Computed tomography angiography vs 3 T black-blood cardiovascular magnetic resonance for identification of symptomatic carotid plaques

Abstract

Background: The purpose of this prospective study was to perform a head-to-head comparison of the two methods most frequently used for evaluation of carotid plaque characteristics: Multi-detector Computed Tomography Angiography (MDCTA) and black-blood 3 T-cardiovascular magnetic resonance (bb-CMR) with respect to their ability to identify symptomatic carotid plaques.

Methods: 22 stroke unit patients with unilateral symptomatic carotid disease and >50% stenosis by duplex ultrasound underwent MDCTA and bb-CMR (TOF, pre- and post-contrast fsT1w-, and fsT2w- sequences) within 15 days of symptom onset. Both symptomatic and contralateral asymptomatic sides were evaluated. By bb-CMR, plaque morphology, composition and prevalence of complicated AHA type VI lesions (AHA-LT6) were evaluated. By MDCTA, plaque type (non-calcified, mixed, calcified), plaque density in HU and presence of ulceration and/or thrombus were evaluated. Sensitivity (SE), specificity (SP), positive and negative predictive value (PPV, NPV) were calculated using a 2-by-2-table.

Results: To distinguish between symptomatic and asymptomatic plaques AHA-LT6 was the best CMR variable and presence / absence of plaque ulceration was the best CT variable, resulting in a SE, SP, PPV and NPV of 80%, 80%, 80% and 80% for AHA-LT6 as assessed by bb-CMR and 40%, 95%, 89% and 61% for plaque ulceration as assessed by MDCTA. The combined SE, SP, PPV and NPV of bb-CMR and MDCTA was 85%, 75%, 77% and 83%, respectively.

Conclusions: Bb-CMR is superior to MDCTA at identifying symptomatic carotid plaques, while MDCTA offers high specificity at the cost of low sensitivity. Results were only slightly improved over bb-CMR alone when combining both techniques.

Figure 1

Shows axial TOF, T1 weighted pre- and post-contrast and T2 weighted high-resolution black-blood CMR and CTA (lower right) images of an ulcerated plaque in the right internal carotid artery of an 87-year old male patient with an acute ischemic stroke in the territory of the right middle cerebral artery. Note the clearly ulcerated plaque surface on both CMR and CTA images as well as the hypersignal of the plaque in TOF and T1 weighted images corresponding to intraplaque hemorrhage (arrow). A lack of contrast enhancement within the plaque indicates the presence of a lipid-rich necrotic core. CTA shows a non-calcified plaque with relatively hypodense plaque interior (mean plaque density = 34,5HUarrowhead).

Figure 2

Shows axial TOF, T1 weighted pre and post contrast and T2 weighted high-resolution black-blood CMR and CTA (lower right) images of a plaque in the left internal carotid artery of a patient with an acute ischemic stroke on the left side. While both CMR and CT images fail to show a distinct surface defect, the fibrous cap is not entirely distinguishable and was therefore by definition classified as thin. The hyperintense signal within the plaque in TOF and T1 weighted images in combination with hypointense signal in the T2 weighted image corresponds to an intraplaque hemorrhage (arrowhead). The relative lack of contrast enhancement within the plaque indicates the presence of a lipid-rich necrotic core (arrow). Correspondingly, the CTA image shows a relatively hypodense plaque interior (mean plaque density = 65,1 HU). The hyperdense area in the dorsal wall of the plaque corresponds to a hypointense signal in the MR images and is consistent with a marginal calcification. * Sternocleid muscle

Figure 3

Shows axial TOF, T1 weighted pre- and post- contrast and T2 weighted high-resolution black-blood CMR and CTA (lower right) images of a stable carotid plaque on the asymptomatic left side in a 66 year old patient who had suffered from right hemispheric stroke. Both CMR and CTA images show the presence of an AHA lesion type 7 plaque in the dorsal wall of the left proximal internal carotid artery. After administration of contrast material the thick fibrous cap is delineated as a hyperintense rim in the T1 weighted contrast enhanced images, separating the plaque from the lumen (white arrowheads). Also note the hypointense signal of the plaque interior in the T1 weighted contrast enhanced image as well as the hypodense area in the CTA image corresponding to a large lipid-rich necrotic core (arrow), measured at 166 HU, probably due to blooming artifacts caused by its calcified portion. The hypointense rim in the peripheral plaque in all MR sequences and the corresponding hyperdense area in the CTA image indicate the presence of a calcification (black arrowheads).

Figure 4

Shows the ROC graph of various variables. Lines are dotted where only one value was present. Note that especially AHA-LT6 with and without ulcer in CTA as well as thin or ruptured fibrous cap show high sensitivity and specificity, while plaque density in HU and volume of calcification are not suitable as predictors of the symptomatic side.