ABSTRACT
OBJECTIVE: To distinguish components of vulnerable atherosclerotic plaque by imaging their energy response using spectral CT and comparing images with histology. METHODS: After spectroscopic calibration using phantoms of plaque surrogates, excised human carotid atherosclerotic plaques were imaged using MARS CT using a photon-processing detector with a silicon sensor layer and microfocus X-ray tube (50 kVp, 0.5 mA) at 38-µm voxel size. The plaques were imaged, sectioned and re-imaged using four threshold energies: 10, 16, 22 and 28 keV; then sequentially stained with modified Von Kossa, Perl's Prussian blue and Oil-Red O, and photographed. Relative Hounsfield units across the energies were entered into a linear algebraic material decomposition model to identify the unknown plaque components. RESULTS: Lipid, calcium, iron and water-like components of plaque have distinguishable energy responses to X-ray, visible on spectral CT images. CT images of the plaque surface correlated very well with histological photographs. Calcium deposits (>1,000 µm) in plaque are larger than iron deposits (<100 µm), but could not be distinguished from each other within the same voxel using the energy range available. CONCLUSIONS: Spectral CT displays energy information in image form at high spatial resolution, enhancing the intrinsic contrast of lipid, calcium and iron within atheroma. KEY POINTS: Spectral computed tomography offers new insights into tissue characterisation. Components of vulnerable atherosclerotic plaque are spectrally distinct with intrinsic contrast. Spectral CT of excised atherosclerotic plaques can display iron, calcium and lipid. Calcium deposits are larger than iron deposits in atheroma. Spectral CT may help in the non-invasive detection of vulnerable plaques.
