HEAD-US-C联合临床指标预测A型血友病患者膝关节出血风险的临床价值  

Clinical value of HEAD-US-C combined with clinical indicators in predicting the risk of knee joint bleeding in patients with type A hemophilia

作　　者：王慧 叶鸣[2] 李明星[1] 赵宇心 周鸿 陈蓉[3] 聂全禹[1,2] 周洋 WANG Hui;YE Ming;LI Mingxing;ZHAO Yuxin;ZHOU Hong;CHEN Rong;NIE Quanyu;ZHOU Yang(Department of Ultrasound,the Affiliated Hospital of Southwest Medical University,Luzhou 646000,China;Department of Ultrasound,the Third People’s Hospital of Chengdu,Chengdu 610031,China;Department of Hematology,the Third People’s Hospital of Chengdu,Chengdu 610031,China)

机构地区：[1]西南医科大学附属医院超声科,四川泸州646000 [2]成都市第三人民医院超声科,四川成都610031 [3]成都市第三人民医院血液科,四川成都610031

出　　处：《临床超声医学杂志》2025年第2期122-128,共7页Journal of Clinical Ultrasound in Medicine

基　　金：四川省区域联合创新重点项目(2024YFHZ0078);四川省医学科研课题计划项目(S18068)。

摘　　要：目的探讨适应于中国人的血友病骨关节早期超声半定量评分系统(HEAD-US-C)联合临床指标预测A型血友病患者膝关节出血风险的临床价值。方法选取于成都市第三人民医院就诊的A型血友病患者78例,其中单侧膝关节出血21例,膝关节未出血57例,采用最小绝对收缩与选择算子(Lasso)回归和十折交叉验证法筛选关节出血的最佳预测因素,将其纳入多因素Logistic回归分析筛选预测A型血友病患者膝关节出血风险的独立影响因素,基于此构建预测模型并绘制列线图;绘制受试者工作特征(ROC)曲线评价列线图模型的区分度,并采用Bootstrap自助抽样法(重复抽样1000次)予以内部验证;采用Hosmer-Lemeshow拟合优度检验评价列线图模型的拟合度,并绘制校准曲线评价其校准度。结果经Lasso回归和十折交叉验证法共筛选出5个最佳预测因素,分别为按需治疗剂量、首次出血年龄及HEAD-US-C评分中的关节积液/积血、滑膜增生和软骨破坏;将其纳入多因素Logistic回归分析,结果显示按需治疗剂量、关节积液/积血、滑膜增生及软骨破坏均为预测A型血友病患者膝关节出血风险的独立影响因素(OR=1.213、4.388、5.334、0.509,均P<0.05)。基于此构建列线图模型,ROC曲线分析显示,列线图模型预测A型血友病患者膝关节出血风险的曲线下面积为0.934(95%可信区间:0.879~0.990),Bootstrap自助抽样法结果显示其一致性指数为0.934,表明其区分度较好。Hosmer-Lemeshow拟合优度检验结果显示,列线图模型拟合度较好(χ^(2)=7.437,P=0.490)。校准曲线分析显示,列线图模型对A型血友病患者膝关节出血风险的预测概率与实际概率的一致性较好,表明其校准度较高。结论HEAD-US-C联合临床指标可用于预测A型血友病患者膝关节出血风险,具有一定的临床指导价值。Objective To investigate the clinical value of haemophilic early arthropathy detection with ultrasound in China(HEAD-US-C)combined with clinical indicators in predicting the risk of knee joint bleeding in patients with type A hemophilia.Methods A total of 78 patients with type A hemophilia who from the Third People’s Hospital of Chengdu were selected,including 21 cases with unilateral knee joint bleeding and 57 cases without knee joint bleeding.The least absolute shrinkage and selection operator(Lasso)regression and cross validation were used to obtain the optimal predictive factors for knee joint bleeding,which were incorporated by multivariate Logistic regression analysis to screen the independent influencing factors for knee joint bleeding in patients with type A hemophilia,and the prediction model was constructed and a nomogram was drawn.Receiver operating characteristic(ROC)curve was drawn to evaluate the discrimination of the nomogram model,and the Bootstrap self-sampling method method(repeated sampling 1000 times)was used for internal validation.The Hosmer-Lemeshow goodness of fit test was used to evaluate the goodness of fit of the nomogram model,and the calibration curve was drawn to evaluate its calibration.Results Totally 5 predictive factors were screened out by Lasso regression and cross validation,namely on-demand treatment dose,age of first bleeding and joint effusion,synovial hyperplasia,cartilage destruction in the HEAD-US-C score.The above factors were included into a multivariate Logistic regression analysis,the results showed that on-demand treatment dose,joint effusion,synovial hyperplasia and cartilage destruction were independent influencing factors for knee joint bleeding in patients with type A hemophilia(OR=1.213,4.388,5.334,0.509,all P<0.05).The nomogram model was constructed based on the results of multivariate Logistic analysis,ROC curve analysis showed that the area under the curve of the nomogram model for predicting the risk of knee joint bleeding in patients with type A hemophil

关 键 词：超声检查 A型血友病 膝关节出血 列线图 

分 类 号：R445.1[医药卫生—影像医学与核医学] R554.1[医药卫生—诊断学]