Ensemble Learning-Based Mortality Prediction After Acute
Myocardial Infarction

doi:10.1007/s12204-023-2611-1

Abstract

Abstract: A mortality prediction model based on small acute myocardial infarction (AMI) patients coherent with low death rate is established. In total, 1 639 AMI patients are selected as research objects who received treatment in seven tertiary and secondary hospitals in Shanghai between January 1, 2016 and January 1, 2018. Among them, 72 patients deceased during the two-year follow-up. Models are established with ensemble learning framework and machine learning algorithms based on 51 physiological indicators of the patient. Shapley additive explanations algorithm and univariate test with point-biserial and phi correlation coefficients are employed to determine significant features and rank feature importance. Based on 5-fold cross validation experiment and external validation, prediction model with self-paced ensemble framework and random forest algorithm achieves the best performance with area under receiver operating characteristic curve (AUROC) score of 0.911 and recall of 0.864. Both feature ranking methods showed that ejection fractions, serum creatinine (admission), hemoglobin and Killip class are the most important features. With these top-ranked features, the simplified prediction model is capable of achieving a comparable result with AUROC score of 0.872 and recall of 0.818. This work proposes a new method to establish mortality prediction models for AMI patients based on self-paced ensemble framework, which allows models to achieve high performance with small scale of patients coherent with low death rate. It will assist in medical decision and prognosis as a new reference.

Key words: acute myocardial infarction (AMI), ensemble learning, machine learning, feature engineering

摘要： 本研究构建了一个基于小规模低死亡率的急性心肌梗死患者群体的死亡预测模型。研究数据来源于2016年1月1日至2018年1月1日间就诊于上海市七家三级及二级医院的急性心肌梗死患者，共计1639例。其中72名患者在2年的随访期间内死亡。预测模型利用集成学习框架采用这些患者的51项生理指标进行构建，并通过SHAP算法和基于二列相关系数与Phi相关系数的单变量检验方法计算这些特征的重要性并进行排序。根据五折交叉检验实验和外部检验，采用自步集成学习框架和随机森林算法构建的预测模型可以取得最优的实验性能，其受试者工作特性曲线线下面积（AUROC）为0.911，召回率为0.864。上述两种特征排序方法均表明射血分数，血肌酐（入院），血红蛋白和Killip分级为与患者死亡最相关的重要特征。基于这些重要特征，本研究进一步构建了一个简化的预测模型，并取得了AUROC为0.872，召回率为0.818的接近使用特征全集构建的完备模型的实验性能。本研究提出了一种基于自步集成学习框架的急性心肌梗死患者死亡预测模型的构建方法，并且可以在小规模低死亡率的患者群体上构建高预测性能的模型。此项工作将在进行医疗决策和诊断的过程中提供参考和辅助。

关键词: 急性心肌梗死，集成学习，机器学习，特征工程

CLC Number:

R542.2+2

YAN Mingxuan¹ (颜铭萱), MIAO Yutong^2，3 (苗雨桐), SHENG Shuqian¹ (盛淑茜), GAN Xiaoying¹ (甘小莺), HE Ben² (何奔), SHEN Lan^2，3* (沈兰). Ensemble Learning-Based Mortality Prediction After Acute Myocardial Infarction[J]. J Shanghai Jiaotong Univ Sci, 2025, 30(1): 153-165.

References

[1] MORROW D A, ANTMAN E M, CHARLESWORTH A, et al. TIMI risk score for ST-elevation myocardial infarction: A convenient, bedside, clinical score for risk assessment at presentation: An intravenous nPA for treatment of infarcting myocardium early II trial substudy [J]. Circulation, 2000, 102(17): 2031-2037.
[2] FOX K A A, DABBOUS O H, GOLDBERG R J, et al. Prediction of risk of death and myocardial infarction in the six months after presentation with acute coronary syndrome: Prospective multinational observational study (GRACE) [J]. BMJ (Clinical Research Ed), 2006, 333(7578): 1091.
[3] FABRIZIO, D’ASCENZO. TIMI, GRACE and alternative risk scores in Acute Coronary Syndromes: A meta-analysis of 40 derivation studies on 216, 552 patients and of 42 validation studies on 31, 625 patients [J]. Contemporary Clinical Trials, 2012, 33(3): 507-514.
[4] HEND, MANSOOR, PHARMD M, et al. Risk prediction model for in-hospital mortality in women with ST-elevation myocardial infarction: A machine learning approach [J]. Heart & Lung, 2017, 46(6): 405-411.
[5] FABRIZIO, D’ASCENZO, MD, et al. Machine learning-based prediction of adverse events following an acute coronary syndrome (PRAISE): A modelling study of pooled datasets [J]. The Lancet, 2021,397(10270): 199-207.
[6] CHAWLA N V, BOWYER K W, HALL L O, et al. SMOTE: Synthetic minority over-sampling technique [J]. Journal of Artificial Intelligence Research, 2002, 16: 321-357.
[7] ISHAQ A, SADIQ S, UMER M, et al. Improving the prediction of heart failure patients’ survival using SMOTE and effective data mining techniques [J]. IEEE Access, 2021, 9: 39707-39716.
[8] TAVARES T R, OLIVEIRA A L I, CABRAL G G, et al. Preprocessing unbalanced data using weighted support vector machines for prediction of heart disease in children [C]//2013 International Joint Conference on Neural Networks. Dallas: IEEE, 2014: 1-8.
[9] POLIKAR R. Ensemble based systems in decision making [J]. IEEE Circuits and Systems Magazine, 2006, 6(3): 21-45.
[10] ROKACH L. Ensemble-based classifiers [J]. Artificial Intelligence Review, 2010, 33(1): 1-39.
[11] WEI X, JIANG F, WEI F, et al. An ensemble model for diabetes diagnosis in large-scale and imbalanced dataset [C]//Computing Frontiers Conference. Siena: ACM, 2017: 71-78.
[12] LIU X Y, WU J X, ZHOU Z H. Exploratory undersampling for class-imbalance learning [J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2009, 39(2): 539-550.
[13] ZHANG J H, XIONG H Y, HUANG Y, et al. M-SEQ: Early detection of anxiety and depression via temporal orders of diagnoses in electronic health data [C]//2015 IEEE International Conference on Big Data. Santa Clara: IEEE, 2015: 2569-2577.
[14] LIU Z N, CAO W, GAO Z F, et al. Self-paced ensemble for highly imbalanced massive data classification [C]//2020 IEEE 36th International Conference on Data Engineering. Dallas: IEEE, 2020: 841-852.
[15] LUNDBERG S M, LEE S I. A unified approach to interpreting model predictions [C]//31st International Conference on Neural Information Processing Systems. Long Beach: NIPS, 2017: 4768-4777.
[16] GUILFORD J P. Psychometric methods [M]. 2nd ed. New York: McGraw-Hill, 1954.
[17] LV H C, YANG X L, WANG B Y, et al. Machine learning–driven models to predict prognostic outcomes in patients hospitalized with heart failure using electronic health records: Retrospective study [J]. Journal of Medical Internet Research, 2021, 23(4): e24996.
[18] KASHIRINA I L, FIRYULINA M A, BONDARENKO Y V, et al. Identification of risk factors for mortality after myocardial infarction using machine learning methods [C]//2021 XXIV International Conference on Soft Computing and Measurements. St. Petersburg: IEEE,2021: 233-236.
[19] MATTHEWS B W. Comparison of the predicted and observed secondary structure of T4 phage lysozyme [J]. Biochimica et Biophysica Acta (BBA) - Protein Structure, 1975, 405(2): 442-451.
[20] BOUGHORBEL S, JARRAY F, EL-ANBARI M. Optimal classifier for imbalanced data using Matthews Correlation Coefficient metric [J]. PLoS One, 2017,12(6): e0177678.
[21] CHICCO D. Ten quick tips for machine learning in computational biology [J]. BioData Mining, 2017, 10:35.
[22] BATISTA G E A P A, PRATI R C, MONARD M C.A study of the behavior of several methods for balancing machine learning training data [J]. ACM SIGKDD Explorations Newsletter, 2004, 6(1): 20-29.
[23] LIU Z N, KANG J, TONG H H, et al. IMBENS: Ensemble class-imbalanced learning in python [DB/OL]. (2021-11-24). https://arxiv.org/abs/2111.12776.
[24] NATEKIN A, KNOLL A. Gradient boosting machines,a tutorial [J]. Frontiers in Neurorobotics, 2013, 7: 21.
[25] CHEN T Q, GUESTRIN C. XGBoost: A scalable tree boosting system [C]//22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco: ACM, 2016: 785-794.
[26] HERNANDEZ-SUAREZ D F, KIM Y, VILLABLANCA P, et al. Machine learning prediction models for in-hospital mortality after transcatheter aortic valve replacement [J]. JACC : Cardiovascular Interventions, 2019, 12(14): 1328-1338.
[27] CHICCO D, JURMAN G. Machine learning can predict survival of patients with heart failure from serum creatinine and ejection fraction alone [J]. BMC Medical Informatics and Decision Making, 2020, 20(1): 16.

Ensemble Learning-Based Mortality Prediction After Acute Myocardial Infarction

基于集成学习的急性心肌梗死死亡预测

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

References

Related Articles 0

Recommended Articles 0

Metrics

Comments