Interpretational Discordance of Arterial Ultrasound and Computed Tomography Angiography for Carotid Artery Disease

Purpose: This quality improvement project was prompted by complaints regarding inconsistencies between carotid artery ultrasound (CAU) and computed tomography angiography (CTA) neck reporting. Our goal was to identify causative agents in reporting inconsistencies. These causative agents would be addressed to result in improved diagnostic quality and better patient care.

Materials and Methods: Over a 22-month period, 233 CTA neck and 301 carotid CAU were performed at our intuition. Patients who underwent both CTA neck and CAU within a 7-day period were selected, excluding patients who underwent surgical intervention in the interim. The studies were analyzed by an attending radiologist and two radiology residents to ascertain the reason for interpretation discordance.

Results: Fifty-two patients underwent both CTA neck and carotid ultrasound. A total of 44 of 52 (84.6%) studies were concordant, meaning the CTA and  CAU impressions were essentially the same, and 8 of 52 (15.4%) were discordant. Of the 8 discordant studies, 2 (3.8%) were considered to be major discrepancies, meaning that the interpretation inconsistencies would result in different patient management based on North American Symptomatic Carotid Endarterectomy guidelines. Six (11.5%) of the discordant studies were considered minor discrepancies, meaning differences in interpretation would not differ patient management based on existing guidelines.

Conclusions: The first major discrepancy was affected by motion artifact on CT imaging, resulting in exaggerated stenosis on axial images. Thorough evaluation of the coronal and or sagittal images would have prevented this misinterpretation. The second major discrepancy was a combination of CAU and CTA misinterpretation. The CAU interpretation relied heavily on the quantitative data, ignoring high-grade internal carotid artery narrowing seen on grayscale images from atherosclerotic plaque. Additionally, motion artifact again caused overestimation of degree of stenosis on axial CT imaging. From our experience, the performing technologists practiced excellent CAU technique. However, two minor discrepancies were the result of underestimated flow velocity caused by error in the angle of Doppler signal acquisition. Other discrepancies were caused by erroneous measurements of the stenosis on CTA. CAU is a very reliable diagnostic modality when compared with CTA if technical, and interpretive errors are minimized by learning from the errors described.