NEW YORK, Nov. 21 (UPI) -- A novel artificial intelligence system could diagnose autism much earlier in children, according to research to be presented this week at the annual meeting of the Radiological Society of North America in Chicago.
The newly developed system that analyzes specialized MRIs of the brain accurately diagnosed children between the ages of 24 and 48 months with autism at a 98.5% accuracy rate, researchers said.
A multi-disciplinary team at the University of Louisville developed the three-stage system to analyze and classify diffusion tensor MRI, or DT-MRI, of the brain. DT-MRI is a special technique that detects how water travels along white matter tracts in the brain, according to a news release.
"The current tools for diagnosing autism are subjective, particularly when evaluating individuals who fall near the borderline between autism and typical development," study co-author Ayman El-Baz, professor and chair of the bioengineering department at the University of Louisville, told UPI via email.
"Consequently, there is an urgent need to develop a new, objective technology for the early diagnosis of autism."
The AI system involves isolating brain tissue images from the DT-MRI scans and extracting imaging markers that indicate the level of connectivity between brain regions.
A machine-learning algorithm compares the marker patterns in the brains of children with autism to those with normally developed brains.
"Autism is primarily a disease of improper connections within the brain," co-author Dr. Gregory N. Barnes, professor of neurology at the University of Louisville and director of the Norton Children's Autism Center in Louisville, said in the release.
"DT-MRI captures these abnormal connections that lead to the symptoms that children with autism often have, such as impaired social communication and repetitive behaviors."
The researchers applied their methodology to the DT-MRI brain scans of 226 children between the ages of 24 and 48 months from the Autism Brain Imaging Data Exchange-II. The dataset included scans of 126 children affected by autism and 100 normally developing children.
Therapeutic intervention before the age of three can lead to better outcomes, including the potential for individuals with autism to achieve greater independence and higher IQs, the researchers noted.
According to the CDC's 2023 Community Report on Autism, fewer than half of children with autism spectrum disorder received a developmental evaluation by age 3, and 30% of children who met the criteria for autism spectrum disorder did not receive a formal diagnosis by age 8.
"Early intensive behavioral intervention within the age range of one to three years can be notably advantageous due to the phenomenon known as neuroplasticity in the infant brain," El-Baz said.
An autism assessment would begin with the researchers' AI system, followed by an abbreviated session with a psychologist to confirm results and guide parents on next steps. It could reduce psychologists' workload by up to 30%, they said.
The investigators want to commercialize and obtain clearance from the Food and Drug Administration for their AI software.
"I am all for any diagnostic technology that can help us diagnose autism earlier and reliably in order to help children access evidence-based intervention sooner," Dr. Leandra Berry, director of the autism program at Texas Children's Hospital in Houston, told UPI in a telephone interview.
However, she added, "Many studies will need to replicate this finding before we really embrace this technology."
With the ages of patients in the study being between 24 and 48 months, she noted that the study can't address whether the technology will be useful in younger children.
Berry also noted that in research studies, MRIs and other scans are typically paid for by grants, but insurers may not cover expensive imaging tests, adding that access to this technology, particularly in rural areas, is limited.
While this technology can provide a diagnosis, an expert will need to convey that information to a family. "There is still a human clinician component that will be critical," Berry said.
Early detection essential
Diana Robins, a professor and director of the A.J. Drexel Autism Institute at Drexel University in Philadelphia, told UPI in a telephone interview "early detection is essential to reduce disability associated with autism and to improve positive outcomes for autistic individuals."
However, "before you draw conclusions about how well a tool identifies autism, you need to include children who have other developmental delays in your sample," said Robins, who has a doctorate in psychology.
It is "important to tease apart" commonly co-occurring diagnoses, Dr. Susan Hyman, a professor in the division of developmental and behavioral pediatrics at the University of Rochester's Golisano Children's Hospital in Rochester, N.Y., told UPI via email.
She added that almost half of autistic children also have attention-deficit/hyperactivity disorder, 40% of autistic individuals with autism also have intellectual disabilities and 25% of autistic individuals have seizures with onset typically in early childhood or adolescence.
"MRIs ordered in the community for autistic children who do not have seizures or neurologic findings rarely provide information that will inform care on an individual level," Hyman said.
It is also difficult for young children to remain still for an MRI.
"You either need to try multiple times, try to scan them at bedtime when they are sleepy, or you need to sedate them, which comes with medical risk," Robins said, adding that parents may not give consent.
Robins noted no data is available to indicate how much this technology will reduce psychologists' workload. "This [technology] is really far from being ready for the general public," she said.