부산대학교 도서관

PURPOSE: To evaluate the agreement and correlation between attenuation values and vascular calcification volume for intra-abdominal structures from true noncontrast (TNC) images and those from virtual noncontrast (VNC) images obtained by dual-source dual-energy computed tomography (CT) using a quadriphasic dynamic protocol. METHODS: Seventy-six patients who underwent quadriphasic abdominal CT were retrospectively reviewed. An arterial, portal venous, and 5-minute delayed phase postcontrast series was obtained using dual-source dual-energy CT. Virtual noncontrast images were processed from the arterial, portal venous, and 5-minute delayed phase series. Attenuation values for the liver, pancreas, kidneys, muscle, fat, vertebra, gallbladder, aorta, inferior vena cava, portal vein, and aortic calcification volumes were recorded. Attenuation values for the liver, pancreas, vertebra, and muscle obtained from VNC were adjusted using linear regression. RESULTS: Repeated-measures analysis of variance and Bonferroni multiple-comparison post hoc correction revealed significant differences between TNC and VNC attenuation values for the organs. There was an excellent correlation between the TNC and VNC attenuation values for the liver, pancreas, muscle, vertebra, and fat. The calcification volume was significantly smaller on VNC than on TNC. The adjusted attenuation values estimated by regression equations afforded better approximation. CONCLUSIONS: Abdominal VNC images obtained with third-generation dual-source dual-energy CT cannot replace TNC images without adjustment of the attenuation values.