CHEO’s AI algorithm makes rare disease diagnoses

OTTAWA – Harnessing the power of artificial intelligence (AI), CHEO researchers have developed a groundbreaking search algorithm that identifies children and youth who may have an undiagnosed rare genetic disease and refers them for genetic testing.

“The ThinkRare algorithm is incredibly exciting and promising because it means we can help families find answers and get the care and support they need sooner,” said Dr. Kym Boycott (pictured), senior scientist at the CHEO Research Institute and Chief of Genetics at CHEO. “This algorithm is a game changer. Using AI to scour CHEO’s electronic health record based on set criteria, ThinkRare can accurately identify kids who may have an undiagnosed rare genetic disease and refer them to our clinic – something that may have never happened without it.”

Ten-year-old Antony Wistaff and his family have spent countless hours at CHEO, calling it a “second home”. Antony was born prematurely in October 2013 and a few days later underwent emergency surgery at CHEO to place a shunt for hydrocephalus.

But that was only the beginning of what would become a decade-long diagnostic journey consisting of more than 100 outpatient appointments across six different specialty clinics at CHEO, and 30 trips to the emergency department for various reasons.

That was until recently, when the ThinkRare algorithm identified Antony as potentially having an undiagnosed rare genetic disease and flagged him for a referral to receive genome-wide sequencing testing – a test that simultaneously analyzes the more than 5,000 genes that have been associated with rare disease – and is now available clinically in Ontario.

The results of the genetic testing discovered that Antony has Chung-Jansen Syndrome – a rare disorder resulting from a pathogenic variant in the PHIP gene. At present, the syndrome has been diagnosed in only about 400 people worldwide and it explained many of Antony’s health and behavioural challenges, including his developmental delays, learning difficulties, and large head size.

“When we found out that Antony was diagnosed with Chung-Jansen Syndrome, it answered so many questions for our family,” said Georges Wistaff, Antony’s dad. “This research brought a kind of peace to our house. Had we known this sooner, it would have meant less questioning as parents, less stress, and more support because we would have had a clear diagnosis for Antony. A little bit of blood and a simple test, answered so many questions.”

To date, Think Rare, which is currently operating as a research project approved by the CHEO Research Ethics Board, is three for three – meaning the first three patients identified by ThinkRare and referred to genetics have received test results and been diagnosed with a rare disease. Genetic testing is underway for many other families identified by ThinkRare.

“Our goal is to flip the diagnostic care journey on its head and start with genetic testing earlier on the care pathway. By incorporating the ThinkRare algorithm into clinical care, we will be able to support CHEO clinicians and frontline workers with the power of machine learning to find the needle in the haystack,” added Dr. Boycott, who is a Tier 1 Canada Research Chair in Rare Disease Precision Health and Professor of Pediatrics at the University of Ottawa.

Work is currently underway at CHEO to transition the ThinkRare project from research into clinical practice, with all the necessary patient privacy mechanisms in place.

“CHEO is uniquely positioned to develop an impactful algorithm such as ThinkRare because of CHEO’s investment in a robust electronic health record system, our commitment to innovation, our close collaboration between clinical and research teams, and because we are the only pediatric healthcare centre in Eastern Ontario serving a wide geographic area. At CHEO, we have brought together all the necessary elements when it comes to making AI advancements in healthcare,” said Dr. Jason Berman, CEO and scientific director, CHEO Research Institute, and vice-president Research, CHEO.

The ThinkRare project was made possible with funding from the CHEO Foundation, the CHAMO Innovation Fund, and Ontario Genomics.