Cloud-Based AI-Enhanced Dual-Mode System For Automatic Coronary Artery Calcification Detection and Quantification

Coronary Artery Disease (CAD) poses a significant global health burden, accounting for nearly a third of all mortalities. Addressing the urgent need for precise and accessible diagnostics, we present a novel framework that applies a dual-network artificial intelligence model for the automatic detection and quantification of coronary artery calcification from low-dose chest CT scans. This framework enhances the efficiency and accuracy of CAD diagnostics by first localizing the heart with an initial specialized network then quantifying calcifications with a subsequent, finely-tuned network. Our approach has demonstrated outstanding performance, achieving a $90 \%$ Dice score for heart localization and offering promising results for calcification quantification. By integrating with the open-source 3D Slicer platform, this framework represents a significant advance in CAD management by providing low-cost, high-precision diagnostic capabilities that are accessible to researchers and clinicians globally, thereby facilitating early intervention and improving patient outcomes.