HypOProto: Hyperbolic Ordinal Prototypes for Left Ventricular Filling Pressure Classification

2026-06-18

AI Summary

Current deep learning–based approaches for left ventricular filling pressure (LVFP) classification suffer from poor interpretability and reliance on operator-dependent E/e′ ratios, limiting their deployment in resource-constrained settings. This work proposes a novel prototype learning framework that freezes an interpretable foundation model as the backbone and arranges class prototypes along a physiologically grounded E/e′ scale within hyperbolic space to model diagnostic certainty gradients. To enhance inter-class separability, we introduce a Hyperbolic Prototype Angular Separation (HyperPAS) loss. Our method is the first to incorporate hyperbolic geometry into LVFP classification, achieving state-of-the-art performance while maintaining high transparency. Visualization of attention maps further highlights clinically relevant regions, substantially improving the model’s trustworthiness and practical utility in real-world clinical decision-making.

Abstract

Echocardiography (echo) is a widely used imaging modality for assessing cardiac function, with Left Ventricular Filling Pressure (LVFP) serving as a critical physiological marker for conditions such as heart failure. Standard LVFP classification into normal \emph{vs} elevated categories relies on the Doppler-derived $E/e'$ ratio, which is operator-dependent and often unavailable in resource-limited settings, motivating methods that infer LVFP directly from B-mode echo. Existing deep learning approaches achieve high performance but remain largely black-box, limiting clinical interpretability. We propose HypOProto, a hyperbolic, ordinal prototype-based framework for interpretable LVFP classification using a frozen, explainable foundation model backbone. HypOProto arranges prototypes along the physiological $E/e'$ scale, placing borderline cases near the hyperboloid root where small angular differences separate similar cases, while normal and elevated cases occupy outward positions reflecting increasing diagnostic certainty. This hyperbolic geometry encodes clinically meaningful ordinal relationships and improves interpretability. We also introduce a novel Hyperbolic Prototype Angular Separation (HyperPAS) loss, enforcing inter-class prototype separation in hyperbolic space. HypOProto achieves SOTA performance while maintaining transparency, and highlights clinically relevant regions in visualizations. This work represents the first prototype-based framework for LVFP classification in echo. Our code can be found at https://github.com/DeepRCL/HypOProto.

Problem

• Left Ventricular Filling Pressure
• Echocardiography
• Interpretable Classification
• B-mode Echo
• LVFP Classification

Innovation

• hyperbolic geometry
• ordinal prototypes
• interpretable deep learning
• LVFP classification
• HyperPAS loss