Exploring a Python-based Semi-Automatic Approach for Coronary Artery Segmentation

Resumo

The segmentation of coronary arteries from Computed Tomography (CT) images is vital for coronary artery diseases (CADs) diagnosis and treatment. Combining thresholding, region growing, and several entity properties, we have implemented an in-house semi-automated method using Python, circumventing commercial software dependence in hospitals. Our technique swiftly isolates coronary arteries in under 2 minutes, enhancing efficiency, accuracy, and reproducibility compared to manual MIMICS® segmentation. Moreover, it is able to detect coronary branches that surge upstream severe stenosis which is usually a major limitation due to lack of
contrasted blood. In sum, our proposed method stands as a transformative stride toward the efficient and accurate segmentation of coronary arteries in the clinical landscape. Anchored in the marriage of computational prowess and clinical imperatives, it emerges as a potent beacon, poised to illuminate a path toward enhanced diagnostic insight and treatment efficacy.