Title:
Artificial Intelligence for Structural Heart Disease: From Image Analysis to Clinical Translation
Abstract:
Structural heart diseases (SHDs), including congenital and acquired cardiac abnormalities, present significant challenges for diagnosis, treatment planning, and long-term management. Recent advances in artificial intelligence (AI) have created new opportunities to improve cardiovascular care through automated image analysis, disease characterization, risk prediction, and personalized treatment support. In this talk, I will present our recent research on AI for structural heart disease, covering image segmentation, disease diagnosis, prognosis prediction, and surgical planning. Drawing on large-scale clinical datasets and real-world applications, I will demonstrate how AI can facilitate anatomical understanding, enhance diagnostic accuracy, support clinical decision-making, and improve the planning of complex cardiovascular interventions. Particular emphasis will be placed on congenital heart disease and other clinically relevant structural heart conditions. I will briefly discuss considerations for translating AI technologies into clinical practice, including data quality, model reliability, interpretability, and integration into existing clinical workflows. I will also outline several future directions for advancing the application of AI in cardiovascular care.
Short Bio:
Dr. Xiaowei Xu is an Associate Professor and PI of the Laboratory of Artificial Intelligence and 3D Technology for Cardiovascular Diseases at Guangdong Provincial People’s Hospital, China, and currently a Humboldt Research Fellow at the Leibniz Institute for Analytical Sciences (ISAS), Germany. His research focuses on artificial intelligence for cardiovascular medicine, including medical image analysis, surgical planning, and clinical decision support. He has authored more than 100 peer-reviewed publications and currently serves as an Associate Editor of Pattern Recognition.
The talk will be held during Lambda colloquim. Online participants can access it through the regular zoom link of the colloquim.