Many of the applications of AI in healthcare have revolved around providing an early diagnosis of a condition, which subsequently allows for a more effective treatment of it. A recent study from McGill University highlights how AI can perform a similar trick for diagnosing dementia.
The researchers believe their algorithms are capable of detecting the signature signs of dementia some two years before its onset. They can do so from analyzing a single amyloid PET scan of the brain. It’s an approach that the team hope will fundamentally change how the condition is managed.
“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,” the team says.
The Hunt for Amyloid
Key to the approach is a protein known as amyloid, which is known to build up in the brains of people with mild cognitive impairment (MCI), which in turn usually leads to dementia. This build-up tends to occur decades before the first signs of dementia occur, however. Whilst this would suggest it’s a useful indicator of the condition, it doesn’t always result in dementia occurring and so is not always that reliable.
The McGill team used data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) to train the algorithm on a few hundred PET scans of MCI patients. It was able to accurately detect 84% of patients before symptoms occurred.
“This is an example how big data and open science brings tangible benefits to patient care,” the authors say.
The team is now working to see if they can spot other biomarkers for dementia that can boost the algorithm still further. Suffice to say, whilst the tool is currently available to the research community, it will need to undergo rigorous testing in order to pass muster with the relevant health authorities.
It’s a nice indication of the progress being made, however, and it will be a fascinating project to track.