| Objective: To estimate the image quality and radiation dose of arterial and portal phase of dual-energy virtual non-contrast(VNC) and investigate advantages and limitations of dual-source CT dual-energy virtual non-contrast in the upper abdomen. Methods: Abdominal CT conventional and enhanced scans were performed in 200 patients with suspected abdominal lesions by using dual-source CT(Siemens Somatom definition flash).The conventional non-contrast(CNC) using the single energy mode(120k V / 350 m As), while the arterial and portal phase using dual-energy mode(100k V / 230 m As and Sn140 k V / 178 m As). Arterial and portal phase virtual non-contrast images were acquired by sending the reconstruction datum(1.5mm in thickness and 1mm in interval) of the arterial and portal phase enhancement images to an application procedure(Liver VNC) available in the workstation(Dual Energy), and changed their window width and window level to bring them correspondence with the conventional scan. Then got the arterial and portal phase VNC images by rebuilting them into images that both thickness and interval are 6mm as CNC through 3D software. Two experienced radiologist were invited to assess the image quality and calculate the number of detected lesions of the three groups of non-contrast images. Then a resident doctor was asked to describe their ROI, measure and record CT values, calculate signal to noise ratio(SNR) and radiation dose. Then their image quality scores, average CT values, SNR, detection rate of lesions and radiation dose were compared, and P<0.05 was considered statistically significant. Results: The image quality score of arterial phase VNC(3.93±0.79) and portal phase VNC(4.01±0.77) was lower than that of CNC(4.43±0.64)(P<0.05), but no significant differences were observed between arterial phase VNC and portal phase VNC(P>0.05). CT value and SNR of the liver was significantly higher in arterial phase VNC and portal phase VNC than in CNC(P<0.05). while the result was significantly opposite in gallbladder, pancreas, kidney, aorta and erector muscle of spine(P<0.05). A signigicant difference was also observed between arterial phase VNC and portal phase VNC in CT value and SNR of the kidney and aorta(P<0.05). No significantly difference was observed in the lesion detection rate between VNC and CNC(P>0.05). In the 200 subjects, only one case VNC image was 2 points in the image quality. There were 7 cases with image artifacts in arterial phase VNC images, and 2 of them were affected in the lesion detection. There were 3 cases with image artifacts in portal phase VNC images, but they were affected in the lesion detection. The arterial phase VNC images of artifacts significantly higher than the portal phase. Arterial phase VNC images can be affected by inhomogeneous iodine subtraction especially in the heart and venae cava inferior. Radiation dose was significantly lower in single enhanced scan than in CNC(P<0.05). No significantly was found in the radiation dose between arterial phase VNC and portal phase VNC(P>0.05). The radiation dose could reduced 38.73% without CNC. Conclusion: Although image quality of VNC is lower than CNC, and CT value and SNR are different in VNC and CNC, but there is no statistical difference in diagnosis accuracy, besides radiation dose can be reduced by using VNC technology, which make it own potential application value. In addition to, the iodine removed more completely in portal phase VNC compared with the arterial phase VNC. Besides, its image quality score and detection rate of lesions are closer to CNC. We recommend that the portal phase are fit for conducting virtual non-contrast. |
PDF Full Text Request