Three-dimensional MRI Bone Models of the Glenohumeral Joint Using Deep Learning: Evaluation of Normal Anatomy and Glenoid Bone Loss.
Three-dimensional MRI Bone Models of the Glenohumeral Joint Using Deep Learning: Evaluation of Normal Anatomy and Glenoid Bone Loss.
复制标题
使用深度学习的盂肱关节三维 MRI 骨模型：正常解剖结构和关节盂骨丢失的评估。
DOI：
10.1148/ryai.2020190116
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
Gyftopoulos,Soterios
中科院分区：
文献类型：
--
作者：
CantarelliRodrigues,Tatiane;Deniz,CemM;Alaia,ErinF;Gorelik,Natalia;Babb,JamesS;Dublin,Jared;Gyftopoulos,Soterios
目的利用卷积神经网络（cnn）对盂肱关节进行全自动MRI分割，并评估用该方法建立的三维（3D） MRI模型的准确性。材料与方法回顾性收集2013年9月至2018年8月期间100例患者（平均年龄44岁，年龄范围14-80岁，男性60例）的肩部MR图像。利用cnn建立了质子密度加权图像的全自动分割模型。2014年5月至2019年4月，回顾性收集了另外50名患者（平均年龄33岁，范围16-65岁，男性35名）的肩部MR图像，利用基于dixon序列的迁移学习创建3D肩关节MRI模型。两名肌肉骨骼放射科医生在全自动和半自动分割3D MRI模型上进行测量，以评估肩关节解剖、肩关节骨丢失（GBL）及其对治疗选择的影响。使用Dice相似系数（DSC）、灵敏度、精度和基于表面的距离测量来评估cnn的性能。测量值比较采用配对Wilcoxon符号秩检验。结果肱骨和关节盂二维CNN模型的DSC分别为0.95和0.86，精度分别为95.5%和87.5%，平均精度分别为98.6%和92.3%，灵敏度分别为94.8%和86.1%。肱骨和关节盂三维CNN模型的DSC分别为0.95和0.86，精度分别为95.1%和87.1%，平均精度分别为98.7%和91.9%，灵敏度分别为94.9%和85.6%。全自动和半自动3D模型测量肩关节和肱骨头宽度无差异（p值范围，0.097 - 0.99）。结论cnn有可能用于临床实践，提供快速准确的三维MRI盂肱骨模型和GBL测量。关键词：计算机应用- 3d，卷积神经网络（CNN），关节，长骨，核磁共振成像，神经网络，分割，肩部，骨骼-阑尾，监督学习，迁移学习©RSNA， 2020
PurposeTo use convolutional neural networks (CNNs) for fully automated MRI segmentation of the glenohumeral joint and evaluate the accuracy of three-dimensional (3D) MRI models created with this method.Materials and MethodsShoulder MR images of 100 patients (average age, 44 years; range, 14–80 years; 60 men) were retrospectively collected from September 2013 to August 2018. CNNs were used to develop a fully automated segmentation model for proton density–weighted images. Shoulder MR images from an additional 50 patients (mean age, 33 years; range, 16–65 years; 35 men) were retrospectively collected from May 2014 to April 2019 to create 3D MRI glenohumeral models by transfer learning using Dixon-based sequences. Two musculoskeletal radiologists performed measurements on fully and semiautomated segmented 3D MRI models to assess glenohumeral anatomy, glenoid bone loss (GBL), and their impact on treatment selection. Performance of the CNNs was evaluated using Dice similarity coefficient (DSC), sensitivity, precision, and surface-based distance measurements. Measurements were compared using matched-pairs Wilcoxon signed rank test.ResultsThe two-dimensional CNN model for the humerus and glenoid achieved a DSC of 0.95 and 0.86, a precision of 95.5% and 87.5%, an average precision of 98.6% and 92.3%, and a sensitivity of 94.8% and 86.1%, respectively. The 3D CNN model, for the humerus and glenoid, achieved a DSC of 0.95 and 0.86, precision of 95.1% and 87.1%, an average precision of 98.7% and 91.9%, and a sensitivity of 94.9% and 85.6%, respectively. There was no difference between glenoid and humeral head width fully and semiautomated 3D model measurements (Pvalue range, .097–.99).ConclusionCNNs could potentially be used in clinical practice to provide rapid and accurate 3D MRI glenohumeral bone models and GBL measurements.Keywords:Computer Applications-3D, Convolutional Neural Network (CNN), Joints, Long Bones, MR-Imaging, Neural Networks, Segmentation, Shoulder, Skeletal-Appendicular, Supervised learning, Transfer learning© RSNA, 2020