中国中西医结合影像学杂志

目的：探讨剪切波弹性成像（SWE）联合多参数的无创预测模型在肝硬化患者食管胃底静脉曲张（EGV）风险分层中的临床应用价值。方法：选取72例肝硬化患者，以胃镜检查结果为金标准，根据EGV情况分为无/轻度组34例和中/重度组38例。比较2组肝脏硬度值（LSM）、脾脏硬度值（SSM）、门静脉内径（PVD）、门静脉流速（PVV）、脾静脉内径（SVD）、脾静脉流速（SVV）、脾厚径、血小板计数（PLT）、凝血酶原时间（PT）、凝血酶原活动度（PTA）、国际标准化比值（INR）、总胆红素、白蛋白及Child-Pugh分级。采用二元logistic回归分析，筛选出EGV严重程度的独立影响因素，构建预测模型。绘制ROC曲线评估模型的预测效能。结果：2组LSM、SSM、PVD、PVV、SVD、SVV、脾厚径、PLT、PT、PTA、INR、总胆红素、白蛋白及Child-Pugh分级比较，差异均有统计学意义（均P<0.05）。logistic回归分析显示，SSM、脾厚径及白蛋白为肝硬化EGV严重程度的独立影响因素（均P<0.05），而包含PLT的预测模型预测效能最佳。将SSM、脾厚径、白蛋白、PLT纳入预测模型。ROC曲线分析显示，SSM、脾厚径、PLT、白蛋白及预测模型预测肝硬化患者发生EGV的AUC分别为0.899、0.842、0.839、0.796及0.959。结论：SWE联合多参数构建的预测模型对肝硬化患者EGV发生风险的预测效能显著，可为临床治疗决策提供重要的参考依据。

Objective:To explore the clinical application value of a predictive model based on shear wave elastography(SWE) and multiple parameters in the risk stratification of esophageal and gastric varices(EGV) in patients with liver cirrhosis. Methods:A total of 72 patients with liver cirrhosis were enrolled. Using the gastroscopy results as the gold standard,all patients were divided into 2 groups,the non/mild EGV group(34 cases) and the moderate/severe EGV group(38 cases). The liver stiffness measurement(LSM),spleen stiffness measurement(SSM),portal vein diameter(PVD),portal vein velocity(PVV),splenic vein diameter(SVD),splenic vein velocity(SVV),spleen thickness,platelet,prothrombin time(PT),prothrombin activity(PTA),international normalized ratio(INR),total bilirubin,albumin,and Child-Pugh classification were compared between the two groups. The binary logistic regression analysis was used to screen for independent influencing factors of the severity of EGV and to construct a predictive model. ROC curve was plotted to evaluate the predictive efficacy of this model. Results:The differences in LSM,SSM,PVD,PVV,SVD,SVV,spleen thickness,platelet,PT,PTA,INR,total bilirubin,albumin and Child-Pugh classification between the two groups were all statistically significant(all P<0.05). Logistic regression analysis showed that SSM,spleen thickness and albumin were independent influencing factors for the severity of EGF in liver cirrhosis(all P<0.05). SSM,spleen thickness,albumin and platelet were included in the prediction model. ROC curve analysis showed that the AUCs for predicting the EGV occurrence were 0.899 for SSM,0.842 for spleen thickness,0.839 for platelet,0.796 for albumin,and 0.959 for the predictive model,and the predictive model demonstrated the best diagnostic efficacy. Conclusion:The predictive model based on SWE and multiple parameters shows good diagnostic efficacy in predicting the risk of EGV in patients with liver cirrhosis,providing an important reference for clinical treatment decisions.