22nd April 2025
The International Progressive MS Alliance has approved funding to support the next phase of a collaborative network led by Prof. Tal Arbel, CIFAR AI Chair and core member at Mila – Quebec AI Institute, and McGill University in Canada. This project will create an image-based causal deep learning model, which will help predict future disability progression in multiple sclerosis, as well as predict individual patient response to different treatments.
The project, titled, “Causal Deep Learning for Progression in MS Based on Magnetic Resonance Imaging (MRI),” has been funded for €675,000 over the next two years and is led by Dr. Tal Arbel, PhD, McGill University, with co-investigator Dr. Douglas Arnold at the Montreal Neurological Institute (MNI) and additional collaborators and partners from Mila, McGill and from around the world. The research team brings extensive combined experience in machine learning, computer vision and medical image analysis.
While existing machine learning models for medical images are only used as prognostic models for untreated patients, this project will combine MRI data from different timepoints – along with clinical and demographic data – to predict how individual patients will respond to different therapies in future timepoints, and how their disease may progress under various treatment scenarios. This project will apply and develop modern artificial intelligence (AI) models to a starting dataset of about 70,000 MRI scans and 200,000 clinical visits from 12,000 people and is anticipated to grow as data from new trials and extension studies is received.
“Recent advancements in deep learning for medicine have been impressive, but learning to predict future individual patient outcomes and treatment effects based on high dimensional medical images requires advances in modern deep learning and causal inference,” said Dr. Tal Arbel. “We will leverage our team’s extensive experience, as well as our interdisciplinary collaborations, to build on recent advances in causal deep learning models in order to change the way clinicians approach an individual’s treatment plans.”
This study is an extension of previous research funded by the International Progressive MS Alliance that was led by Dr. Arnold, focused on speeding up and improving clinical trials in progressive MS. After successfully harmonizing the dataset of scans and clinical visits from multiple clinical trials, the Alliance will support Dr. Arbel and her team to expand this work to advance deep learning models that can predict disease progression and/or understand the heterogeneity of treatment response.
“The significant, specific expertise of the team that is coming together for this project can only be accomplished when we collaborate with the best minds, which often go across borders,” said Dr. Robert Fox of the Cleveland Clinic and Chair of the Alliance Scientific Steering Committee. “By working together with researchers, academic and industry partners around the world, this project utilizes the most recent advances in AI to change the story for people living with MS.”
Learn more about the International Progressive MS Alliance’s priority to Speed Up and Improve Clinical Trials.