All CT datasets were transferred to a commercially available workstation for analysis (Advantage Windows version 4.6; GE Healthcare). Subsequently, two experienced radiologists (with 25 and 5 years of abdominal CT experience, respectively), who were blinded to clinical data and pathological results, analyzed the images. According to a previous study (18 (link)), VMS images captured at 70 keV using 40% adaptive statistical reconstruction techniques exhibited decreased levels of image noise and high contrast-noise-ratio compared with 120 kVp images. The IC value (measured in 100 µg/ml, throughout the text) was then measured using axial iodine-based MD image at 1.25-mm thickness (Figs. 1A-D and 2A-D). The outline of the region of interest (ROI)-lesions encompassed the largest area of the lesion in 2 or 3 consecutive layers, while avoiding areas of necrosis, vessels, air and fat attenuation. ICs in the AP (ICAP) and the VP (ICVP) were measured separately. Circular ROI was assigned to the abdominal aorta in the same layer as the ROI-lesion for calculating the normalized IC (nIC); nIC=IClesion/ICaorta (11 (link)). All of the ICAP/VP and nICAP/VP data obtained from the ROIs of the same tumor were averaged and recorded (Table II).