Segmentation of the thoracic aorta in noncontrast cardiac CT images

Area of Science:

• Medical Imaging
• Cardiovascular Imaging
• Computational Anatomy

Background:

• Aortic calcification is a known indicator of cardiovascular disease.
• Accurate segmentation of the thoracic aorta is crucial for calcification detection.
• Existing methods may lack precision in noncontrast CT imaging.

Purpose of the Study:

• To develop and validate a novel computational method for thoracic aorta localization, centerline extraction, and segmentation.
• To enable the detection of aortic calcification in noncontrast cardiac CT images.
• To improve the assessment of cardiovascular disease risk.

Main Methods:

• A multi-stage approach involving regression for initial localization and optimal path detection for centerline extraction.
• Utilized dynamic programming in Hough space and a fast marching-based minimal path extraction framework.
• Resampled the volume and applied dynamic programming for 2D cross-sectional segmentation, mapped back to the original space.

Main Results:

• The proposed method successfully localized, extracted the centerline, and segmented the thoracic aorta.
• Promising results were achieved when assessed on noncontrast cardiac CT scans.
• The technique demonstrated potential for accurate aortic calcification detection.

Conclusions:

• The developed method offers a robust approach for thoracic aorta segmentation in noncontrast cardiac CT.
• This technique can significantly contribute to the early detection of aortic calcification and cardiovascular disease.
• Further validation on larger datasets is warranted to confirm clinical utility.