机器学习确定特征性生物标志物预测主动脉瓣置换术后不良心血管事件

林锐; 王媛; 周宁; 严昱; 周翔天; 杜杰

doi:10.13201/j.issn.1001-1439.2021.03.013

机器学习确定特征性生物标志物预测主动脉瓣置换术后不良心血管事件

- 1. 温州医科大学附属第二医院心内科(浙江温州, 325035);
- 2. 首都医科大学附属北京安贞医院北京市心肺血管疾病研究所血管生物研究室;
- 3. 温州医科大学眼视光医院省部共建眼视光学和视觉科学国家重点实验室
基金项目:
国家自然科学基金(No:81861128025)

详细信息

通讯作者: 杜杰,E-mail:jiedu@yahoo.com

中图分类号: R542.5

收稿日期: 2021-01-22

Machine learning to identify characteristic biomarkers for predicting adverse cardiovascular events after aortic valve replacement

- 1. Department of Cardiology, The Second Affiliated Hospital of Wenzhou Medical University, Wenzhou, Zhejiang, 325035, China;
- 2. Beijing Anzhen Hospital, Capital Medical University, Beijing Institute of Heart, Lung and Blood Vessel Disease;
- 3. State Key Laboratory of Ophthalmic Optics and Visual Science, Optometry, Eye Hospital, Wenzhou Medical University

More Information

Corresponding author: DU Jie, E-mail: jiedu@yahoo.com

Received Date: 22 January 2021

摘要

摘要: 目的:使用机器学习确定与主动脉瓣置换手术后不良心血管事件相关的特征性生物标志物,并验证其预测不良心血管事件发生的价值。方法:纳入2016-2019年在北京安贞医院住院行主动脉瓣置换手术的患者共1455例,收集其临床特征和实验室检查指标。对所有患者进行1年以上的术后随访,以全因死亡、心力衰竭或心肌梗死发作再入院作为不良心血管事件,使用机器学习筛选特征性生物标志物,进一步验证其预测预后的价值。结果:使用多种机器学习构建模型确定特征性生物标志物为术前B型钠尿肽和术后24 h的超敏肌钙蛋白I。COX比例风险模型结果显示,术前B型钠尿肽[HR(95%CI):1.758(1.191,2.595),P<0.01]与术后超敏肌钙蛋白I[HR(95%CI):1.830(1.137,2.945),P=0.013]是主动脉瓣置换术后不良事件发生的独立危险因素。当术前B型钠尿肽和术后24 h的超敏肌钙蛋白I共同升高时,能在一定程度上预测主动脉瓣置换患者不良事件发生[HR(95%CI):1.937(1.072,3.500),P=0.028)]。结论:基于机器学习方法确定特征性生物标志物为术前B型钠尿肽和术后24 h的超敏肌钙蛋白I,二者联合可用于预测主动脉瓣置换术后的不良心血管事件。
- 主动脉瓣置换术 /
- 机器学习 /
- 血清生物标志物 /
- 不良事件
Abstract: Objective: To use machine learning to identify the characteristic biomarkers related to major adverse cardiovascular events(MACE) after aortic valve replacement, and to verify their prognostic value.Methods: A total of 1455 patients who underwent aortic valve replacement in Beijing Anzhen Hospital from 2016 to 2019 were enrolled in the study. Their clinical characteristics and the value of serum biomarkers were collected. All patients were followed up for more than 1 year. Outcomes were defined as all-cause death, admission to hospital with heart failure or myocardial infarction. Machine learning was used to screen characteristic biomarkers and verify their prognostic value.Results: The machine learning identified the characteristic biomarkers as preoperative natriuretic peptide-B and high-sensitivity troponin I 24 hours after operation. COX regression model showed that preoperative natriuretic peptide-B(HR[95%CI]: 1.758[1.191,2.595], P<0.01) and high-sensitivity troponin I(HR[95%CI]: 1.830[1.137,2.945], P=0.013) were independent risk factors for MACE after aortic valve replacement. The increase of the both natriuretic peptide-B and high-sensitivity troponin I could predict the occurrence of MACE(HR[95%CI]: 1.937[1.072,3.500], P=0.028).Conclusion: The machine learning identifies the characteristic biomarkers are preoperative natriuretic peptide-B and post-operative high-sensitivity troponin I. Combination of the two biomarkers could predict MACE after aortic valve replacement.
- aortic valve replacement /
- machine learning /
- biomarkers /
- adverse events

HTML全文

参考文献(20)

[1]	Durko AP,Kappetein AP.Long-term survival after surgical aortic valve replacement:expectations and reality[J].J Am Coll Cardiol,2019,74(1):34-35.
[2]	Bonow RO,Leon MB,Doshi D,et al.Management strategies and future challenges for aortic valve disease[J].Lancet,2016,387(10025):1312-23.
[3]	Jiang H,Vánky F,Hultkvist H,et al.NT-proBNP and postoperative heart failure in surgery for aortic stenosis[J].Open Heart,2019,6(1):e001063.
[4]	Mazine A,David TE,Rao V,et al.Long-term outcomes of the ross procedure versus mechanical aortic valve replacement:propensity-matched cohort study[J].Circulation,2016,134(8):576-85.
[5]	Steiner J,Rodés-Cabau J,Holmes DR Jr,et al.Mechanical intervention for aortic valve stenosis in patients with heart failure and reduced ejection fraction[J].J Am Coll Cardiol,2017,70(24):3026-3041.
[6]	Wang TJ.The evolution of the cardiovascular biomarker study[J].Circulation,2020,142(15):1422-1424.
[7]	Ibrahim NE,Januzzi JL Jr.Established and emerging roles of biomarkers in heart failure[J].Circ Res,2018,123(5):614-629.
[8]	Shang N,Weng C,Hripcsak G.A conceptual framework for evaluating data suitability for observational studies[J].J Am Med Inform Assoc,2018,25(3):248-258.
[9]	Kilic A.Artificial intelligence and machine learning in cardiovascular health care[J].Ann Thorac Surg,2020,109(5):1323-1329.
[10]	Andell P,Li X,Martinsson A,et al.Epidemiology of valvular heart disease in a Swedish nationwide hospital-based register study[J].Heart,2017,103(21):1696-1703.
[11]	Khan AA,Murtaza G,Khalid MF,et al.Risk stratification for transcatheter aortic valve replacement[J].Cardiol Res,2019,10(6):323-330.
[12]	Wang C,Tang YF,Zhang JJ,et al.Comparison of four risk scores for in-hospital mortality in patients undergoing heart valve surgery:A multicenter study in a Chinese population[J].Heart Lung,2016,45(5):423-428.
[13]	Deeb GM,Reardon MJ,Chetcuti S,et al.3-Year outcomes in high-risk patients who underwent surgical or transcatheter aortic valve replacement[J].J Am Coll Cardiol,2016,67(22):2565-2574.
[14]	Mortazavi BJ,Downing NS,Bucholz EM,et al.Analysis of machine learning techniques for heart failure readmissions[J].Circ Cardiovasc Qual Outcomes,2016,9(6):629-640.
[15]	Zhang B,Xu H,Zhang H,et al.Prognostic value of N-terminal pro-B-type natriuretic peptide in elderly patients with valvular heart disease[J].J Am Coll Cardiol,2020,75(14):1659-1672.
[16]	Gardezi SK,Coffey S,Prendergast BD,et al.Serum biomarkers in valvular heart disease[J].Heart,2018,104(4):349-358.
[17]	Dweck MR,Everett RJ.Multibiomarker strategies in aortic stenosis[J].JACC Cardiovasc Imaging,2018,11(7):948-950.
[18]	Sandoval Y,Nowak R,deFilippi CR,et al.Myocardial infarction risk stratification with a single measurement of high-sensitivity troponin I[J].J Am Coll Cardiol,2019,74(3):271-282.
[19]	Beattie WS,Wijeysundera DN,Chan M,et al.Survival after isolated post-operative troponin elevation[J].J Am Coll Cardiol,2017,70(7):907-908.
[20]	Duchnowski P,Hryniewiecki T,Koźma M,et al.High-sensitivity troponin T is a prognostic marker of hemodynamic instability in patients undergoing valve surgery[J].Biomark Med,2018,12(12):1303-1309.