Development of an AI-Based Website for Automated Pre-Operative Planning in Aortic Valve Replacement Using Deep Learning

Abstract

This paper introduces a web-based software pipeline to help perform minimally invasive aortic valve replacement (AVR) procedures. The complete pipeline contains three main steps: aortic extraction, automatic landmark detection, and calculation of the specific measurements required for valve replacement. For Computed Tomography Angiography (CTA) scans, separate deep neural network models based on the U-Net architecture in the Medical Open Network for AI (MONAI) framework segment out key anatomical landmarks, including the Sinotubular Junction (STJ), Left Coronary Artery (LCA), Right Coronary Artery (RCA), and Annulus Plane. This improves the accuracy of aorta segmentation and landmark detection. A web interface facilitates easy data upload, visualization of segmentation results, and access to key measurements for preoperative planning. Artificial Intelligence (AI) and machine learning in medical imaging lead to more accurate preoperative planning of Transcatheter Aortic Valve Replacement (TAVR) procedures, reducing the likelihood of complications from improperly fitted valves. There is still more work to be done before the model can be used in cardiovascular disease diagnosis, including refining models, enriching datasets, and integrating several features that are beneficial for cardiologists.