Coronary calcified plaque imaging using dual-energy CT: a phantom study / 中华放射学杂志

ABSTRACT

Objective To analyze the impact of dual energy monochromatic reconstructions (50-160 keV) on coronary calcified plaque stenosis quantification in a cardiac phantom with the real stenosis as standard of reference.Methods Signal-to-noise(SNR)and contrast-to-noise ratios(CNR)were calculated. In conventional 120 kV coronary CT angiography (CCTA) images as well as 12 monochromatic series(50-160 keV),luminal narrowing of calcified plaque was measured and compared to the real stenosis. Bland-Altman plots were performed to analyze the correlation of reference standard with conventional 120 kV and 12 monochromatic series (50-160 keV). The sensitivity, specificity, and accuracy of CCTA for detection of≥50% stenosis were calculated and receiver operating characteristic(ROC)curve was analyzed with the real stenosis as reference standard.Results The SNR and CNR were decreased with the increase of keV from 70 keV.The SNR and CNR of monochromatic imaging were lower compared with conventional 120 kV images(SNR:76.4±16.1;CNR:274.7±54.1)(P<0.05).The Bland-Altman plots presented a smaller measurement bias towards 90-160 keV than conventional 120 kV, and smallest measurement bias was revealed in 100-130 keV imaging(100-130 keV:bias 17.2% vs 120 kV:21.4%).Using the stenosis≥50% as cut-off value, the specificity were higher in the monochromatic series (70-160 keV) than conventional 120 kV CCTA.The specificity(75.0%)were the highest in the 120 keV and 130 keV images.The area under the ROC curve(AUC)of 90-130 keV(0.991-0.995)was higher than that in conventional 120 kV imaging (0.990). Conclusions The use of monochromatic imaging improves the overall accuracy of stenosis evaluation in coronary calcified plaques. Reconstructions at 100-130 keV for calcified lesions yielded the optimal results.