论文标题
使用MRI图像深度学习的前交叉韧带损伤的层次严重程度分期
Hierarchical Severity Staging of Anterior Cruciate Ligament Injuries using Deep Learning with MRI Images
论文作者
论文摘要
目的：评估两个卷积神经网络（CNN）的诊断效用，以进行前交叉韧带（ACL）损伤的严重程度分期。 材料和方法：从224名患者（47 +/- 14岁，54％，54％的女性）在2011年至2014年间获得的1243次膝盖MR图像（1008个完整，部分撕裂，完全撕裂和140个重建的ACL）对1243个膝盖MR图像进行了回顾性分析。放射线使用了修改的计数，该测得的计数。为了将ACL损伤与深度学习进行分类，使用了两种类型的CNN，一种具有三维（3D），另一种具有二维（2D）卷积内核。性能指标包括灵敏度，特异性，加权Cohen的Kappa以及整体准确性，然后进行McNemar的测试以比较CNNS性能。 结果：使用2D CNN（准确性：92％（233/254）和Kappa：0.83），报道的ACL损伤分类的总体准确性和加权Cohen的Kappa高于3D CNN（精度：89％（225/254）和KAPPA：KAPPA：0.83）（0.83）（p = .27）。 2D CNN和3D CNN在对完整的ACL分类（2D CNN：93％（188/203）的敏感性和90％（46/51）特异性； 3D CNN：89％（180/203）敏感性和88％（45/51）特异性）方面相似。两个网络对全部泪的分类也可比（2D CNN：82％（14/17）敏感性和94％（222/237）特异性； 3D CNN：76％（13/17）敏感性和100％（236/237）特异性）。 2D CNN正确地对所有重建ACL进行了分类。 结论：应用于ACL病变分类的2D和3D CNN具有较高的灵敏度和特异性，这表明这些网络可用于帮助通过非专家级的ACL损伤。
Purpose: To evaluate the diagnostic utility of two convolutional neural networks (CNNs) for severity staging of anterior cruciate ligament (ACL) injuries. Materials and Methods: This retrospective analysis was conducted on 1243 knee MR images (1008 intact, 18 partially torn, 77 fully torn, and 140 reconstructed ACLs) from 224 patients (age 47 +/- 14 years, 54% women) acquired between 2011 and 2014. The radiologists used a modified scoring metric. To classify ACL injuries with deep learning, two types of CNNs were used, one with three-dimensional (3D) and the other with two-dimensional (2D) convolutional kernels. Performance metrics included sensitivity, specificity, weighted Cohen's kappa, and overall accuracy, followed by McNemar's test to compare the CNNs performance. Results: The overall accuracy and weighted Cohen's kappa reported for ACL injury classification were higher using the 2D CNN (accuracy: 92% (233/254) and kappa: 0.83) than the 3D CNN (accuracy: 89% (225/254) and kappa: 0.83) (P = .27). The 2D CNN and 3D CNN performed similarly in classifying intact ACLs (2D CNN: 93% (188/203) sensitivity and 90% (46/51) specificity; 3D CNN: 89% (180/203) sensitivity and 88% (45/51) specificity). Classification of full tears by both networks were also comparable (2D CNN: 82% (14/17) sensitivity and 94% (222/237) specificity; 3D CNN: 76% (13/17) sensitivity and 100% (236/237) specificity). The 2D CNN classified all reconstructed ACLs correctly. Conclusion: 2D and 3D CNNs applied to ACL lesion classification had high sensitivity and specificity, suggesting that these networks could be used to help grade ACL injuries by non-experts.
