Is the Eye the Window to Alzheimer’s?

The company’s AI model analyzes eye scans for anomalies like the buildup of certain proteins or blood vessels with a twisted shape, that are associated with Alzheimer’s
The company’s AI model analyzes eye scans for anomalies like the buildup of certain proteins or blood vessels with a twisted shape, that are associated with Alzheimer’s

Summary

New AI tools could diagnose the disease with visual scans

Getting tested for Alzheimer’s disease could one day be as easy as checking your eyesight.

RetiSpec has developed an artificial-intelligence algorithm that it says can analyze results from an eye scanner and detect signs of Alzheimer’s 20 years before symptoms develop. The tool is part of broader work by startups and researchers to harness AI to unlock the mysteries of a disease that afflicts more than seven million Americans.

For years, people have studied individual hallmarks of Alzheimer’s, including brain inflammation and neurodegeneration, but the exact causes of the disease remain elusive. AI, researchers say, could open a new era in the diagnosis of a neurological disease that remains difficult to identify, let alone treat.

“There still remains a huge amount we fundamentally don’t understand about the brain and how it works," said Eliav Shaked, co-founder of Toronto-based RetiSpec. “The power in AI is that it can help connect the dots."

Another company, Sacramento, Calif.-based Neurovision, aims to use machine learning to develop retinal scans and blood tests to identify people at risk of developing Alzheimer’s and other forms of dementia. The company’s AI model analyzes eye scans for anomalies like the buildup of certain proteins or blood vessels with a twisted shape, that are associated with Alzheimer’s, said Steven Verdooner, Neurovision’s co-founder.

It can be difficult for people to discern such signs in the scans. Many scans have dark areas, and plaque deposits can be very small. The human eye can’t distinguish them very well, Verdooner said.

“The algorithm does a better job," he said.

At the University of Arizona College of Medicine in Tucson, neurology associate professor Rui Chang built an AI model that aims to identify genetic triggers that are linked to Alzheimer’s. The traditional approach researchers follow is painstakingly slow, said Chang.

“It is like looking at the forest one tree at a time," he said.

AI can absorb the whole forest of information at once and find patterns people can’t. The model took two months to identify 6,000 gene targets that, if knocked down or repressed, might change how Alzheimer’s develops. Chang said the tool has cut a decade off his research.

Chang founded a company, Path-Biotech, which will start clinical trials next year based on his AI research.

Alzheimer’s was the sixth-leading cause of death in the U.S. in 2021 not including Covid-19.

The Food and Drug Administration in July approved a drug, Leqembi, that removes amyloid, a sticky plaque that gathers in the brains of Alzheimer’s patients. But current techniques for identifying the disease are expensive and difficult.

People with symptoms can get a spinal tap or a PET scan to see if they have high levels of amyloid and tangled strands of the protein tau, which is also commonly found in Alzheimer’s patients. The scans are very accurate, which makes them the gold standard of diagnosis, said Catherine Bornbaum, RetiSpec’s chief business officer. Compared with autopsy results—still the only way to tell for certain whether a patient had Alzheimer’s when they died—the PET scan’s diagnostic success rate is close to 90%.

But the machines aren’t widely available, the scans are expensive, and getting a diagnosis can take weeks. Insurers don’t routinely cover the scans and they can cost around $6,000.

Artificial intelligence technologies could speed up the process of getting a diagnosis and make it cheaper. RetiSpec’s AI reads scans from a camera that can be attached to machines already available in most optometrists’ offices, for example. The camera measures a wider range of the spectrum than the human eye can see, which allows the AI to detect unique optical signatures that correspond with the presence of amyloid in the brain. The model, which delivers results instantaneously, was 80% accurate in detecting such signatures in a recent study of 271 patients.

AI tools in medical research can perform well in clinical testing but break down in messier real-life situations, said Matt Leming, a research fellow at Massachusetts General Hospital.

“Biotech AI models are finicky," he said.

AI learns better from huge amounts of data, Leming said. AI models like ChatGPT are good at analyzing and mimicking writing, for example, because they learn from text gathered across the internet.

Medical data is comparatively scarce and proprietary. That means AI in biotech has a more limited sample to learn from and its results can be easily thrown off by wider variation in cases it encounters in a clinic compared with more controlled laboratory settings.

“When it comes to AI fundamentally changing the way we do medicine, I don’t think it is going to happen," said Leming.

Chang at the University of Arizona said he has tried to overcome this problem by using mathematical models that minimize errors and improve prediction accuracy. RetiSpec said the company has taken samples from 14 research partners, from whom it gathers samples from racially and socioeconomically diverse communities. Neurovision said it took samples from diverse data sets and tested them against others to minimize errors.

“Some of the most important work we’ve done is to make sure the AI doesn’t suffer from garbage in and garbage out," RetiSpec’s Shaked said.

Write to Vipal Monga at vipal.monga@wsj.com

Catch all the Business News, Market News, Breaking News Events and Latest News Updates on Live Mint. Download The Mint News App to get Daily Market Updates.
more

topics

MINT SPECIALS

Switch to the Mint app for fast and personalized news - Get App