Detecting Alzheimer’s earlier, and more conclusively

By Diane Luckow

SFU engineering science professor Mirza Faisal Beg is pioneering new brain-imaging research that can diagnose Alzheimer’s dementia earlier, less invasively, more cheaply and more conclusively than current testing methods.

More than 500,000 Canadians are currently living with this progressive, fatal disease that is the most common form of dementia. And within 15 years, the Alzheimer’s Society of Canada estimates that almost one million Canadians will have Alzheimer’s.

To help combat this disease, Beg and his students have developed algorithms that can interpret and quantify images of the brain captured from magnetic resonance imaging (MRI), FDG-PET imaging, and newer retinal imaging.

“We take measurements from various regions in the brain, and convert these to point clouds,” says Beg. “We then train pattern-recognition algorithms to recognize the differences between brains known to have the Alzheimer’s pathology, and those that don’t.”

The algorithms reveal the signature patterns of different dementias based on the brain’s regional volume measurements and glucose metabolism levels. Such measurements can be crucial to studying Alzheimer’s and its progress.

So far, Beg’s team has mapped information from several thousand brain images, discovering that those from individuals who have Alzheimer’s show smaller measurements for volume and metabolism. They found they could display these images as “bands” resembling a barcode pattern. The bands reveal that key regions of the brain, such as the hippocampus, or memory center, are not using as much glucose as in the non-Alzheimer’s brains.

“Clinicians are amazed,” he says, “because these findings agree with our understanding of what Alzheimer’s is, and where it should affect the brain.”

The findings have important ramifications for Alzheimer’s testing, he says.

“Current Alzheimer’s tests are inconclusive, especially in the early stages of Alzheimer’s. But the barcode pattern of the brain’s volume and metabolism reveals how well it is working, and can serve as a valuable marker for establishing Alzheimer’s.”

Tracking Alzheimer’s through the retina?

More recently, Beg has begun working on Alzheimer’s markings that can be found from scanning the retina, which is a quick, non-invasive procedure.

“Early research has shown some evidence that the amyloid protein deposits found in the brains of Alzheimer’s patients may also be present in the retina, since the retina is actually a part of the brain,” he says. “The question is, do the deposits in the retina occur before, after or concurrently with those in the brain, and can we image them accurately?”

To explore this possible connection, his team is studying retinal images from mice that carry the Alzheimer’s gene, examining their brains and retinas to see how they change in each area over time as the disease progresses.

Once the ability to image amyloid deposits and their timing in both the brain and the retina are established, he says it may be possible to track Alzheimer’s progress by scanning the retina and examining the changes. Such as test would be non-invasive and cheap, and could be used regularly to detect both the onset of Alzheimer’s and whether it is worsening.

Beg is the principle investigator of Western Canada’s only team in brain-imaging and retinal imaging to be funded by Brain Canada. The team’s other SFU members are SFU engineering science professor Marinko Sarunic who designs state of art optical imaging systems for retina imaging, SFU computing science professor Greg Mori, who is an expert in machine learning and pattern recognition and statistics professor Jinko Graham who is an expert in imaging genomics. They hold a $1.5-million research grant.

Simon Fraser University