information concerning the clinical value of this use of duplex sonography remains limited.

The peak Doppler-shifted frequency and degree of lumen narrowing were compared in 75 cervical carotid stenoses. Peak frequency was not found to precisely indicate severity of stenosis, but it was possible to divide stenoses into four ranges of severity on the basis of peak frequency. Peak frequencies of less than 5 KHz, in most instances, indicated stenoses of less than 50% decrease in lumen area. Frequencies from 5 to 8 KHz were generally associated with stenoses of 50-75% decrease in lumen area, and frequencies of 8-12 KHz were most often associated with stenoses of 75-90% decrease in area. Frequencies greater than 12 KHz almost invariably occurred in very severe lesions of greater than 90% decrease in lumen area (greater than 70% decrease in diameter). The use of peak frequency as an indicator of severity of stenosis, while not specific, is felt to be of considerable clinical benefit as it provides a more quantitative evaluation of stenosis than auditory assessment of Doppler frequencies.

Duplex ultrasound (US) scans of 110 carotid arteries ipsilateral to hemispheric strokes were compared with scans of 90 asymptomatic vessels in the same patients to determine the relative prevalence of stenotic lesions. In addition, scans of paired carotid arteries in patients with stroke involving only one cerebral hemisphere were compared to determine whether the incriminated side demonstrated a greater degree of stenosis than the asymptomatic side. The duplex US findings demonstrated a positive correlation between stenosis and hemispheric stroke. However, only 20% of carotid arteries ipsilateral to hemispheric stroke showed a reduction in diameter greater than 70%, compared with 5% of asymptomatic vessels. A minimal difference was demonstrated between the symptomatic and asymptomatic groups with respect to lesser degrees of stenosis. In paired carotid arteries, the degree of stenosis of the symptomatic vessel exceeded that of the asymptomatic vessel in only 43% of cases. These results suggest that the prevalence of severe carotid stenosis in stroke patients has been previously overestimated. The findings also emphasize the need for further investigation of other plaque-related risk factors that may enhance stroke prevention through improved selection of surgical or medical therapy. Factors currently under investigation include plaque ulceration, intraplaque hemorrhage, plaque echogenicity, and the effects of sequential stenoses.

Fifty errors with Doppler examination of 356 carotid bifurcations were examined to determine their cause and to establish methods of prevention. Only those errors related to hemodynamically significant stenosis or complete occlusion were considered. The relative frequency of errors in diagnosis of occlusion (30 false-positive or negative versus 31 true-positive) was considerably greater than the rate of inaccuracy for diagnosis of hemodynamically significant stenosis (20 false-positive or-negative versus 89 true-positive). The high error rate in diagnosis of occlusion was attributed to reliance on negative information. The source of error could be established in 48 of the 50 cases. In all but three cases, potential for preventing error existed through use of additional noninvasive techniques such as examination of common carotid resistivity or use of oculoplethysmography. Twenty-two errors of localization of stenosis or occlusion were encountered in addition to the 50 false-positive and -negative errors. In three of these, the errors might have led to patient mismanagement.

Eighteen digital subtraction angiography (DSA) examinations were retrospectively evaluated for factors that led to their erroneous interpretation. Overlapping vessels obscured pathologic conditions in five cases. In four cases the lesions were not adequately profiled by the DSA projections. Eight lesions were rendered inconspicuous by misregistration artifacts attributable to motion, either from swallowing or from pulsation of vessel walls. One diagnostic error was caused by poor opacification from a degraded contrast bolus secondary to low cardiac output.