Simple test predicts Alzheimer's risk, 20 years before symptoms show
Australian National University (ANU) physicists have combined nanotechnology, artificial intelligence and molecular biology to design a novel method that looks for Alzheimer’s disease protein markers in blood. These markers are tell-tale signs of early neurodegeneration, and early detection is so far the best defense we have in order to effectively intervene in Alzheimer’s progression. While there’s no cure for the disease, a 20-year jump on symptoms first appearing has the potential to significantly change health outcomes.
While a large body of research is focused on developing targeted therapeutics for treating advanced Alzheimer’s, there has also been a lot of progress made in the field of advanced diagnostics.
“Currently, Alzheimer’s is mostly diagnosed based on evidence of mental deterioration, by which stage the disease has already seriously damaged the brain,” said co-author Professor Patrick Kluth, from the ANU Research School of Physics. “Early detection, which is vital for effective treatment, normally involves invasive and expensive hospital procedures such as a lumbar puncture, which can be physically and mentally taxing for patients.”
The researchers developed an ultra-thin silicon chip covered in tiny, nanometer-sized holes, or solid-state nanopores. A small amount of blood is then placed on the chip, and through a process of translocation through the nanopores, complex mixes of proteins in the blood could be isolated. Then, the chip was inserted into a phone-sized device, where an artificial intelligence algorithm searched for protein signatures that matched with those associated with early onset Alzheimer’s disease.
And by classifying protein signals into clusters based on signal attributes, the researchers found the model returned both a significant and high degree of accuracy (specificity of 96.4%) in identifying combinations of the four machine-learned proteins. Since proteins contain distinct and individualized genetic blueprints, they can play a much larger role in diagnostic medicine, given the right technology.
“If that person can find out their risk level that far in advance, then it gives them plenty of time to start making positive lifestyle changes and adopt medication strategies that may help slow down the progression of the disease,” said co-author Shankar Dutt, a researcher at ANU.
While Alzheimer’s focused, the researchers point out the algorithm used can be trained up to look for other diseases – and to test for them at the same time. Those include Parkinson’s disease, multiple sclerosis (MS) and amyotrophic lateral sclerosis (ALS).
“Blood is a complex fluid that contains more than 10,000 different biomolecules,” Dutt said. “By employing advanced filtration techniques and harnessing our nanopore platform, combined with our intelligent machine learning algorithms, we may be able to identify even the most elusive proteins.”
The team hopes this screening technique will be available within the next five years, allowing patients to get results “in near real time,” added Kluth.
“The quick and simple test could be done by GPs and other clinicians, which would eliminate the need for a hospital visit and prove especially convenient for people living in regional and remote areas,” he said.
The research is published in the journal Small Methods.
Source: Australian National University
