AI could improve Alzheimer’s diagnosis, study finds

The use of AI in medicine has increased in recent years, especially in the diagnosis of diseases and conditions. A growing number of physicians already rely on deep learning, a machine learning method modeled after artificial neural networks, to learn by example how the human brain does, to detect potentially life-threatening conditions that can easily be overlooked, such as and even asymptomatic cases of COVID-19.

But the next breakthrough for AI in medicine could be the identification of Alzheimer’s, the devastating disease that causes irreversible cognitive decline and dementia, for which medical researchers have eluded treatment and reliable early detection in the century since the disease was discovered.

Researchers at Massachusetts General Hospital recently tested deep learning techniques for detecting Alzheimer’s and found that deep learning was not only more accurate than comparable AI models that weren’t trained to analyze multiple variables together, but also Alzheimer’s cases independently of factors that normally make early detection difficult, e.g. B. the age of a patient. The findings were published in a study published in last week PLUS ONEa scientific and medical journal.

The researchers trained a deep learning model using tens of thousands of brain scan images collected from over 10,000 people with and without Alzheimer’s disease. The study then tested the model against real clinical data on Alzheimer’s diagnoses.

The deep learning model was able to identify Alzheimer’s cases with an accuracy rate of 90.2%, about five percentage points higher than the simpler AI models that didn’t rely on the deep learning system. The AI ​​model performed better regardless of when and where patients were diagnosed with Alzheimer’s and regardless of how old they were at the time.

“This is one of the few studies that has routinely used collected brain MRIs to try to detect dementia,” said Matthew Leming, a research associate at Massachusetts General Hospital and the study’s lead author, in a statement. “Our results — with generalizability across sites, across time, and across populations — make a strong case for the clinical use of this diagnostic technology.”

A 90% accuracy rate in diagnosing Alzheimer’s would be miles ahead of clinical detection rates in humans, which are 77% according to a 2017 study.

The Great Medical Rise of AI

While AI-powered search engines developed by OpenAI, Microsoft and Google have been making most of the headlines lately about artificial intelligence for its promise to disrupt search and how we work, machine learning could potentially have life-saving applications in medicine have.

According to a December study by the US Department of Health and Human Services, more than 7 million people who come to US emergency rooms each year are misdiagnosed. This study found that nearly 3 million ER patients are burdened with side effects from a misdiagnosis, while over 370,000 suffer from permanent disability or death.

Misdiagnosis is also an economic burden, as eliminating incorrect tests and treatments, and the malpractice lawsuits that result from misdiagnosis, could result in savings of around $100 billion per year, according to the nonprofit Society to Improve Diagnosis in Medicine.

Doctors and medical professionals have said that AI shows promise in efforts to improve diagnostic techniques, despite many of the same problems with AI that have been identified elsewhere, such as B. the potential for factual error and racial bias, have also emerged in medical research. A literature review on AI in medical diagnosis published last year found that the technology shows promise in areas such as cancer, diabetes and Alzheimer’s diagnosis, although further research is recommended to improve AI’s accuracy in identifying medical problems.

A major role in Alzheimer’s research

But if future research brings AI and deep learning to diagnostics more widely, it could be a game changer for Alzheimer’s, one of the most difficult diseases to predict and diagnose.

Alzheimer’s is the most common form of dementia in the elderly, affecting approximately 44 million people worldwide. But it’s just one form of a large family of dementia diseases that can easily be misinterpreted as Alzheimer’s.

A 2017 study of over 900 people found that up to one in four people with Alzheimer’s was misdiagnosed, with a roughly even split between false positives and false negatives. Alzheimer’s propensity for misdiagnosis largely depends on how many of its symptoms overlap with other common neurological disorders, including Lewy bodies or frontotemporal dementia. The likelihood of misdiagnosis increases with age, according to the American Academy of Neurology, which says that Alzheimer’s disease and other dementias “can easily be misdiagnosed in older people.”

Predicting that a patient will develop Alzheimer’s is no easier than diagnosing it, as over 90% of Alzheimer’s cases are considered “sporadic” – occurring in patients with no family history of the disease. Because of these difficulties, there are almost no reliable early detection models for Alzheimer’s, with most cases being diagnosed after the first symptoms of brain damage are seen.

The Massachusetts General Hospital study didn’t address whether deep learning could help predict Alzheimer’s, but other studies seem to suggest that AI could play an important role there as well.

An AI model developed at the University of Florida was able to use electronic health records to predict which patients may be at high risk for Alzheimer’s up to five years before a diagnosis, the university said last week. While the researchers recommended more testing before doctors start using AI prediction tools, they found that AI models could help with early detection and reduce the severity of the disease in the long term.

Learn how to navigate and build trust in your organization with The Trust Factor, a weekly newsletter exploring what leaders need to succeed. Login here.