肌骨超声多个声学征象诊断部分性肩袖损伤的Logistic回归分析

Logistic regression analysis of multiple acoustic signs of musculoskeletal ultrasound in diagnosis of partial rotator cuff injury

目的构建肌骨超声多个声学征象诊断部分性肩袖损伤的Logistic回归模型,应用ROC曲线评价各声学征象及其联合应用诊断部分型性肩袖损伤的工作效能。方法回顾性分析我院159例肩部疼痛患者(175个肩关节)的肌骨超声检查资料,其中89例肩痛患者(92个肩关节)经关节镜确诊为部分性肩袖损伤,设为研究组,余70例患者(83个肩关节)设为对照组。观察两组的肩峰下滑囊积液、肱二头肌长头肌腱鞘积液、盂肱关节积液、腱体内低回声、腱体滑囊面凹陷、腱体内钙化、腱体弹性成像,对其行单因素及多因素Logistic回归分析,建立回归方程,并绘制ROC曲线,观察各征象及联合应用的诊断工作效能。结果单因素分析结果显示,两组的肩峰下滑囊积液深度及腱体内低回声、腱体滑囊面凹陷、腱体内钙化检出数比较,差异有统计学意义(P<0.05)。多因素Logistic回归分析筛选出肩峰下滑囊积液、腱体内低回声、腱体滑囊面凹陷为诊断部分性肩袖损伤的重要指标,OR值分别为47.912、32.633、42.388,ROC曲线下面积分别为0.780、0.742、0.607,联合应用诊断部分性肩袖损伤的ROC曲线下面积为0.915,灵敏度为78.3%,特异度为95.2%。结论基于肌骨超声多个声学征象建立的Logistic回归模型对部分性肩袖损伤的诊断有一定的价值,而联合应用有更高的诊断价值。

Objective To establish a Logistic regression model for the diagnosis of partial rotator cuff injury by multiple acoustic signs of musculoskeletal ultrasound,and to evaluate the effectiveness of various acoustic signs and the combination in the diagnosis of partial rotator cuff injury by ROC curve.Methods The musculoskeletal ultrasonography data of 159 patients with shoulder pain(175 shoulder joints)in our hospital were retrospectively analyzed.Among them,89 patients with shoulder pain(92 shoulder joints)were diagnosed as partial rotator cuff injury by arthroscopy and assigned to study group,and the remaining 70 patients(83 shoulder joints)were assigned to control group.The effusion of subacromial bursa,effusion of tendon sheath of long head of biceps brachii,effusion of glenohumeral joint,low echo in tendon,concavity of bursal surface of tendon body,intratendinous calcification and elastography of tendon body were observed in the two groups.Univariate and multivariate Logistic regression analysis were performed to establish the regression process,and ROC curve was drawn to observe the diagnostic efficacy of signs and combined application.Results The univariate analysis showed that there were significant differences in the depth of effusion of subacromial bursa,the detected number of low echo in tendon,concavity of bursal surface of tendon body and intratendinous calcification between the two groups.The multivariate Logistic regression analysis showed that effusion of subacromial bursa,low echo in tendon and concavity of bursal surface of tendon body were the important indicators for the diagnosis of partial rotator cuff injury,the OR values were 47.912,32.633 and 42.388 respectively,and the area under ROC curve was 0.780,0.742 and 0.607 respectively.The area under ROC curve of combined application was 0.915,the sensitivity was 78.3%,and the specificity was 95.2%.Conclusion The Logistic regression model based on multiple acoustic signs of musculoskeletal ultrasound has a certain value in the diagnosis of partial rotator cuff injury,and the combined application has a higher diagnostic value.