Enhancing percutaneous coronary intervention with heuristic path planning and deep-learning-based vascular...

Area of Science:

• Medical Imaging
• Interventional Cardiology
• Computer-Aided Surgery

Background:

• Percutaneous coronary intervention (PCI) relies on hemodynamic parameters for guidance, which are less intuitive than imaging.
• Current guidance methods present challenges for cardiologists' decision-making during PCI procedures.

Purpose of the Study:

• To develop a novel PCI guiding assistance system combining vascular segmentation and path planning.
• To provide cardiologists with clear, visualized information for improved PCI procedures.

Main Methods:

• A novel vascular segmentation network and a heuristic intervention path planning algorithm were developed.
• A dataset of 1077 DSA images from 288 patients was utilized.
• User experiments with a Likert Scale were conducted to evaluate system performance.

Main Results:

• The system generated satisfactory and reasonable paths for PCI.
• The proposed method outperformed state-of-the-art baselines in Jaccard (0.4091), F1 (0.5626), and Accuracy (0.9583).

Conclusions:

• The developed system effectively assists cardiologists by providing clear vascular segmentation and optimal intervention paths.
• The system shows significant potential for advancing robotic PCI autonomy.