Computed Tomographic Attenuation Value of Coronary Atherosclerotic Plaques With Different Tube Voltage

Abstract

To compare the diagnostic performance of computed tomographic (CT) attenuation and CT attenuation ratio at different tube voltages for ex vivo plaque characterization. Human coronary arteries were obtained at the time of autopsy in 15 subjects. The coronary arteries were serially cut into 5-mm-long segments and scanned ex vivo using 4 sets of tube voltages and tube currents (80 kV, 660 mA; 100 kV, 500 mA; 120 kV, 400 mA; and 140 kV, 340 mA). The CT attenuation value at the center of each plaque was obtained, and the ratio of the CT attenuation value at the 80-kV setting divided by that at the 140-kV setting (Hounsfield ratio [HR], 80:140) was calculated. Separate receiver operating characteristic (ROC) analyses were used to assess the usefulness of the CT attenuation value and the 80:140 HR for the differential diagnosis of lipid-rich plaques from other types of plaques. A total of 93 coronary plaques were detected macroscopically. Histological examination revealed 39 lipid-rich, 24 calcified, and 30 fibrotic plaques. At all the tube voltages, the CT attenuation values of the lipid-rich plaques were lower than those of the calcified plaques, whereas the CT attenuation values of the lipid-rich and fibrotic plaques overlapped. An ROC analysis showed that the area under the curve (AUC) for the differential diagnosis of lipid-rich plaques from fibrotic plaques was 0.813 at 80 kV, 0.772 at 100 kV, 0.682 at 120 kV, or 0.651 at 140 kV. Regarding the 80:140 HR, the AUC was 0.952 (0.029). The AUC was significantly larger at 80 and 100 kV and 80:140 HR compared with the AUC at 120 kV. The diagnostic performance of CT analysis for ex vivo plaque characterization was superior at lower energy settings and using the dual-energy method.