Auxílio ao Diagnóstico para Predição de Morte Súbita em Pacientes Chagásicos a Partir de Dados Clínicos: uma Abordagem baseada em Aprendizagem de Máquina

• Pedro E. O. Primo UFC
• Weslley L. Caldas UFC
• Gabriel S. Almeida UFC
• Luan P. L. Brasil UFC
• Carlos H. L. Cavalcante IFCE
• João P. V. Madeiro UFC
• Danielo G. Gomes UFC
• Roberto C. Pedrosa UFRJ

DOI: https://doi.org/10.5753/sbcas.2021.16077

Abstract

Chagas Disease (CD) affects about 7 million people worldwide and one of the main adverse outcomes is the sudden cardiac death (SCD) caused by cardiomyopathy, whose evolution can be controlled with an early diagnosis. At this paper, we use 7 machine learning algorithms over a specific dataset with clinic data from chagasic patients aiming at discrimination among patients with high and patients with low predisposition for SCD, applying feature selection and resampling methods. K-Nearest Neighbors showed the best performance, with AUC:85.35 and F1:75.79. Due to their high weights in the machine learning classifiers, we suggest Non-Sustained Ventricular Tachycardia and Total Ventricular Extrasystoles as important features to identify SCD.

References

Alberto, A. C., Limeira, G. A., Pedrosa, R. C., Zarzoso, V., and Nadal, J. (2017). Ecgbased predictors of sudden cardiac death in chagas’ disease. In 2017 Computing in Cardiology (CinC), pages 1–4. IEEE.

Alberto, A. C., Pedrosa, R. C., Zarzoso, V., and Nadal, J. (2020). Association between circadian holter ecg changes and sudden cardiac death in patients with chagas heart disease. Physiological Measurement, 41(2):025006.

Chawla, N. V., Bowyer, K. W., Hall, L. O., and Kegelmeyer, W. P. (2002). Smote: synthetic minority over-sampling technique. Journal of artificial intelligence research, 16:321–357.

Coura, J. R. and Viñas, P. A. (2010). Chagas disease: a new worldwide challenge. Nature, 465(7301):S6–S7.

de Souza, A. C. J., Salles, G., Hasslocher-Moreno, A. M., de Sousa, A. S., do Brasil, P. E. A. A., Saraiva, R. M., and Xavier, S. S. (2015). Development of a risk score to predict sudden death in patients with chaga’s heart disease. International journal of cardiology, 187:700–704.

Guedes, P. M. M., Gutierrez, F. R. S., Silva, G. K., Dellalibera-Joviliano, R., Rodrigues, G. J., Bendhack, L. M., Rassi Jr, A., Rassi, A., Schmidt, A., Maciel, B. C., et al. (2012). Deficient regulatory t cell activity and low frequency of il-17-producing t cells correlate with the extent of cardiomyopathy in human chagas’ disease. PLoS Negl Trop Dis, 6(4):e1630.

Hernandez, J., Carrasco-Ochoa, J. A., and Martínez-Trinidad, J. F. (2013). An empirical study of oversampling and undersampling for instance selection methods on imbalance datasets. In Iberoamerican Congress on Pattern Recognition, pages 262–269. Springer.

Jovíc, A., Brkíc, K., and Bogunovíc, N. (2015). A review of feature selection methods with applications. In 2015 38th international convention on information and communication technology, electronics and microelectronics (MIPRO), pages 1200–1205. Ieee.

Li, P.-J., Jin, T., Luo, D.-H., Shen, T., Mai, D.-M., Hu, W.-H., and Mo, H.-Y. (2015). Effect of prolonged radiotherapy treatment time on survival outcomes after intensity-modulated radiation therapy in nasopharyngeal carcinoma. PloS one, 10(10):e0141332.

Rassi Jr, A., Rassi, A., Little, W. C., Xavier, S. S., Rassi, S. G., Rassi, A. G., Rassi, G. G., Hasslocher-Moreno, A., Sousa, A. S., and Scanavacca, M. I. (2006). Development and validation of a risk score for predicting death in chagas’ heart disease. New England Journal of Medicine, 355(8):799–808.

Silva, L. E. V., Moreira, H. T., Bernardo, M. M. M., Schmidt, A., Romano, M. M. D., Salgado, H. C., Fazan Jr, R., Tin´os, R., and Marin-Neto, J. A. (2021). Prediction of echocardiographic parameters in chagas disease using heart rate variability and machine learning. Biomedical Signal Processing and Control, 67:102513.

WHO (2021). Chagas disease (american trypanosomiasis). https://www.who.int/healthtopics/chagas-disease.