Between-method correlation in quantifying internal carotid stenosis.

BACKGROUND AND PURPOSE: The degree of internal carotid stenosis has emerged as the most important predictor of ischemic stroke in extracranial carotid artery disease. The purpose of this study was to assess the validity of the noninvasive techniques for quantifying internal carotid stenosis with respect to the accepted standard of intra-arterial angiography. METHODS: We measured the maximum percentage reduction in luminal diameter on the intra-arterial digital subtraction angiograms of 56 symptomatic patients with extracranial internal carotid stenosis (n = 77) or occlusion (n = 20). These data were compared with independent measurements based on continuous-wave Doppler ultrasonography, pulsed-wave Doppler spectrum analysis, color Doppler-assisted duplex imaging, and magnetic resonance angiography. RESULTS: Correlations with intra-arterial angiography were equally strong (r > .90) for magnetic resonance angiography, continuous-wave Doppler, and color duplex analysis. Positive and negative predictive values for (therapeutically relevant) 70% to 99% stenosis were higher for continuous-wave Doppler (.82, .97) and color duplex (.84, .98) than for magnetic resonance angiography (.79, .81). Also, accuracy in quantifying high-grade stenosis was better for both of these ultrasonographic techniques, mainly due to the frequent occurrence of a "flow gap" on the magnetic resonance angiograms. Continuous-wave Doppler and magnetic resonance angiography, but not color duplex, failed to detect slow residual arterial flow in one and two cases of symptomatic "pseudo-occlusion" of the internal carotid, respectively. CONCLUSIONS: (1) Several noninvasive methods compare well with intra-arterial angiography in identifying and quantifying high-grade internal carotid stenosis; (2) the use of these noninvasive methods may suffice for treatment decisions; and (3) because residual between-method disagreement is partly explained by principles of physics, the validity of continuous-wave Doppler and color duplex in quantifying 60% to 99% stenosis is likely to be underestimated by correlation with intra-arterial angiography.