СРАВНИТЕЛЬНЫЙ АНАЛИЗ МЕТОДОВ МАШИННОГО ОБУЧЕНИЯ ДЛЯ ПРОГНОЗИРОВАНИЯ КОРОНАРНОЙ БОЛЕЗНИ СЕРДЦА: НА ОСНОВЕ НАБОРА ДАННЫХ UCI HEART DISEASE

Коронарная болезнь сердца остаётся одной из основных причин смертности и инвалидизации в мире. Своевременная диагностика позволяет снизить частоту осложнений и уменьшить нагрузку на систему здравоохранения. Однако традиционные методы остаются дорогими, инвазивными и ограниченно доступными. Недавние исследования подтверждают перспективность применения методов машинного обучения в клинической практике. В этой связи возникает вопрос: можно ли достоверно прогнозировать наличие заболевания, используя только клинико-демографические данные, без визуализационных методов?
Целью данной работы было оценить точность и практическую ценность таких моделей. На объединённом протоколе данных UCI Heart Disease (n = 920) был обучен алгоритм LightGBM, который показал следующие результаты: accuracy = 0.8696, precision = 0.8679, recall = 0.9020, F1-score = 0.8846. Полученные данные дополняют результаты предыдущих исследований, основанных на визуализации.
Исследование включало сравнение нескольких алгоритмов при единых условиях предварительной обработки и валидации, проверку калибровки вероятностей и интерпретацию с помощью метода SHAP. Анализ показал, что ключевые предикторы (например, депрессия STсегмента) согласуются с клиническими знаниями. Это подтверждает возможность применения модели для первичного скрининга и направления на дополнительную диагностику. В целом, калиброванные и интерпретируемые алгоритмы на основе открытых клинических данных могут стать ценным инструментом маршрутизации пациентов в условиях ограниченных ресурсов.

1. Reducing the Global Burden of Cardiovascular Disease, Part 1: The Epidemiology and Risk Factors / Р. Joseph et al // Circulation research. – 2017. – № 121(6). – Р. 677-694. https://doi.org/10.1161/CIRCRESAHA.117.308903.

2. Clinical, Economical, and Organizational Impact of Chronic Ischemic Cardiovascular Disease in Italy: Evaluation of 2019 Nationwide Hospital Admissions Data / F. Pietrantonio et al // Int. J. Environ. Res. Public Health. – 2025. – № 22. – Р. 530. https://doi.org/10.3390/ijerph22040530.

3. Prevalence and trends of coronary artery disease risk factors and their effect on age of diagnosis in patients with established coronary artery disease: Tehran Heart Center (2005–2015) / K. Hosseini et al // BMC Cardiovasc Disord 21. – 2021. – № 477(2021). https://doi.org/10.1186/s12872-021-02293-y.

4. Economic Analysis of the CADScor System for Ruling Out Coronary Artery Disease in England / М. Javanbakht еt al // PharmacoEconomics. – 2022. – Open 6. – Р. 123-135. https://doi.org/10.1007/s41669-021-00297-0.

5. Roth G. The Global Burden of Cardiovascular Diseases and Risks: A Compass for Global Action / G. Roth, G. Mensah, V. Fuster // JACC. – 2020. – № 76(25). – Р. 2980-2981. https://doi.org/10.1016/j.jacc.2020.11.021.

6. Estimation of the Burden of Ischemic Heart Disease in the Tabasco Population, Mexico, 2013–2021 / J.J. Cárdenas-Anguiano et al // Int. J. Environ. Res. Public Health. – 2025. – № 22. – Р. 423. https://doi.org/10.3390/ijerph22030423.

7. Fadah K. Epidemiology, Pathophysiology, and Management of Coronary Artery Disease in the Elderly / K. Fadah, A. Hechanova, D. Mukherjee // The International journal of angiology: official publication of the International College of Angiology. – 2022. – № 31(4). – Р. 244-250. https://doi.org/10.1055/s-0042-1751234.

8. National prevalence of coronary heart disease and its relationship with human development index: A systematic review / Ke-Fu Zhu et al // European Journal of Preventive Cardiology. – 2016. – Vol. 23, Issue 5. – Р. 530-543. https://doi.org/10.1177/2047487315587402.

9. Prevalence of coronary artery disease and its risk factors in Kerala, South India: a communitybased cross-sectional study / M.N. Krishnan et al // BMC Cardiovasc Disord. – 2016. – № 16. – Р. 12. https://doi.org/10.1186/s12872-016-0189-3.

10. Risk Factor Burden and Long-Term Prognosis of Patients With Premature Coronary Artery Disease / М. Zeitouni et al // Journal of the American Heart Association. – 2020. – № 9(24). – Р. e017712. https://doi.org/10.1161/JAHA.120.017712.

11. Stratification of Patients with Coronary Artery Disease by Circulating Cytokines Profile: A Pilot Study / С. Iside et al // M. J. Clin. Med. – 2023. – № 12. – Р. 6649. https://doi.org/10.3390/jcm12206649.

12. Myocardial Ischemia: Differentiating between Epicardial Coronary Artery Atherosclerosis, Microvascular Dysfunction and Vasospasm in the Catheterization Laboratory / G. Monizzi et al // J. Clin. Med. – 2024. – № 13. – Р. 4172. https://doi.org/10.3390/jcm13144172.

13. Genetics of coronary artery disease and myocardial infarction / X. Dai et al // World journal of cardiology. – 2016. – № 8(1). – Р. 1-23. https://doi.org/10.4330/wjc.v8.i1.1.

14. Non-invasive detection of coronary artery disease in high-risk patients based on the stenosis prediction of separate coronary arteries / R. Alizadehsani et al // Computer methods and programs in biomedicine. – 2018. – № 162. – Р. 119-127. https://doi.org/10.1016/j.cmpb.2018.05.009.

15. Symptom Presentation among Women with Suspected Ischemia and No Obstructive Coronary Artery Disease (INOCA) / Y.K. Taha et al // J. Clin. Med. – 2023. – № 12. – Р. 5836. https://doi.org/10.3390/jcm12185836.

16. Comparison of Outcomes in Patients With Acute Coronary Syndrome Presenting With Typical Versus Atypical Symptoms / Y. Hammer et al // The American journal of cardiology. – 2019. – № 124(12). – Р. 1851-1856. https://doi.org/10.1016/j.amjcard.2019.09.007.

17. Silent Myocardial Ischemia: From Pathophysiology to Diagnosis and Treatment / Р. Theofilis et al // Biomedicines. – 2024. – № 12. – Р. 259. https://doi.org/10.3390/biomedicines12020259.

18. Nelson AJ. Current approach to the diagnosis of atherosclerotic coronary artery disease: more questions than answers / AJ Nelson, M Ardissino, PJ. Psaltis // Therapeutic Advances in Chronic Disease. – 2019. – №10. https://doi.org/10.1177/2040622319884819.

19. Diagnostic accuracy of non-invasive cardiac imaging modalities in patients with a history of coronary artery disease: a meta-analysis / RA Jukema et al // Heart. – 2025. – № 111. – Р. 4-10. https://doi.org/10.1136/heartjnl-2024-324248.

20. Speckle Tracking Echocardiography: Early Predictor of Diagnosis and Prognosis in Coronary Artery Disease / М.С. Pastore et al // BioMed research international. – 2021. https://doi.org/10.1155/2021/6685378.

21. Challenges and Burdens in the Coronary Artery Disease Care Pathway for Patients Undergoing Percutaneous Coronary Intervention: A Contemporary Narrative Review / М. Kodeboina et al // Int. J. Environ. Res. – 2023. – № 20. – Р. 5633. https://doi.org/10.3390/ijerph20095633.

22. Wang L.W. Non-invasive screening for coronary artery disease: current perspectives, patient, public health and ethical considerations in evaluating symptomatic and asymptomatic individuals / L.W. Wang // Internal medicine journal. – 2025. – № 55(4). – Р. 555-563. https://doi.org/10.1111/imj.16585.

23. Advancements in Biomarkers for Early Detection and Risk Stratification of Cardiovascular Diseases-A Literature Review / А. Nazir et al // Health science reports. – 2025. – № 8(5). – Р. e70878. https://doi.org/10.1002/hsr2.70878.

24. Zachariadis C.B. Harnessing Artificial Intelligence for Automated Diagnosis / C.B. Zachariadis, H.C. Leligou // Information. – 2024. – № 15. – Р. 311. https://doi.org/10.3390/info15060311.

25. A new machine learning technique for an accurate diagnosis of coronary artery disease / М. Abdar et al // Computer methods and programs in biomedicine. – 2019. – № 179. – Р. 104992. https://doi.org/10.1016/j.cmpb.2019.104992.

26. Multi-constraints based deep learning model for automated segmentation and diagnosis of coronary artery disease in X-ray angiographic images / М. Algarni et al // PeerJ Computer Science. – 2022. – № 8. – Р. e993. https://doi.org/10.7717/peerj-cs.993.

27. Machine Learning Quantitation of Cardiovascular and Cerebrovascular Disease: A Systematic Review of Clinical Applications / С. Boyd et al // Diagnostics. – 2021. – № 11. – Р. 551. https://doi.org/10.3390/diagnostics11030551.

28. Akella A. Machine Learning Algorithms for Predicting Coronary Artery Disease: Efforts Toward an Open Source Solution / A. Akella, S. Akella // Future Science OA. – 2021. – № 7(6). https://doi.org/10.2144/fsoa-2020-0206.

29. A 3D deep learning classifier and its explainability when assessing coronary artery disease / Cheung, Wing Keung et al/ – 2023. 10.48550/arXiv.2308.00009.

30. Integration of artificial intelligence into clinical patient management: focus on cardiac imaging / F. Loncaric et al // Revista Española de Cardiología (English Edition). – 2021. – Vol. 74, Issue 1. – Р. 72-80. https://doi.org/10.1016/j.rec.2020.07.003.

31. Non-invasive Assessment of Coronary Artery Disease: The Role of AI in the Current Status and Future Directions / F. Ezekwueme et al // Cureus. – 2025. – № 17(2). – Р. e78994. https://doi.org/10.7759/cureus.78994.

32. Using artificial intelligence in the development of diagnostic models of coronary artery disease with imaging markers: A scoping review / X. Wang et al // Front. Cardiovasc. – 2022. – Med. 9:945451. https://doi.org/10.3389/fcvm.2022.945451.

33. Machine learning and deep learning-based approach in smart healthcare: Recent advances, applications, challenges and opportunities / Anichur Rahman et al // [J]. AIMS Public Health. – 2024. – № 11(1). – Р. 58-109. https://doi.org/10.3934/publichealth.2024004.

34. Machine Learning-Based Predictive Models for Detection of Cardiovascular Diseases / А. Ogunpola et al // Diagnostics. – 2024. – № 14. – Р. 144. https://doi.org/10.3390/diagnostics14020144.

35. Implementing Machine Learning in Interventional Cardiology: The Benefits Are Worth the Trouble / Ben Ali W. Et al // Front. Cardiovasc. Med. – 2021. – № 8. – Р. 711401. https://doi.org/10.3389/fcvm.2021.711401.

36. Wong K. Deep learning-based cardiovascular image diagnosis: A promising challenge / K. Wong, G. Fortino, D. Abbott // Future Generation Computer Systems. – 2019. – № 110. – Р. 10. 1016/j.future.2019.09.047.

37. Advances in Diagnosis, Therapy, and Prognosis of Coronary Artery Disease Powered by Deep Learning Algorithms / M. Chu еt al // JACC: Asia. – 2023. – № 3(1). – Р. 1-14. https://doi.org/10.1016/j.jacasi.2022.12.005.

38. The Role of Artificial Intelligence in Improving Patient Outcomes and Future of Healthcare Delivery in Cardiology: A Narrative Review of the Literature / D. Gala et al // Healthcare. – 2024. – № 12. – Р. 481. https://doi.org/10.3390/healthcare12040481.

39. Classification models for assessing coronary artery disease instances using clinical and biometric data: an explainable man-in-the-loop approach / AD. Samaras et al // Sci Rep. – 2023. – № 13. – Р. 6668. https://doi.org/10.1038/s41598-023-33500-9.

40. Heart Disease [Dataset] / А. Janosi et al // UCI Machine Learning Repository. – 1989. https://doi.org/10.24432/C52P4X.

41. Hancock J.T. Survey on categorical data for neural networks / J.T. Hancock, T.M. Khoshgoftaar // J Big Data 7, 28 (2020). https://doi.org/10.1186/s40537-020-00305-w.

42. The choice of scaling technique matters for classification performance / B.V. Lucas de Amorim et al // Applied Soft Computing. – 2023. – Vol. 133. – Р. 109924, ISSN 1568-4946, https://doi.org/10.1016/j.asoc.2022.109924.

43. Conditional mutual information-based feature selection algorithm for maximal relevance minimal redundancy / Х. Gu et al // Appl Intell. – 2022. – № 52. – Р. 1436-1447. https://doi.org/10.1007/s10489-021-02412-4.

44. Evaluation of clinical prediction models (part 1): from development to external validation / GS. Collins et al // BMJ. – 2024. – № 8. – Р. 384. e074819. https://doi.org/10.1136/bmj-2023-074819. PMID: 38191193; PMCID: PMC10772854.

45. Demircioğlu A. Applying oversampling before cross-validation will lead to high bias in radiomics / A. Demircioğlu // Sci Rep. – 2024. – № 14. – Р. 11563. https://doi.org/10.1038/s41598-024-62585-z.

46. On evaluation metrics for medical applications of artificial intelligence / S.A. Hicks et al // Sci Rep. – 2022. – № 12. – Р. 5979. https://doi.org/10.1038/s41598-022-09954-8.

47. Explainable coronary artery disease prediction model based on AutoGluon from AutoML framework / J. Wang et al // Front. Cardiovasc. Med. – 2024. – № 11. – Р. 1360548. https://doi.org/10.3389/fcvm.2024.1360548.

48. Evaluation of clinical prediction models (part 1): from development to external validation / G.S. Collins et al // BMJ. – 2024. – № 384. – Р. e074819. https://doi.org/10.1136/bmj-2023-074819.

49. Predicting COVID-19 mortality with electronic medical records / Н. Estiri et al // npj Digit. Med. – 2021. – № 4. – Р. 15. https://doi.org/10.1038/s41746-021-00383-x.