Why this research matters
Some children have mild forms of arthritis, while others have forms that lead to devastating impacts that last into adulthood. It can be hard to predict which kids should get which treatments at what times, leading to unnecessary side effects. This Arthritis Society-supported study used an artificial intelligence algorithm to help predict how childhood arthritis will behave based on which joints are inflamed. This will shape clinical practice and inform tailored treatment decisions to keep kids healthy.
What is this research about?
Children and infants can get arthritis. Some kids can have a milder disease or may grow out of it, while others have serious forms of the disease and need aggressive treatment. Since it can be hard to predict how arthritis will behave, children might get more treatment than they really need, along with some negative side effects. Doctors have recognized that kids with different patterns of arthritic joints can have different disease severity, but these patterns aren’t usually factored into how the disease is classified or treated.
What did the researchers do?
Dr. Simon Eng worked with Dr. Rae Yeung and Dr. Quaid Morris and their team to develop an algorithm based on artificial intelligence (AI) to predict how childhood arthritis behaves in patients. They used AI to discover patterns in the painful and swollen joints in 640 children with arthritis from 16 hospitals across Canada and assessed how the patterns were connected to disease remission.
What did they find?
The researchers found seven distinct patterns of arthritic joints that presented a new way to classify childhood arthritis into subtypes. They were able to identify children at high risk for poor outcomes who may require aggressive treatment earlier.
How can this research be used?
This research highlights how information doctors routinely collect on which joints are painful and swollen can be used in a new way to predict which children with arthritis will have a more severe form of the disease. This is expected to change how treatment decisions are made, in a big step towards precision medicine – giving the right treatment to the right patient at the right time.
Given the clinical importance of the patterns discovered, they may be integrated into new disease classification and treatment guidelines for childhood arthritis. The successful AI approach could also be used in other forms of arthritis to improve diagnosis and treatment.
When I was diagnosed with JIA at 18 months of age, my parents were given few answers and many unknowns. As a child growing up in the 80’s with JIA, the treatments available to me merely responded to the symptoms that arose. Personalized medicine would not only have offered me more effective treatment but would have helped provide a path towards solutions amidst so much uncertainty. "
— Yvonne Wallace, Arthritis Society Online Consumer Panel Member
What impact could this have?
One child facing the pain of arthritis is one too many. While current medications can help kids with arthritis control their symptoms, many drugs have undesirable side effects like nausea or increased risk of infections. By giving doctors a way to identify kids predicted to go into remission quickly, this research can help kids avoid unnecessary treatment and live their lives to the fullest. It also gives doctors the tools they need to identify kids who should receive aggressive treatment early on to ensure the best possible outcome.
Guiding treatment planning using research evidence like this can give parents the confidence that their child is getting the best treatment for their condition, without compromising their wellness with unnecessary drug side effects.
About the researcher
Dr. Simon Eng completed his PhD under the supervision of Dr. Rae Yeung and Dr. Quaid Morris at The Hospital for Sick Children and the University of Toronto, with the support of an Arthritis Society PhD Salary Award. He continues to work with the Yeung lab as a bioinformatician. This research is being built upon in the large-scale UCAN CURE project led by Dr. Yeung and researchers across Canada, made possible in part by the Arthritis Society’s Stop Childhood Arthritis initiative.
Eng SWM, Aeschlimann FA, van Veenendaal M, Berard RA, Rosenberg AM, Morris Q, Yeung RSM; ReACCh-Out Research Consortium. Patterns of joint involvement in juvenile idiopathic arthritis and prediction of disease course: A prospective study with multilayer non-negative matrix factorization. PLoS Med. 2019;16(2):e1002750.
Research at the Arthritis Society
Through the trust and support of our donors and partners, the Arthritis Society is Canada’s largest charitable funder of cutting-edge arthritis research, investing over $220 million in research projects since our founding. These projects have led to breakthroughs in the diagnosis, treatment and care of people with arthritis. Visit us at arthritis.ca/research.