Overestimation of carotid artery stenosis with magnetic resonance angiography compared with digital subtraction angiography

Background and purpose: Three-dimensional time-of-flight (3D TOF) magnetic resonance angiography (MRA) is generally considered to overestimate the degree of stenosis in the internal carotid artery (ICA) in comparison with the reference standard intraarterial digital subtraction angiography (DSA). We evaluated whether the degree of stenosis was more accurately assessed with 3D TOF MRA if corresponding projections on MRA and DSA were compared instead of comparison of maximal stenosis at MRA with maximal stenosis at DSA. Methods: From February 1997 to December 1999, we included 186 symptomatic and 17 asymptomatic consecutive patients suspected of having carotid artery stenosis on the basis of clinical presentation and screening with duplex ultrasound scan examination. All patients subsequently underwent DSA and MRA imaging. From each ICA, 12 maximum intensity projections with 3D TOF MRA and two or three projections with DSA were obtained. First, we compared the maximal stenosis at MRA with the maximal stenosis at DSA. Subsequently, we used the stenosis at MRA measured on the projection corresponding with the DSA projection that showed the maximal stenosis. For both strategies, the mean differences in stenosis and sensitivity and specificity for assessment of severe stenosis (70% to 99%) were calculated and compared. Results. The MRA and DSA images of 354 ICAs could be compared. The sensitivity and specificity of MRA with the projection that showed the maximal stenosis were 92.6% (95% CI, 85.3% to 97.0%) and 82.7% (95% CI, 78.1% to 87.3%), respectively. The sensitivity and specificity with the MRA projection, corresponding with the DSA projection showing the maximal stenosis, were 88.3% (95% CI, 81.8% to 94.8%) and 89.6% (95% CI, 85.9% to 93.3%), respectively. The mean difference between maximal stenosis at MRA and DSA was 7.5% (95% CI, 5.2% to 9.9%). The mean difference between stenosis at MRA and DSA in corresponding projections was 0.4% (95% CI, -2.0% to 2.7%). Conclusion: If corresponding MRA and intraarterial DSA projections are compared, 3D TOF MRA does not overestimate carotid stenosis.