ISENet: a deep learning model for detecting ischemic ST changes in long-term ECG monitoring.
Saved in:
| Title: | ISENet: a deep learning model for detecting ischemic ST changes in long-term ECG monitoring. |
|---|---|
| Authors: | Lin CC; Department of Electrical Engineering, National Chin-Yi University of Technology, Taichung, Taiwan. cclin@ncut.edu.tw., Yeh CY; Department of Electrical Engineering, National Chin-Yi University of Technology, Taichung, Taiwan., Lin JH; Department of Electrical Engineering, National Chin-Yi University of Technology, Taichung, Taiwan. |
| Source: | Medical & biological engineering & computing [Med Biol Eng Comput] 2025 Dec; Vol. 63 (12), pp. 3589-3609. Date of Electronic Publication: 2025 Jul 19. |
| Publication Type: | Journal Article |
| Language: | English |
| Journal Info: | Publisher: Springer Country of Publication: United States NLM ID: 7704869 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1741-0444 (Electronic) Linking ISSN: 01400118 NLM ISO Abbreviation: Med Biol Eng Comput Subsets: MEDLINE |
| Imprint Name(s): | Publication: New York, NY : Springer Original Publication: Stevenage, Eng., Peregrinus. |
| MeSH Terms: | Deep Learning* , Electrocardiography*/methods , Myocardial Ischemia*/diagnosis , Myocardial Ischemia*/physiopathology, Humans ; Signal Processing, Computer-Assisted ; Neural Networks, Computer ; Databases, Factual ; Algorithms ; Heart Rate/physiology |
| Abstract: | Long-term ECG monitoring is crucial for detecting asymptomatic or intermittent myocardial ischemia, as it mitigates irreversible cardiac damage and prevents disease progression. Myocardial ischemia appears on ECG as transient ST-segment level and morphology alterations, known as ischemic ST change events (ISE). However, automatically identifying ISE based on ECG signals is challenging, as its recognition is highly susceptible to interference from non-ischemic ST change events, including heart rate-related ST change events (HRE), axis shift events (ASE), and conduction change events (CCE). To address this challenge, this study proposes ISENet, a lightweight deep learning-based neural network for ISE detection. The model was trained and evaluated using ECG signals and annotations from the PhysioNet long-term ST database, with tenfold cross-validation to ensure robustness and generalizability. Experimental results show that ISENet achieves an average ISE detection accuracy of 83.5%, surpassing benchmark models like VGG19 and ResNet50 while significantly reducing model complexity. This study is the first to apply a deep learning-based neural network for ISE detection using ECG signals from the long-term ST database. Compared to previous feature-engineering and feature-learning approaches, ISENet addresses key limitations in experimental design and methodology, representing a significant advancement in automated myocardial ischemia detection. (© 2025. International Federation for Medical and Biological Engineering.) |
| References: | Deanfield JE, Shea M, Kensett M, Horlock P, Wilson RA, de Landsheere CM et al (1984) Silent myocardial ischaemia due to mental stress. Lancet 2(8410):1001–1005. https://doi.org/10.1016/s0140-6736(84)91106-1. (PMID: 10.1016/s0140-6736(84)91106-16149394) Andrews TC, Fenton T, Toyosaki N, Glasser SP, Young PM, MacCallum G et al (1993) Subsets of ambulatory myocardial ischemia based on heart rate activity. Circadian distribution and response to anti-ischemic medication. The Angina and Silent Ischemia Study Group (ASIS). Circulation 88(1):92–100. https://doi.org/10.1161/01.cir.88.1.92. (PMID: 10.1161/01.cir.88.1.928319361) Ansari S, Farzaneh N, Duda M, Horan K, Andersson HB, Goldberger ZD, Nallamothu BK, Najarian K (2017) A review of automated methods for detection of myocardial ischemia and infarction using electrocardiogram and electronic health records. IEEE Rev Biomed Eng 10:264–298. https://doi.org/10.1109/rbme.2017.2757953. (PMID: 10.1109/rbme.2017.2757953290352259044695) Xiong P, Lee SM, Chan G (2022) Deep learning for detecting and locating myocardial infarction by electrocardiogram: a literature review. Front Cardiovasc Med 9:860032. https://doi.org/10.3389/fcvm.2022.860032. (PMID: 10.3389/fcvm.2022.860032354025638990170) Zhao X, Zhang J, Gong Y, Xu L, Liu H, Wei S, Wu Y, Cha G, Wei H, Mao J, Xia L (2022) Reliable detection of myocardial ischemia using machine learning based on temporal-spatial characteristics of electrocardiogram and vectorcardiogram. Front Physiol 13:854191. https://doi.org/10.3389/fphys.2022.854191. (PMID: 10.3389/fphys.2022.854191357070129192098) Wang CL, Wei CC, Tsai CT, Lee YH, Liu LY, Chen KY, Lin YJ, Lin PL (2023) Early detection of myocardial ischemia in resting ECG: analysis by HHT. Biomed Eng Online 22(1):23. https://doi.org/10.1186/s12938-023-01089-9. (PMID: 10.1186/s12938-023-01089-9368949849999640) Gragnaniello M, Borghese A, Marrazzo VR, Maresca L, Breglio G, Irace A, Riccio M (2024) Real-time myocardial infarction detection approaches with a microcontroller-based Edge-AI device. Sensors (Basel) 24(3):828. https://doi.org/10.3390/s24030828. (PMID: 10.3390/s2403082838339545) Bourque JM, Beller GA (2015) Value of exercise ECG for risk stratification in suspected or known CAD in the era of advanced imaging technologies. JACC Cardiovasc Imaging 8(11):1309–1321. https://doi.org/10.1016/j.jcmg.2015.09.006. (PMID: 10.1016/j.jcmg.2015.09.006265638614646721) Garzillo CL, Hueb W, Gersh B, Rezende PC, Lima EG, Favarato D, Franchini Ramires JA, Kalil Filho R (2019) Association between stress testing-induced myocardial ischemia and clinical events in patients with multivessel coronary artery disease. JAMA Intern Med 179(10):1345–1351. https://doi.org/10.1001/jamainternmed.2019.2227. (PMID: 10.1001/jamainternmed.2019.2227313292216647357) Thorild P, Mourtzinis G (2023) The value of exercise electrocardiography in outpatients with stable chest pain and low pre-test probability of significant coronary artery disease. J Clin Med 12(14):4670. https://doi.org/10.3390/jcm12144670. (PMID: 10.3390/jcm121446703751078510380256) Huang W, Huang Z, Koh NSY, Ho JS, Chua TSJ, Tan SY (2020) Adding coronary calcium score to exercise treadmill test: an alternative to refine coronary artery disease risk stratification in patients with intermediate risk chest pain. Glob Heart 15(1):22. https://doi.org/10.5334/gh.766. (PMID: 10.5334/gh.766324897957218771) Heimann J, Schramm T, Grauduszus M, Joisten C (2025) Examining the effects of different exertion levels and reference standards on the diagnostic accuracy of exercise ECG for detecting coronary artery disease - a pilot study. Ann Noninvasive Electrocardiol 30(1):e70042. https://doi.org/10.1111/anec.70042. (PMID: 10.1111/anec.7004239707636) Gul Z, Shams P, Makaryus AN (2024) Silent myocardial ischemia. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2025 Jan. Bonow RO, Mann DL, Zipes DP, Libby P (2011) Braunwald’s heart disease: a textbook of cardiovascular medicine, 9th edn. Elsevier Science, Philadelphia. Hampton JR (2003) The ECG in practice, 4th edn. Churchill Livingstone, Oxford. Taddei A, Distante G, Emdin M, Pisani P, Moody GB, Zeelenberg C et al (1992) The European ST-T database: standard for evaluating systems for the analysis of ST-T changes in ambulatory electrocardiography. Eur Heart J 13:1164–1172. https://doi.org/10.1093/oxfordjournals.eurheartj.a060332. (PMID: 10.1093/oxfordjournals.eurheartj.a0603321396824) Cerutti S, Mainardi L, Bianchi A, Signorini MG, Bertinelli M (1992) Time-variant autoregressive spectral estimation in acute ischemic episodes. Proc Comput Cardiol 315–318. https://doi.org/10.1109/CIC.1992.269368. Presedo J, Fernandez EA, Vila J, Barro S (1996) Cycles of ECG parameter evolution during ischemic episodes. Proc Comput Cardiol 23:489–492. https://doi.org/10.1109/CIC.1996.542580. (PMID: 10.1109/CIC.1996.542580) Emdin M, Taddei A, Varanini M, Raciti M, Pola S, Marchesi C et al (1997) Electrocardiographic and signal monitoring in ischemic heart disease: state of the art and perspective. J Med Eng Technol 21(5):162–165. https://doi.org/10.3109/03091909709016222. (PMID: 10.3109/030919097090162229350595) Taddei A, Emdin M, Varanini M, Nassi G, Bertinelli M, Picano E, Marchesi C (1997) Imaging-documented cardiovascular signal database for assessing methods for ischemia analysis. J Med Eng Technol 21(5):169–173. https://doi.org/10.3109/03091909709016224. (PMID: 10.3109/030919097090162249350597) Jager F, Taddei A, Moody GB, Emdin M, Antolic G, Dorn R et al (2003) Long-term ST database: a reference for the development and evaluation of automated ischemia detectors and for the study of the dynamics of myocardial ischemia. Med Biol Eng Comput 41:172–182. https://doi.org/10.1007/bf02344885. (PMID: 10.1007/bf0234488512691437) Vilakazi CB, Tarassenko L, Clifford GD (2011) Distinguishing between supply ischemic and non-supply ischemic ST events using a relevance vector machine. Comput Cardiol 38:633–636 ( https://ieeexplore.ieee.org/document/6164645 ). Amon M, Jager F (2016) Electrocardiogram ST-segment morphology delineation method using orthogonal transformations. PLoS ONE 11(2):e0148814. https://doi.org/10.1371/journal.pone.0148814. (PMID: 10.1371/journal.pone.0148814268631404749300) Fernández Biscay C, Arini PD, Rincón Soler AI, Bonomini MP (2020) Classification of ischemic and non-ischemic cardiac events in Holter recordings based on the continuous wavelet transform. Med Biol Eng Comput 58(5):1069–1078. https://doi.org/10.1007/s11517-020-02134-8. (PMID: 10.1007/s11517-020-02134-832157593) Xiao R, Xu Y, Pelter MM, Mortara DW, Hu X (2018) A deep learning approach to examine ischemic ST changes in ambulatory ECG recordings. AMIA Jt Summits Transl Sci Proc 2017:256–262 ( http://www.ncbi.nlm.nih.gov/pmc/articles/pmc5961830/ ). (PMID: 29888083) Xiao R, Xu Y, Pelter MM, Fidler R, Badilini F, Mortara DW et al (2018) Monitoring significant ST changes through deep learning. J Electrocardiol 51(6S):S78–S82. https://doi.org/10.1016/j.jelectrocard.2018.07.026. (PMID: 10.1016/j.jelectrocard.2018.07.026300820876261793) Simonyan K, Zisserman A (2015) Very deep convolutional networks for large-scale image recognition. Proc 3rd Int Conf Learn Represent (ICLR 2015) 1–14. https://ora.ox.ac.uk/objects/uuid:60713f18-a6d1-4d97-8f45-b60ad8aebbce . Accessed 15 Jul 2025. He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. Proc IEEE Conf Comput Vis Pattern Recognit (CVPR) 770–778. https://doi.org/10.1109/CVPR.2016.90. Martínez JP, Laguna P, Olmos S, Pahlm O, Pettersson J, Sörnmo L (2007) Assessment of QT-measurement accuracy using the 12-lead electrocardiogram derived from EASI leads. J Electrocardiol 40(2):172–179. https://doi.org/10.1016/j.jelectrocard.2006.08.089. (PMID: 10.1016/j.jelectrocard.2006.08.08917027840) Jager F, Moody GB, Taddei A, Antolic G, Emdin M, Smrdel A et al (1998) A long-term ST database for development and evaluation of ischemia detectors. Comput Cardiol 25:301–304. https://doi.org/10.1109/CIC.1998.731793. (PMID: 10.1109/CIC.1998.731793) Jager F, Moody GB, Taddei A, Emdin M, Antolic G, Smrdel A et al (2000) The long-term ST database: a research resource for algorithm development and physiologic studies of transient myocardial ischemia. Comput Cardiol 27:841–844. https://doi.org/10.1109/CIC.2000.898656. (PMID: 10.1109/CIC.2000.898656) Foret P, Kleiner A, Mobahi H, Neyshabur B (2021) Sharpness-aware minimization for efficiently improving generalization. arXiv2021, arXiv:2010.01412v3 [cs.LG]. https://doi.org/10.48550/arXiv.2010.01412. Zweig MH, Campbell G (1993) Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clin Chem 39:561–577. https://doi.org/10.1093/clinchem/39.4.561. (PMID: 10.1093/clinchem/39.4.5618472349) |
| Grant Information: | MOST 111-2637-E-167-003- National Science and Technology Council; NSTC 112-2637-E-167-003- National Science and Technology Council |
| Contributed Indexing: | Keywords: Convolutional neural network; Deep learning; Ischemic ST change event; Myocardial ischemia; Residual neural network |
| Entry Date(s): | Date Created: 20250719 Date Completed: 20251203 Latest Revision: 20251203 |
| Update Code: | 20251204 |
| DOI: | 10.1007/s11517-025-03416-9 |
| PMID: | 40682722 |
| Database: | MEDLINE |
Full Text Finder 
Nájsť tento článok vo Web of Science | Abstract: | Long-term ECG monitoring is crucial for detecting asymptomatic or intermittent myocardial ischemia, as it mitigates irreversible cardiac damage and prevents disease progression. Myocardial ischemia appears on ECG as transient ST-segment level and morphology alterations, known as ischemic ST change events (ISE). However, automatically identifying ISE based on ECG signals is challenging, as its recognition is highly susceptible to interference from non-ischemic ST change events, including heart rate-related ST change events (HRE), axis shift events (ASE), and conduction change events (CCE). To address this challenge, this study proposes ISENet, a lightweight deep learning-based neural network for ISE detection. The model was trained and evaluated using ECG signals and annotations from the PhysioNet long-term ST database, with tenfold cross-validation to ensure robustness and generalizability. Experimental results show that ISENet achieves an average ISE detection accuracy of 83.5%, surpassing benchmark models like VGG19 and ResNet50 while significantly reducing model complexity. This study is the first to apply a deep learning-based neural network for ISE detection using ECG signals from the long-term ST database. Compared to previous feature-engineering and feature-learning approaches, ISENet addresses key limitations in experimental design and methodology, representing a significant advancement in automated myocardial ischemia detection.<br /> (© 2025. International Federation for Medical and Biological Engineering.) |
|---|---|
| ISSN: | 1741-0444 |
| DOI: | 10.1007/s11517-025-03416-9 |