Risk prediction of cerebrovascular events with carotid plaque magneitc resonance analysis: A meta-analysis

Background and purpose. - It is not conclusive that magnetic resonance (MR)-based carotid atherosclerotic plaque assessment identifies high-risk features associated with cerebrovascular events. We aimed to systematically summarize the association of MR imaging (MRI)-determined intraplaque hemorrhage (IPH), lipid-rich necrotic core (LRNC), and thinning/rupture of the fibrous cap (TRFC) with subsequent ischemic events. Materials and methods. - We performed a comprehensive literature search evaluating the association of MRI-based carotid plaque composition with ischemic outcomes. We included cohort studies examining IPH, LRNC, or TRFC with mean follow-up of >= 6 months and an outcome measure of ipsilateral ischemic events. A meta-analysis was done according to the Cochrane guideline. Results. - We identified 13 studies including 1.150 patients and 1.208 analyzed carotid arteries, with mean follow-up of 21.1 months. The hazard ratios (FIR) for IPH, LRNC, and TRFC as predictors of subsequent ischemic events were 4.41 (95% CI: 2.87, 6.79), 3.00 (95% CI: 1.51, 5.95), and 5.94 (95% CI: 2.66, 13.28), respectively. The predictive value of carotid plaque MRI for ischemic events was acceptable, with sensitivity of 0.80 (95% CI: 0.66, 0.90) and specificity of 0.63 (95% CI: 0.57, 0.68). However, it was limited to confirm or exclude future ischemic events in clinical context, with positive likelihood ratio (LR) of 2.2 (95% CI: 1.9, 2.5) and negative LR of 0.31 (95% CI: 0.18, 0.55). No statistically significant heterogeneity or publication bias was observed. Conclusion. - The presence of IPH, LRNC, and TRFC determined by MRI is associated with increased risk of future ischemic events, but its predictive value is moderate and should not be used for confirmation or exclusion of future ischemic events in clinical context. (C) 2018 Elsevier Masson SAS. All rights reserved.