Characterising cardiac tissue properties with graph neural networks
Authors: Ching-En Chiu, Yoo Ri Kim, Magdi Saba, Danilo Mandic, Marta Varela
Organizations: Department of Electrical and Electronic Engineering, Imperial College London, London, United Kingdom · Department of Internal Medicine, Chonnam National University College of Medicine, Gwangju, Republic of Korea · St. George’s University Hospital, London, United Kingdom · Cardiovascular & Genomics Research Institute, City St George’s University of London, United Kingdom · National Heart & Lung Institute, Imperial College London, London, United Kingdom
Characterising electrophysiological properties of cardiac tissue efficiently and accurately from spatially sparse intracardiac measurements is clinically important for localising ablation targets and improving arrhythmia treatment. We developed a graph neural network-based framework trained on synthetic electrogram signals on 2D flat surfaces to identify areas of interest in the context of cardiac ablation for premature ventricular complexes (PVCs). Our method achieved an average precision of 0.96, 0.97, and 0.95 for the detection of single-patch fibrosis, rapid depolarisation and high excitability, respectively. The trained model can then be applied to 2D curved surfaces with few-shot fine-tuning, demonstrating its generalisation capability. Future work will develop this framework further for clinical use in PVC ablation.
Figures & tables
Figure 1: A schematic of our proposed framework. The temporal encoder module based on a CNN learns latent representations of the EGM waveforms. Then, the GNN based spatial module propagates the latent information from sparsely distributed electrodes to high-resolution tissue nodes. A small classification head then takes the final embeddings and performs binary classification.
Average
Macro
Precision
Recall
ROC
precision
F1
AUC
Low Resolution (LR) Model (CNN only)
Fibrosis ( D↓ ), single patch
0.76
0.80
0.45
0.96
0.990
Fibrosis ( D↓ ), multiple patches
0.78
0.82
0.57
0.90
0.960
Rapid depolarisation ( k↑ )
0.74
0.86
0.61
0.92
0.991
High excitability ( a<0 )
0.75
0.85
0.59
0.93
0.990
Table 1: Performance evaluated by average precision, macro F1, precision/recall (of the abnormal class at a decision threshold of 0.5), and ROC AUC for low resolution (LR) and GNN-based high resolution (HR) model.
Figure 2: Visualisation of the low resolution model and our proposed GNN-based high resolution model for detecting single & multi-patch lower D , higher k , and negative a . The performance with sparse electrodes and noisy signals are also shown. Examples of healthy signals and signals from the pathological regions (in the locations indicated by a matching colour star) with and without noise are shown in the rightmost column.
Figure 3: Average precision of the HR model with increasing electrode sparsity and signal noise. Performance for both random and regular downsampling of electrodes are shown. Error bars are computed over 3 different rounds of random sampling.
Metric
Lower curvature
Higher curvature
(max K = 2.23×10−3mm−2 )
(max K = 1.19×10−2mm−2 )
Zero-shot
Few-shot
Zero-shot
Few-shot
Fibrosis ( D↓ )
Average Precision
0.857
0.986
0.632
0.958
ROC-AUC
0.966
0.998
0.944
0.994
Max-F1 score
0.822
0.937
0.704
0.880
Table 2: Zero-shot and few-shot performance on curved surfaces.
Figure 4: Visualisation of focal pathology detection results after few-shot (10 graphs) fine-tuning on curved surfaces.
Sudden cardiac death mostly originates from post-infarct ventricular tachycardia. This reentry is sustained by propagation around scar tissue, but its risk remains difficult to predict from infarct imaging alone. While in silico induction studies can reveal these mechanisms, exploring many substrates and pacing locations remains computationally expensive. We propose a graph attention classifier for early prediction of sustained ventricular reentry from simulated cardiac electrophysiology. Each simulation is represented as a graph with early nodal voltage samples and edge attributes encoding geometry, stimulation frequency, tissue labels, amplitude, and diastolic interval. These descriptors are computed only from the first two paced activations, before sustained reentry is observed. On a balanced independent test set of 1000 simulations, the classifier achieves 94.6% accuracy, 95.4% sensitivity, 93.8% specificity, 94.6% F1-score, 0.987 ROC-AUC, and 0.985 PR-AUC. Ablation experiments show that diastolic interval is the most critical descriptor. Qualitative attention inspection suggests that different heads focus on border-zone, stimulation location, and early propagation path. These results support graph attention as a fast early-prediction approach for large-scale in silico ventricular reentry studies.
Ma{ë}lis Morier, Jairo Rodr{í}guez Padilla, Patrick Gallinari +1
Left ventricular ejection fraction (LVEF) is a primary marker of cardiac function. However, routine estimation from endocardial measurements requires manual delineation at end-diastole (ED) and end-systole (ES), a process that is time-consuming and subject to inter-observer variability. Reliable automation is especially valuable for point-of-care ultrasound (POCUS), where computational resources are limited and acquisition quality varies. We propose Echo-E3Net, an anatomy-guided spatio-temporal network that explicitly embeds cardiac anatomy into LVEF prediction. A dual-phase Endocardial Border Detector (E2CBD) uses phase-specific cross-attention to localize ED/ES endocardial landmarks and produce phase-aware landmark embeddings, while an Endocardial Feature Aggregator (E2FA) fuses these embeddings with global statistical descriptors of deep feature maps to refine EF regression. Training is guided by a lightweight geometric loss that uses ED and ES endocardial landmarks to regularize EF prediction. On EchoNet-Dynamic and a PSAX subset of EchoNet-Pediatric, Echo-E3Net attains competitive performance using only 1.55M parameters and 8.05 GFLOPs, an order-of-magnitude compute reduction versus recent baselines, supporting real-time deployment. Our code is publicly available at https://github.com/moeinheidari7829/Echo-E3Net.
University of British Columbia, Vancouver, BC, Canada · University of Regensburg, Regensburg, Germany · Iran University of Science and Technology, Tehran, Iran +1
Cardiac arrhythmogenesis is governed by complex electromechanical interactions that are not directly observable in vivo, motivating the development of non-invasive computational approaches for reconstructing three-dimensional activation dynamics. We present a physics-informed neural network framework for recovering cardiac activation patterns, active tension propagation, deformation fields, and hydrostatic pressure from measurable deformation data in simplified left ventricular geometries. Our approach integrates nonlinear anisotropic constitutive modeling, heterogeneous fiber orientation, weak formulations of the governing mechanics, and finite-element-based loss functions to embed physical constraints directly into training. We demonstrate that the proposed framework accurately reconstructs spatiotemporal activation dynamics under varying levels of measurement noise and reduced spatial resolution, while preserving global propagation patterns and activation timing. By coupling mechanistic modeling with data-driven inference, this method establishes a pathway toward patient-specific, non-invasive reconstruction of cardiac activation, with potential applications in digital phenotyping and computational support for arrhythmia assessment.
Nathan Dermul, Hans Dierckx
Department of Cardiovascular Sciences, Cardiovascular Imaging and Dynamics, KU Leuven, Belgium. · Department of Cardiology, Leiden University Medical Centre, Albinusdreef 2, Leiden, 2333 ZA, the Netherlands. · Mathematical Institute, Leiden University, Einsteinweg 55, Leiden, 2333 CC, the Netherlands. +1