From the Department of Radiology (A.M., A.B., D.M., S.S.S., R.C.N.) and Carl E. Ravin Advanced Imaging Laboratories (K.R.C., J.M.W.), Duke University Medical Center, Box 3808, Erwin Rd, Durham, NC 27710.
Radiology. 2016 Apr;279(1):269-77. doi: 10.1148/radiol.2015150919. Epub 2015 Nov 4.
To determine the variance in virtual monochromatic computed tomography (CT) numbers from the same lesion, comparing the two clinically available dual-energy multidetector CT hardware implementations (single-source projection-based and dual-source image-based), in a phantom-based simulated abdominal environment.
This phantom-based study was exempt from institutional review board oversight. Polyethylene terephthalate spheres (15 and 18 mm) with two iodine-to-saline dilutions (0.8 and 1.2 mg of iodine per millilliter) were serially suspended in a cylindrical polypropylene bottle filled with diluted iodinated contrast material. The bottle was placed into a 36-cm-wide torso-shaped water phantom simulating the abdomen of a medium-sized patient. Dual-energy (80/140 kVp) and single-energy (100 and 120 kVp) scans were obtained with single-source and dual-source multidetector CT implementations. Virtual monochromatic images were reconstructed at energy levels of 40-140 keV (in 10-keV increments) in either the projection-space or image-space domain. A multivariate regression analysis approach was used to investigate the effect of energy level, lesion size, lesion iodine content, and implementation type on measured CT numbers.
There were significant differences in the attenuation values measured in the simulated lesions with the single-source projection-based platform and the dual-source image-based implementation (P < .001 for all comparisons). The magnitude of these differences was greatest at lower monochromatic energy levels and at lower iodine concentrations (average difference at 40 keV: 25.7 HU; average difference at 140 keV: 7 HU). The monochromatic energy level and the lesion iodine concentration had a significant effect on the difference in the measured attenuation values between the two implementations, which indicates that the two imaging platforms respond differently to changes in investigated variables (P < .001 for all comparisons).
There is a statistically significant variance in virtual monochromatic CT numbers from the same lesion examined with single-source projection-based and dual-source image-based implementations. The magnitude of the variance is a function of the selected energy level and the lesion iodine content.
在模拟腹部环境的体模中，比较两种临床可用的双能多排 CT 硬件实现方式（单源投影式和双源图像式），确定同一病灶的虚拟单能 CT 数值的差异。
本基于体模的研究获得机构审查委员会豁免。将聚对苯二甲酸乙二醇酯球（15 和 18 毫米）与两种碘盐稀释液（0.8 和 1.2 毫克碘/毫升）串联悬浮在一个充满稀释碘对比剂的圆柱形聚丙烯瓶中。将该瓶置于一个 36 厘米宽的躯干形状水模中，模拟中等体型患者的腹部。使用单源和双源多排 CT 实现方式，进行双能（80/140 kVp）和单能（100 和 120 kVp）扫描。在投影空间或图像空间域，以 10 keV 为增量，在 40-140 keV 的能量水平上重建虚拟单能图像。采用多元回归分析方法，研究能量水平、病灶大小、病灶碘含量和实现方式对测量 CT 值的影响。
在单源投影式平台和双源图像式实现方式中，模拟病灶的衰减值存在显著差异（所有比较的 P 值均<.001）。在较低的单能能量水平和较低的碘浓度时，这些差异的幅度最大（40 keV 时的平均差异：25.7 HU；140 keV 时的平均差异：7 HU）。单能能量水平和病灶碘浓度对两种实现方式之间测量衰减值的差异有显著影响，这表明两种成像平台对所研究变量的变化有不同的响应（所有比较的 P 值均<.001）。
使用单源投影式和双源图像式实现方式检查同一病灶时，虚拟单能 CT 数值存在具有统计学意义的差异。差异的幅度是所选能量水平和病灶碘含量的函数。
