Objectives: Risk of thrombo-embolic stroke is thought to be better reflected by carotid plaque composition than by luminal stenosis. We set out to determine whether high resolution MRI was a valid method of quantifying plaque components in vivo.
Design: A prospective cohort study validating in vivo MRI against histological analysis of excised carotid plaques.
Materials: Twenty-five recently symptomatic patients with severe internal carotid artery stenosis underwent pre-operative in vivo multi-sequence MRI of the carotid artery using a 1.5 T system.
Methods: Individual plaque constituents were characterized on axial MR images according to net signal intensities. Analysis of fibrous cap and lipid core content was quantified proportional to overall plaque area. Bland-Altman plots were generated, and intra-class coefficients computed to determine the level of agreement between the two methods and inter-observer variability.
Results: The intra-class correlation coefficients between two MR readers were 0.94 and 0.88 for quantifying fibrous cap and lipid core components, respectively. There was good agreement between MR and histology derived quantification of both fibrous cap and lipid core content; the mean % difference for fibrous cap was 0.75% (+/-2.86%) and for lipid core was 0.86% (+/-1.76%).
Conclusion: High resolution carotid MRI can be used to quantify plaque components and may prove useful in risk stratification.