Hanai Kouzou, Horiuchi Tetsuya, Sekiguchi Junko, Muramatsu Yoshihisa, Kakinuma Ryutaro, Moriyama Noriyuki, Tuchiiya Ryosuke, Niki Noboru
Division of Diagnosis Radiology, National Hospital Organization Kanagawa Hospital, Kanagawa, Japan.
J Comput Assist Tomogr. 2006 Nov-Dec;30(6):955-61. doi: 10.1097/01.rct.0000230011.16468.0e.
The purpose of this study is to assess the relative influence of noise and artifact in detecting lung nodules on low dose computed tomographic (CT) screening.
We develop the computer-simulation technique that allows tube current simulation and virtual nodule insertion in any CT images. The tube current simulation uses a reduction model that adds random Gaussian noise distribution to existing projection data. The virtual nodules are generated using a dedicated CT simulation tool with same scanner geometry.
The coefficient of the correlations between the contrast-to-noise ratio of the actual scan and simulated tube current images was 0.98. There was no difference in CT number between virtual nodules and actual nodules [t test results = 0.60, t50(0.01) = 2.70 at 10 mA] and the coefficient of the correlations of the image noise was 0.99.
Our technique is useful for systematic evaluation of radiation dose reduction and structure visibility in low-dose CT screening.
本研究旨在评估噪声和伪影对低剂量计算机断层扫描（CT）筛查中肺结节检测的相对影响。
我们开发了一种计算机模拟技术，该技术允许在任何CT图像中进行管电流模拟和虚拟结节插入。管电流模拟使用一种还原模型，该模型将随机高斯噪声分布添加到现有的投影数据中。虚拟结节使用具有相同扫描仪几何形状的专用CT模拟工具生成。
实际扫描与模拟管电流图像的对比度噪声比之间的相关系数为0.98。虚拟结节与实际结节之间的CT值无差异[t检验结果 = 0.60，在10 mA时t50(0.01) = 2.70]，图像噪声的相关系数为0.99。
我们的技术有助于系统评估低剂量CT筛查中的辐射剂量降低和结构可见性。
