• Posted July 15, 2024

AI Better at Predicting Progression to Alzheimer's Than Standard Care

An AI program has proven better than doctors at sifting through the telltale signs that indicate who with early dementia will progress to Alzheimer’s disease, a new study says.

AI predicted in 4 cases out of 5 when early dementia would either remain stable or worsen into Alzheimer’s, according to a report in the journal eClinical Medicine.

Overall, AI was about three times more accurate at predicting the progression to Alzheimer’s than doctors armed with brain scans, cognitive testing and analysis of spinal fluid for telltale proteins like tau and amyloid, results show.

“We’ve created a tool which, despite using only data from cognitive tests and MRI scans, is much more sensitive than current approaches at predicting whether someone will progress from mild symptoms to Alzheimer’s – and if so, whether this progress will be fast or slow,” senior researcher Zoe Kourtzi, a professor of experimental psychology and computational cognitive neuroscience at the University of Cambridge in the U.K., said in a news release.

For the study, researchers built an AI model using brain scans and cognitive tests from 400 people participating in a U.S. research project.

The research team then tested the model by feeding it real-world cognitive test results and brain scans from another 600 U.S. patients from the same project, as well as 900 people who sought treatment at memory clinics in the U.K. and Singapore.

The AI was able to point out who would develop Alzheimer’s within three years 82% of the time, and who would have mild but stable dementia in 81% of cases, results show.

The AI also allowed researchers to put patients into groups based on their risk, results show. About 50% of participants had dementia that would remain stable, while 35% would progress to Alzheimer’s slowly and 15% more rapidly.

As a result, the AI could help people at most risk for Alzheimer’s receive early treatment and close monitoring, researchers said.

At the same time, the 50% of people with symptoms like memory loss who are otherwise stable could be examined for problems other than dementia that might be contributing to those symptoms, researchers said.

For example, they might be suffering from a mood disorder like anxiety or depression rather than dementia, researchers said.

These results show that the AI program could be implemented in real-world settings to help steer patients to the best care, researchers argue.

“Memory problems are common as we get older. In clinic I see how uncertainty about whether these might be the first signs of dementia can cause a lot of worry for people and their families, as well as being frustrating for doctors who would much prefer to give definitive answers,” researcher Dr. Ben Underwood, an assistant professor of psychiatry at the University of Cambridge, said in a news release.

“The fact that we might be able to reduce this uncertainty with information we already have is exciting and is likely to become even more important as new treatments emerge,” Underwood added.

The team next plans to extend the AI model to other forms of dementia. They also would like the AI to consider additional types of data when assessing dementia, such as blood test results.

“Our vision is to scale up our AI tool to help clinicians assign the right person at the right time to the right diagnostic and treatment pathway,” Kourtzi said. “Our tool can help match the right patients to clinical trials, accelerating new drug discovery for disease modifying treatments.”

More information

The Alzheimer’s Association has more on the difference between dementia and Alzheimer’s disease.

SOURCE: University of Cambridge, news release, July 12, 2024