Accuracy of MDCT in the determination of supraaortic artery stenosis using DSA as the reference standard

Purpose: Accurate stenosis quantification in the carotid arteries is of great clinical importance. We aimed to compare the diagnostic accuracy of multi-slice computed tomography angiography (CTA) to digital subtraction angiography (DSA) for the detection and grading of atherosclerotic lesions involving the supraaortic arteries.

Materials and methods: We retrospectively analyzed 30 patients (10 women; mean age, 67 years). CTA was performed after administration of 100 ml Ultravist 370 (Bayer Schering, Germany), at a flow of 5 ml/s, using a Philips Brilliance 16MDCT scanner (Philips, Best, Netherland) at a collimation of 16 mm×0.75 mm prior to DSA. The supraaortic arteries were divided into 17 segments, and, within each segment, the presence and severity of stenotic or occlusive lesions was determined, based on a four-point scale (0-49%, 50-69%, 70-99%, occlusion), by four independent readers using the NASCET criteria. Sensitivity and specificity of MDCT was calculated for the detection of moderate (50-69%) versus significant stenoses (70-99%) and occlusion.

Results: There were 291 segments assessed with both methods. Thirteen lesions were "not assessable" on CTA. DSA identified 53 significant lesions, and CTA 56 significant lesions. With regard to significant lesions, CTA overrated six lesions and underestimated six lesions, resulting in a sensitivity, specificity, and accuracy of 86.4%, 97.6%, and 95.9%, respectively. For the detection of stenoses greater than 50%, sensitivity, specificity, and accuracy were 90.2%, 95.8%, and 94.8%, respectively.

Conclusions: Compared to DSA, CTA shows high accuracy in the detection and grading of lesions involving the supraaortic arteries enabling its use in the detection and treatment planning for stenoses of the supraaortic vessels.