GRC-ProbNet: Uncertainty-aware Feature Extraction for Cardiovascular Disease Classification

2026-07-11

AI Summary

This work addresses the limitation of existing cardiovascular disease classification methods that rely on deterministic segmentation masks and neglect the inherent anatomical uncertainty in cardiac structures, thereby constraining classification performance. To overcome this, the study proposes the first approach to explicitly model segmentation uncertainty by leveraging deep ensembles to generate diverse segmentation masks, from which uncertainty-aware features are extracted. These features are then integrated with radiomic and geometric information for downstream classification. Experimental results demonstrate that the proposed method achieves a classification AUROC of 92.92% on the MM-WHS and ASOCA datasets, outperforming the baseline GRC-Net (91.25%) and revealing the differential impact of various uncertainty metrics on both segmentation quality and classification accuracy.

Abstract

The automatic detection and classification of cardiovascular disease (CVD) from computed tomography (CT) images plays an important role in clinical practice. Recently, a hybrid pipeline (GRC-Net) for CVD classification was proposed, which leverages a deep-learning-based segmentation and registration method to extract radiomic and geometric features. However, GRC-Net relies on a deterministic segmentation mask, without considering the inherent ambiguity associated with cardiac anatomy. In this paper, we propose GRC-ProbNet, which takes advantage of a deep ensemble to produce multiple segmentation masks for a given input. From these masks, we extract multiple uncertainty features. We analyze these uncertainty features for both their correlation with segmentation error and their propagation effects on downstream CVD classification performance. Our experiments on the publicly available MM-WHS and ASOCA datasets show that the uncertainty measure that best reflects segmentation quality is not necessarily the one that provides the strongest signal for downstream CVD classification. Overall, our results demonstrate that GRC-ProbNet utilizing uncertainty features substantially improves CVD classification AUROC (92.92\) compared to the baseline GRC-Net model (91.25%). Our code is publicly available: https://github.com/biomedia-mira/GRC-ProbNet.

Problem

• cardiovascular disease classification
• uncertainty-aware feature extraction
• segmentation ambiguity
• computed tomography
• radiomic features

Innovation

• uncertainty-aware
• deep ensemble
• probabilistic segmentation
• radiomic features
• cardiovascular disease classification