Loading...
Publication
Deteção de insuficiência cardíaca em pacientes com doenças de chagas a partir de algoritmos de extração automática de parâmetros e de aprendizagem de máquina em sinais ECG
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 968.06 KB | Adobe PDF |
Abstract(s)
A Doença de Chagas é uma infeção parasitária crónica causada pelo Trypanosoma cruzi, sendo uma das principais causas de insuficiência cardíaca na América Latina. A deteção precoce da cardiomiopatia chagásica pode permitir uma intervenção mais eficaz, reduzindo complicações graves. Neste estudo, propõe-se uma nova abordagem na utilização de algoritmos de Machine Learning (ML) para a deteção automática de insuficiência cardíaca em pacientes com Doença de Chagas, utilizando características espectrais extraídas do eletrocardiograma (ECG). A nova abordagem recorre ao uso de características espectrais, como a SEF95 (Spectral Edge Frequency), MF (Mean Frequency), SE (Spectral Entropy), potências e rácios extraídos do Espectro de potências do sinal ECG. Posteriormente, classificadores de Machine Learning foram aplicados para prever o grau de insuficiência cardíaca. Os resultados indicaram que os modelos conseguem distinguir bem entre estágios extremos da doença (Exatidão: 83,3%; AUC ROC (Área sob a Curva ROC): 0.80-0.88), mas enfrentam dificuldades na estratificação da insuficiência cardíaca moderada, devido à heterogeneidade clínica da doença.
Chagas disease is a chronic parasitic infection caused by Trypanosoma cruzi and is one of the leading causes of heart failure in Latin America. Early detection of Chagasic cardiomyopathy can enable more effective intervention, reducing the risk of severe complications. This study proposes a novel approach using Machine Learning algorithms for the automatic detection of heart failure in patients with Chagas disease, based on spectral features extracted from the electrocardiogram (ECG). The proposed method focuses on the use of spectral features such as SEF95 (Spectral Edge Frequency), MF (Mean Frequency), SE (Spectral Entropy), power values, and ratios. These features were then used as input to various Machine Learning classifiers to predict the degree of heart failure. The results showed that the models performed well in distinguishing between extreme stages of the disease (Accuracy: 83.3%; AUC-ROC: 0.80–0.88), but had difficulties in correctly identifying moderate heart failure, mainly due to the clinical heterogeneity of the condition.
Chagas disease is a chronic parasitic infection caused by Trypanosoma cruzi and is one of the leading causes of heart failure in Latin America. Early detection of Chagasic cardiomyopathy can enable more effective intervention, reducing the risk of severe complications. This study proposes a novel approach using Machine Learning algorithms for the automatic detection of heart failure in patients with Chagas disease, based on spectral features extracted from the electrocardiogram (ECG). The proposed method focuses on the use of spectral features such as SEF95 (Spectral Edge Frequency), MF (Mean Frequency), SE (Spectral Entropy), power values, and ratios. These features were then used as input to various Machine Learning classifiers to predict the degree of heart failure. The results showed that the models performed well in distinguishing between extreme stages of the disease (Accuracy: 83.3%; AUC-ROC: 0.80–0.88), but had difficulties in correctly identifying moderate heart failure, mainly due to the clinical heterogeneity of the condition.
Description
Keywords
Doença de Chagas Insuficiência cardíaca Machine learning Eletrocardiograma Análise espectral Chagas disease ECG Spectral analysis Heart failure detection
Pedagogical Context
Citation
Publisher
CC License
Without CC licence