AI helps to predict arthritis in children
Some children are unfortunate to have arthritis, which occurs when the immune system mistakes the body’s own cells for foreign invaders, typically attacking the lining of the joints to cause swelling, pain and long-lasting damage. There is no known cure, and treatment is costly and increasingly aggressive – potential side effects include risks of infection.
Recently, a machine learning tool to identify arthritic patients has been developed by Quaid Morris and Dr Rae Yeung of the University of Toronto, who are professors of computer science at the Donnelly Centre for Cellular and Biomolecular Research and of Paediatrics, Immunology and Medical Science respectively.
The researchers describe an approach using a form of artificial intelligence (AI), in which the computer learned to recognise recurrent patterns of joint pain. Clinical data from 640 Canadian children were classified into seven distinct categories of joint activity, such as in the fingers and knees. The AI also predicted disease course and severity in those children. While the majority fell into a single category, about one third had painful (or active) joints that belonged to multiple categories. These children with non-localised joint involvement generally had worse outcomes and took longer to go into remission.
The current patient classification for childhood arthritis only accounts for the overall number of affected joints. However, it was unusual that the data pointed out the minority of children with different outcomes. Identifying this particular group of children early will prevent unnecessary pain and disability from the disease.
The AI system still needs further development for a successful study; Morris said that greater understanding allows for better grouping of children in response to diagnosis and treatment of arthritis.
Category: Features, Technology & Devices
