New AI tool predicts disease progression and evolution in neurodegenerative disorders

By Sally Robertson, B.Sc.Jan 28 2020

Researchers at the Montreal Neurological Institute, Canada have found that testing blood using an artificial intelligence (AI) algorithm successfully predicted disease progression in neurodegenerative disorders, as well as identifying underlying molecular pathways that were predictive of disease evolution.

Image Credit: Zapp2Photo/Shutterstock.com

The findings have important implications for personalizing the treatment of diseases such as Alzheimer’s and Huntington’s, by helping doctors to choose more targeted and effective treatments.

This test could one day be used by doctors to evaluate patients and prescribe therapies tailored to their needs… It could also be used in clinical trials to categorize patients and better determine how experimental drugs impact their predicted disease progression."

Yasser Iturria-Medina

Overcoming the limitations of previous studies

The most common neurodegenerative diseases take decades to develop, but a lack of data on longitudinal gene expression over time has been a barrier to understanding the underlying molecular mechanisms involved as the conditions evolve.

Previous research has used "snapshot" data that is limited in what it can reveal about these slowly progressing disorders.

The current study aimed to overcome this limitation by obtaining chronological information from a large dataset, shedding light on how changes in gene expression that occur in neurodegenerative patients over time, are related to changes in their condition.

“Here, we overcome this key limitation by introducing a novel gene expression contrastive trajectory inference (GE-cTI) method that reveals enriched temporal patterns in a diseased population,” writes Iturria-Medina and colleagues from the institute’s McConnell Brain Imaging Center.

The researchers used an AI algorithm to analyze blood samples and post-mortem brain tissue samples taken from 1,969 people who had neurogenerative disorders, identifying molecular pathways that were specific to the disease in this population.

More specifically, the team used datasets held for people in the spectrum of late-onset Alzheimer’s and Huntington’s disease that was made available by the Harvard Brain Tissue Resource Center, the Religious Orders Study and the Rush Memory and Aging Project Alzheimer's Disease Neuroimaging Initiative.

The algorithm strongly predicted disease stage and progression to advanced disease

As recently reported in the journal Brain, the algorithm strongly predicted the various stages and neuropathological severity of the disease. When applied to blood samples taken at baseline, it also strongly predicted clinical deterioration and progression to advanced stages of the disease, suggesting its potential as a minimally invasive technique for early screening.

Furthermore, the AI tool identified genes and molecular pathways in both blood and brain tissue that were strong predictors of disease evolution. Between 85 and 90% of the most highly predictive molecular pathways identified in the blood were the same as those identified in the brain, suggesting that the underlying molecular changes are similar between the brain and the peripheral body.

“These pathways support the importance of studying the peripheral-brain axis,” writes the team.

How could the test be used in the future?

Iturria-Medina says the test could one day be used in the clinic to assess patients and prescribe treatments that are tailored to their needs.

It could also be used in clinical trials to categorize patients and better determine how experimental drugs impact their predicted disease progression,"

Iturria-Medina says the team now plans to test the approach in other neurodegenerative diseases, including Parkinson's disease and amyotrophic lateral sclerosis.

“The GE-cTI is a promising tool for revealing complex neuropathological mechanisms, with direct implications for implementing personalized dynamic treatments in neurology,” concludes the team.