肝硬化患者脾切除术后急性期门静脉血栓形成预测模型建立及验证

Development and validation of a predictive model for portal static thrombosis in the acute phase after splenectomy in patients with liver cirrhosis

目的建立肝硬化患者脾切除术后急性期门静脉血栓形成(PVT)的列线图预测模型,评估临床应用价值。方法回顾性分析2013年1月至2021年11月在丽水市人民医院肝胆外科完成的92例肝硬化行脾切除患者的临床资料,根据术后1周内是否发生门静脉系统血栓分成,采用单因素及多因素Logistic回归分析独立危险因素,根据结果构建Nomogram模型,再随机选取浙江大学医学院附属第一医院同期肝硬化行脾切除病例43例,进行外部验证。采用ROC曲线、校准曲线、Hosmer-Lemeshow拟合优度检验对模型的预测价值进行验证,采用临床决策曲线(DCA)对模型的临床获益进行评估。结果多因素Logistic回归分析显示,门静脉直径、脾静脉直径是肝硬化患者脾切除术后急性期门静脉血栓发生的独立危险因素(OR值分别是1.311和1.345,P<0.05)。ROC分析显示,本研究建立的发生风险预测模型区分度较好(AUC=0.824,95%CI:0.739~0.910)。加强Bootstrap校准图形中标准曲线与校准预测曲线相贴合,表明模型的校准度良好。DCA曲线显示模型具有潜在的临床应用价值。结论术前门静脉直径、脾静脉直径是肝硬化患者行脾切除术后急性期门静脉血栓形成的独立危险因素,基于这两项因素建立的Logistic回归模型对预测脾切除术后急性期门静脉血栓形成的风险有一定价值。

Objective To develop a columnar line graph prediction model for portal vein thrombosis(PVT)in the acute phase after splenectomy in patients with cirrhosis and to assess the value of clinical application.Methods The clinical data of 92 patients with cirrhosis undergoing splenectomy completed in our hospital from January 2013 to November 2021 were retrospectively analyzed,and independent risk factors were analyzed by single-factor and multi-factor logistic regression according to whether portal system thrombosis occurred within 1 week after surgery,and Nomogram models were constructed according to the results,and then randomly selected 43 cases of liver cirrhosis undergoing splenectomy at the First Affiliated Hospital of Zhejiang University during the same period were selected for external validation.The ROC curve,calibration curve,and Hosmer-Lemeshow goodness-of-fit test were used to validate the predictive value of the model,and the clinical benefit of the model was evaluated using the clinical decision curve(DCA).Results Multi-factor logistic regression analysis showed that portal vein diameter and splenic vein diameter were independent risk factors for the development of portal vein thrombosis in the acute phase after splenectomy in patients with cirrhosis(OR 1.311 and 1.345,respectively,P<0.05).ROC analysis showed that the risk of occurrence prediction model established in this study had good discrimination(AUC=0.824).The standard curve in the enhanced Bootstrap calibration graph fitted the calibrated prediction curve,indicating that the model was well calibrated.The DCA curves showed that the model has potential clinical application.Conclusion Preoperative portal vein diameter and splenic vein diameter are independent risk factors for acute portal vein thrombosis after splenectomy in patients with cirrhosis,and a logistic regression model based on these two factors can effectively predict the risk of acute portal vein thrombosis after splenectomy.