Algorithm accurately predicts dementia risk using a single MRI scan
Researchers from Duke have developed an algorithm that can predict a person’s age based on a single MRI scan of their brain.
The tool can also be used to predict the risk of a myriad of chronic conditions and cognitive impairment as well. Experts involved in its development are hopeful that it can help providers identify patients’ risks of chronic diseases sooner, enabling them to manage their health more effectively.
“The way we age as we get older is quite distinct from how many times we’ve traveled around the sun,” said Ahmad Hariri, PhD, professor of psychology and neuroscience at Duke University. “What‘s really cool about this is that we‘ve captured how fast people are aging using data collected in midlife. And it’s helping us predict diagnosis of dementia among people who are much older.”
This algorithm differs from other similar algorithms designed to predict age and cognitive decline because it was trained on longitudinal data from the same group of individuals over the entirety of their life thus far. The data were derived from the Dunedin study, which collects extensive medical metrics—blood pressure, body mass index, glucose and cholesterol levels, lung and kidney function, imaging exams and more—on participants every few years.
Researchers combined all of this data (more than 20 years’ worth) to calculate a general score for how fast each individual was aging. They then incorporated that information alongside the MRI brain scans of 860 of the Dunedin participants when they were 45 years old in order to estimate the accuracy of the aging scores.
The team tested the algorithm, dubbed DunedinPACNI (Dunedin Pace of Aging Calculated from NeuroImaging), on the Alzheimer’s Disease Neuroimaging Initiative, UK Biobank and BrainLat datasets to determine the accuracy of its predictions. This revealed that faster DenedinPACNI scores, which are indicative of accelerated aging, ably predicted cognitive impairment, accelerated brain atrophy on MRI and eventual dementia diagnoses. It also was able to accurately predict frailty, poor health, future chronic diseases and mortality in older adults.
During an analysis specific to Alzheimer’s disease, researchers determined that individuals with the fastest aging scores when they joined the study were 60% more likely to develop the neurodegenerative condition in the years that followed.
“It seems to be capturing something that is reflected in all brains,” Hariri said. “We really think of it as hopefully being a key new tool in forecasting and predicting risk for diseases, especially Alzheimer's and related dementias, and also perhaps gaining a better foothold on progression of disease.”
Learn more about the algorithm here.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.