基于深度学习的腰椎稳定性自动诊断工具的研发及验证  

Development and validation of an automatic diagnostic tool for lumbar stability based on deep learning

作　　者：胡厚民 王贤帝[2] 杨恒[1] 张劲[1] 李康 曾建成[2] HU Houmin;WANG Xiandi;YANG Heng;ZHANG Jin;LI Kang;ZENG Jiancheng(College of Electrical Engineering,Sichuan University,Chengdu Sichuan,610041,P.R.China;Orthopedic Research Institute,Department of Orthopedics,West China Hospital,Sichuan University,Chengdu Sichuan,610041,P.R.China;West China Biomedical Big Data Center,West China Hospital,Sichuan University,Chengdu Sichuan,610041,P.R.China)

机构地区：[1]四川大学电气工程学院,成都610041 [2]四川大学华西医院骨科骨科研究所,成都610041 [3]四川大学华西医院生物医学大数据中心,成都610041

出　　处：《中国修复重建外科杂志》2023年第1期81-90,共10页Chinese Journal of Reparative and Reconstructive Surgery

基　　金：国家重点研发计划项目(2020YFB1711500);四川大学华西医院学科卓越发展1·3·5工程项目(ZYYC21004)。

摘　　要：目的 基于深度学习研发腰椎稳定性自动诊断工具,并验证其诊断精度。方法 收集153例腰椎疾病患者术前腰椎过屈、过伸位X线片,由3名骨科医师标注5个关键点,分别为L4后下角、前下角以及L5后上角、前上角、后下角,共获得3套标注结果。将306张腰椎X线片按照3∶1∶2比例随机分为训练集(n=156)、验证集(n=50)和测试集(n=100)。提出一种新的神经网络结构Swin-PGNet,使用已标注的X线片图像对其进行训练,使其能自动定位腰椎椎体关键点,并通过关键点测算L4、5椎间Cobb角和L4椎体滑移距离。对于关键点定位、Cobb角测量和椎体滑移距离测量,以平均误差、组内相关系数(intra-class correlation coefficient,ICC)比较医师标注与Swin-PGNet之间的差异。椎间Cobb角变化超过11°作为腰椎不稳判断标准,腰椎滑移距离超过3 mm作为腰椎滑脱判断标准,比较医师和Swin-PGNet判断腰椎稳定性的准确率。结果 (1) Swin-PGNet关键点定位平均误差为(1.407±0.939)mm,医师间平均误差为(3.034±2.612)mm。(2) Cobb角标注:Swin-PGNet平均误差为(2.062±1.352)°,医师间平均误差为(3.580±2.338)°;Swin-PGNet与3名医师间误差比较,差异均无统计学意义(P>0.05),但不同医师间误差比较差异有统计学意义(P<0.05)。(3)椎体滑移距离:Swin-PGNet平均误差为(1.656±0.878)mm,医师标注平均误差为(1.884±1.612)mm;Swin-PGNet与3名医师间误差比较以及不同医师间误差比较,差异均无统计学意义(P>0.05)。Swin-PGNet腰椎不稳判断准确率为84.0%、医师为75.3%,腰椎滑脱判断准确率分别为71.3%、70.7%,Swin-PGNet与3名医师间误差比较以及不同医师间误差比较,差异均无统计学意义(P>0.05)。(4)腰椎稳定性判定一致性分析:3名医师标注椎间Cobb角的ICC为0.913 [95%CI(0.898,0.934)](P<0.05),椎体滑移距离为0.741 [95%CI(0.729,0.796)](P<0.05),说明3名医师间标注具有一致性。SwinPGNet-所有医师间椎间Cobb角ICC为Objective To develop an automatic diagnostic tool based on deep learning for lumbar spine stability and validate diagnostic accuracy. Methods Preoperative lumbar hyper-flexion and hyper-extension X-ray films were collected from 153 patients with lumbar disease. The following 5 key points were marked by 3 orthopedic surgeons: L4posteroinferior, anterior inferior angles as well as L5posterosuperior, anterior superior, and posterior inferior angles. The labeling results of each surgeon were preserved independently, and a total of three sets of labeling results were obtained. A total of 306 lumbar X-ray films were randomly divided into training(n=156), validation(n=50), and test(n=100) sets in a ratio of 3∶1∶2. A new neural network architecture, Swin-PGNet was proposed, which was trained using annotated radiograph images to automatically locate the lumbar vertebral key points and calculate L4, 5intervertebral Cobb angle and L4 lumbar sliding distance through the predicted key points. The mean error and intra-class correlation coefficient(ICC)were used as an evaluation index, to compare the differences between surgeons’ annotations and Swin-PGNet on the three tasks(key point positioning, Cobb angle measurement, and lumbar sliding distance measurement). Meanwhile, the change of Cobb angle more than 11° was taken as the criterion of lumbar instability, and the lumbar sliding distance more than 3 mm was taken as the criterion of lumbar spondylolisthesis. The accuracy of surgeon annotation and Swin-PGNet in judging lumbar instability was compared. Results(1) Key point: The mean error of key point location by Swin-PGNet was(1.407±0.939) mm, and by different surgeons was(3.034±2.612) mm.(2) Cobb angle: The mean error of Swin-PGNet was(2.062±1.352)° and the mean error of surgeons was(3.580±2.338)°. There was no significant difference between SwinPGNet and surgeons(P>0.05), but there was a significant difference between different surgeons(P<0.05).(3) Lumbar sliding distance: The mean error of Swin-PGNet was(1.656

关 键 词：人工智能 医学图像处理 关键点定位 腰椎不稳 辅助诊断 

分 类 号：R687.3[医药卫生—骨科学]