Dr. James White, MD, director of the Libin Institute’s Stephenson Cardiovascular Imaging Centre and a professor at UCalgary’s Cumming School of Medicine, has spent over a decade developing novel ways that medical images can help predict patient outcomes.
His expanding research program, called the Personalized Diagnostics Program (PDP), focuses on how data collected at the time of heart imaging tests can be used to predict an individual’s risk of developing cardiovascular disease or its complications. These tools allow physicians to tailor care plans to each individual patient, a concept called Personalized Medicine.
“While today’s electronic health records are excellent for monitoring patient care and patterns of health, they lack the raw data created by diagnostic tests,” says White. “Imaging tests produce thousands of data points that can be used to describe each patient’s disease. By starting with this data and adding relevant data from the medical record, we believe we can make much more accurate predictions for patients.”
According to White, the team is looking to predict such things as whether an individual will be admitted to hospital with heart failure, develop an arrhythmia or experience sudden cardiac death.
Getting to this point hasn’t been easy as generating predictive models requires a large amount of data.
White and his team have asked more than 18,000 patients with heart disease to contribute by participating in the Cardiovascular Imaging Registry of Calgary (CIROC), a study of patients with heart disease undergoing imaging tests in Calgary. On average, eight out of 10 patients have consented.
The registry is one of the largest patient-consented databases for cardiac imaging in the world. Every set of images has been linked to standardized disease descriptions, patient-reported health questionnaires, echocardiograms, laboratory tests and health outcomes.
Data scientists on White’s team analyze the data using artificial intelligence methods, such as machine learning and deep learning. They look for patterns in the data that are difficult to see without these more advanced techniques. These patterns are helping researchers discover new ways of diagnosing certain types of disease; cluster individuals that have similar characteristics; and develop computer models that can predict when future events will occur.
The CIROC project has already led to powerful demonstrations of personalized medicine. For example, from 6,000 patients with no prior history of arrhythmia (heart rhythm abnormalities), White’s team was able to build a computer model that can predict when and if someone will develop atrial fibrillation, a common arrhythmia that is linked to strokes.
They are also refining a model to predict future heart failure hospital admissions and have developed a method of automatically diagnosing heart disease directly from MRI images.
White is now expanding his research program to a global network of hospitals. His team recently combined data from 12 hospitals in Canada, the US, Spain and Italy to show the value of cardiac MRI data in predicting cardiac events in patients with dilated cardiomyopathy. The study, recently published in the peer-reviewed journal Circulation: Cardiovascular Imaging, found that cardiac MRI can predict life-threatening arrhythmias in this group of patients.
This emerging research network, called MINICOR, will now adopt White’s established approach to research, implementing it at sites across Canada, US, South America and Europe. The plan is to launch the network at 10 institutions this year, with the goal of enrolling 60,000 to 70,000 patients by 2025. MINICOR will greatly expand the team’s potential to build and validate relevant risk-prediction models for patients across a broad range of cardiovascular diseases.
“It has been a tremendous challenge, but we are well on our way and very much looking forward to getting this important project off the ground,” said White of MINICOR, adding his team has been working on getting the project launched for more than two years.
This project is being funded in part by donations raised at the Libin Institute’s 2019 gala, The Beat Goes On. The Institute is grateful for its community partners.