Prospective evaluation of carotid artery stenosis: elliptic centric contrast-enhanced MR angiography and spiral CT angiography compared with digital subtraction angiography.

BACKGROUND AND PURPOSE Although digital subtraction angiography (DSA) is the reference standard for assessing carotid arteries, it is uncomfortable for patients and has a small risk of disabling stroke and death. These problems have fueled the use of spiral CT angiography and MR angiography. We prospectively compared elliptic centric contrast-enhanced MR angiography and spiral CT angiography with conventional DSA for detecting carotid artery stenosis. METHODS Eighty carotid arteries (in 40 symptomatic patients) were assessed. Elliptic centric MR and spiral CT angiographic data were reconstructed with maximum intensity projection and multiplanar reconstruction techniques. All patients had been referred for DSA evaluation on the basis of findings at Doppler sonography, which served as a screening method (degree of stenosis > or = 70% or inconclusive results). Degree of carotid stenosis estimated by using the three modalities was compared. RESULTS Significant correlation with DSA was found for stenosis degree for both elliptic centric MR and spiral CT angiography; however, the correlation coefficient was higher for MR than for CT angiography (r = 0.98 vs r = 0.86). Underestimation of stenoses of 70-99% occurred in one case with elliptic centric MR angiography (a 70% stenosis was underestimated as 65%) and in nine cases with spiral CT angiography, in comparison to DSA findings. Overestimation occurred in two cases with MR angiography (stenoses of 65-67% were overestimated as 70-75%). With CT, overestimation occurred in seven cases; a stenosis of 60% in one case was overestimated as 70%. Both techniques confirmed the three cases of carotid occlusion. With elliptic centric MR angiography, carotid stenoses of 70% or greater were detected with high sensitivity, 97.1%; specificity, 95.2%; likelihood ratio (LR) for a positive test result, 20.4; and ratio of LR(+) to LR(-), -0.3. With spiral CT angiography, sensitivity, specificity, LR(+), and LR(+):LR(-) were 74.3%, 97.6%, 31.2, and 0.3, respectively. CONCLUSION Elliptic centric contrast-enhanced MR angiography is more accurate than spiral CT angiography to adequately evaluate carotid stenosis. Furthermore, elliptic centric contrast-enhanced MR angiography appears to be adequate to replace conventional DSA in most patients examined.