AI programme spots Alzheimer's years before confirmed diagnosis

Donna Miller
November 9, 2018

Research has linked the disease process to changes in metabolism, as shown by glucose uptake in certain regions of the brain, but these changes can be hard to recognize. "People are good at finding specific biomarkers of disease, but metabolic changes represent a more global and subtle process".

There is no cure for Alzheimer's and the effectiveness of the best treatments for the disease diminishes as it progresses.

The "deep learning algorithm" developed by the team was trained by showing it 2109 F-FDG PET (or fluorine 18 (18F) fluorodeoxyglucose positron emission tomography) scans from 1,002 patients. These PET scans can detect metabolic activities from all parts of the brain by showing uptake of a radioactive glucose compound that is administered by injection into the blood stream.

Despite the fact that Alzheimer's is among the top 10 deadliest diseases in the world, we've only begun to have a rudimentary understanding of the condition in recent years.

Specifically, the new system's machine learning system was able to learn patterns in almost 2,000 brain scans taken of 1,000 patients.

Through deep learning, the Alzheimer's algorithm was able to teach itself to recognise metabolic patterns in brain scans that indicated disease.

Dolphins, Gase fined for violating injury policy
The Dolphins and Gase pretty much received a slap on the wrist for their part in not reporting Tannehill's injury correctly. Gase, the former Bears offensive coordinator who is in his third year as the Dolphins' head coach, was fined $15,000.

However, they cautioned that the study was very small - involving just 40 people - and that much more work is needed to test and refine the technique before it could be used in the health service.

Diagnosing Alzheimer's is no easy task and, so far, research has only managed to link the disease process to metabolism changes shown by glucose uptake in certain regions of the brain but even knowing this, the changes are, more often than not, hard to recognize.

Dr Jae Ho Sohn, a member of the team from the University of California at San Francisco, said: "We were very pleased with the algorithm's performance". Results revealed that the programme could detect signs of Alzheimer's disease in 40 patients six years (average 75.8 months) before symptoms of the disease manifested themselves and the disease could be formally diagnosed. "It was able to predict every single case that advanced to Alzheimer's disease". In the future, the team wants to train the algorithm to look for other patterns associated with accumulation of beta-amyloid and tau proteins, both markers specific to Alzheimer's disease.

"If we only diagnose Alzheimer's when all the symptoms have appeared, the brain volume loss is so significant that it's too late to intervene", he added.

Early detection of Alzheimer's could open the door to new ways of slowing down or even halting the progression of the disease.

Other reports by

Discuss This Article