Discontinuity-Aware Coronary Artery Segmentation on CCTA Image

Accurate segmentation of coronary arteries from Coronary Computed Tomography Angiography (CCTA) images is crucial in the diagnosis of Coronary Artery Disease (CAD). However, existing vascular segmentation algorithms often struggle with discontinuity problems (i.e., vessel segmentation with breakpoints), making it challenging to obtain segmentation results with complete topological structure. To address this issue, our study introduces a novel algorithm to detect these discontinuities using knowledge of vascular anatomy and further incorporate it into automated segmentation process to improve performance. Specifically, our model starts with a coarse segmentation network to get preliminary segmentation results, and then employs a discontinuity detector to identify discontinuity from these preliminary results. Following these steps, a refinement network further improves the segmentation with guidance from discontinuity findings. Finally, a merging module is used to combine coarse and refined segmentation for final segmentation results. Extensive experiments on our collected dataset demonstrate that our method outperforms state-of-the-art approaches. The code is available at Link.