冠状动脉慢性完全闭塞病人PCI开通后近期MACE发生风险的列线图模型构建

Construction of Nomogram Model for the Risk of MACE in Patients with Chronic Total Occlusion of Coronary Arteries after PCI

目的:构建预测冠状动脉慢性完全闭塞(CTO)病人经皮冠状动脉介入(PCI)开通后近期主要心血管不良事件(MACE)发生风险的列线图模型,并评价该模型的预测准确性与临床适用性。方法:选取2018年1月—2021年6月于我院行PCI介入开通治疗的冠状动脉CTO病人,所有病人均随访1年,根据1年随访结果将病人分为MACE发生组和MACE未发生组。从整体研究对象中随机选取70%病人作为训练集,30%作为测试集,在训练集中利用Logistic回归分析病人发生MACE的独立危险因素,基于分析得到的独立危险因素构建列线图模型,对构建的列线图模型从3方面进行验证:通过受试者工作特征(ROC)曲线评价列线图模型的预测效能,绘制校准曲线评价及验证风险预测模型的一致性,利用临床决策曲线验证模型的获益性。结果:共纳入病人280例,发生MACE病人96例;其中训练集病人196例,发生MACE病人67例,测试集病人84例,发生MACE病人29例。通过Logistic回归分析训练集病人发生MACE的独立危险因素包括PCI介入开通失败、年龄较大、吸烟史、合并糖尿病、左室射血分数(LVEF)较低和低密度脂蛋白胆固醇(LDL-C)较高。构建的列线图模型ROC曲线验证后显示,训练集和测试集均具有良好的预测效能,训练集ROC曲线下面积(AUC)为0.854,测试集AUC为0.832。绘制的校准曲线显示,训练集和测试集均具有良好的拟合度,训练集一致性系数(C-index)为0.879,测试集C-index为0.856;临床决策曲线在训练集中的最佳适用区间为0.04~0.95,测试集为0.06~0.91,提示均具有较大的临床获益性。结论:PCI介入开通失败、年龄较大、吸烟史、合并糖尿病、LVEF较低和LDL-C较高是导致冠状动脉CTO病人PCI开通治疗后近期MACE发生的独立危险因素,基于上述独立危险因素构建的列线图模型具有良好的预测准确性和临床适用性。

Objective:To construct the nomogram model for predicting the risk of recent major adverse cardiovascular events(MACE)after percutaneous coronary intervention(PCI)in patients with chronic total occlusion(CTO)of the coronary arteries,and to evaluate the predictive accuracy.Methods:The CTO of coronary arteries patients who underwent PCI interventional treatment were selected,and followed up for 1 year.Based on the results of the 1-year follow-up,the patients were divided into the MACE occurrence group and the MACE non-occurrence group.There were the patients(70%)from the overall study were randomly selected as the training set and 30%patients as the test set Logistic regression was used to analyze the independent risk factors for MACE in the training set,and nomogram model was constructed based on the independent risk factors obtained from the analysis.The constructed column chart model was validated in 3 aspects:evaluating the predictive efficacy of the column chart model by receiver operating characteristic(ROC)curve,plotting calibration curves to evaluate and validate the consistency of the risk prediction model,and validating the model′s benefit by the clinical decision curve.Results:A total of 280 patients were included,including 96 patients with MACE.Among the 196 patients in the training set,and 67 patients with MACE,there were 84 patients in the test set,29 patients with MACE.Independent risk factors for MACE in the training set analyzed by Logistic regression were PCI intervention failure,older age,smoking,diabetes mellitus,lower left ventricular ejection fraction(LVEF),and higher low-density lipoprotein cholesterol(LDL-C).ROC curve verification of the constructed nomogram model showed that both the training set and the test set showed better prediction efficiency,and the area under ROC curve(AUC)of the training set was 0.854,and that of the AUC of the test set was 0.832.The calibration curves showed better predictive performance for both the training and test sets,with concordance index(C-index)of 0.879 for the training set and concordance index(C-index)of 0.856 for the test set.The optimal range of clinical decision curve in the training set was 0.04-0.95,that in the test set was 0.06-0.91,suggesting that both of them showed great clinical benefit.Conclusion:PCI intervention failure,older age,smoking,diabetes mellitus,low LVEF and high LDL-C were independent risk factors for MACE with CTO of coronary arteries patients.The nomogram model constructed on the basis of the above independent risk factors had good predictive accuracy and clinical applicability.