Characterization of coronary atherosclerosis by dual-source computed tomography and HU-based color mapping: a pilot study

Abstract

To assess HU-based color mapping for characterization of coronary plaque, using intravascular ultrasound virtual histology (IVUS-VH) as a standard of reference. Dual-source computed tomography and IVUS-VH were prospectively performed in 13 patients. In five lesions, HU thresholds of the color-coding software were calibrated to IVUS-VH. In a 15-lesion verification cohort, volumes of vessel, lumen and plaque or percentages of lipid, fibrous and calcified components were obtained through use of pre-set HU cut-offs as well as through purely visual adjustment of color maps. Calibrated HU ranges for fatty or fibrous plaque, lumen and calcification were -10-69, 70-158, 159-436 and 437+. Using these cut-offs, HU-based analysis achieved good agreement of plaque volume with IVUS (47.0 vs. 51.0 mm(3)). Visual segmentation led to significant overestimation of atheroma (61.6 vs. 51.0 mm(3); P = 0.04) Correlation coefficients for volumes of vessel, lumen and plaque were 0.92, 0.87 and 0.83 with HU-based analysis or 0.92, 0.85 and 0.71 with visual evaluation. With both methods, correlation of percentage plaque composition was poor or insignificant. HU-based plaque analysis showed good reproducibility with intra-class correlation coefficients being 0.90 for plaque volume and 0.81, 0.94 or 0.98 for percentages of fatty, fibrous or calcified components. With use of optimized HU thresholds, color mapping allows for accurate and reproducible quantification of coronary plaque.

Fig. 1

Schematic description of the data analysis procedure

Fig. 2

Illustration of plaque analysis tool. HU distribution in plaque and vessel is shown with thresholds for fatty (dark green), fibrous (light green) and calcified atheroma (purple). Modification of thresholds results in change of plaque volume and volumes of its various components

Fig. 3

Bland-Altman analysis comparing total plaque volumes as obtained with DSCT and IVUS. Data are shown for visual adjustment of color maps to plaque (a) and evaluation based on HU thresholds (b). The X axis denotes the average of DSCT and IVUS; the point of intersection with the Y axis indicates the bias of DSCT. Dotted lines show 1 standard deviation of the bias

Fig. 4

Correlation of total plaque volumes in DSCT with IVUS. Data are shown for visual adjustment of color maps to plaque (a) and evaluation based on HU thresholds (b)

Fig. 5

Linear regression plot of optimal IVUS calibrated lower and upper HU cut-offs for lumen against attenuation in the distinct vessel segment

Fig. 6

Virtual plaque histology in DSCT and IVUS. In an axial MIP, the arrow indicates a non-calcified plaque in segment 6 of the LAD (a). Cross-sectional images of the vessel with color coding are presented for visual adjustment of the maps (b) or HU-based segmentation (c) as well as for grey-scale IVUS (d) and IVUS-VH (e). For IVUS-VH, red indicates necrotic core, light green the fatty-fibrous, dark green the fibrous and white the calcified plaque. In this lesion, total plaque volumes with visual CT evaluation, HU-based analysis or IVUS were 66.6 mm³, 55.8 mm³ or 53.1 mm³ respectively