AI in Medicine: Some New Results and Next Steps to the Future, Ge Wang

Speaker: Ge Wang, PhD | Clark & Crossan Endowed Chair Professor and Director of Biomedical Imaging Center, Rensselaer Polytechnic Institute

Date: Monday, August 19th, 2024

Time: 10:00 AM Central Time

Location: Zoom

Title: AI in Medicine: Some New Results and Next Steps to the Future

Abstract: The advent of foundation models in AI has paved the way for great advancements in medical applications. This talk will delve into the development of Specialty-Oriented Generalist Medical AI (SOGMAI) models and physics-inspired generative models in medical imaging. First, we report a SOGMAI model for low-dose CT lung cancer screening, known as the Medical Multimodal Multitask Foundation Model (M3FM). This model integrates multimodal data, including 3D CT scans and clinical records, to perform multiple tasks such as lung nodule detection, lung cancer risk prediction, and cardiovascular disease diagnosis. Then, we discuss the role of physics-inspired generative models, specifically diffusion models, in medical imaging. These models offer a novel approach to reconstructing high-quality images from low-dose CT scans. By incorporating physical principles into the generative modeling process, these models significantly improve the quality and stability of reconstructed images. Through these studies, we highlight the transformative potential of sophisticated large AI models for medicine and discuss future directions of research.

Bio: Ge Wang, the Clark & Crossan Endowed Chair Professor and Director of the Biomedical Imaging Center at Rensselaer Polytechnic Institute, is a renowned leader in medical imaging. His current interests include not only classic physics-based tomographic methods but also modern generative AI models, multimodal foundation models, solutions to AI-specific problems in medical imaging, as well as their medical applications. He has published over 600 journal papers (1 in Nature, 5 in Nature Machine Intelligence, 1 in Nature Communications, 3 in PNAS, and >100 in IEEE Transactions) and >150 issued/published patents. Supported by GE, his team developed medical AI software, which is incorporated into CT scanners and in clinical translation. He is a Fellow of IEEE, SPIE, AAPM, OSA, AIMBE, AAAS, and National Academy of Inventors (NAI), further distinguished by the RPI Wiley Distinguished Faculty Award, IEEE R1 Outstanding Teaching Award, EMBS Career Achievement Award, SPIE Meinel Technology Award, Sigma Xi Chubb Award for Innovation, IEEE NMISC Hoffman Medical Imaging Scientist Award, among others.