基于多模态数据构建急性ST段抬高型心肌梗死老年患者院内主要不良心血管事件风险预测模型OA
Construction of a risk prediction model for in-hospital major adverse cardiovascular events in elderly patients with acute ST-segment elevation myocardial infarction based on multimodal data
目的 基于多模态临床数据分析急性ST段抬高型心肌梗死(STEMI)老年患者经皮冠状动脉介入治疗(PCI)术后发生院内主要不良心血管事件(MACE)的影响因素并构建预测模型,并评估该模型预测效能.方法 回顾性选取2020年5月至2025年5月于长沙市第四医院接受PCI的老年STEMI患者为研究对象,按7∶3的比例随机分为训练集和测试集,根据住院期间是否发生MACE将训练集患者分为MACE组和非MACE组,采用Lasso回归及多因素Logistic回归筛选独立预测因素,并构建列线图模型.采用受试者工作特征曲线及其曲线下面积(AUC)、校准曲线及决策曲线评估模型性能,并采用Bootstrap法(重复抽样1 000次)进行内部验证.结果 共纳入515例老年STEMI患者,其中训练集361例,测试集154例,院内MACE发生率为26.02%.多因素Logistic回归显示,Killip分级≥Ⅱ级[OR=4.186,95%CI(1.967,8.907)]、肌钙蛋白[OR=2.415,95%CI(1.690,3.451)]、N末端脑钠肽前体(NT-proBNP)[OR=1.233,95%CI(1.129,1.347)]、冠状动脉血运重建术与心脏外科手术协作(SYNTAX)评分升高[OR=1.214,95%CI(1.129,1.304)]是院内 MACE 的独立危险因素,而左心室射血分数(LVEF)[OR=0.889,95%CI(0.843,0.938)]、预后营养指数(PNI)升高[OR=0.843,95%CI(0.784,0.905)]是院内MACE的独立保护因素.训练集和测试集的AUC分别为0.941和0.934;校准曲线与理想对角线基本重合(Hosmer-Lemeshow检验P>0.05);决策曲线显示临床净收益优于极端策略.结论 基于Killip分级、肌钙蛋白、NT-proBNP、LVEF、SYNTAX评分及PNI构建的列线图模型对老年STEMI患者PCI术后院内MACE风险表现出较好的预测效能和临床实用性,初步提示其可用于临床早期识别高危患者及个体化干预,但仍需多中心、大样本的外部独立队列进一步验证.
Objective To analyze the influencing factors of in-hospital major adverse cardiovascular events(MACE)after percutaneous coronary intervention(PCI)in elderly patients with acute ST-segment elevation myocardial infarction(STEMI)based on multimodal clinical data,to establish a predictive model,and to evaluate the predictive performance of the model.Methods Elderly STEMI patients who underwent PCI at the Fourth Hospital of Changsha from May 2020 to May 2025 were retrospectively enrolled as research subjects.The patients were randomly divided into a training set and a validation set at a ratio of 7∶3.Patients in the training set were further assigned to the MACE group and the non-MACE group according to whether in-hospital MACE occurred.Lasso regression and multivariate Logistic regression were applied to screen independent predictive factors,and a nomogram model was constructed.The model performance was assessed using the receiver operating characteristic curve and the area under the curve(AUC),as well as calibration curve and decision curve analysis.Bootstrap resampling(1,000 repetitions)was performed for internal validation.Results A total of 515 elderly STEMI patients were included,with an in-hospital MACE incidence of 26.02%,of whom 361 were assigned to the training set and 154 to the validation set.Multivariate Logistic regression revealed that elevodbed Killip classification(≥Grade Ⅱ)[OR=4.186,95%CI(1.967,8.907)],troponin[OR=2.415,95%CI(1.690,3.451)],N-terminal pro-brain natriuretic peptide(NT-proBNP)[OR=1.233,95%CI(1.129,1.347)],and Synergy between Percutaneous Coronary Intervention with TAXUS and Cardiac Surgery(SYNTAX)score[OR=1.214,95%CI(1.129,1.304)]were independent risk factors for in-hospital MACE,whereas eleuated left ventricular ejection fraction(LVEF)[OR=0.889,95%CI(0.843,0.938)]and prognostic nutritional index(PNI)[OR=0.843,95%CI(0.784,0.905)]were independent protective factors for in-hospital MACE.The AUC values of the training set and validation set were 0.941 and 0.934,respectively.The calibration curve largely overlapped with the ideal diagonal line(Hosmer-Lemeshow test,P>0.05).Decision curve analysis demonstrated superior clinical net benefit compared with extreme strategies.Conclusion The nomogram model constructed based on Killip classification,troponin,NT-proBNP,LVEF,SYNTAX score and PNI exhibits favorable predictive performance and clinical practicability for in-hospital MACE risk in elderly STEMI patients after PCI.It preliminarily suggests that this model can be used for early clinical identification of high-risk patients and individualized intervention.Further validation with multicenter,large-sample external independent cohorts is still required.
刘芷静;彭希;金静
长沙市第四医院/长沙市中西医结合医院心血管内科(长沙 410006)湖南省人民医院/湖南师范大学附属第一医院心血管内科(长沙 410001)长沙市第四医院/长沙市中西医结合医院心血管内科(长沙 410006)
医药卫生
急性ST段抬高型心肌梗死老年主要不良心血管事件经皮冠状动脉介入治疗多模态数据预测模型
Acute ST-segment elevation myocardial infarctionElderlyMajor adverse cardiovascular eventsPercutaneous coronary interventionMultimodal dataPrediction model
《医学新知》 2026 (9)
980-988,中插7-中插8,11
湖南省自然科学基金联合基金项目(2025JJ80465)长沙市卫生健康科研课题(KJ-202548)
评论