基于超声造影与中医证素的股骨头坏死塌陷风险的预测模型构建与验证OA

Objective:To construct a collapse risk prediction model for osteonecrosis of the femoral head(ONFH)based on contrast-enhanced ultrasound(CEUS)combined with traditional Chinese medicine(TCM)syndrome elements,providing a basis for the early identification and diagnosis of high-risk ONFH patients.Methods:A total of 234 ONFH patients(307 hips)who underwent CEUS examination were included.The patients were randomly divided into a training set(214 hips)and a validation set(93 hips)at a ratio of 7∶3.The least absolute shrinkage and selection operator(LASSO)algorithm and the Boruta algorithm were used to screen for collapse risk variables in ONFH patients.Multivariate logistic regression analysis was then performed on the identified risk variables to construct a collapse risk prediction model,and a nomogram was constructed.The model was evaluated using ROC curve,Hosmer-Lemeshow test,calibration curve,and decision curve analyses.Results:LASSO regression identified 12 collapse risk variables,while the Boruta algorithm identified 10 collapse risk variables.The intersection of variables identified by both methods included 7 variables,mean transit time(MTT),D-dimer,descending slope(DS),peak intensity(PI),low-density lipoprotein cholesterol(LDL-C),and the TCM syndrome elements of kidney and blood stasis.Multivariate logistic regression analysis showed that the prolonged MTT(OR=1.452,95%CI 1.004-2.037),elevated PI(OR=1.461,95%CI 1.056-2.045),elevated D-dimer(OR=1.586,95%CI 1.145-2.227),elevated LDL-C(OR=1.528,95%CI 1.106-2.141),blood stasis(OR=3.039,95%CI 1.535-6.132),and kidney(OR=2.715,95%CI 1.387-5.476)were independent risk factors for ONFH collapse.Conversely,an increase in the absolute value of DS(OR=0.680,95%CI 0.472-0.964)was an independent protective factor for ONFH collapse.A prediction model for ONFH collapse was constructed based on these seven variables.AUC values for predicting collapse risk in the training and validation sets were 0.816(95%CI 0.756-0.876)and 0.753(95%CI 0.648-0.859),respectively.The calibration curves were generally consistent with the ideal curves.The Hosmer-Lemeshow test showed good model fit for both the training set(χ²=6.940,P=0.543)and the validation set(χ²=6.742,P=0.564).Decision curve analysis showed that the prediction model provided net benefit across a wide range of probability thresholds,indicating certain clinical value.Conclusion:The prediction model for ONFH collapse risk incorporating CEUS parameters and TCM syndrome elements,provides a reference for predicting the occurrence and development of ONFH collapse in clinical practice.