DOI: https://doi.org/10.31071/kit2017.13.05

Inventory reference

ISSN 1812-7231 Klin.inform.telemed. Volume 12, Issue 13, 2017, Pages 35–41

Author(s)

A. V. Frolov¹, M. A. Martsennyuk², T. G. Vaikhanskaya¹, V. B. Poljakov²

Institution(s)

¹Republican Scientific and Practical Centre "Cardiology", Minsk, Belarus ²Perm State National Research University, Perm, Russia

Article title

Fuzzy logic in prediction of adverse scenarios of the cardiac diseases

Abstract (resume)

Introduction. Prediction of adverse scenarios of the diseases of the cardiovascular system has an important clinical significance. A number of effective models based on Cox proportional hazard method have been created. However, with dichotomous variables and hardware errors in the neighborhood of the threshold values, the probability of an erroneous forecast is high. The method of fuzzy logic is able to reduce this disadvantage.

Purpose. Development of the model for individualized risk-stratification of patients based on fuzzy logic and its testing in patients with chronic heart failure.

Material and methods of the study. The risk-stratification model is based on the Cox logit-regression method in combination with the fuzzy logic. The model was tested in 240 patients with chronic heart failure. The digital ECG-study with an assessment of markers of the myocardium electrical instability, 24-hour holter monitoring and echocardiography were performed.

Results of the study. The probability function of adverse cardiovascular events in dichotomous variables was investigated. Its spasmodic character is revealed. The Cox logit regression is proposed to be supplemented with fuzzy sets in the form of sigmoidal membership functions. The model was tested in 240 patients with chronic heart failure. The sensitivity of the developed model of individual risk-stratification was 81%, and its specificity 99%.

Conclusions. Fuzzy logic allowed to reduce the jumps in assessing the probability of adverse events near threshold values, which favorably affected on the accuracy of the prognosis. The personal prognostic model developed by us has 94% predictive accuracy.

Keywords

ECG, electrical instability of the myocardium, risk-stratification, logit regression, fuzzy logic

References

1. Sulimov V., Okisheva E., Tsaregorodsev D. Non-invasive risk stratification for sudden cardiac death by heart rate turbulence and microvolt T-wave alternans in patients with myocardial infarction. Europace, 2012, vol. 14, iss. 12, pp. 1786-1792.
https://doi.org/10.1093/europace/eus238
PMid:22849973

2. O'Machoni C., Jichi F., Pavlou M., Menserrat L., Anastasakis A. and oth. A novel clinical risk prediction model for sudden cardiac death in hypertrophic cardiomyopathy. European Heart J., 2014, vol. 35, pp. 2010-2020.
https://doi.org/10.1093/eurheartj/eht439
PMid:24126876

3. Frolov A. V., Vaikhanskayay T. G., Melnikova O. P., Vorobiev A. P., Guel L. M. Ot risk stratifikacii pacientov c dilatacionnoj kardiomiopatiej k optimalnoj taktike lechenia [From the risk stratification in patients with dilated cardiomyopathy to the optimal treatment strategy]. Annaly Aritmologii [Annals of Arrhythmology]. 2016, vol. 13, iss.2, pp. 80 – 86. (in Russ.)
https://doi.org/10.15275/annaritmol.2016.2.3

4. Frolov A. V., Vaikhanskaya T. G., Melnikova O. P., Vorobiev A. P., Guel L. M. Risk stratification personalized model for prediction life-threatening ventricular tachyarrhythmias in patients with chronic heart failure. Kardilogia Polska [Polish Heart J.], 2017, vol. 75, iss. 7, pp. 682 - 688.

5. Jehn M., Guallar E., Clark J., Couper D., Duncan B, and oth. A prospective study of plasma ferritin level and incident diabetes. The atherosclerosis risk in communities study. American J. of Epidemiology, 2017, vol. 165, iss.9, pp. 1047 – 1054.
https://doi.org/10.1093/aje/kwk093
PMid:17284722

6. Lan D., and Wei L. The robust interference the Cox proportional hazard model. J. of American Statistician Association, 1989, vol. 84, pp. 1074 – 1078.
https://doi.org/10.1080/01621459.1989.10478874

7. Vittinghoff E., Glidden D., Shiborski S., McCaloch C. Regression methods in biostatistics: linear, logistic, and repeated measures models. Ed. By Dietz K. Springer Science-Business Media, NY, 2005, 338 p.

8. Zade L. Soft computing and fuzzy logic. IEEE Software J., 1994, iss. 11, pp. 48 – 56.
https://doi.org/10.1109/52.329401

9. Piegat A. Nechetkoe modelirovanie i ypravlenie [Fuzzy modeling and control]. Moscow, Binom Publ, 2013, 798 p. (in Russ.)

10. Suarla E., Syrimpeis V., Stanatopoulou K., Merekoulias G., Tsakalidis A., Tzimas G. Exploiting fuzzy expert systems in cardiology. Proc. Int. Conf. On Engineering and Applications of Neural Networks (EANN), 2013, pp. 80 - 89.
https://doi.org/10.1007/978-3-642-41016-1_9

11. Merlin P., and Prado-Arechiga G. New hybrid intelligent systems for diagnosis and risk evaluating of arterial hypertension, SpringerBriefs, NY, 2017 (e-book).
https://doi.org/10.1007/978-3-319-61149-5

12. Rydenko T. A., Vlasenko M. A. Sistemy nechetkoi logiki v diagnostike decinchronii miokarda [Fuzzy logic systems in desynchrony of the myocardium diagnostics]. Medichni nayki Publ., 2015, vol.4, iss. 5, pp. 52 – 61.
https://doi.org/10.15587/2313-8416.2015.43286

13. Nofal M., and Fouad K. Developing web-based semantic and fuzzy expert systems using proposed tool. Intern. J. of Computer Applications, 2015, vol. 112, iss. 7, pp. 39 - 45.

14. Yunda L., Pacheco D., and Millan J. A web-based fuzzy interference system based tool for cardiovascular diseases risk assessment. Nova J. of Medical and Biological Sciences, 2015, vol. 13, iss. 24, pp. 7 - 16.
https://doi.org/10.22490/24629448.1712

15. Varnavsky A., and Antonenko A. Determining of human cardiac pacemaker using fuzzy logic. Proc. Intern. Conf. On Information Technologies on Business and Industry, 2016. 7 p.
https://doi.org/10.1088/1742-6596/803/1/012168

Full-text version http://kit-journal.com.ua/en/viewer_en.html?doc/2017_13/004.pdf