Italian researchers developed a machine learning algorithm that can analyze MRI scans to detect Alzheimer’s 10 years before its symptoms show up.
Characteristic symptoms of Alzheimer’s disease don’t appear abruptly, as they’re the result of a slow process whose early stages can go unnoticed.
Until a definitive cure for Alzheimer’s is found, detecting its symptoms as early as possible is crucial so that the patient enjoys an independent, high quality of living for as long as possible.
It’s been shown recently that Alzheimer’s doesn’t necessarily destroy memories and it doesn’t prevent patients from developing new ones. To rediscover these memories and enable new ones to stick around would have a great impact on the lives of millions of people on the dementia spectrum and those caring for them.AI algorithm detects Alzheimer's 10 years before symptoms show up.Click To Tweet
Now, with this algorithm, AI could help spot subtle neurological manifestations of Alzheimer’s in the brain years before doctors can diagnose the disease with conventional methods.
AI to Give Patients and Doctors Greater Leeway in the Fight Against Alzheimer’s
Affecting an estimated 46.8 million people worldwide, dementia-related diseases pose a real challenge for patients, caregivers, healthcare community, and society as a whole. Not to mention their economic burden, expected to reach $1 trillion USD by 2018 and $2 trillion by 2030.
Alzheimer’s, accounting for 60 to 80% of all dementia-related cases, is the most common type of dementia, for which there’s currently no cure. Even with laser treatment, the likelihood of total eradication is slim, at least in the short term.
However, the emergence of AI has given a boost to research dedicated to understanding, in a much more precise way, the neurodegenerative processes of the disease to fight it better, and there’s been remarkable progress made recently.
Human doctors can diagnose Alzheimer’s using an MRI scan, but usually that’s only when looking at the disease at an advanced stage–often there’s little to be done. By then, the disease would’ve wrecked too much havoc.
In 2015, researchers from the VU University Medical Centre in Amsterdam (Netherlands) explored the path of combing AI and MRI to detect early forms of dementia, with promising results.
MRI-Reading AI Diagnoses Alzheimer’s, 10 Years Before Doctors can
Italian researchers have provided more proof of the potential of machine learning algorithms as a tool against Alzheimer’s.
Researchers at the University of Bari and Istituto Nazionale di Fisica Nucleare have developed a machine learning algorithm that can spot dementia-related brain changes, up to 10 years before the onset of the first symptoms that doctors usually recognize.
Researchers used 67 publically available MRI scans (38 were from people with Alzheimer’s disease and 29 as healthy controls) to train their algorithm.New #Alzheimer's detecting AI 86% accurate in early diagnostics.Click To Tweet
The algorithm was then tested independently using 148 subjects, including 48 MRI scans from people with AD, 48 with mild cognitive impairment (MCI), and 52 healthy patients.
The results are rather impressive: the AI algorithm was accurate 86% of the time in telling a healthy brain from one with AD.
But it’s the AI’s 84% accuracy in differentiating between healthy brains and those with MCI that’s the most promising, as MCI will eventually lead to Alzheimer’s over the course of a decade.
A paper detailing the process and results of the study can be found on the arXiv.org open library.
While a very early diagnosis of Alzheimer’s doesn’t ensure healing, it does give doctors the opportunity to initiate treatment as soon as possible and thus slow the effects and the progression of the disease. At least patients will have time to prepare properly by adopting an adequate lifestyle.