A simple blood test could be enough to detect the onset of Parkinson's 7 years before the first symptoms appear. The revolutionary test was developed by a team of researchers from University College London and University Medical Center Goettingen: it uses artificial intelligence to analyze 8 biomarkers and has so far proven to be 100% accurate.
The blood test to discover Parkinson's
As explained by the experts in the study published in the journal “Nature Communications”, the test uses artificial intelligence to early predict the neurodegenerative pathology which currently affects almost 10 million people worldwide and is among the fastest growing neurodegenerative diseases. It is a progressive disorder caused by the death of nerve cells in the part of the brain called the 'substantia nigra', which controls movement. These nerve cells die or deteriorate, losing the ability to produce dopamine, an important chemical, due to the accumulation of the protein alpha-synuclein.
Today, people with Parkinson's are treated with dopamine replacement therapy after they have already developed symptoms such as tremor, slow movement and gait, and memory problems. But researchers believe that prediction and early diagnosis would be valuable in finding treatments that can slow or stop Parkinson's by protecting brain cells targeted by the disease. “As new therapies for Parkinson's become available, we need to diagnose patients with the condition before they develop symptoms,” notes the study's senior author, Kevin Mills, from the UCL Great Ormond Street Institute of Child Health. This is because, he points out, “we can't regrow our brain cells, so we have to protect the ones we have.”
At the moment, however, “we are closing the stable door after the horse has bolted – adds the expert – and we have to start experimental treatments sooner. Therefore we have decided to use cutting-edge technology to find new and better biomarkers for the disease Parkinson's disease and develop them into a test that we can translate into any large National Health Service laboratory, hopefully within two years.” Scientists found that the branch of artificial intelligence called machine learning, by analyzing a panel of 8 biomarkers in the blood whose concentrations are altered in patients with Parkinson's disease, managed to provide a diagnosis with 100% accuracy.
The accuracy of the blood test
To see if the test could predict a person's likelihood of developing the disease, the team analyzed the blood of 72 patients with REM sleep behavior disorder. This disorder causes patients to physically act out their dreams without knowing it (having vivid or violent dreams). And it is now known that approximately 75-80% of people with this disorder will develop synucleinopathy (a type of brain disorder caused by the abnormal accumulation of a protein called alpha-synuclein in brain cells), including Parkinson's. When the machine learning tool analyzed the blood of these patients, it identified that 79% of them had the same profile as a person with Parkinson's.
The patients were followed over the course of 10 years and the AI's predictions have so far matched the clinical conversion rate: the team correctly predicted that 16 patients would develop Parkinson's and was able to do so up to 7 years earlier the onset of any symptoms. The team is now continuing to follow patients who may develop the disease to further verify the accuracy of the test. “By determining 8 proteins in the blood – explains one of the first authors of the study, Michael Bartl, University Medical Center Goettingen – we can identify potential Parkinson's patients several years in advance. This means that drug therapies could be administered in a phase previous, which could slow the progression of the disease or even prevent its onset,” he hypothesizes. “Not only have we developed a test, but we can diagnose the disease based on markers directly linked to processes such as inflammation and the degradation of non-functional proteins. So these markers also represent possible targets for new drug treatments.”
Co-author Kailash Bhatia, Ucl Queen Square Institute of Neurology, and his team are currently examining the accuracy of the test by analyzing samples from people who are part of the population at high risk of developing Parkinson's, for example those who have mutations in particular genes such as Lrrk2 or GBa. The researchers hope to also obtain funding to create a test that can be performed more easily by placing a drop of blood on a card to send to the laboratory. The aim is to understand whether it can predict the disease even before 7 years before the onset of symptoms.