Chris McIntosh

At a Glance

• The focus of my lab is on advancing the theory and application of artificial intelligence (AI) in medicine from bench-to-bedside
• Methodologies include transfer learning, meta learning, computer vision, and explainable AI to build clinically usable models
• Close collaborations with clinical stake holders yield relevant clinical problems, and large data sets with tens of thousands of images
• Developed technologies are used to prospectively provide patient care at UHN and globally
• Diseases studied include prostate cancer, lung cancer, cardiovascular disease, and associated interventions

Short Bio

Chris McIntosh is a Senior Scientist at the University Health Network’s Peter Munk Cardiac Centre Research Institute, and Associate Professor in the Department of Medical Biophysics at the University of Toronto, holding additional appointments in Computer Science, and Medical Imaging. His research focuses on the theory, development, and clinical application of AI in medicine for improving patient care, encompassing multimodal learning, meta learning, and explainable AI. His past work on AI in radiation therapy has received regulatory approvals and is used directly in patient care around the world. His current investigations include AI and wearable technologies for heart failure management, and using multimodal learning to build foundational models in healthcare. He has authored publications in clinical and technical venues including Nature Medicine, IEEE Transactions on Medical Imaging, the International Conference on Computer Vision, the European Conference on Computer Vision, and Medical Image Computing and Computer Assisted Intervention.

Research Synopsis

Machine Learning for Healthcare from Bench to Bedside: The McIntosh Lab develops and deploys artificial intelligence methods that directly improve patient care. Embedded within the University Health Network and the University of Toronto, our group works at the intersection of computer science, medical biophysics, imaging, cardiology, oncology, and clinical translation. We focus on AI systems that can learn from diverse medical data including medical imaging, physiological signals, wearable sensors, structured reports, and clinical text to support earlier detection, better treatment planning, improved workflow efficiency, and more equitable deployment across healthcare settings.

Multimodal and Generative Medical AI: Modern healthcare decisions require integrating many complementary data sources. We develop multimodal and generative AI models that jointly learn from imaging, electrocardiograms, echocardiography, wearable data, and clinical text. These models are designed to support cross-modal reasoning, uncertainty-aware prediction, and generalizable representations across medical tasks. Current projects include medical foundation models for multimodal binding, ECG-to-echocardiography prediction, and models that can integrate patient data across modalities to improve diagnosis, monitoring, and clinical decision support.

Wearable Biomarkers and Longitudinal Patient Monitoring: In collaboration with clinical partners in heart failure and industry partners including Apple, we develop AI methods that transform wearable sensor data into clinically meaningful biomarkers. Our recent work demonstrated that smartwatch data can estimate daily functional capacity in heart failure patients and that longitudinal declines in these estimates can predict subsequent healthcare events. This research aims to shift healthcare from episodic assessment toward continuous, patient-centered monitoring that can identify deterioration earlier and support timely intervention.

Recent Publications

View recent publications of Dr. McIntosh on Google Scholar.

Graduate Students

Siham Amara Belgadi
Aly Sherif Khalifa
Cathy Ong Ly