Department of Diagnostic Radiology, Keio University School of Medicine, Shinjuku-ku, Tokyo, Japan.
Invest Radiol. 2012 May;47(5):292-8. doi: 10.1097/RLI.0b013e318240a874.
The purpose of this study was to select the optimal monochromatic level for virtual monochromatic spectral (VMS) imaging to minimize the image noise of the liver parenchyma and to acquire a high contrast-to-noise ratio (CNR) of hypovascular hepatic metastases in the portal-dominant phase.
This study was conducted with the approval of our institutional review board, and written informed consent was obtained from all the participating patients. Ninety patients with hepatic metastases were scanned by fast kilovoltage switching dual-energy computed tomography in the portal-dominant phase. One hundred one sets of VMS images in the range of 40 to 140 keV at 1-keV intervals were reconstructed. The image noise of the liver parenchyma in each patient and the CNR of each metastasis (n = 303) were measured on all the 101 VMS image sets. Data were analyzed by the paired t test and mixed-effects model. Bonferroni's method was used for multiple comparisons.
The lowest noise of the liver parenchyma was obtained in 6, 15, 31, 29, 7, 1, and 1 patient at 67, 68, 69, 70, 71, 72, and 73 keV, respectively. The mean noise of the liver parenchyma on the 69-keV VMS images in all 90 patients was significantly lower than that on the 67-, 68-, 71-, 72-, and 73-keV VMS images (P < 0.001); however, there was no significant difference in the mean noise of the liver parenchyma between the 69-keV and 70-keV VMS images (P = 0.279). For 95% of the hepatic metastases (288/303 metastases), the highest metastasis-to-liver CNR was obtained in the 66- to 71- keV VMS images. The mean metastasis-to-liver CNR for the 303 metastases was numerically highest at 68 keV; however, there was no significant difference in the mean metastasis-to-liver CNR between the 68-keV and 69-keV images (P = 0.737) or between the 68-keV and 70-keV images (P = 0.103).
VMS imaging at approximately 70 keV (69-70 keV) yielded the lowest image noise of the liver parenchyma and a high CNR for hypovascular hepatic metastases in the portal-dominant phase.
本研究旨在选择最佳的虚拟单能量光谱（VMS）成像单能量水平，以最小化肝脏实质的图像噪声，并在门静脉期获得低血供肝转移的高对比噪声比（CNR）。
本研究经机构审查委员会批准，并获得所有参与患者的书面知情同意。90 例肝转移患者在门静脉期行快速千伏切换双能 CT 扫描。以 1keV 为间隔重建 40 至 140keV 范围内的 101 组 VMS 图像。在所有 101 组 VMS 图像上测量每位患者的肝脏实质图像噪声和每个转移灶的 CNR（n=303）。采用配对 t 检验和混合效应模型进行数据分析。采用 Bonferroni 法进行多重比较。
分别在 67、68、69、69、70、71 和 73keV 下，有 1、6、15、31、29、7 和 1 位患者的肝脏实质噪声最低。90 例患者的 69keV VMS 图像肝脏实质平均噪声明显低于 67、68、71、72 和 73keV VMS 图像（P<0.001）；然而，69keV 和 70keV VMS 图像的肝脏实质平均噪声无显著差异（P=0.279）。95%（288/303 个转移灶）的肝转移灶与肝脏的 CNR 最高出现在 66-71keV VMS 图像中。303 个转移灶的平均转移灶与肝脏 CNR 在 68keV 时数值最高；然而，68keV 和 69keV 图像之间（P=0.737）或 68keV 和 70keV 图像之间（P=0.103）的平均转移灶与肝脏 CNR 无显著差异。
在门静脉期，VMS 成像约 70keV（69-70keV）可获得最低的肝脏实质图像噪声和低血供肝转移的高 CNR。
