Contemplative older man, maybe with Alzheimer's, at home
(Credit: FG Trade / Getty Images)

A new online tool could offer a detailed prediction of how people with early-stage dementia or mild cognitive impairment are expected to develop worsening conditions, including Alzheimer’s. 

The tool is part of the continued push to develop better ways to predict the path of cognitive impairment, to create better treatment plans or, thanks to new drugs, forestall further decline.

The creators of the model, named DETree, suggested that it can give a more detailed picture than previous prediction models of where someone falls within a timeline of dementia progression. 

More nuanced data analysis, for Alzheimer’s and other diseases, now is possible thanks to the rapidly evolving artificial intelligence analytics at researchers’ disposal.

“The continuous nature of Alzheimer’s development and transition states between successive Alzheimer’s related stages have been typically overlooked,” the study authors wrote. “How to effectively predict the individual patient’s status within a wide spectrum of continuous Alzheimer’s progression has been largely understudied.”

The DETree model came out of data collected from the Alzheimer’s Disease Neuroimaging Initiative, a multi-center research project that pools a wide variety of patient disease data, including PET and MRI scans, cognitive tests and biomarkers.

Based on the data sets they used, the DETree was able to correctly identify disease progress in 78% of cases, the researchers said.

A similar software tool, BrainSee, is designed to predict whether mild cognitive impairment will progress into full-blown dementia such as Alzheimer’s. BrainSee received market clearance from the US Food and Drug Administration last month.

The DETree researchers, who work at the University of Texas at Arlington, said that they hope to create similar models for other diseases common to older adults, including Parkinson’s, Huntington’s and Creutzfeldt-Jakob disease.

The findings on DETree were published last month in the open-access journal Pharmacological Research.