MRI Research Today is a free monthly online journal that collates and summarizes the latest research about MRI, including details on magnetic resonance imaging, neuroimaging, brain tumors.

Diagnostic accuracy of MRI for middle cerebral artery stenosis: a postmortem study.

Chen XY, Lam WW, Ng HK, Zhao HL, Wong KS

Department of Anatomical and Cellular Pathology, Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, Hong Kong.

BACKGROUND AND PURPOSE: Magnetic resonance imaging (MRI) has been widely applied in detecting intracranial large artery stenosis, but there have been few validation studies to compare with histopathology. The aim of the postmortem study is to assess the accuracy of MRI in identifying middle cerebral artery (MCA) stenosis. METHODS: We recruited, consecutively, Chinese postmortem autopsies in our hospital during 19 months. MRI was performed in the postmortem brains to scan the cross-sections of MCAs with barium expanding the artery lumen. The MCAs were then removed for histopathologic studies. With histopathology as a reference standard, the accuracy of MRI was evaluated, and the correlation between MCA stenosis identified by MRI and radiologically or histopathologically evident brain infarcts was investigated. RESULTS: Seventy-six consecutive autopsies were recruited. The sensitivity and specificity of MRI in detecting more than 30% MCA stenosis were 38.6% and 92.2%, with a positive predictive value of 87.2% and negative predictive value of 52.2%, and the corresponding values of MRI in identifying more than 50% MCA stenosis were 57.1%, 90.8%, 50%, and 83.0%, respectively. Both more than 30% and more than 50% MCA stenosis identified by MRI were found to be associated with infarctions in the corresponding MCA territory (P= 0.001, odds ratio = 4.365, 95% CI: 1.684-11.313; and P= 0.039, odds ratio = 2.694, 95% CI: 1.139-6.377). CONCLUSIONS: Our study demonstrates the agreement between ex vivo MRI and histopathology in identifying MCA stenosis, and the correlation between the MCA stenosis identified by MRI and radiologically or histopathologically evident brain infarcts.

Published 11 October 2006 in J Neuroimaging, 16(4): 318-22.
Full-text of this article is available online (may require subscription).

© 2005-2008 MRI Research Today. All Rights Reserved.