Deep Learning Based Automatic Segmentation of the Thoracic Aorta from Chest Computed Tomography in Healthy Korean Adults

This study aimed to establish reference values for aortic size using a fully automated deep learning-based segmentation method on chest computed tomography (CT) scans. The study included 704 healthy Korean adults (mean age 50.6 years; 57.8% males) who underwent contrast-enhanced chest CT for health screening. A convolutional neural network (CNN) was trained and applied to 3D CT images to automatically segment the thoracic aorta based on the Society for Vascular Surgery/Society of Thoracic Surgeons classification. The CNN-generated masks were reviewed and corrected by expert cardiac radiologists. Results showed that aortic size was significantly larger in males across all zones (zones 0-8) and increased with age, approximately 1 mm per decade. For example, aortic size in zone 2 increased from 25.4 mm in men aged 30-39 to 29.8 mm in men aged 70 and older. The deep learning algorithm provided reliable aortic measurements, which have clinical implications for detecting aortic aneurysms and other aortic diseases.