PPCAR-Net reconstructs 3D coronary arteries from sparse X-rays fast
PPCAR-Net: Projection-Refined Parametric 3D Coronary Artery Reconstruction from Sparse X-ray Angiographic Views
Computer Vision and Pattern Recognition
Summary
Reconstructing 3D models of coronary arteries from limited X-ray views is challenging because vessels can overlap or look distorted. The authors present PPCAR-Net, a deep learning system that predicts the artery’s center paths and widths directly from segmented X-ray images without complicated matching steps. PPCAR-Net uses a two-stage method: it first makes a rough prediction, then refines it by comparing projections with input images. The system works quickly and shows good accuracy on simulated data, particularly for right coronary arteries.
What this means in practice
- •For cardiovascular imaging teams: Generate accurate 3D artery models from limited X-ray angiography views to improve diagnosis and intervention planning.
- •For medical device developers: Integrate fast artery reconstruction technology into imaging systems to provide real-time 3D visualization during procedures.$Commercial implications: Enables development of imaging products with real-time artery modeling for interventional cardiology using limited X-ray data.
Authors
Yu Ren, Hwee Kuan Lee, Tat-Jen Cham, Jonathan Yap, Khung Keong Yeo
Abstract
Sparse-view 3D coronary reconstruction commonly relies on cross-view correspondence and triangulation, which are vulnerable to vessel overlap and foreshortening, or on volumetric prediction followed by vascular-graph extraction, which does not directly provide centrelines and radii. We introduce PPCAR-Net, a projection-refined parametric coronary artery reconstruction network that directly predicts a branch-structured centreline-and-radius representation without explicit point matching, triangulation, or an intermediate volume. Given a variable number of segmented views, a coarse predictor combines frozen VGGT features with learned branch queries to estimate branch presence, B-spline centreline trajectories, and dense radius profiles. Projection-guided geometry and radius refiners then sample local evidence from the input views and apply residual corrections learned with 3D supervision. We evaluate representation fidelity and sparse-view reconstruction quantitatively and qualitatively. On simulated angiographic masks generated from CT-derived coronary anatomy, PPCAR-Net produces better connected artery reconstructions and achieves strong centreline accuracy, particularly for RCA, while maintaining competitive volumetric overlap. Coarse-to-fine inference takes 121 ms, enabling real-time reconstruction.