Best Life: AI locates brain lesions in seconds

SAN ANTONIO, Texas (Ivanhoe Newswire) – Brain lesions occur in 90% of patients over 65, and they are often among the first signs of dementia.

The key is knowing how many lesions are lodged in the brain, and where they are.

Counting them is nearly impossible, so, researchers are turning to artificial intelligence to do that in mere seconds.

Director of neuroimaging at UT Health in San Antonio, Mohamad Habes, studies MRI scans to find lesions in the brain deeply lodged in perivascular folds.

“They are tiny and because of their small size, it could be difficult for a clinician to really count all of them,” said Habes.

Because people over 70 could have hundreds of lesions in the perivascular space, counting them becomes next to impossible.

“We have developed a sophisticated system that is able to detect and quantify those lesions one by one throughout the whole brain,” said Habes.

Lesions are formed from damaged brain tissue, triggered by vascular disease or injury.

“We see them a lot in people who have traumatic brain injuries. A hit on the brain, on the frontal lobe, for example, could cause a lesion,” said Habes.

And that affects how a person walks or talks. Doctors know that when an older person changes their gait, or way of walking, it can be an early indicator of dementia.

“It doesn’t mean that the moment you develop these lesions, you will develop dementia, but it means your risk of developing dementia would be greater,” said Habes.

Habes says this groundbreaking AI algorithm allows people to improve lifestyle choices and stave off dementia.

He also says this breakthrough study should lead to clinical application within a year, so doctors can easily uncover cerebral small-vessel disease in the blink of an eye by using AI.

Contributors to this news report include: Donna Parker, Producer; Bruce Maniscalco, Videographer; Roque Correa, Editor.

Click here to sign up for our newsletter!

Click here to report a spelling or grammar error. Please include the headline.

Copyright 2023 WMC. All rights reserved.