Predicting the development of atrial fibrillation in patients with sinus rhythm by some parameters of standard transthoracic echocardiography using a trained neural network

Authors: Kotlyarov S.N., Lyubavin A.V.

Company: 1 Ryazan State Medical University named after academician I.P. Pavlov, Ryazan, Russian Federation
2 Lipetsk City Hospital No. 4 “Lipetsk-Med”, Lipetsk, Russian FederationГУЗ «Липецкая городская больница № 4 “Липецк-Мед”», Липецк, Российская Федерация

For correspondence:  Sign in or register.

Type:  Original articles


DOI: https://doi.org/10.24022/1997-3187-2023-17-4-481-490

For citation: Kotlyarov S.N., Lyubavin A.V. Predicting the development of atrial fibrillation in patients with sinus rhythm by some parameters of standard transthoracic echocardiography using a trained neural network. Creative Cardiology. 2023; 17 (4): 481–90 (in Russ.). DOI: 10.24022/1997-3187-2023-17-4-481-490

Received / Accepted:  18.09.2023 / 24.10.2023

Keywords: atrial fibrillation neural networks transthoracic echocardiography



Subscribe 🔒

 

Abstract

Objective. Purpose of the study was to determine the possibilities of predicting the development of atrial fibrillation in patients with sinus rhythm by some parameters of standard transthoracic echocardiography using a trained neural network.

Material and methods. The study includes data from transthoracic echocardiography protocols of 695 patients treated in the cardiology department Lipetsk city Hospital No. 4 in 2022–2023. On the basis of the size of the heart chambers, the thickness of the walls of the left ventricle, the fraction of the left ventricle ejection and the severity of valvular regurgitation, the neural network was trained. Then, using a trained neural network echocardiography data of 30 patients, who underwent transthoracic echocardiography on sinus rhythm for one year before the detection of atrial fibrillation, were processed.

Results. The trained neural network successfully differentiated between patients with atrial fibrillation and those without a history of atrial fibrillation in 81.0% of cases (area under the curve (AUC) AUC 0.799, 95% confidence interval (CI) 0.738–0.932, р<0.0001) and correctly predicted the development of atrial fibrillation in patients with sinus rhythm during 3.0±0.45 months before manifestation in 80% of cases with method sensitivity of 88.9% and specificity 81.1% (AUC 0.854, 95% CI 0.738–0.932, р<0.0001).

Conclusion. Transthoracic echocardiography data identify patients with high probability of atrial fibrillation. The use of a trained neural network to predict the development of atrial fibrillation may be useful in determining indication for long-term electrocardiography-monitoring in patients with sinus rhythm in order to detect latent atrial fibrillation early.

References

  1. Kalinin R.E., Suchkov I.A., Boguto O.N. Embolism of main arteries in atrial fibillation. I.P. Pavlov Russian Medical Biological Herald. 2012; 20 (2): 234–41. DOI: 10.17816/PAVLOVJ20122234-241
  2. Tatarintseva Z.G., Kosmacheva E.D., Porhanov V.A., Kruchinova S.V. Echo-cardiographic parameters for atrial fibrillation in combination with acute coronary syndrome in real clinical practice according to register of acute coronary syndrome in the Krasnodar region. Cardiovascular Therapy and Prevention. 2019; 18 (2): 20–5 (in Russ.). DOI: 10.15829/1728-8800-2019-2-20-25
  3. Sergeev Yu.A., Sterleva E.A., Niazyan D.A., Application of neural networks in medicine. Comparison of methods of neural network and group analysis of pathologies. StudNet. 2021; 9 (in Russ.).
  4. Mazaev V.P., Komkov A.A., Ryazanova S.V., Samochatov D.N., Drapkina O.M. The identification of latent atrial fibrillation, decreased ejection fraction of the left ventricle and potassium metabolism disorders on the conventional electrocardiogram using deep convolutional neural network as an element of artificial intelligence. Modern Problems of Science and Education. 2020; 6 (in Russ.).
  5. Yaroslavskaya E.I., Dyachkov S.M., Gorbatenko E.A. Artificial neural networks in prediction of atrial fibrillation in men with coronary artery disease. The Siberian Journal of Clinical and Experimental Medicine. 2020; 35 (4): 119–27. DOI: 10.29001/2073-8552-2020-35-4-119-127
  6. Python. https://www.python.org (accessed 15 august, 2023).
  7. TensorFlow. https://www.tensorflow.org (accessed 15 august, 2023).
  8. MedCalc. https://www.medcalc.org (accessed 15 august, 2023).
  9. GitHub. https://github.com/alexlubavin/tte_af_predict_network (accessed 23 august, 2023).
  10. Sieweke J.T., Hagemus J., Biber S., Berliner D., Grosse G.M., Schallhorn S. et al. Echocardiographic parameters to predict atrial fibrillation in clinical routine-the EAHsy-AF risk score. Front. Cardiovasc. Med. 2022; 9: 851474. DOI: 10.3389/fcvm.2022.851474
  11. Eren H., Acar R.D., Demir S., Omar M.B., Öcal L., Kalkan M.E. Speckle-tracking echocardiography can predict atrial fibrillation in patients with supraventricular tachycardia. Pacing Clin. Electrophysiol. 2021; 44 (8): 1387–96. DOI: 10.1111/pace.14304
  12. Magne J., Salerno B., Mohty D., Serena C., Rolle F., Piccardo A. et al. Echocardiography is useful to predict postoperative atrial fibrillation in patients undergoing isolated coronary bypass surgery: a prospective study. Eur. Heart J. Acute Cardiovasc. Care. 2019; 8 (2): 104–13. DOI: 10.1177/2048872616688419
  13. Ntaios G., Perlepe K., Lambrou D., Sirimarco G., Strambo D., Eskandari A. et al. Identification of patients with embolic stroke of undetermined source and low risk of new incident atrial fibrillation: the AF-ESUS score. Int. J. Stroke. 2021; 16 (1): 29–38. DOI: 10.1177/1747493020925281
  14. Atta-Fosu T., LaBarbera M., Ghose S., Schoenhagen P., Saliba W., Tchou P.J. et al. A new machine learning approach for predicting likelihood of recurrence following ablation for atrial fibrillation from CT. BMC Med. Imaging. 2021; 21 (1): 45. DOI: 10.1186/s12880-021-00578-4
  15. Himmelreich J.C.L., Lucassen W.A.M., Harskamp R.E., Aussems C., van Weert H.C.P.M., Nielen M.M.J. CHARGEAF in a national routine primary care electronic health records database in the Netherlands: validation for 5-year risk of atrial fibrillation and implications for patient selection in atrial fibrillation screening. Open Heart. 2021; 8 (1): e001459. DOI: 10.1136/openhrt-2020-001459. Erratum in: Open Heart. 2021; 8 (2): 1.
  16. Tseng A.S., Noseworthy P.A. Prediction of atrial fibrillation using machine learning: a review. Front. Physiol. 2021; 12: 752317. DOI: 10.3389/fphys.2021.752317

About Authors

  • Stanislav N. Kotlyarov, Cand. Med. Sci., Chief of Chair; ORCID
  • Aleksander V. Lyubavin, Cardiologist; ORCID

Chief Editor

Leo A. Bockeria, MD, PhD, DSc, Professor, Academician of Russian Academy of Sciences, President of Bakoulev National Medical Research Center for Cardiovascular Surgery