Prediction of carotid disease by ultrasound and digital subtraction angiography.

Ultrasound (US) technology (B-mode real-time imaging and Doppler signal analysis) and digital subtraction angiography (DSAV) were prospectively compared in 90 consecutive patients. Twenty-four endarterectomy specimens became available. The DSAV was uninterpretable in 9.4% of vessels and US, in 8.3%. Estimation of degree of stenosis by both modalities agreed well with endarterectomy specimens. Neither predicted the presence or absence of ulceration accurately. Not considering occlusions (where DSAV is probably more accurate), no clear superiority as a screening test compared against a common standard was demonstrated for either. In all 148 vessels with adequate studies by both, the technologies agreed in categorizing stenosis (less than or equal to 50% v greater than 50% diameter stenosis) in 89%. Correlation was poor for occlusions and ulceration. If DSAV were definitive, US would be an effective screen for surgically relevant stenosis with a negative predictive value of 95%.