Carotid duplex criteria: What have we learned in 40 years?

Before the development of the first prototype duplex ultrasound scanner at the University of Washington in the late 1970s, the only noninvasive tests available for extracranial carotid artery disease were indirect methods, such as the periorbital Doppler examination and oculoplethysmography. The duplex scanner combined real-time two-dimensional B-mode imaging and pulsed-Doppler flow detection in a single instrument and provided Doppler spectral waveforms from discrete sites within the vessel lumen. Spectral waveforms allowed characterization of the flow patterns and velocity changes associated with normal and diseased arteries. In a series of validation studies, Dr. D. Eugene Strandness, Jr. and colleagues compared various spectral waveform parameters obtained from internal carotid arteries to independently read carotid arteriograms and established quantitative threshold criteria for classification of carotid artery disease. These criteria were based on peak systolic velocity and end-diastolic velocity, as well as features such as spectral broadening and flow separation. Internal carotid arteries were classified as normal, 1% to 15% diameter reduction, 16% to 49% diameter reduction, 50% to 79% diameter reduction, 80% to 99% diameter reduction, and occluded. Since the 1980s, the University of Washington carotid duplex criteria have been widely used and modified in vascular laboratories throughout the world. Additional clinically relevant criteria have also been developed, such as a threshold for the 70% to 99% North American Symptomatic Carotid Endarterectomy Trial (NASCET) stenosis. Validation of carotid criteria has always depended on comparing spectral waveform parameters to the “gold standard” of contrast arteriography. However, experience has shown that the relationship between velocity and arteriographic stenosis is subject to significant variability. Based on these observations, standardization of carotid duplex criteria should lead to more consistent reporting among vascular laboratories, but it is unlikely to result in improved correlation with arteriography.