Prototype Learning to Create Refined Interpretable Digital Phenotypes from ECGs
This work addresses the need for interpretable digital phenotyping from physiological data in healthcare, offering transferable intermediate phenotypes that could aid clinical decision-making, though it is incremental in applying existing prototype methods to new clinical validation.
The study tackled the problem of whether prototype-based neural networks trained for ECG classification capture clinically meaningful phenotypes, and found that individual prototypes showed stronger associations with hospital discharge diagnoses than class predictions or NLP-extracted concepts, with AUCs ranging from 0.89 to 0.91 for conditions like atrial fibrillation and heart failure.
Prototype-based neural networks offer interpretable predictions by comparing inputs to learned, representative signal patterns anchored in training data. While such models have shown promise in the classification of physiological data, it remains unclear whether their prototypes capture an underlying structure that aligns with broader clinical phenotypes. We use a prototype-based deep learning model trained for multi-label ECG classification using the PTB-XL dataset. Then without modification we performed inference on the MIMIC-IV clinical database. We assess whether individual prototypes, trained solely for classification, are associated with hospital discharge diagnoses in the form of phecodes in this external population. Individual prototypes demonstrate significantly stronger and more specific associations with clinical outcomes compared to the classifier's class predictions, NLP-extracted concepts, or broader prototype classes across all phecode categories. Prototype classes with mixed significance patterns exhibit significantly greater intra-class distances (p $<$ 0.0001), indicating the model learned to differentiate clinically meaningful variations within diagnostic categories. The prototypes achieve strong predictive performance across diverse conditions, with AUCs ranging from 0.89 for atrial fibrillation to 0.91 for heart failure, while also showing substantial signal for non-cardiac conditions such as sepsis and renal disease. These findings suggest that prototype-based models can support interpretable digital phenotyping from physiologic time-series data, providing transferable intermediate phenotypes that capture clinically meaningful physiologic signatures beyond their original training objectives.