Background: To evaluate the ability of ultrahigh-field magnetic resonance imaging (MRI) to accurately depict the composition of the human arterial vessel wall ex vivo and to detect early atherosclerotic lesion formation in comparison to histology.
Methods: Eight iliac artery specimens with low-grade atherosclerotic lesions obtained from human organ donors were studied. Three-dimensional, high-resolution MRI (spatial resolution: 79 x 79 x 109 microm) was performed using T1-, T2- and proton density (PD)-weightings (7.0 Tesla MR system, Bruker Pharmascan). A total of 36 MR slices and corresponding histological sections were matched for comparative evaluation of area measurements of lumen, media and adventitia and--if present--plaque size. Statistical correlation between histology and MR measurements was tested and a ROC-analysis was performed to determine the plaque size being predictive of correctly identifying atherosclerotic lesions with MRI.
Results: The areas of vessel lumen and media as measured on T1-, T2- and PD-weighted MR images showed a strong correlation with the corresponding histological measurements (r = 0.84 to r = 0.89; P < 0.01), however, a systematic overestimation of 34-41% was found. For the area of adventitia, only a moderate, though significant, correlation (r = 0.55 to r = 0.62; P < 0.01) could be demonstrated with a similar overestimation by MRI (38-43%). With T1-weighted MRI, sensitivity and specificity for the detection of plaques > 4.0 mm(2) were 79% and 91%, respectively. With T2- and PD-weighted MRI, however, sensitivity and specificity for the detection of plaques > 0.4 mm2 were 93% and 89%.
Conclusions: In an experimental ex vivo setting, ultrahigh-field MRI of the human arterial vessel wall resulted in an accurate visualization of vessel wall composition when compared to histology and, thus, allowed for a quantitative assessment. T2- and PD-weighted MRI proved capable of reliably detecting submillimeter atherosclerotic lesions.