基于心脏磁共振成像构建扩张型心肌病患者1年内发生主要不良心血管事件预测模型OA
Construction of a predictive model for MACE within one year in patients with dilated cardiomyopathy based on CMR
目的 基于心脏磁共振成像(CMR)数据构建扩张型心肌病(DCM)患者1年内发生主要不良心血管事件(MACE)的预测模型,探讨延迟钆增强(late gadolinium enhancement,LGE)在DCM患者发生MACE中的预测价值.方法 连续性收集2023年9月至2024年12月在新疆维吾尔自治区人民医院心内科住院确诊为DCM患者92例的基本资料、CMR数据、实验室检测结果、超声心动图资料等.通过定期随访,根据是否发生MACE分为MACE组及非MACE组.利用Lasso回归筛选出DCM患者1年内发生MACE相关的最佳预测因子,并构建预测列线图模型.采用Kaplan-Meier曲线分析LGE与患者预后的关系.结果 92例患者平均随访7.5个月(0.3～12个月)后分为MACE组32例,非MACE组60例.使用患者临床资料、CMR数据、实验室检测结果、超声心动图数据等制作数据集,使用Lasso回归及10折交叉验证筛选出LGE、脑钠肽(brain natriuretic peptide,BNP)及基于超声心动图测量的超声左心室射血分数(LVEF-Ultrasound)是最佳预测因子.使用以上3个预测因子构建出DCM患者1年内发生MACE的列线图模型.模型在全部研究对象中的AUC值为0.83(95％CI 0.68～0.97).校准曲线评估模型的预测概率与实际发生的概率偏差较小.决策曲线揭示了模型在不同风险阈值下提供的净收益.Kaplan-Meier生存曲线分析显示LGE阳性患者的预后显著差于LGE阴性患者(log-rank,P＜0.001).结论 在DCM患者中,LGE阳性、BNP及LVEF-Ultrasound是DCM患者发生MACE的独立预测因子,且预测模型在识别患者发生MACE中具有一定的临床价值;LGE阳性提示DCM患者预后更差.
Objective To construct a predictive model for major adverse cardiovascular events(MACE)within one year in patients with dilated cardiomyopathy(DCM)based on cardiac magnetic resonance imaging(CMR)data and to explore the value of late gadolinium enhancement in predicting MACE in DCM patients.Methods The basic data,CMR data,laboratory test results and echocardiography data of 92 DCM patients admitted to the Department of Cardiology,People's Hospital of Xinjiang Uygur Autonomous Region from September 2023 to December 2024 were collected continuously.According to occurrence of MACE during the follow-up,DCM patients were divided into MACE group and non-MACE group.Lasso regression was used to screen for the best predictors of MACE occurrence within 1 year,and the predictive nomogram model was constructed.Kaplan-Meier curve was used to analyze the relationship between LGE and prognosis.Results Totally 92 patients were followed up for an average of 7.5 months(0.3-12 months).After that,they were divided into the MACE group(32 cases)and the non-MACE group(60 cases).A dataset was made using patients' clinical data,CMR data,laboratory test results,echocardiography data,etc.,and Lasso regression and 10-fold cross-validation were used to screen out that LGE,BNP and left ventricular ejection fraction based on echocardiography(LVEF-Ultrasound)were the best predictors.A nomographic model of MACE within 1 year in DCM patients was constructed using the above 3 predictors.The AUC value of the model in all the research subjects was 0.83(95％CI:0.68-0.97).The difference in predictive probability by the model was relatively small compared with the actual probability.The decision curve revealed the net profit of the model under different risk threshhold.Kaplan-Meier survival curve analysis showed that the prognosis of LGE positive patients was significantly worse than that of LGE negative patients.Conclusion Positive LGE,BNP and LVEF-Ultrasound are independent predictors of MACE in DCM patients,and predictive models have some clinical value in identifying MACE events.Positive LGE indicates a poorer prognosis for DCM patients.
赛亚热木·西尔扎提;周航宇;王婷婷;王甜甜;李杰
新疆维吾尔自治区人民医院心脏及泛血管医学诊疗中心,新疆维吾尔自治区乌鲁木齐 830001新疆维吾尔自治区人民医院心脏及泛血管医学诊疗中心,新疆维吾尔自治区乌鲁木齐 830001新疆维吾尔自治区人民医院心脏及泛血管医学诊疗中心,新疆维吾尔自治区乌鲁木齐 830001新疆维吾尔自治区人民医院心脏及泛血管医学诊疗中心,新疆维吾尔自治区乌鲁木齐 830001新疆维吾尔自治区人民医院心脏及泛血管医学诊疗中心,新疆维吾尔自治区乌鲁木齐 830001
医药卫生
扩张型心肌病心脏磁共振成像风险预测模型
dilated cardiomyopathycardiac magnetic resonance imagingrisk prediction model
《中国实用内科杂志》 2026 (5)
406-412,7
新疆维吾尔自治区自然科学基金(2025D01C154)
评论