Parkinson’s disease is a very common neurodegenerative disease caused by the progressive degeneration of certain nerve cells in the brain, resulting in a deficit of the neurotransmitter dopamine. The balance of dopamine and other neurotransmitters in the brain is essential for the proper functioning of the musculoskeletal system, and if dopamine is lacking, motor disorders occur and patients present slowness of movement, increased muscle tension and tremors.
Early diagnosis of Parkinson’s disease could help to find treatments that can slow or stop its progression by protecting dopamine-producing brain cells. Now, a simple blood test using artificial intelligence (AI) developed by a team of researchers led by scientists from University College London (UCL) and the University Medical Center Goettingen in Germany has been able to predict Parkinson’s disease up to seven years before its first symptoms appear.
“As new therapies become available to treat Parkinson’s, we need to diagnose patients before they develop symptoms. We can’t regenerate our brain cells, so we need to protect the ones we have,” said Professor Kevin Mills (UCL Great Ormond Street Institute of Child Health) and senior author of the study.
People with Parkinson’s are treated with dopamine replacement therapy after they have already developed symptoms such as tremors, slowness of movement, balance and memory problems. “We are acting too late, and we need to start experimental treatments before patients develop symptoms,” Mills said. “So we set out to use cutting-edge technology to find new and better biomarkers for Parkinson’s and turn them into a test that we can use in any large NHS laboratory.” [Servicio Nacional de Salud de Reino Unido]“With sufficient funding, we hope to make this possible within two years.”
Blood proteins that predict the development of Parkinson’s
The results of the research have been published in Nature Communications and show that using a branch of AI called machine learning to analyse a panel of eight blood biomarkers whose levels are altered in Parkinson’s patients could provide a diagnosis with 100% accuracy.
The researchers conducted experiments to see if the test could predict a person’s likelihood of developing Parkinson’s. They analysed blood from 72 patients with REM sleep behaviour disorder, in which muscle atonia is absent, resulting in vigorous twitching when having vivid or violent dreams. It is known that around 75-80% of people with this sleep disorder will develop synucleinopathy (a type of brain disorder caused by abnormal build-up of the protein alpha-synuclein in brain cells), including Parkinson’s.
“We can identify potential Parkinson’s patients several years earlier and drug therapies could be administered at an earlier stage and slow the progression of the disease, or even prevent it.”
When the machine learning tool analysed the blood of these patients, it found that 79% of them had the same profile as someone with Parkinson’s. The study identified 23 proteins in the blood samples as potential biomarkers of Parkinson’s disease using mass spectrometry. With the help of machine learning, eight of these proteins were able to predict Parkinson’s disease up to seven years in advance in 79% of at-risk patients with REM sleep behaviour disorder.
Patients were followed for 10 years and the AI predictions matched the clinical conversion rate, correctly predicting that 16 patients would develop Parkinson’s up to seven years before the onset of any symptoms. The team is continuing to monitor those in whom Parkinson’s development was predicted to further verify the test’s accuracy.
Dr Michael Bartl (University Medical Centre Goettingen), co-senior author who conducted the research with Dr Jenny Hällqvist (UCL Great Ormond Street Institute of Child Health), commented: “By determining eight proteins in the blood, we can identify potential Parkinson’s patients several years earlier. This means that drug therapies could be given at an earlier stage, potentially slowing the progression of the disease or even preventing it.”
“Not only have we developed a test, but we can diagnose the disease based on markers directly related to processes such as inflammation and the degradation of non-functional proteins. These markers represent potential targets for new drug treatments,” he concludes.
