New Blood Test for Predicting Parkinson’s Disease With A.I. Shows Promise, Study Suggests

In preliminary research, scientists identified eight protein anomalies in the blood of patients with Parkinson’s, which they say can help diagnose the disease up to seven years before symptoms appear

A close-up of a young nurse holding the wrinkled hands of an older person.
The new blood test, researchers hope, will help doctors diagnose Parkinson's disease years before symptoms occur, helping them provide more proactive treatments. Visoot Uthairam via Getty Images

Preliminary testing of a newly developed, A.I.-enhanced blood test has shown promise in being able to accurately predict if someone will develop Parkinson’s disease in the coming years.

In research published last week in the journal Nature Communications, a team of scientists in Europe identified eight blood-based biomarkers that might alert doctors to a high potential for Parkinson’s development in a patient—up to seven years before the onset of symptoms.

The test, while still in its early stages, offers hope for a disorder that has no cure, affects nearly nine million people worldwide, kills more than 300,000 people annually and is becoming more prevalent, according to the World Health Organization. The disease is caused by protein buildup in certain dopamine-producing neurons in the brain, which ultimately kills off the cells.

Currently, treatment for Parkinson’s tends to be reactive, focused on controlling symptoms—which include tremors, slow movements, stiffness and a loss of balance—after a diagnosis is made.

“At the moment, we’re shutting the stable door after the horse has bolted,” senior author Kevin Mills, a biochemist at University College London’s Great Ormond Street Institute of Child Health, tells the Guardian’s Ian Sample. “We need to get to people before they develop symptoms. It’s always better to do prevention rather than cure.”

With this goal in mind, the team began their work by collecting blood samples from 99 people who have Parkinson’s disease and 36 people who do not. Analyzing a selection of 70 percent of these samples, a machine learning algorithm identified eight proteins that appeared in different concentrations in the blood of those with the disease.

This pattern “could provide a diagnosis with 100 percent accuracy,” according to a statement from University College London. In a follow-up evaluation, the algorithm was given the remaining 30 percent of blood samples that it hadn’t been trained on—30 from people currently with Parkinson’s disease and 11 from people without it. The tool aced the test, correctly diagnosing every patient.

“This means that drug therapies could potentially be given at an earlier stage, which could possibly slow down disease progression or even prevent it from occurring,” Michael Bartl, a neurologist at University Medical Center Goettingen in Germany and the co-first author of the study, says in the statement.

In another trial, the researchers worked long-term with 54 people who currently have isolated rapid eye movement sleep behavior disorder (iRBD), a neurological disorder that tends to precede Parkinson’s disease, foreshadowing a diagnosis of Parkinson’s or a similar condition between 75 percent and 80 percent of the time. They took one to five blood samples from each patient and tested for the eight biomarkers.

Based on their protein patterns, the tool found that 79 percent of the iRBD patients had blood profiles consistent with someone who would go on to develop Parkinson’s disease. Following up with the patients over a ten-year period, the researchers found 16 of them have been diagnosed with Parkinson’s.

The team correctly predicted these diagnoses, on average, 3.5 years before symptoms presented themselves, with the earliest prediction coming 7.3 years before symptom onset. They’re continuing to check in with the other iRBD patients to confirm the blood test’s accuracy.

“We’ve seen tremendous progress in the development of exciting new tests for Parkinson’s in the last year alone,” Katherine Fletcher, research communications lead at the nonprofit Parkinson’s U.K. who was not involved in the study, tells Live Science’s Michael Schubert. “We are hopeful that these new tests will start being used within the next few years.”

Still, some scientists point out the challenges that remain.

“Parkinson’s is not a single disease but a syndrome and can present in various different ways,” Ray Chaudhuri, the medical director of the Parkinson Foundation International Center of Excellence who was also not involved with the research, tells the Guardian. “As such, management differs and one size does not fit all. The prediction is unlikely to signpost these subgroups at this stage.”

What is crucial about early diagnosis, researchers say, is that it can allow patients to enroll in experimental trials of preventative treatments for Parkinson’s disease as a proactive measure.

“People are diagnosed when neurons are already lost,” Jenny Hällqvist, a biochemist at University College London and a co-author of the study, tells BBC News Philippa Roxby. “We need to protect those neurons, not wait till they are gone.”

Get the latest stories in your inbox every weekday.