Background and Purpose-High-resolution, multicontrast magnetic resonance imaging (MRI) has developed into an effective tool for the identification of carotid atherosclerotic plaque components, such as necrotic core, fibrous matrix, and hemorrhage/thrombus. Factors that may lead to plaque instability are lipid content, thin fibrous cap, and intraplaque hemorrhage. Determining the age of intraplaque hemorrhage can give insight to the history and current condition of the biologically active plaque. The aim of this study was to develop criteria for the identification of the stages of intraplaque hemorrhage using high-resolution MRI. Methods-Twenty-seven patients, scheduled for carotid endarterectomy (CEA), were imaged on a 1.5-T GE SIGNA scanner (sequences: 3-dimensional time of flight, double-inversion recovery, T1-weighted (T1W), PDW and T2W). Two readers, blinded to histology, reviewed MR images and grouped hemorrhage into fresh, recent, and old categories using a modified cerebral hemorrhage criteria. The CEA specimens were serially sectioned and graded as to presence and stage of hemorrhage. Results-Hemorrhage was histologically identified and staged in 145/189 (77%) of carotid artery plaque locations. MRI detected intraplaque hemorrhage with high sensitivity (90%) but moderate specificity (74%). Moderate agreement in classifying stages occurred between MRI and histology (Cohen ϭ0.7, 95% CI: 0.5 to 0.8 for reviewer 1 and 0.4, 95% CI: 0.2 to 0.6 for reviewer 2), with moderate agreement between the 2 MRI readers (ϭ0.4, 95% CI: 0.3 to 0.6). Conclusion-Multicontrast MRI can detect and classify carotid intraplaque hemorrhage with high sensitivity and moderate specificity.

Background-Intraplaque hemorrhage and juxtaluminal hemorrhage/thrombus may differ in cause and clinical implications. This study tested the hypothesis that MRI can distinguish between intraplaque hemorrhage and juxtaluminal hemorrhage/thrombus and investigated the association between hemorrhage and underlying lesion types. Methods and Results-Twenty-six patients scheduled for carotid endarterectomy were imaged with a 1.5-T GE scanner by a multicontrast-weighted MRI technique. Hemorrhages were identified with previously established MRI criteria, and differentiations were made between intraplaque and juxtaluminal hemorrhage/thrombus. Corresponding histology was used to confirm the magnetic resonance findings. Tissues underlying areas of hemorrhage/thrombus were histologically categorized according to modified American Heart Association criteria. Of 190 matched sections, 140 contained areas of hemorrhage by histology, of which MRI correctly detected 134. The sensitivity and specificity for MRI to correctly identify cross sections that contained hemorrhage were 96% and 82%, respectively. Furthermore, MRI was able to distinguish juxtaluminal hemorrhage/thrombus from intraplaque hemorrhage with an accuracy of 96%. The distribution of lesion types underlying hemorrhages differed significantly (Pϭ0.004). Intraplaque hemorrhage had an underlying lipid-rich type IV/V lesion in 55% of histological sections, whereas juxtaluminal hemorrhage/thrombus had an underlying calcified lesion type VII in 70% of sections. Conclusions-In vivo high-resolution MRI can detect and differentiate intraplaque hemorrhage from juxtaluminal hemorrhage/thrombus with good accuracy. The association of hemorrhage and lesion types suggests potential differences in origin. Noninvasive MRI therefore provides a possible tool for prospectively studying differences in origin of plaque hemorrhage and the association of plaque progression and instability.

In vivo MRI is capable of quantifying plaque volume and plaque composition, such as %lipid-rich/necrotic core and %calcification, in the clinical setting of a multi-center trial with low inter-scan variability. This study provides the basis for sample size calculation of future MRI trials.