Potential new imaging method to rapidly differentiate similar neurodegenerative diseases

A new Canadian study¹ reported that investigators are developing a non-invasive, affordable diagnostic imaging tool that, for the first time, can rapidly differentiate retinal deposits in Alzheimer disease and those in amyotrophic lateral sclerosis (ALS) and frontotemporal lobular dementia (FTLD-TDP), according to first author Melanie Campbell, MD. She and other investigators from the University of Waterloo, Waterloo, Ontario, Canada, were joined in this study by researchers from the University of British Columbia, Vancouver.

The researchers published their findings in the journal Alzheimer’s & Dementia.

As they explained in a press release issued by the University of Waterloo, no objective diagnostic test is currently available for ALS or FTLD-TDP, in which the protein TDP-43 forms deposits in the spinal cord and brain, respectively.

In their previous studies, the investigators discovered that the interactions with polarized light differed significantly between retinal amyloid beta deposits in patients with Alzheimer disease and deposits of TDP-43 found in patients with FTLD and ALS. “Our non-invasive retinal imaging could be the first differential diagnostic of these neurodegenerative diseases,” they explained.

Neurodegenerative disease study methodology

In the study under discussion, the researchers studied eyes and brains from two deceased patients with ALS, one of whom also had FTLD, and four individuals with FTLD, including one with type C. TDP-43 was present in the brain in the FTLD cases, and some had age-related tau.

The researchers used polarized light to image the protein deposits in donated retinal samples from the patients with Alzheimer and compared them to samples from patients with FTLD-TDP and ALS.

They then uploaded the data from the light interactions into two artificial intelligence models, Random Forest, an ensemble learning method, and convolutional neural networks, an image-based method, they explained, to see if they could learn to differentiate amyloid beta from TDP-43 deposits, according to the press release.

The investigators reported that the differences in the two types of deposits allowed them to predict the correct disease 86% of the time using random forest and 96% of the time using convolutional neural networks.

The results allowed the researchers to conclude that “machine learning, using the averages and standard deviations of polarized light properties in Random Forest or images showing the distribution of these properties across the deposits (convolutional neural networks), can differentiate retinal deposits associated with Alzheimer disease from those associated with ALS and FTLD, with a relatively high accuracy. This first differential diagnostic of Alzheimer’s disease from TDP-43-related diseases is early, non-invasive, and inexpensive and would reach underserved populations.”

Another hope is that this new technology can facilitate earlier diagnoses of these neurodegenerative disorders with the ultimate goal of slowing disease progression and promoting new targeted treatments.

Reference

1. Campbell MCW, Acheson L, Mason EL, et al. Retinal deposits of TDP-43 and amyloid beta and associated neurodegenerative diseases are accurately classified using measured interactions with polarized light in machine learning algorithms. Alzheimers Dement. 2025 Dec 23;21(Suppl 7): e108465. doi:10.1002/alz70861_108465