Bridge Christopher P, Best Till D, Wrobel Maria M, Marquardt J Peter, Magudia Kirti, Javidan Cylen, Chung Jonathan H, Kalpathy-Cramer Jayashree, Andriole Katherine P, Fintelmann Florian J
Massachusetts General Hospital and Brigham and Women's Hospital Center for Clinical Data Science (C.P.B., J.K.C., K.P.A.); Martinos Center for Biomedical Imaging, Department of Radiology (C.P.B, K.P.A.); Division of Thoracic Imaging and Intervention (T.D.B., M.M.W., J.P.M., F.J.F.), Department of Radiology, Massachusetts General Hospital; and Department of Radiology, Brigham and Women's Hospital, (K.P.A.), 55 Fruit St, Boston, MA 02114; Charité-Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Department of Radiology, Berlin, Germany (T.D.B.); Department of Radiology, Berlin Institute of Health, Berlin, Germany (T.D.B.); Department of Radiology, Ludwig Maximilian University, Munich, Germany (M.M.W.); Department of Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, Calif (K.M.); Mallinckrodt Institute of Radiology, School of Medicine, Washington University, St Louis, Mo (C.J.); and Departments of Medicine and Radiology, University of Chicago, Chicago, Ill (J.H.C.).
Radiol Artif Intell. 2022 Jan 5;4(1):e210080. doi: 10.1148/ryai.210080. eCollection 2022 Jan.
Body composition on chest CT scans encompasses a set of important imaging biomarkers. This study developed and validated a fully automated analysis pipeline for multi-vertebral level assessment of muscle and adipose tissue on routine chest CT scans. This study retrospectively trained two convolutional neural networks on 629 chest CT scans from 629 patients (55% women; mean age, 67 years ± 10 [standard deviation]) obtained between 2014 and 2017 prior to lobectomy for primary lung cancer at three institutions. A slice-selection network was developed to identify an axial image at the level of the fifth, eighth, and 10th thoracic vertebral bodies. A segmentation network (U-Net) was trained to segment muscle and adipose tissue on an axial image. Radiologist-guided manual-level selection and segmentation generated ground truth. The authors then assessed the predictive performance of their approach for cross-sectional area (CSA) (in centimeters squared) and attenuation (in Hounsfield units) on an independent test set. For the pipeline, median absolute error and intraclass correlation coefficients for both tissues were 3.6% (interquartile range, 1.3%-7.0%) and 0.959-0.998 for the CSA and 1.0 HU (interquartile range, 0.0-2.0 HU) and 0.95-0.99 for median attenuation. This study demonstrates accurate and reliable fully automated multi-vertebral level quantification and characterization of muscle and adipose tissue on routine chest CT scans. : Skeletal Muscle, Adipose Tissue, CT, Chest, Body Composition Analysis, Convolutional Neural Network (CNN), Supervised Learning © RSNA, 2022.
胸部CT扫描的身体成分包含一系列重要的影像生物标志物。本研究开发并验证了一种用于在常规胸部CT扫描上对多椎体水平的肌肉和脂肪组织进行评估的全自动分析流程。本研究回顾性地在来自三个机构的629例患者（55%为女性；平均年龄67岁±10[标准差]）于2014年至2017年间在进行原发性肺癌肺叶切除术前获得的629例胸部CT扫描上训练了两个卷积神经网络。开发了一个切片选择网络来识别第五、第八和第十胸椎椎体水平的轴向图像。训练了一个分割网络（U-Net）以在轴向图像上分割肌肉和脂肪组织。由放射科医生指导进行手动水平选择和分割以生成真值。然后作者在一个独立测试集上评估了他们的方法对横截面积（CSA）（平方厘米）和衰减（亨氏单位）的预测性能。对于该流程，两种组织的CSA的中位数绝对误差和组内相关系数分别为3.6%（四分位间距，1.3%-7.0%）和0.959-0.998，中位数衰减的分别为1.0 HU（四分位间距，0.0-2.0 HU）和0.95-0.99。本研究证明了在常规胸部CT扫描上对肌肉和脂肪组织进行准确且可靠的全自动多椎体水平定量和特征分析。：骨骼肌、脂肪组织、CT、胸部、身体成分分析、卷积神经网络（CNN）、监督学习 © RSNA，2022
