Objective: We investigated the accuracy of color-flow Doppler (CD) scanning, power Doppler (PD) scanning, and peak systolic Doppler frequency shift (PSF) in assessment of carotid artery stenosis with angiography used as gold standard, including the measurement techniques used in the North American Symptomatic Carotid Surgery Trial (NASCET) and the European Carotid Surgery Trial (ECST).
Methods: Fifty-eight consecutive patients diagnosed for carotid artery surgery underwent color-coded duplex sonography and angiography. The duplex examination included the assessment of PSF and the videotaping of sagittal images in CD and PD mode from the proximal common carotid artery to the distal internal carotid artery. Two experienced examiners performed the studies, but once one examiner had done the taping, the other examiner was allowed only to review the tape. Separately, each examiner reviewed the tapes and determined by cursor settings each stenosis according to NASCET and ECST. For interobserver agreement kappa statistic was used. To compare with angiography (degree of stenosis 40%, 50%, 60%, 70%, and 80%) sensitivity, specificity, positive and negative predictive values, and overall accuracy were calculated. PSF cut-off frequencies were based on receiver operator curve analysis.
Results: Because interobserver agreement in CD and PD was good (chance-corrected kappa > 0.6), further analysis used the between-observer mean value for each stenosis. With the NASCET measurement technique, accuracy of Doppler techniques to distinguish a 50% stenosis was 89% for PSF, 91% for CD, and 93% for PD; for a 70% stenosis it was 83% for PSF, 84% for CD, and 81% for PD. With the ECST measurement technique, accuracy to distinguish a 70% stenosis was 86% for PSF, 88% for CD, and 86% for PD; for an 80% stenosis it was 87% for PSF, 87% for CD, and 77% for PD.
Conclusion: CD and PD carotid artery stenosis measurements are highly reproducible, and in our hands provided accuracy equal to PSF.