Swiss Medical Weekly (Feb 2020)
Plaque vulnerability in patients with high- and moderate-grade carotid stenosis – comparison of plaque features on MRI with histopathological findings
Abstract
BACKGROUND Plaque vulnerability plays an important role in determining the risk of subsequent cerebrovascular events in patients with carotid stenosis. Plaque morphology magnetic resonance imaging (MRI) can be used to assess plaque vulnerability. We therefore set out to examine the diagnostic accuracy of plaque morphology MRI compared with histopathological findings as gold standard in moderate- to high-grade carotid stenosis at our centre. METHODS A total of 36 patients with moderate- to high-grade carotid stenosis underwent plaque morphology MRI with a multisequence protocol (time of flight sequence, dark blood T1- [native and post-gadolinium] and T2-weighted sequence with fat suppression). The status of the fibrous cap, calcification, lipid-rich necrotic core (LRNC) and intraplaque haemorrhage (IPH) were assessed by means of qualitative MR analysis of plaque characteristics and compared with the histopathological findings. Detection statistics (sensitivity, specificity), chi-squared test, Cohen’s kappa (κ), percentage of agreement and phi coefficient (φ) were determined. RESULTS Carotid stenosis was symptomatic (transient ischaemic attack, amaurosis fugax or ischaemic stroke in the territory of the stenosed carotid artery) in 25 patients (69.5%). Twenty-eight patients (77.8%) had a high-grade and eight patients (12.2%) a moderate-grade stenosis. Significant congruence between MRI and histology was found for plaque calcification (89% histology, 75% MRI, κ = 0.364, p = 0.013), for LRNC (89% histology, 53% MRI, κ = 0.245, p = 0.025) and IPH (75% histology, 53% MRI, κ = 0.314, p = 0.035). In a subgroup of patients with symptomatic stenosis, the agreement for LRNC and IPH was slightly better (LRNC κ = 0.390, p = 0.014; IPH κ = 0.386, p = 0.045). Status of the fibrous cap, essentially ulceration, did not show any significant agreement (κ = 0.032, p = 0.842). There was significant correlation between LRNC on MRI and symptomatic carotid stenosis (φ = 0.339, p = 0.042). CONCLUSION Plaque morphology MRI is capable of identifying the main components of atherosclerotic plaques with moderate to good accuracy as compared with histopathological findings as gold standard. LRNC seems to be a useful marker of plaque vulnerability and warrants its use in clinical decision making.
Keywords