A recent study led by researchers Massachusetts General Hospital offers a potential breakthrough leveraging artificial intelligence to identify potential diagnoses of Alzheimer’s and early onset Alzheimer’s disease.
The team is part of the Massachusetts Alzheimer’s Disease Research Center, in conjunction with the National Institutes of Health and the Ministry of Trade, Industry, and Energy of the Republic of Kora. Matthew Leming, PhD, the study’s lead analyzed previous Magnetic Resonance Images (MRI) of the brain to train a deep computer intelligence learning program - a combination of machine learning and artificial intelligence – to better conduct accurate analysis for the early detection of Alzheimer’s.
The first images given to the deep learning technology come from patients of Mass General from prior to 2019, both diagnosed with and without Alzheimer’s.
The next step was to test if the computer A.I. would be as accurate when digesting images from varying sources and time periods. To do so, researchers provided five datasets from different hospitals and across different years.
As Alzheimer’s disease is typically associated with older patients, researchers determined it was important to test if the program would have the same accuracy in diagnosing early-onset Alzheimer’s, when evaluating scans from patients of varying ages.
Dr. Leming, noted that “we addressed this by making the deep learning model ‘blind’ to features of the brain that it finds to be overly associated with the patient’s listed age.”
In total, researchers fed the deep learning program 11,103 different MRIs from 2,348 patients, classified as ‘at risk’ for Alzheimer’s, and 26,892 MRIs from 8,456 patients without and not ‘at risk’ for Alzheimer’s disease.
Of the images provided, the program was able to detect patient’s risk of developing Alzheimer’s disease with an impressive 90.2% accuracy rate.
Researchers were highly encouraged by the detection result, the first real-world studies of Alzheimer’s disease and dementia.
Typically, such studies are conducted in controlled laboratory environments that don’t translate perfectly to outside settings; however, the scientists noted that “this study made substantial steps towards actually performing this in real-world clinical settings.”
These results provide an optimistic outlook on the future of earlier Alzheimer’s diagnoses and the use of deep learning A.I. programs in healthcare settings.
Recently, there has been a considerable amount of attention paid to A.I. technologies as they become more enmeshed in daily life, for better or worse. Many question what the future of our world deploying these technologies may look like.
This study exemplifies a true fantastical use-case of A.I. technology in healthcare, able to detect this horrific disease with 90% accuracy.
Perhaps these results will encourage further research into the ways in which A.I can be used to advance medicinal science.
The Fisher Center for Alzheimer's Research Foundation is sponsoring a Brain Awareness Week in mid-March.
