According to the (SEN), between 3 and 4% of the Spanish population between 75 and 79 years old is diagnosed with Alzheimer’s. Data that increase to 34% in people over 85 years of age.
The most worrying thing, warns the SEN, is that 80% of the cases of Alzheimer’s disease that are still mild are undiagnosed. This situation prevents pharmacological and non-pharmacological treatments from being started in a earlywhich would slow down cognitive decline.
However, the advancement of science could allow its early detection with 98% efficiency.
This emerges from a new study, led by a team of researchers from the imperial college londonwhich has shown that a brain scan and artificial intelligence can diagnose the disease “even before obvious brain shrinkage occurs.”
It is the leading cause of neurodegenerative dementia whose prevalence increases after the age of 65. Every year, about 40,000 new cases are diagnosed. The SEN also estimates that around 15% of the population over 65 years of age suffers from mild cognitive impairment. In 50% of cases, it would be due to this disease.
The research, which has been published this Monday in the scientific journal , would change the way in which Alzheimer’s is diagnosed.
Currently, they are used cognitive tests and memory, as well as brain scans (to check if the hippocampus, a region of the brain linked to long-term memory, has shrunk). They can be prolonged in time and, consequently, increase the symptoms of the disease.
Patients could be identified at an early stage of the disease for clinical trials of new drug treatments
But, the great advantage of this new technique, as its authors have highlighted, is its simplicity. This type of scanner is available in most hospital centers around the world.
-
“Today, no other simple and widely available method can predict Alzheimer’s disease with this level of accuracy, so our research is an important step forward,” explains Professor Eric Aboagye, from Imperial’s Department of Surgery and Cancer. College of London.
The drugs currently used to treat this neurodegenerative disease They are only effective if diagnosed early.
The use of the scanner together with artificial intelligence has already been tested in 420 patients in the United Kingdom with very promising efficacy: in 98% of cases it was detected at a very early stage”.
-
“This method could make diagnosis a simpler process and reduce some of the uncertainty. It could also identify patients early in the disease for clinical trials of new drug treatments or lifestyle changes, which is currently very difficult to do.”
“It is estimated that half of the cases of Alzheimer’s disease can be attributed to nine potentially modifiable risk factors: diabetes mellitus, arterial hypertension in middle age, obesity in middle age, smoking, physical inactivity, depression, cognitive inactivity or low educational level, hearing loss and social isolation”, explains the doctor Juan Forteaof the Spanish Society of Neurology (SEN).
How does this test work?
The test is done using a magnetic resonance of the brain with a standard 1.5 Tesla scanner (there is also a 3 Tesla scanner, with a higher magnetic field power), which is usually part of the technology of any hospital.
Previously, the researchers adapted an algorithm that was created to classify cancerous tumors. They examined each part of the brain and divided it into 115 regions. To these regions, they assigned 660 different characteristics (size, shape, and texture).
The experts trained this algorithm so that it could identify in which regions these characteristic changes occurred and predict the presence of the disease in advance. And even in areas of the brain that had not previously been associated with Alzheimer’s.
We talk about cerebellumposterior part of the brain and which is responsible for muscle coordination, and the ventral diencephalonwhich plays an essential role in the processing of motor information.
An algorithm “capable of selecting the texture and subtle structural features of the brain that are affected by Alzheimer’s disease and improving the information we can get from the standard imaging techniques”, emphasizes the neurologist Paresh Malhotra.
And most importantly, it does not require a subject matter expert.
-
“This method uses preset software for both brain segmentation and radiomic analysis (which converts medical images into quantifiable data). The algorithm calculates the manually designed features for very easy clinical interpretation and translation.”