跳转至

利用图神经网络表征心脏组织特性

文章背景与核心概要

准确表征心脏组织的电生理特性对于定位消融靶点及治疗室性早搏（PVC）等心律失常至关重要。然而，临床上往往只能获得空间稀疏的心内测量数据，这给精准诊断带来了挑战。

本文提出了一种基于图神经网络（GNN）的框架，通过在映射到二维平面的合成电图信号上进行训练，实现了对关键心脏组织区域的识别。该模型在检测单斑块纤维化、快速去极化和高兴奋性方面表现出极高的精度。此外，该框架通过少样本微调（few-shot fine-tuning）展现了在二维曲面上的强大泛化能力，为未来在临床PVC消融手术中的应用奠定了基础。

摘要 (Summary)

准确表征心脏组织的电生理特性对于定位消融靶点及改善心律失常治疗具有重要的临床意义。本文介绍了一种基于图神经网络（GNN）的框架，该框架在映射到二维平面的合成电图信号上进行训练，旨在识别心脏消融手术中与室性早搏（PVC）相关的关键区域。

Accurately characterising the electrophysiological properties of cardiac tissue from sparse intracardiac measurements is vital for localising ablation targets and treating arrhythmias such as premature ventricular complexes (PVCs). This paper introduces a graph neural network (GNN) framework trained on synthetic electrogram signals mapped to 2D flat surfaces to identify critical tissue regions. The model achieved high precision in detecting single-patch fibrosis, rapid depolarisation, and high excitability. Furthermore, the framework demonstrates strong generalisation capabilities when applied to 2D curved surfaces using few-shot fine-tuning, paving the way for future clinical deployment in PVC ablation procedures.

文章元数据 (Article Metadata)

  • arXiv 标识符: arXiv:2608.15843
  • 主要学科: 定量方法 (q-bio.QM)
  • 次要学科: 人工智能 (cs.AI), 信号处理 (eess.SP)
  • 提交日期: 2026年8月16日
  • 会议/研讨会: 已被 2026年心脏统计图谱与计算建模 (STACOM) 研讨会 录用
  • DOI: 10.48550/arXiv.2608.15843
  • arXiv Identifier: arXiv:2608.15843
  • Primary Subject: Quantitative Methods (q-bio.QM)
  • Secondary Subjects: Artificial Intelligence (cs.AI), Signal Processing (eess.SP)
  • Submission Date: August 16, 2026
  • Conference/Workshop: Accepted at The Statistical Atlases and Computational Modeling of the Heart (STACOM) workshop 2026
  • DOI: 10.48550/arXiv.2608.15843

作者 (Authors)

  • Ching-En Chiu
  • Yoo Ri Kim
  • Magdi Saba
  • Danilo Mandic
  • Marta Varela
  • Ching-En Chiu
  • Yoo Ri Kim
  • Magdi Saba
  • Danilo Mandic
  • Marta Varela

论文摘要 (Abstract)

从空间稀疏的心内测量数据中高效且准确地表征心脏组织的电生理特性，对于定位消融靶点和改善心律失常治疗具有重要的临床意义。我们开发了一种基于图神经网络的框架，该框架在二维平面上的合成电图信号上进行训练，以识别心脏消融背景下与室性早搏（PVC）相关的感兴趣区域。我们的方法在检测单斑块纤维化、快速去极化和高兴奋性方面分别达到了 0.96、0.97 和 0.95 的平均精度。训练后的模型可以通过少样本微调应用于二维曲面，证明了其泛化能力。未来的工作将进一步开发该框架，以用于临床PVC消融手术。

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.