Abstract 681: Three Dimensional Color-coded Maps of Multi-detector Row Computed Tomography for Tissue Characterization of Coronary Plaques
Background: Multi-detector row Computed Tomography (MDCT) permits non-invasive assessment of the degree of coronary stenosis. However, tissue characterization of coronary plaques using MDCT has not been established well.
Methods: One hundred fifty sites in 50 patients were visualized by both MDCT and integrated backscatter intravascular ultrasound (IB-IVUS) at the same site. To determine the threshold for discrimination of plaque components, we compared CT values of coronary plaques with IB values as a gold standard. Optimal thresholds between calcification and fibrosis, and between fibrosis and lipid pools were determined from the analysis of receiver operating characteristic (ROC) curves. After determining the thresholds, we constructed three dimensional (3D) color-coded maps of MDCT for plaque tissue characterization.
Results: Plaques were classified as lipid pool (n =50), fibrosis (n =65) and calcification (n = 35) by IB-IVUS. Hounsfield density of lipid pool, fibrosis and calcification were 17±17 HU (range: −19 to 48 HU), 96±23 HU (51 to 154 HU) and 378±99 HU (188 to 605 HU), respectively. Using the ROC curve, threshold of 50 HU and 180 HU were the most reliable predictor of lipid pools, fibrosis and calcification. %lipid and %fibrosis volume measured by 3D-MDCT Color-coded map was correlated with % lipid and %fibrosis volume measured 3D-IVUS (r=0.76, P<0.001 and r=0.404,P=0.01).
Conclusion: 3D color-coded maps of MDCT were promising methods for non-invasive volumetric assessment for tissue characterization of coronary plaques.