
A new device that makes use of machine learning to diagnose autism spectrum disorders (ASD) in children was approved for being marketed by the US FDA last week. It can help diagnose the condition among kids between the age of 18 months and 5 years.
The device called Canvas Dx does the detection of ASD in three steps. It first asks about the general behaviour patterns of the kids in a questionnaire on a mobile app. Then, it uses videos of the child uploaded by parents to analyse their motor skills. The final step is a review given by the healthcare provider in the form of answers to pre-loaded questions.
In a press release announcing the approval, director of the FDA’s Center for Devices and Radiological Health Jeff Shuren, M.D., J.D. said, “The earlier ASD can be diagnosed, the more quickly intervention strategies and appropriate therapies can begin. Today’s marketing authorization provides a new tool for helping diagnose children with ASD.”
The release also noted that while it is possible to diagnose ASD symptoms as early as 18 months of age, many people get their diagnosis well into adolescence or adulthood. In cases where parents or caregivers suspect developmental delays among young children, Canvas Dx can be helpful in the detection of ASD. It was found to be 98% accurate in diagnosing kids with ASD as positive during an assessment test, and 78.9% accurate in diagnosing those without not as negative. Results shown by the device matched those of an expert panel in 81% of cases.