Alan B. McMillan, PhD, Principal Investigator of the Molecular Imaging/Magnetic Resonance Technology Lab (MIMRTL), recently delivered two talks emphasizing the growing impact of foundation models in medical imaging.
On April 9, 2025, Dr. McMillan presented at the “Clinical Practice Enhanced by Artificial Intelligence Grand Rounds 2025-2026: Artificial Intelligence Applications in Radiology.” His talk, “Artificial Intelligence (AI) Applications in Radiology – Harnessing the Power of Foundation Models,” highlighted how foundational AI models streamline radiological workflows by reducing the need for specialized training for each individual application. foundation models can generalize across multiple imaging tasks, improving efficiency in image interpretation, enhancing diagnostic accuracy, and fostering broader adoption of AI in clinical radiology practice. The session, co-presented with colleagueJoshua Warner, MD, PhD, CIIP, focused on practical implementations and future integration of advanced AI tools in everyday radiological workflows.
On April 10, 2025, Dr. McMillan continued this discussion at the University of Wisconsin Carbone Cancer Center’s Annual Research Retreat, held at the Health Science Learning Center Atrium. With the event themed “AI for Patient Care,” his presentation, titled “Foundation Models and Generalist Artificial Intelligence in Medical Imaging,” explored how generalist AI and foundational models can transform medical imaging, particularly in cancer detection, diagnosis, and treatment planning. Dr. McMillan showcased ongoing work at MIMRTL, highlighting advances in foundation models for disease classification, multimodal data search, and model drift detection to maintain accuracy over time.
Looking forward, foundation models hold significant potential to revolutionize medical imaging by providing more flexible, scalable, and robust AI solutions. By facilitating rapid adaptation to new clinical challenges, these models can substantially improve patient outcomes, streamline clinical decision-making processes, and support personalized medicine initiatives.