A Cascading Multi-Stage Deep Learning Approach for Detecting Chagas Disease from Electrocardiograms
Year: 2025
Type: conference-paper
Abstract: Aims: For the PhysioNet Challenge 2025, our team ”bug busters” developed an approach to detect Chagas disease from electrocardiograms. This parasitic infection can be life-threatening when untreated, and electocardiogram based screening could direct limited resources more efficiently. Methods: We implemented a novel multi-stage cascading approach using five deep learning models: two ResNet18 variants with attention mechanisms, two SimpleCNN models, and an AttentionCNN. Our key innovation is a progressive filtering pipeline that ranks healthy samples by their prediction scores and removes those most confidently classified as healthy, creating increasingly focused training sets. Results: Our approach scored 0.369 in the official stage on the validation dataset and 0.224 in the test set. Our team was ranked 14th out of 41. Conclusion: The cascading multi-stage methodology shows promise for Chagas disease detection, overcoming the limitations of single-model approaches. Future work should investigate performance across diverse patient populations and explore interpretability of model decisions.
Source: Computing in cardiology
Language: English
Topic: ECG Monitoring and Analysis
Subfield: Cardiology and Cardiovascular Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W4417306847
FWCI:
Cites:
Cited by:
Related to: