基于冠脉易损斑块特征构建冠心病患者主要不良心脏事件的Nomogram预测模型  

Construction of nomogram prediction model for major adverse cardiovascular events based on the characteristics of coronary vulnerable plaques in patients with coronary heart disease

作　　者：汪慧娟 吴继雄[2] 储岳峰[3] 陈宗洋 WANG Huijuan;WU Jixiong;CHU Yuefeng;CHEN Zongyang(Department of Cardiovascular Medicine,The Fourth People's Hospital of Lu'an,Lu'an 237006,Anhui,China;Department of Cardiovascular Medicine,The Second Affiliated Hospital of Anhui Medical University,Hefei 230000,Anhui,China;Department of Cardiovascular Medicine,Lu'an People's Hospital,Lu'an 237000,Anhui,China Corresponding author:WU)

机构地区：[1]六安市第四人民医院心血管内科,安徽六安237006 [2]安徽医科大学第二附属医院心血管内科,安徽合肥230000 [3]六安市人民医院心血管内科,安徽六安237000

出　　处：《中国现代手术学杂志》2024年第4期283-290,共8页Chinese Journal of Modern Operative Surgery

基　　金：安徽医科大学校科研基金项目【2020xkj228】。

摘　　要：目的分析行经皮冠状动脉介入手术(percutaneous coronary intervention,PCI)后冠心病患者发生主要不良心脏事件(major adverse cardiovascular events,MACE)的影响因素,构建基于冠脉易损斑块特征的Nomogram预测模型。方法选取2021年5月至2022年3月住院行PCI手术治疗的180例冠心病患者(作为训练集)进行回顾性分析。根据MACE发生与否分为MACE组(n=58)和无MACE组(n=122),采用单因素和多因素二元logistic回归分析MACE发生的影响因素,应用R4.3.1软件构建冠心病患者术后发生MACE的风险预测列线图模型,绘制受试者工作特征曲线(receiver operating characteristic curve,ROC),以曲线下面积(area under curve,AUC)、校准曲线、Hosmer-Lemeshow拟合优度检验和决策曲线(decision curve analysis,DCA)等评价该模型预测效能和临床获益情况。收集2022年5月至2023年3月符合纳入标准的90例冠心病患者作为验证集进行外部验证。结果单因素分析显示,MACE的发生与年龄、吸烟史、高血压史、甘油三酯、斑块负荷、脂质斑块体积及非钙化斑块体积相关(P<0.05)。多因素临床特征分析显示,甘油三酯、吸烟史和高血压史是MACE发生的独立影响因素(P<0.05),其风险预测模型为Logit(p)=-4.227+1.494×甘油三酯+0.767×吸烟史+0.920×高血压史。血管内超声影像学特征分析显示,斑块负荷、脂质斑块体积及非钙化斑块体积是MACE发生的独立影响因素(P<0.05),其风险预测模型为Logit(p)=-13.723+0.056×斑块负荷+0.178×脂质斑块体积+0.044×非钙化斑块体积。联合分析显示,甘油三酯、斑块负荷、脂质斑块体积、非钙化斑块体积和高血压史是MACE发生的独立影响因素(P<0.05),其风险预测模型为:Logit(p)=-17.395+1.399×甘油三酯+0.056×斑块负荷+0.182×脂质斑块体积+0.043×非钙化斑块体积+0.858×高血压史。基于上述因素构建Nomogram风险预测模型,并以ROC曲线与校准图对模型进行内部验证,结果�Objective To analyze the influencing factors of major adverse cardiovascular events(MACE)in the patients with coronary heart disease(CHD)performed percutaneous coronary intervention(PCI),and construct a nomogram prediction model based on the characteristics of coronary vulnerable plaques.Methods A total of 180 CHD patients performed PCI admitted from May 2021 to March 2022 were selected as the training set and retrospectively analyzed.According to the occurrence of MACE,the patients were divided into MACE group(n=58)and non-MACE group(n=122).Univariate analysis and multivariate logistic regression was used to analyze the risk factors of MACE,and a prediction nomogram model of postoperative MACE was constructed by the R software.ROC curve was drawn,and area under curve(AUC),calibration curve,Hosmer-Lemeshow goodness-of-fit test and decision curve analysis(DCA)were applied to evaluate the predictive efficiency and clinical benefit of the prediction model.And 90 CHD patients who met the inclusion criteria from May 2022 to March 2023 were enrolled as the validation set for external validation.Results Univariate analysis showed that MACE was related with the factors of age,smoking history,hypertension history,triglyceride level,plaque load,lipid plaque volume and noncalcified plaque volume(P<0.05).Multivariate analysis revealed:①Clinical characteristic analysis showed that triglyceride,smoking history and hypertension history were the independent influencing factors of MACE(P<0.05),with the risk prediction model:Logit(p)=-4.227+1.494×triglyceride+0.767×smoking history+0.920×hypertension history.②Imaging characteristic analysis of intravascular ultrasound showed that plaque load,lipid plaque volume and non-calcified plaque volume were the independent influencing factors of MACE(P<0.05),with the risk prediction model:Logit(p)=-13.723+0.056×plaque load+0.178×lipid plaque volume+0.044×non-calcified plaque volume.③Conjoint analysis indicated that triglyceride level,plaque load,lipid plaque volume,non-calcified

关 键 词：冠脉易损斑块特征 冠心病 主要不良心脏事件 列线图 预测模型 

分 类 号：R541.4[医药卫生—心血管疾病]