Please cite this article as: SHANG Z, LIU Y, YUAN YY, WANG XY, YU HY, GAO W. Early identification of STEMI patients with emergency chest pain using lipidomics combined with machine learning. J Geriatr Cardiol 2022; 19(9): 685−695. DOI: 10.11909/j.issn.1671-5411.2022.09.003

Citation: Please cite this article as: SHANG Z, LIU Y, YUAN YY, WANG XY, YU HY, GAO W. Early identification of STEMI patients with emergency chest pain using lipidomics combined with machine learning. J Geriatr Cardiol 2022; 19(9): 685−695. DOI: 10.11909/j.issn.1671-5411.2022.09.003 doi: 10.11909/j.issn.1671-5411.2022.09.003 1.

Department of Cardiology, Peking University Third Hospital, NHC Key Laboratory of Cardiovascular Molecular Biology and Regulatory Peptides, Key Laboratory of Molecular Cardiovascular Science, Ministry of Education, Beijing, China

2.

Institute of Systems Biomedicine, Department of Pathology, School of Basic Medical Sciences, Peking-Tsinghua Center for Life Sciences, Peking University Health Science Center, Beijing, China

More Information Abstract

 OBJECTIVES To analyze the differential expression of lipid spectrum between ST-segment elevated myocardial infarction (STEMI) and patients with emergency chest pain and excluded coronary artery disease (CAD), and establish the predictive model which could predict STEMI in the early stage. METHODS We conducted a single-center, nested case-control study using the emergency chest pain cohort of Peking University Third Hospital. Untargeted lipidomics were conducted while LASSO regression as well as XGBoost combined with greedy algorithm were used to select lipid molecules. RESULTS Fifty-two STEMI patients along with 52 controls were enrolled. A total of 1925 lipid molecules were detected. There were 93 lipid molecules in the positive ion mode which were differentially expressed between the STEMI and the control group, while in the negative ion mode, there were 73 differentially expressed lipid molecules. In the positive ion mode, the differentially expressed lipid subclasses were mainly diacylglycerol (DG), lysophophatidylcholine (LPC), acylcarnitine (CAR), lysophosphatidyl ethanolamine (LPE), and phosphatidylcholine (PC), while in the negative ion mode, significantly expressed lipid subclasses were mainly free fatty acid (FA), LPE, PC, phosphatidylethanolamine (PE), and phosphatidylinositol (PI). LASSO regression selected 22 lipids while XGBoost combined with greedy algorithm selected 10 lipids. PC (15: 0/18: 2), PI (19: 4), and LPI (20: 3) were the overlapping lipid molecules selected by the two feature screening methods. Logistic model established using the three lipids had excellent performance in discrimination and calibration both in the derivation set (AUC: 0.972) and an internal validation set (AUC: 0.967). In 19 STEMI patients with normal cardiac troponin, 18 patients were correctly diagnosed using lipid model. CONCLUSIONS The differentially expressed lipids were mainly DG, CAR, LPC, LPE, PC, PI, PE, and FA. Using lipid molecules selected by XGBoost combined with greedy algorithm and LASSO regression to establish model could accurately predict STEMI even in the more earlier stage.

References [1]

Roth GA, Mensah GA, Johnson CO, et al. Burden of Cardiovascular Diseases Writing Group. Global burden of cardiovascular diseases and risk factors, 1990-2019: update from the GBD 2019 study. J Am Coll Cardiol 2020; 76: 2982−3021.

[2]

Gulati M, Levy PD, Mukherjee D, et al. 2021 AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR Guideline for the Evaluation and Diagnosis of Chest Pain: Executive Summary: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation 2021; 144: e368−e454.

[3] Hsia RY, Hale Z, Tabas JA. A National study of the prevalence of life-threatening diagnoses in patients with chest pain. JAMA Intern Med 2016; 176: 1029−1032. doi: 10.1001/jamainternmed.2016.2498 [4]

Aronson JK, Ferner RE. Biomarkers-a general review. Curr Protoc Pharmacol 2017; 76: 9.23.1−9.23.17.

[5] Poss AM, Maschek JA, Cox JE, et al. Machine learning reveals serum sphingolipids as cholesterol-independent biomarkers of coronary artery disease. J Clin Invest 2020; 130: 1363−1376. doi: 10.1172/JCI131838 [6] Mudrick DW, Chen AY, Roe MT, et al. Changes in glycoprotein IIb/IIIa inhibitor excess dosing with site-specific safety feedback in the Can Rapid risk stratification of Unstable angina patients Suppress ADverse outcomes with Early implementation of the ACC/AHA guidelines (CRUSADE) initiative. Am Heart J 2010; 160: 1072−1078. doi: 10.1016/j.ahj.2010.08.008 [7] van’t Hof AW, Rasoul S, van de Wetering H, et al; On-TIME study group. Feasibility and benefit of prehospital diagnosis, triage, and therapy by paramedics only in patients who are candidates for primary angioplasty for acute myocardial infarction. Am Heart J 2006; 151: 1255.e1−5. doi: 10.1016/j.ahj.2006.03.014 [8] Collins GS, Reitsma JB, Altman DG, et al. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ 2015; 350: g7594. doi: 10.1136/bmj.g7594 [9]

Thygesen K, Alpert JS, Jaffe AS, et al. Executive Group on behalf of the Joint European Society of Cardiology (ESC)/American College of Cardiology (ACC)/American Heart Association (AHA)/World Heart Federation (WHF). Task Force for the Universal Definition of Myocardial Infarction. Fourth Universal Definition of Myocardial Infarction (2018). J Am Coll Cardiol 2018; 72: 2231−2264.

[10] Ibanez B, James S, Agewall S, et al. 2017 ESC Guidelines for the management of acute myocardial infarction in patients presenting with ST-segment elevation: The Task Force for the management of acute myocardial infarction in patients presenting with ST-segment elevation of the European Society of Cardiology (ESC). Eur Heart J 2018; 39: 119−177. doi: 10.1093/eurheartj/ehx393 [11] Knuuti J, Wijns W, Saraste A, et al. 2019 ESC Guidelines for the diagnosis and management of chronic coronary syndromes. Eur Heart J 2020; 41: 407−477. doi: 10.1093/eurheartj/ehz425 [12]

Zhou J, Yin Y. Use of liquid chromatography-mass spectrometry-based metabolomics to identify biomarkers of tuberculosis. Methods Mol Biol 2019; 1859: 241−251.

[13] Fahy E, Subramaniam S, Brown HA, et al. A comprehensive classification system for lipids. J Lipid Res 2005; 46: 839−861. doi: 10.1194/jlr.E400004-JLR200 [14] Bunea F, She Y, Ombao H, et al. Penalized least squares regression methods and applications to neuroimaging. Neuroimage 2011; 55: 1519−1527. doi: 10.1016/j.neuroimage.2010.12.028 [15]

Li J, Gong M, Joshi Y, et al. Machine learning prediction model for acute renal failure after acute aortic syndrome surgery. Front Med (Lausanne) 2022; 8: 728521.

[16] Ruan Y, Bellot A, Moysova Z, et al. Predicting the risk of inpatient hypoglycemia with machine learning using electronic health records. Diabetes Care 2020; 43: 1504−1511. doi: 10.2337/dc19-1743 [17] Wang G, Yao H, Gong Y, et al. Metabolic detection and systems analyses of pancreatic ductal adenocarcinoma through machine learning, lipidomics, and multi-omics. Sci Adv 2021; 7: eabh2724. doi: 10.1126/sciadv.abh2724 [18] Wang G, Qiu M, Xing X, et al. Lung cancer scRNA-seq and lipidomics reveal aberrant lipid metabolism for early-stage diagnosis. Sci Transl Med 2022; 14: eabk2756. doi: 10.1126/scitranslmed.abk2756 Proportional views