Joshua Levy

Dr. Joshua Levy is the Director of Digital Pathology Research at Cedars-Sinai and Assistant Professor in the Departments of Pathology and Computational Biomedicine. He formerly served as an Assistant Professor of Pathology, Dermatology, and Epidemiology at Dartmouth College Geisel School of Medicine. Dr. Levy's research focuses on integrating AI and biostatistics in anatomic pathology, developing computational methods for histopathology, cytopathology, text and image processing, and analyzing spatial molecular variations in tumors through spatial genomics technologies, amongst numerous other application areas. His academic path commenced with a physics degree from the University of California, Berkeley, followed by computational biology and software development roles at Lawrence Berkeley National Laboratories and in industry, leading to a doctorate in quantitative biomedical sciences from Dartmouth College Geisel School of Medicine. At Cedars-Sinai, his goal is to establish a nationally recognized digital pathology research program and create student recruitment pathways to bring new talent into the field. Dr. Levy's mentorship efforts are highlighted by the establishment and directorship of a national internship program, through which he has guided over 150 students from diverse educational backgrounds. Dr. Levy also serves as Associate Director of the Cedars Sinai AI Campus program. Dr. Levy also holds various consulting roles at the Biomedical National Elemental Imaging Resource and Veteran Affairs.

‍

‍

Digital Pathology and Artificial Intelligence for Spatial Molecular Inference and Multimodal Integration

This talk focuses on the pivotal role of artificial intelligence in enhancing digital pathology by enabling detailed spatial molecular analysis of tissues, crucial for biomarker identification and the development of clinical decision support tools. We will discuss how AI models are used to predict spatial gene expression patterns directly from whole slide images through paired spatial transcriptomics data to facilitate spatial molecular assessment at scale. The presentation will also cover the application of data valuation approaches for impact the specimen processing and image quality on spatial genomics applications, highlighting the balance between cost and algorithmic performance. We will introduce unpublished work on spatial multimodal profiling to uncover additional indicators of disease metastasis. This session is designed to provide a comprehensive overview of the latest spatial biology AI applications in digital pathology for both experts and newcomers to the field.