Alzheimer's disease can be diagnosed with a single examination of the brain, shows a new British scientific study, which uses artificial intelligence (machine learning technology) to detect even invisible but characteristic structural changes in the brain that indicate neurodegenerative.
The advantage of the new technique is its simplicity and the fact that it can diagnose the disease at an early stage, when the diagnosis is usually considered very difficult due to the absence of noticeable symptoms. Although the disease is incurable so far, early diagnosis helps patients find the right support, better manage their symptoms, and plan for the future.
Alzheimer's disease is the most common form of dementia and occurs in most patients after the age of 65, but younger people can also develop it. The most common symptoms are memory loss and difficulties in thinking, problem solving and language.
The disease is usually diagnosed with various memory and cognitive tests, as well as tests that reveal the accumulation of toxic proteins in the brain and the shrinkage of the hippocampus, the brain-related area of memory. All of these tests, however, can take weeks to complete.
Researchers at Imperial College London, led by Professor Eric Abogaye of the Department of Surgery, who published the paper in the medical journal Communications Medicine, have developed a new approach that requires only a test, a magnetic resonance imaging 1,5 Tesla, available in most hospitals.
Scientists have adapted for Alzheimer's an artificial intelligence algorithm that was originally developed to categorize cancerous tumors. After dividing the brain into 115 regions and distinguishing 660 different characteristics for each region (such as size, shape, etc.), they trained the algorithm to detect the changes associated with Alzheimer's disease. It is important, according to the researchers, that the new system "catches" changes in the brain that until now had not been associated with Alzheimer's disease, e.g. in the cerebellum.
The new method was tested in more than 400 patients with early-stage or advanced Alzheimer's disease, as well as in two control groups for comparison purposes: in healthy people and in patients with other neurological conditions such as Parkinson's and frontal dementia.
It was found that in almost all cases (98%) the "smart" system could predict whether the patient had Alzheimer's or not. He was also able to distinguish with satisfactory accuracy (in 79% of patients) whether it was Alzheimer's disease onset or advanced.
Dr Abogay said: "At the moment no other simple and widespread method can predict Alzheimer's disease with this level of accuracy, so our research is an important step forward." He added that "waiting for a diagnosis can be a terrifying experience for patients and their families. "If we could shorten the time they have to wait, we would make the diagnosis a simpler procedure and partially reduce their uncertainty, which would be a great help."
Source: RES-EAP