Blood test for Alzheimer’s shows 100 percent accuracy in early trials

[Kyla]

http://www.gizmag.com/blood-test-alzheimers-100-percent/43767/


excerpt:

Blood test for Alzheimer’s shows 100 percent accuracy in early trials
Nick Lavars June 9, 2016


A reliable blood test for Alzheimer’s may one day become a vital tool for early diagnosis (Credit:National Eye Institute/Creative Commons)

Alzheimer's can be quite a stealthy foe, at times causing damage in a sufferer long before any symptoms start to show. Naturally, this makes it tricky to detect early on, which is problematic because treatment options only narrow as the disease progresses. But researchers may have now uncovered what could become a hugely valuable diagnostics tool, developing a blood test capable of picking up early stage Alzheimer's with "unparalleled accuracy." ...

 

[Bob]

I'd like to know if this test, or similar, has been used to test ME patients for auto-antibodies...

2.2 Human protein microarrays
To detect and identify autoantibodies in human sera, we used Invitrogen's ProtoArray v5.0 Human Protein Microarrays (Cat. No. PAH0525020, Invitrogen, Carlsbad, CA, USA), each containing 9486 unique human protein antigens (www.invitrogen.com/protoarray).

http://www.sciencedirect.com/science/article/pii/S235287291630015X

 

[Simon]

Rough and grumpy today, so sounding off accordingly:

The blood samples were screened with human protein microarrays, each with 9,486 unique proteins, as a way of drawing out autoantibodies that were indicative of the different diseases...

The researchers identified the top 50 autoantibody biomarkers that were capable of detecting early-stage Alzheimer's. Across a number of tests, this set of biomarkers could distinguish between MCI sufferers with and without Alzheimer's with 100 percent precision. The test could also tell the difference between the early Alzheimer's and more advanced with 98.7 percent accuracy, early Parkinson's (98 percent) and both multiple sclerosis and breast cancer (100 percent).

With 9,486 different proteins, that's very, very likely to be a false positive. No sign of appropiate statistical corrections in the paper's abstract Detection of Alzheimer's disease at mild cognitive impairment and disease progression using autoantibodies as blood-based biomarkers
(though see this quote on the approach they used, from wiki "Random decision forests correct for decision trees' habit of overfitting to their training set.[3]:587–588" edit, correction: but still a small sample and 9,000+ variables to pick from...).

"Our results show that it is possible to use a small number of blood-borne autoantibodies to accurately diagnose early-stage Alzheimer's," says DeMarshall.

No, they don't.

The researchers are now setting their sights on reproducing these promising results in a larger study.

IF it gets past that stage, it might be interesting.

 

[panckage]

@Simon does this work have a hypothesis?