Abstract

BACKGROUND: Multidetector computed tomography (MDCT) has recently emerged as a potential noninvasive alternative for high-resolution imaging of coronary arteries.
OBJECTIVE: In this study, we evaluated 64-slice MDCT for detection, quantification, and characterization of atherosclerotic plaque burden in nonculprit lesions.
METHODS: Data from 11 patients who underwent both MDCT and intravascular ultrasound (IVUS) for suspected coronary artery disease were collected, and a total of 17 coronary segments and 122 cross-sectional slices were analyzed by MDCT and IVUS. Coronary segments on MDCT were coregistered to IVUS, and each obtained slice was scored by 2 blinded observers for presence and type of plaque. Quantitative measurements included cross-sectional vessel area, lumen area, wall area, plaque volume, and plaque burden. Mean and standard deviation of Hounsfield units (HUs) were recorded for plaque when present.
RESULTS: Overall sensitivity for plaque detection was 95.0%, and specificity, positive predictive value, negative predictive value were 88.7%, 89.1%, and 94.8%, respectively. Spearman's correlation coefficients were 0.85, 0.75, 0.70, 0.89, and 0.54 for cross-sectional vessel area, lumen area, wall area, plaque volume, and plaque burden, respectively. The interobserver variability for plaque burden and plaque volume measurements between readers on 64-MDCT was high at 32.7% and 30.4%, respectively. Combined noncalcified plaque had a mean MDCT density significantly different from that of calcified plaque. Soft and fibrous plaques were not able to be distinguished based on their HU values.
CONCLUSION: Sixty-four-slice MDCT had good correlation with IVUS but with high interobserver variability. Plaque characterization remains a challenge with present MDCT technology.