What is the clinical evidence from this study?

Study design: Other. Population: Trypanosoma cruzi infection (Chagas disease) (n=72). Intervention: Graph Attention Networks (GAT) with Variational Graph Autoencoder (VGAE) data augmentation vs. Traditional machine learning classifiers (RF, ETC, DT, SVM). Primary outcome: Area under the ROC curve (AUROC) for identification of infection stage.

July 18, 2026Medical & Biological Engineering & Computing0 citationsOpen Access

Multimodal graph learning for chagas disease classification

GCGabriel Carcedo-Rodríguez EMErik Molino Minero JPJorge Pérez-González

Key Result

The proposed Graph Attention Network architecture with Variational Graph Autoencoder data augmentation achieved an AUROC of 100% in identifying Trypanosoma cruzi infection stages using multimodal features.

Key Points

This research aims to enhance the classification of Chagas disease stages based on Trypanosoma cruzi infection using advanced graph learning techniques.
Developed a Graph Attention Network architecture to integrate multimodal features including Electrocardiogram, Echocardiogram, Doppler, and ELISA.
Employed a Variational Graph Autoencoder for data augmentation to mitigate overfitting and improve model generalization.
Validated model interpretations using GNNExplainer to identify critical clinical features.
Achieved an AUROC of 100% for accurate infection stage identification through multimodal fusion.
Demonstrated significant performance optimization for complex modalities, particularly Doppler.
Established pathophysiological consistency in model decisions using interpretability schemes.

Structured PICO

Population

72 female ICR mice experimentally infected with Trypanosoma cruzi to model acute and chronic stages of Chagas disease.

Intervention

Multimodal graph learning architecture based on Graph Attention Networks (GATs) with Variational Graph Autoencoder for data augmentation

Comparator

State-of-the-art classifiers

Outcome

Classification of Chagas disease stages (AUROC)

A multimodal graph learning approach combining ECG, echocardiogram, Doppler, and ELISA data can highly accurately classify Chagas disease stages in experimental models.

Limitations

Scarcity of labeled biological samples (Small Data)
High dimensionality of diagnostic sources

Abstract

The accurate classification of Chagas disease stages based on Trypanosoma cruzi infection in experimental models faces challenges due to the scarcity of data and the high dimensionality of diagnostic sources. This study proposes an architecture based on Graph Attention Networks on fully connected graphs to combine multimodal features (Electrocardiogram, Echocardiogram, Doppler, and ELISA). A Variational Graph Autoencoder was used to implement a data augmentation strategy that reduces overfitting and optimizes generalization, producing synthetic subjects with a biological covariance structure. The results demonstrate that the proposed methodology is effective compared to state-of-the-art classifiers, achieving an AUROC of 100% in the identification of infection stage through multimodal fusion and significantly optimizing performance in complex modalities such as Doppler. The pathophysiological consistency of the model's decisions was also validated using an interpretability scheme (GNNExplainer), which allowed the detection of important clinical features and established GATs as a robust method in the classification of Trypanosoma cruzi infection stages.

问 AI

Bookmark

View Full Paper