Improvement in accuracy of diagnosis of carotid artery stenosis with duplex ultrasound scanning with combined use of linear array 7.5 MHz and convex array 3.5 MHz probes: validation versus 489 arteriographic procedures.

Objective: Validity of a method to improve the accuracy of carotid artery duplex scanning was tested in comparison with arteriography.

Study Design: In 489 patients who had not previously undergone arteriography, 978 carotid arteries were examined with duplex ultrasound scanning. In method A, a linear array 7.5 MHz transducer with pulsed-wave 4.7 MHz Doppler scanning was used. For the diagnosis and grading of carotid stenosis, peak systolic and end-diastolic velocity of the Doppler waves were recorded. Method B consisted of complete ultrasound imaging and color-flow mapping with a convex array 3.5 MHz transducer with pulsed-wave 2.8 MHz Doppler scanning in all patients who had previously undergone method A. Further velocity measurements were performed at the sites of stenosis. The results of methods A and B were compared with data from neurologic assessment and arteriographic studies.

Results: Method B showed significantly higher diagnostic agreement with arteriography than did method A (K 95% confidence interval [CI], 0.87-0.93 vs 0.79-0.85; P <.05), and the number of mistakes in grading stenosis was significantly lower (primarily because of decreased overestimation) in patients with internal carotid kinking (>60 degrees of angulation) (P <.05), distal stenosis (>20 mm from bifurcation) (P <.01), or wide acoustic shadowing (>1 cm) (P <.01) and in those without these conditions (P <.05). Compared with arteriography, diagnostic accuracy with the new method proved higher for carotid stenoses 50% or greater, 60% or greater, 70% or greater, and 80% or greater; no statistically significant difference was found for carotid stenosis 96% or greater or for carotid occlusion. Compared with data from neurologic assessment and arteriography, method B proved more accurate than method A in designating patients for carotid endarterectomy (P =.014).

Conclusions: The new method significantly improved diagnostic reliability of duplex ultrasound scanning, especially in carotid arteries with kinking, distal stenosis, or wide acoustic shadowing (32.2% of all arteries studied). In clinical practice, we suggest additional use of a lower frequency transducer in cases in which these three conditions are found or suspected at first scanning.