BerlinEarly detection of Alzheimer's disease is very important. If an already existing patient is detected early in the dementia, it may delay the medication at least.
"If you diagnose Alzheimer's disease clearly, the loss of brain size is so good that it is usually too late for effective intervention," explains Jou Ho Sokh.
Together with the team at the University of California in San Francisco, the doctor has developed a new tool for early detection of Alzheimer's disease: an adaptive algorithm that accurately predicts dementia to a physician's diagnosis.
Researchers have evaluated the development of subtle metabolic changes in the brain due to the onset of the disease. These changes can be visualized by the so-called positron emission tomography (PET).
Seven factors of Alzheimer's disease
However, traces of early stages of the disease are weak, so it is not even recognized for experienced doctors. Sokh explains: "It's easier for people to find biomarkers for a particular illness." "But metabolic changes are far more subtle processes."
Researchers taught artificial intelligence using data from the Alzheimer's Disease Neuroematology Initiative (ADNI). In addition, these data contain thousands of PET images of Alzheimer's patients. 90 percent of these records were studied by the researchers and the remaining 10 percent were tracked for success.
For the final test, the AI had to analyze 40 images that had not been submitted before. As a result, the son said, "Algorithm has been able to reliably identify any situation that Alzheimer's disease has started."
Physicians were impressed by the early detection of cases other than the 100 percent percussion. On average, the system has detected symptoms more than six years before the actual diagnosis. "We're happy with this outcome," said the boy. The physician also knows that the series of tests is less and should be further tested.
However, he sees the potential of Alzheimer's important instrument in the algorithm: "If you have previously detected the disease, it will allow researchers to find the best way to slow down or even stop the process. «