Histological validation of carotid plaque characterization by in-vivo T2 mapping in patients with recent cerebrovascular events: preliminary results

Background Multicontrast CMR is the conventional method for invivo characterization of carotid plaques. However, its non-quantitative nature and the need for extensive postacquisition interpretation limit its clinical application. We have recently proposed using in-vivo quantitative T2 mapping for atherosclerotic plaque characterization [Biasiolli et al. JCMR 2013, 15:69]. T2 maps have the advantage of providing an absolute physical measure of plaque components that can be standardized among different MR systems and widely adopted in multi-centre studies. This pilot study investigates the agreement between T2 mapping and histology using AHA plaque type classification.


Background
Multicontrast CMR is the conventional method for invivo characterization of carotid plaques. However, its non-quantitative nature and the need for extensive postacquisition interpretation limit its clinical application. We have recently proposed using in-vivo quantitative T2 mapping for atherosclerotic plaque characterization [Biasiolli et al. JCMR 2013, 15:69]. T2 maps have the advantage of providing an absolute physical measure of plaque components that can be standardized among different MR systems and widely adopted in multi-centre studies. This pilot study investigates the agreement between T2 mapping and histology using AHA plaque type classification.

Methods
CMR: 19 symptomatic patients scheduled for endarterectomy (14 males, 73 ± 11 years, range 54-89 years, IRB approved, written consent obtained) were imaged at 3T (Siemens Verio) with a pair of 2-channel phased-array carotid coils (Machnet). In addition to the conventional multicontrast CMR protocol, a novel black-blood multislice DANTE-prepared [Li et al. MRM 2012, 68:1423-1438 multiecho-spin-echo sequence was used to acquire 5 slices with 14 echoes (TR = 2000 ms, TE = 9-127 ms) in <4 minutes (partial Fourier = 5/8, FOV = 128 × 128, matrix size = 384 × 384, slice thickness = 2 mm, slice gap = 2 mm). DANTE parameters: flip angle = 8°, number of pulses = 120, time duration between pulses = 0.5 ms, Gz = 18 mT/m and gradient duration = 0.4 ms. T2 maps were generated by monoexponential fitting using non-linear least-squares regression. A reviewer (L.B.) classified plaque types following the CMR-modified AHA scheme using T2 maps + TOF only. Histology: Carotid plaques were freshly obtained at the time of endarterectomy, divided at the level of maximal stenosis into two 4 mm-segments and processed for formalin-fixed paraffin-embedded (FFPE) sections and cryosections, respectively. FFPE sections were stained for H&E, Masson's trichrome, while cryosections were used for Oil-Red-O/adipophilin (foam cells marker) staining to visualize lipid. AHA type of plaque histology was determined by a reviewer (J.T.C.) blinded to the T2 map results. CMR-histology slice location matching was performed for each 4 mm segments using the carotid bifurcation as the common anatomical landmark and cross-sectional T1-weighted images of the vessel wall.

Results
6 of the 19 patients scheduled for endarterectomy were excluded due to severe patient motion artefacts on CMR. The table 1 presents the AHA type classification of the remaining 13 plaques using T2 maps + TOF vs.  histology. The 2 cases of misclassification were due to the difficulty of staging intraplaque haemorrhage (IPH) accurately. T2 maps were able to differentiate lipid-rich necrotic core, fibrous tissue, calcification and recent IPH, as illustrated in the Figure 1 for a type VI plaque.

Conclusions
These preliminary results show the potential of in-vivo T2 mapping for atherosclerotic plaque characterization. The ability of T2 maps to discriminate plaque components was confirmed by histology.

Funding
This study was supported by EPSRC and Oxford Biomedical Research Centre.