Reproducibility of 3.0T High-Resolution Magnetic Resonance Imaging for the Identification and Quantification of Middle Cerebral Arterial Atherosclerotic Plaques.

Abstract Objective: To assess the reproducibility of 3.0T high-resolution magnetic resonance imaging for the identification and quantification of atherosclerotic plaques in the middle cerebral artery. Methods: Sixty-nine consecutive patients with ischemic stroke or asymptomatic stenosis (>30%) of the middle cerebral artery underwent 3.0T high-resolution magnetic resonance imaging examinations. Two independent investigators reviewed all images with 1 investigator re-evaluating all images 4 weeks later. Wall characteristics of the middle cerebral artery, including plaque surface morphology, plaque location, plaque components, and burden were identified and measured. Results: Intraobserver and interobserver agreement were all substantial in identifying plaque surface irregularity (k = 0.741, 0.555-0.897; k = 0.685, 0.490-0.843; respectively) and intraplaque hemorrhage (k = 0.654, 0.446-0.838; k = 0.605, 0.369-0.792; respectively). Intraobserver agreement was substantial (k = 0.654) and interobserver agreement was moderate (k = 0.553) for the identification of plaque fibrous caps. The total intraobserver and interobserver reproducibility was almost excellent for the identification of plaque position. With regards to vessel area measurement at the site of maximal lumen narrowing, intraobserver and interobserver reproducibility was excellent (intraclass correlation coefficient was 0.886 and 0.885, respectively) and moderate for lumen area at the site of maximal lumen narrowing (intraclass correlation coefficient was 0.695 and 0.558, respectively). In addition, intraobserver and interobserver reproducibility was excellent for vessel area and lumen area measurements at the reference sites. Conclusions: The reproducibility of 3.0T high-resolution magnetic resonance imaging for the identification and quantification of artery wall characteristics was overall acceptable. However, the reliability for lumen area measurement at the maximum narrowing site and identification of the fibrous cap needs to be improved.