基于LASSO回归的塔城地区人群冠状动脉狭窄程度预测模型构建及验证  

Construction and validation of a prediction model for coronary artery stenosis based on LASSO regression

作　　者：徐以康 刘蕾 刘丽敏[1] 马晶茹[1] 王嘉钰 马军 甄紫伊 XU Yikang;LIU Lei;LIU Limin;MA Jingru;WANG Jiayu;MA Jun;ZHEN Ziyi(Department of Cardiology,The Second Affiliated Hospital of Shenyang Medical College,Shenyang 110001,China;Department of Community Nursing,School of Nursing,Liaoning University of Tranditional Chinese Medicine,Shenyang 110033,China;Department of Cardiology,Tacheng People’s Hospital,Tacheng 834700,China;Department of Epidemiology and Health Statistics,School of Public Health,Shenyang Medical College,Shenyang 110034,China)

机构地区：[1]沈阳医学院附属第二医院心内科,沈阳110001 [2]辽宁中医药大学护理学院社区护理教研室,沈阳110033 [3]塔城市人民医院心内科,新疆塔城834700 [4]沈阳医学院公共卫生学院流行病与卫生统计学教研室,沈阳110034

出　　处：《中国医科大学学报》2025年第2期139-143,149,共6页Journal of China Medical University

基　　金：辽宁省科学技术计划科技援疆、援藏医疗专项(2023-MS-19);辽宁省教育厅基本科研项目(LJKMZ20221798)。

摘　　要：目的 分析新疆塔城地区人群冠状动脉中重度狭窄的危险因素,构建并验证冠状动脉狭窄程度列线图风险预测模型。方法 回顾性收集2021年1月至2023年6月于塔城市人民医院心内科住院治疗的629例患者的临床资料。用R语言软件将所有患者的临床资料纳入LASSO回归进行危险因素初筛。按7∶3的比例将629例患者随机分为训练组(440例)和验证组(189例)。训练组数据用于模型构建,以冠状动脉狭窄程度为因变量,将LASSO回归筛选出的变量作为自变量纳入logistic回归建模。验证组用于模型验证。基于logistic分析结果,用R语言软件构建冠状动脉狭窄程度预测的可视化列线图。应用曲线下面积(AUC)、临床决策曲线分析(DCA)及校准曲线评价模型的区分度、临床效用和校准度。结果 年龄、非汉族、高血压、高脂血症、脑血管病史是发生冠状动脉中重度狭窄的危险因素,纳入风险预测模型。训练组和验证组AUC分别为0.905 (95%CI:0.790~0.863)和0.864(95%CI:0.744~0.861),校准曲线预测值与实际值一致度较高(训练组和验证组Brier得分分别为0.03和0.14),模型的预测性能好,DCA结果提示本模型具有临床净获益。结论 本研究所构建塔城地区人群冠状动脉狭窄程度风险预测模型具有良好的预测性能,可为筛查冠状动脉中重度狭窄患者提供简便易行、经济、易推广的评估工具。Objective To analyze the risk factors for moderate-to-severe coronary artery stenosis in the population of Tacheng,Xinjiang Uygur Autonomous Region,and to construct and verify a nomogram prediction model for the degree of coronary artery stenosis.Methods We retrospectively selected 629 patients who were hospitalized in the Cardiovascular Department of Tacheng People's Hospital from January 2021 to June 2023.Using R language software,the sociodemographic data,disease-related data,and various laboratory indicators of the 629 patients were included in the initial screening of risk factors for use in the LASSO regression analysis using a random number table method.The 629 patients were divided into a training group(n = 440) and a validation group(n = 189) in a 7 ∶ 3 ratio.Data from the training group were used for model construction,with the degree of coronary artery stenosis as the dependent variable,and the variables selected by LASSO regression as independent variables in the logistic regression model.The validation group was used for model validation.Based on the results of the logistic regression analysis,a visual nomogram for predicting the degree of coronary artery stenosis was constructed using R language software.The discriminability,calibration,and clinical utility of the model were evaluated using the area under the receiver operating characteristic curve(AUC),a calibration curve,and decision curve analysis(DCA).Results Age,non-Han ethnicity,hypertension,hyperlipidemia,and a history of cerebrovascular disease were risk factors for moderate-to-severe coronary artery stenosis and were included in the risk prediction model.The AUC of the training group and the validation group were 0.905(95%CI:0.790-0.863) and 0.864(95%CI:0.744-0.861),respectively.The predicted values of the calibration curve were consistent with the actual values(Brier scores of the training and validation group:0.03 and 0.14,respectively).The predictive performance of the model was good,and the DCA results indicated that the model had net

关 键 词：冠状动脉狭窄程度 列线图 预测模型 LASSO回归 

分 类 号：R543.3[医药卫生—心血管疾病]