GVM-Net: A GNN-Based Vessel Matching Network for 2D/3D Non-Rigid Coronary Artery Registration

• Published in: IEEE transactions on medical imaging Vol. 44; no. 6; pp. 2617 - 2630
• Main Authors: Chen, Yankai, Li, Guanyu, Li, Chunming, Yu, Wei, Fan, Zehao, Bai, Jingfeng, Tu, Shengxian
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.06.2025
• Subjects: 2D/3D non-rigid registration; Accuracy; Algorithms; Arteries; Computed Tomography Angiography - methods; Coronary Angiography - methods; Coronary Vessels - diagnostic imaging; Databases, Factual; Deformation; Feature extraction; feature matching; graph attention network; Humans; Image segmentation; Imaging, Three-Dimensional - methods; Iterative methods; Light sources; Neural Networks, Computer; Three-dimensional displays; Topology; Training; vessel graph matching
• ISSN: 0278-0062; 1558-254X; 1558-254X
• DOI: 10.1109/TMI.2025.3540906

The registration of coronary artery structures from preoperative coronary computed tomography angiography to intraoperative coronary angiography is of great interest to improve guidance in percutaneous coronary interventions. However, non-rigid deformation and discrepancies in both dimensions and topology between the two imaging modalities present a challenge in the 2D/3D coronary artery registration. In this study, we address this problem by formulating it as a centerline feature matching task and propose a GNN-based vessel matching network (GVM-Net) to establish dense correspondence between different image modalities in an end-to-end manner. GVM-Net considers centerline points as nodes in graphs and effectively models the complex topological relationships between them through attention mechanisms and message passing. Furthermore, by incorporating redundant rows and columns into the matching matrix, GVM-Net can effectively handle inconsistencies in vascular structures. We also introduce the query-based nodes grouping module, which clusters nodes in the feature space to further explore the topological relationships. GVM-Net achieves an average F1-score of 89.74% with a mean pixel distance of 0.48 pixels on the synthetic dataset with 276 data pairs and an average F1-score of 83.35% with a mean error of 1.52 mm in 55 manually labeled clinical cases, both exceeding existing feature matching methods.