Dual-energy computed tomography angiography: virtual calcified plaque subtraction in a vascular phantom

Background Material decomposition of dual-energy computed tomography (DECT) enables subtraction of calcified plaque. Purpose To evaluate the accuracy of lumen area measurement in calcified plaque by subtraction of DECT and to determine the effect of contrast material concentration, lumen diameter, density, and thickness of calcified plaque for the measurement. Material and Methods Vessel phantoms were made with six lumen diameters (5.7, 4.9, 3.9, 3.0, 1.9, and 1.3 mm) and six types of calcified plaques with three densities and two thicknesses were attached. CT scans were performed with three contrast material concentrations (62, 111, and 170 mg iodine/mL). Lumen area discrepancy (AD) was calculated by subtracting the measured lumen area from a reference value. The lumen area underestimation percentage (AU), defined as (AD/reference value) × 100, was calculated. General linear model analysis was used to test the effect of variables for log-transformed AU (ln_AU). Results The AD and AU was calculated to be 6.1 ± 4.8 mm2 and 69.8 ± 29.4%, respectively. Ln_AU was significantly affected by contrast material concentration (P < 0.001), calcium density (P = 0.001), plaque thickness (P = 0.010), and lumen diameter (P < 0.001). Ln_AU was significantly higher in 62 mg iodine/mL than in 111 or 170 mg iodine/mL (P < 0.001 for both). Ln_AU was significantly lower at a lumen diameter of 5.7 mm than 3.9 mm (P = 0.001) or 3.0 (P < 0.001). Conclusion Calcified plaque subtraction in DECT substantially underestimates measurements of lumen area. Higher enhancement in larger vessels ensures more accurate subtraction of calcified plaque.