The measurement of internal carotid artery stenosis: comparison of duplex with digital subtraction angiography.

Objectives: To compare the subjective (eyeballed) method for measuring internal carotid artery (ICA) stenosis by non-selective intra-arterial digital subtraction angiography (IA-DSA) with objective and duplex methods.
Design: Retrospective study.
Materials and Methods: Fifty-three consecutive patients underwent IA-DSA prior to carotid endarterectomy providing 103 carotid angiograms. Objective assessment of ICA stenosis was by the North American Symptomatic Carotid Endarteectomy Trial (NASCET) criterion and the Carotid Stenosis Index (CSI). Duplex estimation of stenosis was derived from peak systolic and diastolic velocities in the ICA and common carotid artery (CCA).
Results: The coefficient of repeatability was poorest for NASCET stenosis estimates (40%), whilst the improved values for CSI (20%) were consistent with the lower variability recorded for measuring the CCA diameter. Correlation and agreement levels between subjective (r = 0.80, -41% to +33%) or objective assessments and duplex (NASCET: r = 0.76; -52 to +28%; CSI: r = 0.72; -27 to +39%) showed similar values.
Conclusions: We conclude that the inter-observer variability for assessing angiograms obtained by arch injection is considerable and precludes high agreement when IA-DSA is compared with other methods. As the agreement of duplex ultrasound with IA-DSA is similar to the agreement between DSA methods, duplex can be offered as the first stage assessment of ICA stenosis, with the proviso that the duplex assessment is performed consistently by an experienced operator.