基于深度学习的心肌延迟强化后处理 CT 去噪。
Deep Learning-based Post Hoc CT Denoising for Myocardial Delayed Enhancement.
机构信息
From the Department of Radiology, National Cerebral and Cardiovascular Center, 6-1 Kishibe-shinmachi, Suita 564-8565, Japan (T.N., T.K., H.T., A.K., Y.O., Y.M., H.H., T.F.); Department of Medical Physics and Engineering, Graduate School of Medicine, Osaka University, Suita, Japan (T.K., K.U., J.O., T.I.); Medical Informatics Section, QST Hospital (K.U., J.O.), and Applied MRI Research, Department of Molecular Imaging and Theranostics, Institute for Quantum Medical Science (K.U., J.O.), National Institutes for Quantum Science and Technology, Chiba, Japan.
出版信息
Radiology. 2022 Oct;305(1):82-91. doi: 10.1148/radiol.220189. Epub 2022 Jun 28.
Background To improve myocardial delayed enhancement (MDE) CT, a deep learning (DL)-based post hoc denoising method supervised with averaged MDE CT data was developed. Purpose To assess the image quality of denoised MDE CT images and evaluate their diagnostic performance by using late gadolinium enhancement (LGE) MRI as a reference. Materials and methods MDE CT data obtained by averaging three acquisitions with a single breath hold 5 minutes after the contrast material injection in patients from July 2020 to October 2021 were retrospectively reviewed. Preaveraged images obtained in 100 patients as inputs and averaged images as ground truths were used to supervise a residual dense network (RDN). The original single-shot image, standard averaged image, RDN-denoised original (DL) image, and RDN-denoised averaged (DL) image of holdout cases were compared. In 40 patients, the CT value and image noise in the left ventricular cavity and myocardium were assessed. The segmental presence of MDE in the remaining 40 patients who underwent reference LGE MRI was evaluated. The sensitivity, specificity, and accuracy of each type of CT image and the improvement in accuracy achieved with the RDN were assessed using odds ratios (ORs) estimated with the generalized estimation equation. Results Overall, 180 patients (median age, 66 years [IQR, 53-74 years]; 107 men) were included. The RDN reduced image noise to 28% of the original level while maintaining equivalence in the CT values ( < .001 for all). The sensitivity, specificity, and accuracy of the original images were 77.9%, 84.4%, and 82.3%, of the averaged images were 89.7%, 87.9%, and 88.5%, of the DL images were 93.1%, 87.5%, and 89.3%, and of the DL images were 95.1%, 93.1%, and 93.8%, respectively. DL images showed improved accuracy compared with the original images (OR, 1.8 [95% CI: 1.2, 2.9]; = .011) and DL images showed improved accuracy compared with the averaged images (OR, 2.0 [95% CI: 1.2, 3.5]; = .009). Conclusion The proposed denoising network supervised with averaged CT images reduced image noise and improved the diagnostic performance for myocardial delayed enhancement CT. © RSNA, 2022 See also the editorial by Vannier and Wang in this issue.
背景 为了改善心肌延迟强化（MDE）CT，开发了一种基于深度学习（DL）的后处理去噪方法，该方法由平均 MDE CT 数据进行监督。目的 评估去噪后的 MDE CT 图像的图像质量，并通过使用晚期钆增强（LGE）MRI 作为参考来评估其诊断性能。材料与方法 回顾性分析了 2020 年 7 月至 2021 年 10 月期间，在注射对比剂后 5 分钟内单次屏气采集的患者的 MDE CT 数据，取 3 次采集的平均值。将 100 例患者的预平均图像作为输入，平均图像作为ground truth 用于监督残差密集网络（RDN）。比较了来自留取病例的原始单幅图像、标准平均图像、RDN 去噪原始图像（DL）和 RDN 去噪平均图像（DL）。在 40 例患者中，评估了左心室腔和心肌的 CT 值和图像噪声。对其余 40 例接受参考 LGE MRI 的患者进行了节段性 MDE 评估。使用广义估计方程估计的比值比（OR）评估每种 CT 图像的敏感性、特异性和准确性，以及 RDN 提高的准确性。结果 共有 180 例患者（中位年龄 66 岁[IQR，53-74 岁]；107 例男性）纳入研究。RDN 将图像噪声降低至原始水平的 28%，同时保持 CT 值的等效性（均<.001）。原始图像的敏感性、特异性和准确性分别为 77.9%、84.4%和 82.3%，平均图像分别为 89.7%、87.9%和 88.5%，DL 图像分别为 93.1%、87.5%和 89.3%，DL 图像分别为 95.1%、93.1%和 93.8%。与原始图像相比，DL 图像的准确性提高（OR，1.8[95%CI：1.2，2.9]； =.011），与平均图像相比，DL 图像的准确性提高（OR，2.0[95%CI：1.2，3.5]； =.009）。结论 该研究提出的基于平均 CT 图像监督的去噪网络降低了图像噪声，提高了心肌延迟强化 CT 的诊断性能。
Zotero 插件