Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

4D-CAT: Synthesis of 4D Coronary Artery Trees from Systole and Diastole

Daosong Hu, Ruomeng Wang, Liang Zhao, Mingyue Cui, Song Ding, Kai Huang

TL;DR

This work tackles the challenge of motion-induced discrepancies in static 3D coronary models by proposing 4D-CAT, a method to synthesize 4D coronary trees from systole and diastole. It develops a centerline-centric pipeline that first segments the vessel centerlines using a cube-based test, then predicts non-rigid deformation fields via two pathways: a PointNet-based deep-learning model trained with soft-DTW loss and a cube-based sorting approach to align centerlines and derive the deformation field. Vessel points are propagated from centerline deformations through cuboid-based clustering and interpolation, yielding intermediate frames along the cardiac cycle. Validation on clinical CTA data and synthetic scenarios demonstrates improved registration and interpolation accuracy over baselines, with ablations confirming the importance of sorting and loss design. The approach offers a pathway to realistic 4D cardiovascular models and may enhance digital organ representations and clinical workflows.

Abstract

The three-dimensional vascular model reconstructed from CT images is widely used in medical diagnosis. At different phases, the beating of the heart can cause deformation of vessels, resulting in different vascular imaging states and false positive diagnostic results. The 4D model can simulate a complete cardiac cycle. Due to the dose limitation of contrast agent injection in patients, it is valuable to synthesize a 4D coronary artery trees through finite phases imaging. In this paper, we propose a method for generating a 4D coronary artery trees, which maps the systole to the diastole through deformation field prediction, interpolates on the timeline, and the motion trajectory of points are obtained. Specifically, the centerline is used to represent vessels and to infer deformation fields using cube-based sorting and neural networks. Adjacent vessel points are aggregated and interpolated based on the deformation field of the centerline point to obtain displacement vectors of different phases. Finally, the proposed method is validated through experiments to achieve the registration of non-rigid vascular points and the generation of 4D coronary trees.

4D-CAT: Synthesis of 4D Coronary Artery Trees from Systole and Diastole

TL;DR

This work tackles the challenge of motion-induced discrepancies in static 3D coronary models by proposing 4D-CAT, a method to synthesize 4D coronary trees from systole and diastole. It develops a centerline-centric pipeline that first segments the vessel centerlines using a cube-based test, then predicts non-rigid deformation fields via two pathways: a PointNet-based deep-learning model trained with soft-DTW loss and a cube-based sorting approach to align centerlines and derive the deformation field. Vessel points are propagated from centerline deformations through cuboid-based clustering and interpolation, yielding intermediate frames along the cardiac cycle. Validation on clinical CTA data and synthetic scenarios demonstrates improved registration and interpolation accuracy over baselines, with ablations confirming the importance of sorting and loss design. The approach offers a pathway to realistic 4D cardiovascular models and may enhance digital organ representations and clinical workflows.

Abstract

The three-dimensional vascular model reconstructed from CT images is widely used in medical diagnosis. At different phases, the beating of the heart can cause deformation of vessels, resulting in different vascular imaging states and false positive diagnostic results. The 4D model can simulate a complete cardiac cycle. Due to the dose limitation of contrast agent injection in patients, it is valuable to synthesize a 4D coronary artery trees through finite phases imaging. In this paper, we propose a method for generating a 4D coronary artery trees, which maps the systole to the diastole through deformation field prediction, interpolates on the timeline, and the motion trajectory of points are obtained. Specifically, the centerline is used to represent vessels and to infer deformation fields using cube-based sorting and neural networks. Adjacent vessel points are aggregated and interpolated based on the deformation field of the centerline point to obtain displacement vectors of different phases. Finally, the proposed method is validated through experiments to achieve the registration of non-rigid vascular points and the generation of 4D coronary trees.
Paper Structure (12 sections, 10 equations, 4 figures, 2 tables)

This paper contains 12 sections, 10 equations, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Proposed method
  3. Centerline segmentation
  4. Deformation field prediction
  5. Deep-learning method
  6. Cube-based sort
  7. Vascular deformation and interpolation
  8. Experiments
  9. Dataset
  10. Synthetic data
  11. Results
  12. Conclusion

Figures (4)

  • Figure 1: The overview of our proposed method.
  • Figure 2: Comparison with state-of-the-art methods for registering point clouds during systole and diastole.
  • Figure 3: Comparison results of interpolation.
  • Figure 4: Visualization results of interpolation. Blue is the systole, green represents the diastole, and orange denotes the interpolation result.