top of page
Image by Ali Hajiluyi
Image by Pin Adventure Map
LCI Logo_Colour-01.png

Cardiac Magnetic Resonance Imaging (MRI) is a non-invasive imaging test that uses a strong magnetic field to generate high quality images of the body's organs without need for radiation.  It has become widely used for the diagnosis of heart muscle problems called "cardiomyopathies", these being a collection of conditions that can lead to heart failure and abnormal heart rhythms. Cardiomyopathies are challenging to diagnose, typically requiring the use of contrast agents (i.e., gadolinium) to highlight specific disease patterns and requiring physicians with highly specialized training to interpret these images. Unique approaches that can minimize this need for contrast and specialized physician input are therefore desired.

Artificial Intelligence (AI) is increasingly a part of our lives and helps us do tasks every day.  It is a branch of computer science that aims to build machines capable of learning how to perform a task, and often referred to as "machine learning". In healthcare, we already use this to perform time-consuming measurements of hearts on our images. However, we are hopeful it can help us diagnose if a patient has a disease, and how best to treat it.  However, this requires careful testing to make sure it is equally helpful for all patients, regardless of their sex or ethnicity. This is why studies, such as the International AID-MRI study, are so important: making sure that these tools help patients from all around the globe irrespective of whether they are female, male, African or Indian.

AID-MRI is an international study testing innovative software tools developed at the University of Calgary to construct beating heart models (digital twins) of each patient's heart from routinely captured, non-contrast cardiac MRI images. Each of these models elegantly describes the unique shape and movement of a patient's heart, providing data that is ideally suited to AI-based analyses of 3-dimensional objects, an approach  called geometric deep learning. Using this approach, the International AID-MRI study aims to describe the ability of AI-informed digital twins to diagnose cardiomyopathy and to predict each individual patient's future risk of major adverse cardiovascular events. 

A total of 10,000 patients with a broad range of heart diseases have already been recruited in Alberta, Canada, providing their consent for images of their hearts to be used for training these models. We are now recruiting 1,000 patients across 10 international sites to test if these models provide similar value in different hospital settings and in patients with different sex and ethnicity. The AID-MRI study is publicly funded by the Canadian Institutes of Health Research (CIHR). 

bottom of page