The contribution of spectral analysis to the diagnosis of carotid artery disease.

Abstract

Pulsed and continuous-wave Doppler examinations with spectral analysis were performed on 258 carotid arteries in 220 patients and compared with multiplane contrast angiography. Each artery was examined for stenosis and assigned to one of four groups: normal to 19%, 20% to 49%, 50% to 99%, and occlusion. The incidence of disease was 70%. Diagnostic sensitivity with all noninvasive techniques varied from 76% to 90% depending on the degree of stenosis, with an overall accuracy rate of 86%. In an attempt to evaluate the diagnostic contribution of spectral analysis, stenosis classification by peak frequency alone was compared with that of the full laboratory profile, including spectral analysis. In the 147 carotid arteries in which such data were available, diagnostic sensitivity was improved with spectral analysis by 8% to 29% depending on the degree of stenosis. Detection of occlusions was unchanged. The major contribution of spectral analysis was in detecting non-flow-limiting stenoses of 20% to 49%. Spectral analysis added little to the diagnostic ability of peak frequency to detect lesions with greater than 50% stenosis. Spectral analysis of either pulsed Doppler or continuous-wave Doppler signals is an accurate noninvasive method for evaluating carotid bifurcation disease, which particularly improves the detection of non-flow-limiting stenoses.