Learning objectives

The data-driven AI approaches in pathology have shown huge promise in enhancing clinical decision-making by integrating complex, high-dimensional patient. In particular, pathology AI foundation models are trained on a large collection of tissue images and paired molecular data from diverse diseases, allowing them to learn highly informative patient representations. In this talk, I will highlight some of our recent works on multi-modal pathology foundation models that integrate tissue images, spatial omics, and EHR data. The goal is to understand what common recipes there are for different successful pathology foundation models and where these models can make impact.

Speaker Bio

Andrew H. Song is an Assistant Professor in the Department of Translational Molecular Pathology at MD Anderson Cancer Center and an Adjunct Professor in Computer Science at Rice University. His research centers on computational healthcare/biomedicine, with a focus on developing multimodal foundation models and agentic AI frameworks that integrate tissue imaging, multi-omics data, and clinical reports, with aims to improve clinical outcome prediction for cancer patients. He was previously a postdoctoral fellow in the Department of Pathology at Brigham and Women’s Hospital/Harvard Medical School with Dr. Faisal Mahmood. Before then, he received his BS and PhD in EECS at MIT.