Finding the Needle in the Haystack
July 21, 2026
Genetic testing paves the way to personalized care in many rare diseases. Now CHEO is harnessing the power of AI to help more patients get the testing – and treatment – they need.
By definition, a rare disease affects fewer than 1 in 2000 people in their lifetime. But with more than 7000 rare diseases out there – and more being added to the list – that adds up to a lot of people: an estimated 3 million in Canada, or about one in every 12.10 About 80 percent of rare diseases have a genetic cause, making genetic and genomic testing a crucial piece of the diagnosis and treatment arc.
Many of these diseases first show up in infants and children, and a prompt and accurate diagnosis gives these young patients their best chance of thriving. But the rarity of these conditions, coupled with their often nonspecific or overlapping symptoms, can make it challenging for health professionals to notice patterns, make connections, and pin down a diagnosis. And without a solid diagnosis, patients miss out on treatment options tailored to their specific condition.35
We now have a new tool to help us figure out which patients could benefit from genetic testing and personalized treatment: artificial intelligence. ThinkRare, a cutting-edge AI-based initiative at the CHEO Research Institute in Ottawa, is pointing the way.
“Rare diseases can often hide in plain sight. ThinkRare helps identify children with a potential undiagnosed rare disease who might otherwise spend years without being referred for genetic testing,” says Dr. Kym Boycott, Senior Scientist at the CHEO Research Institute, Chief of Genetics at CHEO, and Canada Research Chair in Rare Disease Precision Health.36 “Our goal is to flip the diagnostic care journey on its head and start with genetic testing earlier in the care pathway,” thus helping health professionals “find the needle in the haystack.”37
“Rare diseases can often hide in plain sight. ThinkRare helps identify children who might otherwise spend years without being referred for genetic testing.”
Dr. Kym Boycott
Senior Scientist, CHEO Research Institute
ThinkRare uses a rules-based AI algorithm that automatically scans through routinely collected information from young patients’ visits to their family doctors and specialists, seeking out clues that a patient might have an undiagnosed rare genetic disease. These patients are flagged so their care team can discuss next steps, which may include genetic testing or referral to a relevant specialist for further discussion of diagnostic and treatment options.38
In 2024, CHEO integrated ThinkRare into its clinical workflow, making it the world’s first AI algorithm to be used in pediatric clinical care.36 Since then, 41 children have been flagged and referred to CHEO Genetics for further testing; of those who completed the testing, 70% have been diagnosed with a rare genetic disease.36 Given that conventional screening and testing algorithms usually only yield a solid diagnosis in about one-third of cases, that means more than twice as many families are getting an answer – often much sooner than via conventional channels.36,38 And the earlier the diagnosis, the more quickly patients and their healthcare team can identify and access the appropriate personalized treatment.35
Soon, patients and families in other parts of Canada will be able to benefit from ThinkRare’s insights. The CHEO Research Institute has recently announced collaborations with McMaster Children’s Hospital in Hamilton, Ontario, and two Alberta hospitals (Alberta Children’s Hospital in Calgary and Stollery Children’s Hospital in Edmonton) to integrate the program in their centres.36
And the innovation doesn’t stop there: CHEO plans to work with additional centres across the country to implement ThinkRare, helping even more patients get their diagnosis and treatment sooner. According to Dr. Boycott, “scaling this AI tool across Canada will help shorten the diagnostic odyssey for countless families, which was my ultimate goal when I first envisioned the impact of ThinkRare.” 36
“ThinkRare shows what happens when clinicians, data scientists, and families come together to build meaningful, responsible AI,” says Dr. Jason Berman, CEO and Scientific Director of the CHEO Research Institute, and Vice-President of Research at CHEO. “Its remarkable success at CHEO – and now its national expansion – means more children and families across Canada will get the answers they need sooner.”36
References
10. Canadian Organization for Rare Diseases (CORD). About CORD. https://www.raredisorders.ca/about/about-cord
35. University Health Network. How genetic testing guides treatment for patients with rare diseases. https://www.uhn.ca/corporate/News/Pages/genetic-testing-guides-rare-disease-treatment.aspx
36. CHEO Research Institute. CHEO Research Institute’s breakthrough ThinkRare algorithm launches national expansion. https://www.cheoresearch.ca/about-us/media/news/cheo-research-institutes-breakthrough-thinkrare-algorithm-launches-national-expansion/
37. Canadian Rare Disease Network. World-first AI algorithm developed by CHEO researchers leads to rare disease diagnosis for families. https://canadianrdn.ca/world-first-ai-algorithm-developed-by-cheo-researchers-leads-to-rare-disease-diagnosis-for-families/
38. University of Ottawa. World-first AI algorithm at CHEO speeds rare-disease diagnoses for kids. https://www.uottawa.ca/faculty-medicine/research-and-innovation/omari-newsletter-blogs/world-first-ai-algorithm-cheo-speeds-rare-disease-diagnoses-kids