This was, to me, a novel approach. NPR had better coverage than this news article; SurroMed was able to measure thousands of paremeters and compare them to a control group. A fairly large number seemed statistically different. The theory is this can work like digital 'fingerprinting' or 'dna matching'; without regard to mechanism of action it may be possible to define an "autism profile" of lab results. It's a very statistical approach to screening -- similar to factor or regression analysis; the correlated findings may be predictive but not "causal". Fascinating. I'd read predictions of this particular approach, but this is the first study where I've seen it at work.
The next step will be to try to identify a practical 'profile subset' and then test it in children to see if it's predictive, using blood samples from a group other than the group used in this study. That work has probably been done and may be published shortly.
Amaral along with pediatric neuropsychologist Blythe Corbett and other M.I.N.D. Institute colleagues took blood samples from 70 children with autism who were between 4 and 6 years old and from 35 children of the same age who didn't have the disorder. The samples were then analyzed by a biotech company, SurroMed, LLC, Menlo Park, Calif., which has developed technology that can identify differences in the number and types of immune cells, proteins, peptides and metabolites in small amounts of blood.
The study has generated an enormous amount of data and M.I.N.D. Institute researchers say it will take months before all of the information has been fully evaluated. But initial findings clearly demonstrate differences in the immune system, as well as proteins and other metabolites in children with autism:
-- The antibody producing B cells are increased by 20 percent in the autism group
-- Natural killer cells are increased by 40 percent
-- More than 100 proteins demonstrated significant differential expression between the autism and typically developing groups
-- Other small molecules (metabolites) also show many differences
No comments:
Post a Comment