Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Blood proteins can predict dementia up to 15 years before clinical diagnosis, scientists using machine learning techniques have found, boosting research into how to prevent the debilitating condition that afflicts more than 55mn people worldwide.
The analysis, the largest of its kind to date, bolsters the findings of smaller studies suggesting certain proteins are “biomarkers” of susceptibility to Alzheimer’s and other neurodegenerative diseases, says the paper by scientists from China’s Fudan university and the UK’s Warwick university.
Effective screening methods for early identification of dementia risk would enable the use of drugs that slow or even reverse its onset, greatly decreasing the costs for health systems.
“We can quite reliably predict dementia 15 years before the diagnosis of the disease,” said Jianfeng Feng, the paper’s lead author and a computer science department professor at the University of Warwick. “We expect that our result will open up an avenue to develop new approaches to slow down the progression of the disease.”
The study, published in Nature Aging on Monday, used blood from more than 52,000 people collected and frozen between 2006 and 2010 by the UK Biobank genetic database. The researchers analysed the samples between April 2021 and February 2022.
More than 1,400 members of the research cohort developed dementia — and showed abnormal levels of some blood proteins. The researchers analysed 1,463 proteins using machine learning and identified 11 that proved to be accurate predictors of future dementia.
The combination of protein analysis and artificial intelligence techniques such as large-language models could provide a precise way of screening middle-aged and older people for dementia risk, Feng said. The results were “relatively ready” to be used in clinical practice by national health systems, he added.
The results come weeks after a study suggesting that a commercially available blood test showed high levels of accuracy in detecting Alzheimer’s, even in its early stages. The test examined concentrations of tau, a protein found in toxic tangles in the brains of Alzheimer’s sufferers.
The emergence of potential dementia diagnostics follows progress towards possible treatments. In October, pharmaceuticals companies Eisai and Eli Lilly unveiled research showing the benefits of using new Alzheimer’s drugs very early in the development of the disease.
The research published in Nature Aging further advanced “fantastic progress in the development of blood tests for Alzheimer’s”, said Dr Sheona Scales, director of research at the Alzheimer’s Research UK charity.
“This new study adds to the growing body of evidence that looking at levels of certain proteins in the blood of healthy people could accurately predict dementia, before symptoms develop.”
Further evaluation of the study’s predictive models would be required, scientists said. The “next steps” should be to “show how these protein markers perform in other cohorts that are less healthy and wealthy than UK Biobank [participants]”, said Charles Marshall, a professor of clinical neurology at Queen Mary University of London, who was not involved in the study.
Another important follow-up would be to explore if predictive accuracy could be further improved by combining the study’s protein marker analysis with other techniques such as blood tests and brain scans, Marshall added.
