Abstract：
BACKGROUND: Prior studies focused on utilization of dual-energy computed tomography (DECT) to better detect intracranial pathology and to reduce artifacts. It is still unclear whether virtual non-contrast (VNC) images of DECT can replace true non-contrast (TNC) images.
OBJECTIVE: To compare attenuation values and image quality of VNC images to TNC images of the brain, obtained using spectral detector CT (SDCT).
METHODS: We retrospectively evaluated patients that underwent head CT with and without contrast material, on a SDCT scanner at our institution (n = 33). The attenuation values of different brain structures were obtained from TNC images, the conventional images of the post-contrast exams (n = 16) or the CT angiography (CTA) (n = 17), and the derived VNC images. In total, 591 regions of interest were obtained, including white and gray matter. Two neuroradiologists independently evaluated the image quality of the VNC and TNC images, using a 5-point Likert scale.
RESULTS: The mean difference between the attenuation values on the VNC versus the TNC images was <4 HU for almost all the structures. The difference reached statistical significance (P < 0.05) for the deep gray structures but not for the white matter. The image quality score of the TNC images was 5 in all the patients (excellent gray-white matter differentiation). The scores of the VNC images differed between post-contrast and CTA examinations, with means of 4.9 ± 0.3 (excellent) and 3.2 ± 0.4 (fair), respectively (P < 0.001).
CONCLUSIONS: Our results show minor differences between attenuation values of different brain structures on VNC versus TNC images of SDCT.
OBJECTIVE: To compare attenuation values and image quality of VNC images to TNC images of the brain, obtained using spectral detector CT (SDCT).
METHODS: We retrospectively evaluated patients that underwent head CT with and without contrast material, on a SDCT scanner at our institution (n = 33). The attenuation values of different brain structures were obtained from TNC images, the conventional images of the post-contrast exams (n = 16) or the CT angiography (CTA) (n = 17), and the derived VNC images. In total, 591 regions of interest were obtained, including white and gray matter. Two neuroradiologists independently evaluated the image quality of the VNC and TNC images, using a 5-point Likert scale.
RESULTS: The mean difference between the attenuation values on the VNC versus the TNC images was <4 HU for almost all the structures. The difference reached statistical significance (P < 0.05) for the deep gray structures but not for the white matter. The image quality score of the TNC images was 5 in all the patients (excellent gray-white matter differentiation). The scores of the VNC images differed between post-contrast and CTA examinations, with means of 4.9 ± 0.3 (excellent) and 3.2 ± 0.4 (fair), respectively (P < 0.001).
CONCLUSIONS: Our results show minor differences between attenuation values of different brain structures on VNC versus TNC images of SDCT.
摘要：
背景：先前的研究集中于利用双能计算机断层扫描（DECT）来更好地检测颅内病理并减少伪影。目前尚不清楚DECT的虚拟非对比（VNC）图像是否可以替代真实的非对比（TNC）图像。
目的：比较大脑的VNC图像和TNC图像的衰减值和图像质量，使用光谱探测器CT(SDCT)获得。
方法：我们对接受头颅CT检查的患者进行了回顾性评估，在我们机构的SDCT扫描仪上(n=33)。从TNC图像中获得不同大脑结构的衰减值，造影后检查（n=16）或CT血管造影（CTA）（n=17）的常规图像，和导出的VNC图像。总的来说,获得了591个感兴趣的区域，包括白色和灰色物质。两名神经放射科医生独立评估了VNC和TNC图像的图像质量，用5分的李克特量表.
结果：对于几乎所有结构，VNC与TNC图像上的衰减值之间的平均差为<4HU。深层灰色结构差异有统计学意义（P<0.05），而白质无统计学意义。所有患者的TNC图像图像质量评分为5分(灰白质分化优秀)。对比后和CTA检查的VNC图像得分不同，平均值为4.9±0.3(优秀)和3.2±0.4(一般)，分别(P<0.001)。
结论：我们的结果显示，SDCT的VNC与TNC图像上不同脑结构的衰减值之间存在微小差异。
目的：比较大脑的VNC图像和TNC图像的衰减值和图像质量，使用光谱探测器CT(SDCT)获得。
方法：我们对接受头颅CT检查的患者进行了回顾性评估，在我们机构的SDCT扫描仪上(n=33)。从TNC图像中获得不同大脑结构的衰减值，造影后检查（n=16）或CT血管造影（CTA）（n=17）的常规图像，和导出的VNC图像。总的来说,获得了591个感兴趣的区域，包括白色和灰色物质。两名神经放射科医生独立评估了VNC和TNC图像的图像质量，用5分的李克特量表.
结果：对于几乎所有结构，VNC与TNC图像上的衰减值之间的平均差为<4HU。深层灰色结构差异有统计学意义（P<0.05），而白质无统计学意义。所有患者的TNC图像图像质量评分为5分(灰白质分化优秀)。对比后和CTA检查的VNC图像得分不同，平均值为4.9±0.3(优秀)和3.2±0.4(一般)，分别(P<0.001)。
结论：我们的结果显示，SDCT的VNC与TNC图像上不同脑结构的衰减值之间存在微小差异。
