A graph-based approach for spatio-temporal segmentation of coronary arteries in X-ray angiographic sequences

Area of Science:

• Medical Imaging
• Cardiovascular Technology
• Computational Cardiology

Background:

• Accurate segmentation and tracking of coronary arteries (CAs) are essential for biophysical measurements in pediatric interventional cardiology.
• Existing methods often focus on either lumen segmentation or centerline tracking, not simultaneously addressing both for specific CAs.

Purpose of the Study:

• To introduce a novel algorithm for simultaneous segmentation and tracking of CAs from 2D X-ray angiography sequences.
• To improve the accuracy and precision of CA analysis in pediatric patients.

Main Methods:

• A new algorithm, Temporal Vessel Walker with superpixels (SP-TVW), was developed.
• The algorithm combines graph-based formulation, temporal priors, and superpixel groups for enhanced CA extraction.
• It processes 2D X-ray angiography sequences.

Main Results:

• The SP-TVW algorithm achieved a mean recall of 84% across 12 patient sequences.
• A Dice index of 70% was obtained for segmenting and tracking right and circumflex coronary arteries.
• The method demonstrated superior performance over the polyline method, with a reduced centerline distance error (0.23mm vs. 0.94mm).

Conclusions:

• The SP-TVW algorithm provides accurate segmentation and tracking of coronary arteries in pediatric angiography.
• This advancement offers improved localization and measurement capabilities for interventional cardiology applications.
• The method shows significant potential for enhancing diagnostic and therapeutic procedures in pediatric cardiac care.