Castiglione James, Somasundaram Elanchezhian, Gilligan Leah A, Trout Andrew T, Brady Samuel
Department of Radiology, Cincinnati Children's Hospital Medical Center, 3333 Burnet Ave, MLC 5031, Cincinnati, OH 45229-3026 (J.C., E.S., L.A.G., A.T.T., S.B.); and Departments of Radiology (E.S., A.T.T., S.B.) and Pediatrics (A.T.T.), University of Cincinnati College of Medicine, Cincinnati, Ohio.
Radiol Artif Intell. 2021 Jan 6;3(2):e200130. doi: 10.1148/ryai.2021200130. eCollection 2021 Mar.
To automate skeletal muscle segmentation in a pediatric population using convolutional neural networks that identify and segment the L3 level at CT.
In this retrospective study, two sets of U-Net-based models were developed to identify the L3 level in the sagittal plane and segment the skeletal muscle from the corresponding axial image. For model development, 370 patients (sampled uniformly across age group from 0 to 18 years and including both sexes) were selected between January 2009 and January 2019, and ground truth L3 location and skeletal muscle segmentation were manually defined. Twenty percent (74 of 370) of the examinations were reserved for testing the L3 locator and muscle segmentation, while the remaining were used for training. For the L3 locator models, maximum intensity projections (MIPs) from a fixed number of central sections of sagittal reformats (either 12 or 18 sections) were used as input with or without transfer learning using an L3 localizer trained on an external dataset (four models total). For the skeletal muscle segmentation models, two loss functions (weighted Dice similarity coefficient [DSC] and binary cross-entropy) were used on models trained with or without data augmentation (four models total). Outputs from each model were compared with ground truth, and the mean relative error and DSC from each of the models were compared with one another.
L3 section detection trained with an 18-section MIP model with transfer learning had a mean error of 3.23 mm ± 2.61 standard deviation, which was within the reconstructed image thickness (3 or 5 mm). Skeletal muscle segmentation trained with the weighted DSC loss model without data augmentation had a mean DSC of 0.93 ± 0.03 and mean relative error of 0.04 ± 0.04.
Convolutional neural network models accurately identified the L3 level and segmented the skeletal muscle on pediatric CT scans.See also the commentary by Cadrin-Chênevert in this issue.© RSNA, 2021.
利用卷积神经网络实现儿科人群骨骼肌的自动分割，该网络可在CT图像上识别并分割出L3椎体水平。
在这项回顾性研究中，开发了两组基于U-Net的模型，用于在矢状面识别L3椎体水平，并从相应的轴位图像中分割出骨骼肌。在模型开发过程中，选取了2009年1月至2019年1月期间的370例患者（年龄在0至18岁之间均匀抽样，包括男性和女性），并手动定义了L3椎体的真实位置和骨骼肌分割。将20%（370例中的74例）的检查保留用于测试L3椎体定位器和肌肉分割，其余用于训练。对于L3椎体定位器模型，矢状面重建的固定数量中心层面（12或18个层面）的最大强度投影（MIP）用作输入，使用或不使用在外部数据集上训练的L3定位器进行迁移学习（共四个模型）。对于骨骼肌分割模型，在使用或不使用数据增强训练的模型上使用两种损失函数（加权骰子相似系数[DSC]和二元交叉熵）（共四个模型）。将每个模型的输出与真实情况进行比较，并将每个模型的平均相对误差和DSC相互比较。
使用迁移学习的18层面MIP模型训练的L3椎体层面检测的平均误差为3.23 mm±2.61标准差，在重建图像厚度（3或5 mm）范围内。使用加权DSC损失模型且不进行数据增强训练的骨骼肌分割的平均DSC为0.93±0.03，平均相对误差为0.04±0.04。
卷积神经网络模型能够准确识别儿科CT扫描中的L3椎体水平并分割骨骼肌。另见本期Cadrin-Chênevert的评论。© RSNA，2021。
