Speaker: Pingkun Yan, PhD | P.K. Lashmet Career Development Chair Professor of Biomedical Engineering, Rensselaer Polytechnic Institute
Date: Tuesday, June 17th, 2025
Time: 10:00 AM Central Time
Location: Zoom
Title: “Medical Image Foundation Models for Cardiovascular Disease Risk Prediction”
Abstract: Cardiovascular disease (CVD) remains a leading cause of mortality worldwide, making early risk prediction essential for timely intervention. This talk presents a series of recent efforts to develop medical image foundation models for CVD risk assessment using chest imaging. We began by exploring whether AI can opportunistically extract CVD risk indicators from low-dose CT scans acquired for lung cancer screening, enabling risk prediction without additional imaging. Building on this success, we expanded our work to multimodal foundation model frameworks to enhance performance and generalizability across chest imaging tasks. Recognizing the ubiquity and accessibility of chest X-rays (CXRs), we developed CheXFound, a foundation model that integrates global and local features from multi-view CXRs, achieving strong performance across diverse clinical tasks, including CVD and mortality risk prediction. To overcome the inherent limitations of 2D CXRs, we further introduced a novel world model that captures 3D volumetric representations from CXRs by learning the transition dynamics of CT volume projections. The trained model, Xray2Xray, synthesizes volumetric context from frontal and lateral views and outperforms existing approaches in CVD risk estimation. Together, these advances demonstrate the promise of medical image foundation models in enabling scalable, accurate, and cost-effective cardiovascular risk stratification from widely available imaging data.
Bio: Dr. Pingkun Yan is the P.K. Lashmet Career Development Chair Professor of Biomedical Engineering at Rensselaer Polytechnic Institute (RPI). Prior to joining RPI in 2017, he served as a Senior Scientist at Philips Research, collaborating clinically at the National Institutes of Health (NIH). His research centers on artificial intelligence and machine learning applications in medical imaging and healthcare informatics, with a focus on translational solutions developed in close partnership with clinicians. Dr. Yan has authored over 110 peer-reviewed journal articles and holds 12 issued patents. His work has received over 12,500 citations, reflecting his contributions to the field. He is a recipient of the NSF CAREER Award and the NIH Trailblazer Award, and his research has been recognized with several best paper awards at international conferences. Dr. Yan also contributes to the international research community as a board member of the MICCAI Society. He is the Chair of the IEEE EMBS Technical Committee on Biomedical Imaging and Image Processing, and was named an IEEE EMBS Distinguished Lecturer for 2025–26. He serves as an associate editor for multiple international journals, including IEEE TMI and Neurocomputing. He is a senior member of both the National Academy of Inventors (NAI) and IEEE.