New technology may be able to predict a certain type of autism with 100% accuracy using biomarkers in the mothers’ blood plasma, suggests a study published in Molecular Psychiatry. The innovation is based on a machine-learning algorithm that identifies key antibodies thought to be linked with maternal autoantibody-related autism spectrum disorder (MAR ASD), which accounts for around 20% of all autism diagnoses, and could become an important screening tool for autism in the future.
MAR ASD is thought to result from the production of antibodies that are reactive with the developing brain of the unborn fetus, called autoantibodies. These autoantibodies react with specific proteins involved in brain development, and how much these autoantibodies react with the brain likely alters how challenging the resulting condition is. Therefore, screening for each may have predictive value for how likely a child will be to develop MAR ASD.
To test this hypothesis, researchers from UC Davis MIND Institute took plasma samples from 450 mothers with children with MAR ASD and 342 mothers with children without MAR ASD. The goal was to use machine learning to identify patterns of autoantibody reactivity that are strongly associated with ASD, and any that are specific for ASD. The mothers were sourced from the CHARGE study, which aims to understand the genetic and environmental factors involved in autism and has been running since 2003.
The researchers discovered three main patterns associated with MAR ASD, including one that significantly increased the chances of the child being on the autism spectrum, called CRMP1 reactivity.
“For example, if the mother has autoantibodies to CRIMP1 and GDA (the most common pattern), her odds of having a child with autism is 31 times greater than the general population, based on this current dataset. That’s huge,” said Professor Judy Van de Water, lead author of the study, in a statement. “There’s very little out there that is going to give you that type of risk assessment.”
This is the first report of machine learning identifying biomarkers that are associated with MAR-ASD with 100% accuracy.
Recognition of these autoantibody patterns could be integral for screening either before or during pregnancy, offering insight into the development of the child for parent preparation.
“We can envision that a woman could have a blood test for these antibodies prior to getting pregnant. If she had them, she’d know she would be at very high risk of having a child with autism. If not, she has a 43% lower chance of having a child with autism as MAR autism is ruled out,” said Van de Water.
Whilst some of the reactive mechanisms are characterized by the authors, understanding the role of all the predictive autoantibodies could be important for not just screening, but therapeutics too. However, the study remains as an early proof-of-concept and will need extensive trials before the technology sees clinical use.
