An artificial intelligence tool trained partially on an Australian bank of dementia cases is set to be trialled in hospitals across the United States after a study found the advanced model could predict 10 different causes of the regressive illness.
Neurologists across Australia and the globe are struggling to keep up with a rapidly growing cohort of people living with dementia, which the World Health Organisation expects to grow to 78 million people by 2030.
MRI brain scans showing different dementia types.Credit: Boston University
Aiming to make dementia diagnosis faster and more accurate, an international research team led by scientists at Boston University developed an artificial intelligence (AI) model that can identify specific causes of dementia based on common clinical data such as MRI scans, individual and family medical histories, medication use, and neurological test scores.
In a study published in Nature Medicine on Thursday, researchers measured the accuracy of neurologists and neuroradiologists using the AI tool and compared it with their accuracy without it.
They found the AI tool boosted the doctor’s ability to diagnose dementia by more than 26 per cent.
Vijaya Kolachalama, an associate professor of medicine at Boston University and the study’s lead author, said the tool’s ability to distinguish among 10 different types of dementia, including Alzheimer’s disease, Lewy body dementia and vascular dementia, would pave the way for earlier diagnosis.
“In the real world, there could be multiple reasons someone has dementia. It could be Alzheimer’s and depression, it could be Alzheimer’s and vascular injury,” he said. “We believe AI can help by identifying these disorders early and assisting doctors in managing their patients more effectively, preventing the diseases from getting worse.”
Kolachalama said the tool was not designed to replace specialists but to help them make informed and efficient decisions.
Professor Carolyn Sue, a neurologist and neurodegeneration researcher at Neuroscience Research Australia (NeuRA), said AI diagnostic tools would not replace the need to see a specialist but could make timely dementia diagnosis “more accessible to a greater part of the population”.
“Some of these forms of dementias – unfortunately not all – can be treated, but our best treatments come when diagnosis is made early,” she said.
Dr Tim Wang, part of a team developing AI neuroimaging tools at the University of Sydney’s Brain and Mind Centre, said it was often difficult to find enough data to train accurate algorithms, and so the size and diversity of the cohort in the study (50,000 people across nine separate datasets) was “really exciting for the field”.
Kolachalama’s team has begun working on a pilot study in a hospital in Arizona. While that might be years away from being used by Australian doctors, Wang and his colleagues have developed another tool that can monitor brain diseases such as Alzheimer’s and multiple sclerosis by analysing MRI scans.
The Therapeutic Goods Administration on Thursday approved the technology for clinical use, days after the federal government invested $20 million in the technology.
Start the day with a summary of the day’s most important and interesting stories, analysis and insights. Sign up for our Morning Edition newsletter.
correction
This article has been updated to correct the spelling of Vijaya Kolachalama’s name.