• Vincent C Cappendijk, Sylvia Heeneman, Alfons G H Kessels, Kitty B J M Cleutjens, Geert Willem H Schurink, Rob J Th J Welten, Werner H Mess, Robert-Jan van Suylen, Tim Leiner, Mat J A P Daemen, Jos M A van Engelshoven, and M Eline Kooi.
• Department of Radiology, Maastricht University Hospital, Maastricht, The Netherlands. vca@rdia.azm.nl
• J Magn Reson Imaging. 2008 Jun 1; 27 (6): 1347-55.

PurposeTo prospectively determine the accuracy of semiquantitative analysis of the amount of lipid-rich necrotic core (LRNC) in atherosclerotic plaque using multi- as well as single-sequence T1-weighted (w) turbo field echo (TFE) MRI. Histology served as a reference standard.Materials And MethodsSixty-four symptomatic patients with carotid stenosis > or =70% were included and underwent endarterectomy after an MRI scan. Two MRI readers classified relative signal intensities in regions of interest in the vessel wall. The amount of LRNC was determined semiquantitatively using an algorithm based on fixed combinations of multiple MR pulse sequences as well as solely based on T1w TFE images. Interreader agreement was expressed by intraclass correlation coefficients (ICCs). Agreement between MRI and histology was determined by linear regression (R).ResultsInterreader reproducibility for quantification of LRNC was high (ICC, 95% confidence interval [CI]): multisequence 0.86 (0.77-0.94), and single sequence 0.91 (0.85-0.95). There was good agreement between MRI and histology for both MR readers for quantification based on multisequence as well as single sequence MRI, 0.80 < or = R < or = 0.85 (P < 0.0001).ConclusionThe amount of LRNC using single-sequence T1w TFE MRI is a reproducible, accurate, and fast way to quantify LRNC in carotid atherosclerotic plaque.2008 Wiley-Liss, Inc

35 keywords...

hide…