Acoustic shadowing impairs accurate characterization of stenosis in carotid ultrasound examinations

Abstract

Objective: Duplex ultrasonography (DUS) has been the mainstay for diagnosing carotid artery stenosis and is often the sole diagnostic modality used prior to intervention. Highly calcified plaque, however, results in an acoustic shadow (AcS) that obscures the vessel lumen and inhibits the sonographer's ability to obtain Doppler velocity measurements. It is unknown whether DUS can accurately determine the degree of carotid stenosis in these settings.

Methods: From July 2012 to December 2013, all patients with AcS on DUS measuring ≥5 mm in the longitudinal axis were cross-referenced with multidetector computed tomographic angiography (MD-CTA) images of the neck to define the study population. After standardizing the MD-CTA windows, percent stenosis was determined by cross-sectional area reduction using two separate previously described methods based on North American Symptomatic Carotid Endarterectomy Trial (NASCET) and European Carotid Surgery Trial (ECST) criteria. DUS waveform parameters in the internal carotid artery near the AcS were then compared with these MD-CTA measurements to determine the accuracy of DUS in characterizing the severity of carotid stenosis.

Results: During this period, 8517 DUS studies were performed at the Massachusetts General Hospital, 550 of which had AcS, for an incidence of 6.45%. There were 92 lesions with a concomitant MD-CTA; however, seven were excluded because of poor study quality, because ≥6 months had elapsed between DUS and MD-CTA, or because the patient had undergone carotid reconstruction between studies. Of the 85 remaining lesions, DUS characterized 17 as severe (peak systolic velocity [PSV] >250 cm/s), 31 as moderate (PSV = 151-250 cm/s), and 37 as mild (PSV ≤150 cm/s) stenoses using PSV criteria. PSV weakly correlated with CTA-NASCET (r = 0.361; P = .004) and CTA-ECST (r = 0.306; P = .004) percent stenosis. Using PSV >250 cm/s as the predictor of >70% stenosis, and a ≥70% cutoff by both CTA-ECST and CTA-NASCET methods as the reference measure, DUS sensitivity ranged from 22.7% to 32.5%, specificity from 89.4% to 91.1%, positive predictive value from 88.2% to 76.4%, and negative predictive value from 25% to 60.2%. A subgroup analysis of lesions identified as non-severe by DUS revealed that waveforms with lower deceleration were associated with severe stenosis on CTA.

Conclusions: In the presence of AcS, DUS alone is inadequate to accurately determine the degree of carotid stenosis with sensitivity, specificity, and negative predictive values far below that needed for clinical decision-making. MD-CTA may be necessary for improved characterization of plaque in these AcS lesions. Further studies are needed to determine DUS parameters that may identify patients who should undergo further evaluation with MD-CTA to characterize the true severity of the stenosis.