Abstract
Clinical decision-making for vascular occlusion is often compromised by incomplete patient data and time constraints, leading to delayed interventions and an elevated risk of complications. To address this challenge, we propose an artificial intelligence (AI)-powered control and decision-support system (CDSS) that establishes a closed-loop platform for predictive intervention. Our framework integrates multi-modal data—from medical imaging and electronic health records— using a personalized computational model powered by machine learning (ML) and computational fluid dynamics (CFD). This integration facilitates real-time risk stratification and hemodynamic simulation of potential interventions, such as stent placement. Key outputs include optimized stent sizing and positioning parameters derived from anatomical and physiological constraints. Functioning as an intelligent controller within a clinical feedback loop, the system provides physicians with data-driven, actionable recommendations, significantly reducing decision-making time and improving long- term vascular outcomes. Validation on retrospective clinical data demonstrated a diagnostic accuracy of 96.5%, a 60% reduction in flow turbulence, and a 91% reduction in time-to-decision within a controlled research environment.. This approach bridges data- driven AI prediction with physics-based simulation, offering a scalable paradigm for personalized cardiovascular medicine.