The causes of autism remain mysterious, and LinusBio enters an ongoing and heated debate about what role a tangle of environmental and genetic factors might play. Researchers have uncovered myriad risk factors associated with autism, including infections during pregnancy, air pollution, and maternal stress. Some pollution from metals, which are known to cause neurodevelopmental problems, has also been linked.
“These risk factors all operate against a background of genetic risk,” said Heather Volk, associate professor in the Department of Mental Health at the Johns Hopkins Bloomberg School of Public Health. She added that over the past 15 years, more researchers have turned their attention to environmental factors.
NBC News spoke to six independent experts from different scientific backgrounds about LinusBio’s test. Although many were excited by the potential of the underlying science, most said caution was needed and that more research was needed. Everyone agreed that the results should be replicated by other teams.
“Certainly there is still much to be done before concluding that this test is a valid measure of risk for autism spectrum disorder,” wrote Dr. Scott Myers, a neurodevelopmental pediatrician at the Geisinger Autism & Developmental Medicine Institute, in an email.
How the test works
The LinusBio test analyzes the history of metabolism and tells the story of what substances or toxins the child was exposed to over time, according to Manish Arora, the company’s co-founder and CEO, who is also a professor of environmental medicine and public health at the Icahn School of Medicine at Mount Sinai, the academic arm of the Mount Sinai Health System. The technology was developed from research work completed at Mount Sinai.
For an infant, hair can provide insight into the stresses of developmentally critical moments like the third trimester of pregnancy.
The test runs a laser along a hair and uses its energy to turn it into a plasma for analysis. A centimeter — less than half an inch — of hair captures about a month’s worth of exposure data, Arora said.
Because a tree’s annual rings tell scientists about growing conditions each year, hair growth allows researchers to understand what was happening in a person’s body at specific points in time. According to LinusBio, its test can reveal metal metabolism in 4-6 hour increments.
“It’s almost like having a security camera where you can go back and look at four frames a day,” Baccarelli said.
The technology generates huge amounts of data. This is where a machine learning algorithm takes over – it’s trained to look for patterns of dysregulation in metals, which the researchers believe are biomarkers of autism.
“We can see the clear rhythm of autism with just about an inch of hair,” Arora said.
Time of autism diagnosis
Arora and his team hope their technology could help young children, even newborns, to get treatment for autism earlier than previously possible.
“The problem with autism is that, on average, it is diagnosed at the age of 4. So much Brian has already developed at this point,” he said. “We want to enable early intervention.”
There is still no biological test for autism spectrum disorders. Rather, children are often diagnosed after parents have noticed behavioral differences, such as such as avoiding eye contact, delaying speaking, or not pointing. However, these behaviors are very different, and autism can also coexist with other conditions, such as attention deficit/hyperactivity disorder (ADHD), anxiety, and mood disorders.
Specialists use neurological exams, speech tests, behavioral observations, and other methods to diagnose a child. The American Academy of Pediatrics recommends autism screenings at 18 months and 24 months.
According to Annette Estes, director of the University of Washington Autism Center in Seattle, early intervention for autism typically involves one-on-one instruction with a trusted teacher. These programs are implemented when symptoms arise to address specific development needs, and often look like a game.
“Babies are little scientists. They’re trying things out and looking for feedback,” she said. “You can really accelerate development in all children.”
However, little is known about the impact of pre-symptomatic intervention on young children at higher risk for autism.
“We have theories about what we could do,” Estes said, “but it hasn’t been explored extensively.”
Next steps, more data
The Food and Drug Administration has given LinusBio’s test a Breakthrough designation, intended to expedite the regulatory approval process for new technologies when there are no alternatives on the market. The designation does not change regulatory standards, and the company faces regulatory hurdles before its device can be considered for widespread use in the United States
In the published study, the researchers trained and tested their technology by evaluating hair samples from 486 children in three countries: Japan, Sweden and the United States.
In an analysis of 97 hair samples, the algorithm identified cases where autism spectrum disorder was diagnosed more than 96% of the time. In about 75% of the cases, negative cases were correctly identified. The group tested included 28 cases of autism, a much higher proportion than in the general population.
Manish Arora. Linus Bio
“This needs to be repeated with larger sample sizes and a larger dataset,” Volk said.
The company, which has raised more than $16 million in venture capital investments, is working on an expanded study and has collected samples and data from a group of about 2,000 people.
Because the predictive value of a test depends on the prevalence of a disorder in the group being studied, the accuracy of the test would be lower in a general population, where autism rates are around 2%.
That’s one of the reasons why the LinusBio team sees the tool purely as an aid to clinicians in making a diagnosis.
“No clinician should make a decision about whether a child has autism based on that alone,” Arora said. “This provides crucial information, but not the only information.”
The test may be most useful in groups at higher risk of autism, such as B. Children who have missed developmental markers or have siblings with autism.
Researchers also believe accuracy could be improved with repeated testing – analyzing and comparing multiple strands of a child’s hair.
However, Estes believes that no test or technology can overcome the biggest and most important barrier facing families with children seeking treatment for autism: finding trained clinicians who can make a specialized diagnosis and building a care team for the child . Many parents can’t get help even when they notice developmental delays, she said.
“On-time procedures are something that most children don’t have access to right now,” Estes said. “We know how to help children. It’s really hard to get access.”
In the future, Arora hopes the new technology could also provide clues as to what changes in a child’s body when autism manifests itself. Perhaps this information could eventually open new avenues for developing drugs or therapies for autism, he said.
LinusBio said it also plans to apply the approach to other health conditions with known links to environmental factors, including Lou Gehrig’s disease, or ALS, stomach diseases and certain cancers.