Assessment of Low-Contrast Resolution for the American College of Radiology Computed Tomographic Accreditation Program: What Is the Impact of Iterative Reconstruction?
Assessment of Low-Contrast Resolution for the American College of Radiology Computed Tomographic Accreditation Program: What Is the Impact of Iterative Reconstruction?
复制标题
DOI：
10.1097/rct.0000000000000245
复制
发表时间：
2015-07-01
影响因子：
1.3
通讯作者：
McCollough, Cynthia H.
中科院分区：
文献类型：
--
作者：
Kofler, James M.;Yu, Lifeng;McCollough, Cynthia H.
目的比较滤波反投影 (FBP) 和迭代重建 (IR) 计算机断层扫描 (CT) 图像中的对比度噪声比 (CNR) 阈值与低对比度分辨率 (LCR) 的视觉评估。方法使用 2 个扫描仪模型在 8、12 和 16 mGy 的 CTDIvol 水平下采集美国放射学会 (ACR) CT 认证体模 LCR 图像，并使用 1 个 FBP 和 2 个 IR 进行重建内核。采集重复100次。三名经过委员会认证的医学物理学家盲目审查了 LCR 切片图像。比较了以前和当前 ACR CT 认证标准的通过百分比 (PPR)。结果 FBP 图像的观察者 PPR 低于 32%。对于红外图像，18 种设置/剂量/模型配置中的 5 种 PPR 大于 32%（最大 76.3%）。对于 FBP 图像的 CNR 评估，15 种配置的 PPR 大于 70%。对于红外图像，所有 PPR 至少为 96%。结论 ACR CT 认证计划使用的 CNR 阈值比 LCR 的视觉评估产生更高的 PPR，可能导致通过 ACR CNR 标准的图像质量较低。
ObjectiveTo compare contrast-to-noise ratio (CNR) thresholds with visual assessment of low-contrast resolution (LCR) in filtered back projection (FBP) and iteratively reconstructed (IR) computed tomographic (CT) images.MethodsAmerican College of Radiology (ACR) CT accreditation phantom LCR images were acquired at CTDIvol levels of 8, 12, and 16 mGy using 2 scanner models and reconstructed using one FBP and 2 IR kernels. Acquisitions were repeated 100 times. Three board-certified medical physicists blindly reviewed the LCR section images. Pass-percentage rates (PPRs) using previous and current ACR CT accreditation criteria were compared.ResultsObserver PPRs for FBP images were less than 32%. For IR images, 5 of 18 settings/dose/model configurations had PPRs greater than 32% (maximum 76.3%). For CNR evaluation of FBP images, PPRs for 15 configurations were greater than 70%. For IR images, all PPRs were at least 96%.ConclusionsThe CNR threshold used by the ACR CT accreditation program yields higher PPRs than visual assessment of LCR, potentially resulting in lower-quality images passing the ACR CNR criteria.