AGFA-Net: Attention-Guided Feature-Aggregated Network for Coronary Artery Segmentation Using Computed Tomography Angiography

coronary artery disease (CAD) remains a prevalent cardiovascular condition, posing significant health risks worldwide. Accurate segmentation of coronary arteries from coronary computed tomography angiography (CCTA) images is crucial for diagnosis and treatment planning. Despite advancements, coronary artery segmentation (CAS) remains a significant challenge due to the considerable variability in the scale of coronary arteries, the complex anatomical structure and morphology, and the low contrast between blood vessels and the surrounding tissue. To effectively tackle these issues, we propose an attention-guided feature-aggregated network (AGFA-Net) specifically designed for CAS using CCTA images. AGFA-Net leverages attention mechanisms, feature refinement and hierarchical feature aggregation to capture salient features, acquire a more distinctive semantic representation and enhance segmentation accuracy. Evaluation on a dataset comprising 1,000 CCTAs demonstrates AGFA-Net's superior performance, achieving an average Dice coefficient similarity of 86.74 % and a 95% Hausdorff distance of 0.82 mm during 5-fold cross-validation. Ablation studies further validate the effectiveness of the proposed modules, highlighting their contributions to improved segmentation accuracy. Overall, AGFA-Net offers a robust and reliable solution for CAS, addressing challenges posed by varying arterial sizes, complex anatomies, and low image contrast.