A.I., Big Data Project Predicts Dementia 2 Years Before Symptoms Onset, Researchers Show

Charles Moore avatar

by Charles Moore |

Share this article:

Share article via email
predicting dementia before symptoms

New artificial intelligence (AI) research conducted at McGill University in Canada suggests that doctors may soon have the tools to predict an individual’s likelihood of developing dementia several years before the onset of symptoms. Such extended prognostic capability would give future dementia patients and their families more time to plan strategies for living with the disorder and to arrange for treatment and care.

Scientists at McGill’s flagship mental health research center, the Douglas Mental Health University Institutes Translational Neuroimaging Laboratory, used artificial intelligence techniques and big data to develop an algorithm that can, with a single amyloid PET scan, detect dementia signatures in the brains of patients at risk of developing Alzheimer’s disease two years before symptoms.

Their findings were recently published in the journal Neurobiology of Aging, in a study titled “Identifying incipient dementia individuals using machine learning and amyloid imaging.”

The coauthors note in the study abstract that identifying individuals who will develop Alzheimer’s dementia within time frames acceptable for clinical trials is an important challenge in designing studies for testing new disease-modifying therapies. And, while amyloid protein is Alzheimer’s disease’s core feature, neuronal degeneration biomarkers are the only ones believed to provide satisfactory reliable prediction of clinical progression within short time frames.

In the study, the researchers propose a machine learning-based predictive method designed to assess the progression to dementia within a 24-month period, based on regional information obtained from a single amyloid positron emission tomography (PET) scan. The investigators’ novel algorithm demonstrated an accuracy rate of 84 percent, outperforming the existing algorithms using the same biomarker measures.

“With its high accuracy, this algorithm has immediate applications for population enrichment in clinical trials designed to test disease-modifying therapies aiming to mitigate the progression to Alzheimer’s disease dementia,” the authors wrote.

Dr. Pedro Rosa-Neto, study co-lead author and associate professor in McGill’s departments of neurology, neurosurgery, and psychiatry, says in a press release that he expects this technology will change the way physicians manage patients and greatly accelerate Alzheimer’s treatment research.

“By using this tool, clinical trials could focus only on individuals with a higher likelihood of progressing to dementia within the time frame of the study. This will greatly reduce the cost and the time necessary to conduct these studies,” said Dr. Serge Gauthier, co-lead author and McGill professor of neurology, neurosurgery, and psychiatry .

According to the McGill release, scientists have long been aware that amyloid protein accumulates in the brains of individuals with mild cognitive impairment (MCI), often a precursor of dementia. But because amyloid begins accumulating decades before dementia symptoms manifest, its presence couldn’t be used as a reliable predictive biomarker because not all MCI patients go on to develop Alzheimer’s disease.

Sulantha Mathotaarachchi, a computer scientist with Rosa-Neto’s and Gauthier’s research team, used hundreds of amyloid PET scans of MCI patients from the Alzheimers Disease Neuroimaging Initiative (ADNI) database to train the algorithm developed by the research team to identify before symptoms onset which patients would develop dementia with the above-noted 84 percent accuracy.

The ADNI, a public/private global research effort, has since 2004 been validating the use of biomarkers, including blood tests, tests of cerebrospinal fluid, and MRI/PET imaging, for Alzheimer’s disease clinical trials and diagnosis in which participating patients agree to complete a variety of imaging and clinical assessments.

Research at McGill to identify other dementia biomarkers that could be incorporated into the algorithm to improve its prediction capabilities is ongoing.

“This is an example how big data and open science brings tangible benefits to patient care,” said Rosa-Neto, who is also director of the McGill University Research Centre for Studies in Aging.

While the McGill team’s new software has been made available online to scientists and students, physicians will not be able to use it in clinical practice before it receives approval by health authorities. The team is currently conducting additional research to validate the algorithm in different dementia patient groups, particularly those who suffer from concurrent age-associated conditions such as small strokes.

The research was funded by the Canadian Consortium on Neurodegeneration in Aging (CCNA) and the Canadian Institutes of Health Research (CIHR), Canada’s federal funding agency for health research.