Computer Vision News - January 2022

59 Coronary 3D Reconstruction from X-ray and x-ray exposures and reducing the risk for contrast over-dose and radiation-associated complications. Finally, embedding this technology in the PCI workflow shortens the procedure and adds confidence to the physician’s decision-making. Overall, this tool allows better anatomical and pathological visualization and appreciation, as well as AI-based algorithms for coronary lesion characterization, andaugmentstheprocedural success, decreases the complication rate and (based on a very broad scientific literature) has the potential of increasing stent patency rate and patients’ long-term survival. This AI-based technology requires no additional hardware installation and can be executed using the existing setup of the cath-lab. It is a quick and efficient method to provide better healthcare for patients undergoing PCI. Read about this and more pioneering solutions in AI for Cardiology . algorithms combined with classic computer vision methods, a 3D model of the coronary arteries is reconstructed accurately, rapidly, and automatically. This model can be used to better visualize artery structure and measure vessel dimensions in points-of- interest. More advanced capabilities, enabled using deep learning, are stenosis detection , 3D quantitative coronary angiography (3D QCA); they can even be the baseline for computerized fractional flow reserve (FFR) measurement . Additional modules can show artery modification due to stent placement or place a virtual stent in the desired position within the coronary artery. These uses replace invasive cardiac measurements and give the physician a better pre-procedure planning tool for stent selection. Using a 3D model for visualization is more intuitive than using the 2D view currently available, resulting in fewer contrast injections