OBJECTIVETo assess the reproducibility of 3.0 T high-resolution magnetic resonance imaging (HR MRI) for evaluation of atherosclerotic stenosis in the middle cerebral artery (MCA).

METHODSFrom February, 2011 to December, 2013, 66 consecutive patients with MCA-M1 atherosclerotic stenosis (50%-99%) confirmed by digital subtractive angiography (DSA) received examinations with 3.0 T HR MRI for measurement of the vessel area (VA) and lumen area (LA) at the maximum narrow site (VA(narrow) and LA(narrow)) and the reference site (VA(reference) and LA(reference)) as well as the plaque distribution (ventral, dorsal, superior, and inferior). Two independent readers reviewed all the images and one reader reevaluated these images 4 weeks later. The inter- and intra-observer reproducibility was evaluated using the intraclass correlation coefficient (ICC).

RESULTSThe measurements of VA(narrow), VA(reference), and LA(reference) using HR MRI showed excellent inter- (ICC=0.801, 0.843, and 0.808, respectively) and intra-observer reproducibility (ICC=0.811, 0.916, and 0.958, respectively), but the measurement of LA(narrow) had only moderate inter- and intra-observer reproducibility (ICC=0.584 and 0.625, respectively). For plaque distribution analysis (ventral, dorsal, superior, and inferior plaques), HR MRI also showed excellent inter- (ICC=0.856, 0.836, 0.791, and 0.905, respectively) and intra-observer reproducibility (ICC=0.876, 0.827, 0.825, and 0.950, respectively).

CONCLUSIONHR MRI shows good inter- and intra-observer reproducibility in identifying MCA-M1 atherosclerotic plaque distribution and vessel and lumen measurements, but its reliability for lumen area measurement at the maximum narrowing site needs to be improved.