Written by Ana Sandoiu
Autism spectrum disorder affects a huge number of children both globally and in the United States. Experts have long acknowledged the importance of detecting autism early, but current diagnosis tools are purely behavioral and not entirely accurate. New research, however, proposes a biological method for accurately predicting whether a child will go on to develop autism.
Worldwide, autism spectrum disorder (ASD) is estimated to affect 1.5 percent of all children, and 1 in 68 U.S. children were diagnosed with ASD in 2014.
The number of ASD diagnoses has drastically increased over the past few decades, and in the U.S., the estimates show a 30 percent increase in the number of children with ASD compared with previous years.
The Centers for Disease Control and Prevention (CDC) highlight the importance of early ASD detection. However, most of the current diagnosis practices and psychometric tools rely purely on the detection of behavioral signs.
Research from the Rensselaer Polytechnic Institute in New York identifies a new method for predicting whether a child is on the ASD spectrum based on substances that are detectable in the blood.
The study, led by Juergen Hahn and Daniel Howsmon, was published in the journal PLOS Computational Biology.
The team analyzed data from the blood samples of 83 children with autism and 76 neurotypical children – that is, they were not affected by ASD. The data was initially collected as part of the IMAGE study carried out by the Arkansas Children’s Hospital Research Institute.
The children were aged between 3 and 10. The scientists set out to measure metabolite concentrations resulting from two metabolic processes: the folate-dependent one-carbon (FOCM) metabolism and transsulfuration (TS) pathways.
Both of these substances have previously been shown to become altered in people with an increased risk of ASD.
New tool predicts almost 98 percent of children with ASD
The researchers also developed multivariate statistical models that accurately classified children with autism based on their neurological status.
The authors note that their models “have much stronger predictability than any existing approaches from the scientific literature.”
Using these tools, Hahn and team correctly identified 97.6 percent of the children that had autism, and 96.1 percent of those who were neurotypical.
“This level of accuracy for classification as well as severity prediction,” the authors conclude, “far exceeds any other approach in this field and is a strong indicator that the metabolites under consideration are strongly correlated with an ASD diagnosis.”
“The method presented in this work is the only one of its kind that can classify an individual as being on the autism spectrum or as being neurotypical. We are not aware of any other method, using any type of biomarker that can do this, much less with the degree of accuracy that we see in our work.”
However, Hahn also concedes that more research is needed to confirm the results. In the future, the researchers aim to investigate the possibility of developing FOCM and TS-based treatments that could alleviate ASD symptoms.