Effect of ultrasonography in the assessment of carotid artery stenosis.

Although ultrasonography (US) advantageously portrays lumen and wall thickness, velocity criteria have been used primarily to interpret carotid artery stenosis. The relationship of US and velocity measurements was investigated. Peak-systolic and end-diastolic velocities (PSV, EDV) increase exponentially as the lumen of the internal carotid artery narrows and the percent stenosis (%S) increases. We tested the consistency of the, relationship between carotid velocities and US %S in two distinct data sets. One data set was used to obtain regression equations relating velocity parameters and %S based on US. Validation of these equations was conducted using a separate, independent data set. US measurements were classified in 12 %S intervals, PSV, EDV, the ratio of the internal carotid artery to the common carotid artery PSV, and %S were entered consecutively until 10 records for each %S interval were obtained. Regression equations obtained in the first data set were used to predict %S in the second data set. Predicted %S was then compared with actual US %S. The highest correlation in the first data set (r = .89) was between %S and the natural logarithm (In) of PSV. This In PSV -%S equation was then applied to a second data set of an additional 120 carotid duplex images. In the second data set, actual %S and PSV-predicted %S differed by >10% in 38 cases (32%). When all velocity-%S regression equations were used for comparison, differences between actual and at least one velocity-predicted %S were >10% in 19% of the arteries. Conversely, actual %S matched at least one prediction of %S based on velocity data in 81% of the cases. US %S differed significantly from single velocity-based estimates of %S in at least one-third of the cases. On the other hand, four of five US measurements were confirmed by at least one velocity parameter. Emphasis on US, in addition to velocity data, is recommended for the interpretation of duplex US carotid examinations.