BerlinThe early detection of Alzheimer's is particularly important. If still unavoidable dementia is discovered earlier, it can at least slow down the course with medicines.
"If we diagnose Alzheimer's disease only when there are clear symptoms, the brain volume loss is so high that effective intervention is usually too late," Jae Ho Sohn explains.
Along with his team from the University of San Francisco, the doctor has developed a new tool for the early diagnosis of Alzheimer's disease: an adaptive algorithm that predicts reliably dementia years before a doctor's diagnosis.
The researchers focused their development on the subtle metabolic changes in the brain caused by the onset of the disease. Such changes can be visualized using an imaging technique known as Positron Emission Tomography (PET).
Seven Factors That Promote Alzheimer's Disease
However, the early stages of the disease are so obscure that even experienced physicians find it hard to recognize. "It's easier for humans to find specific biomarkers for the disease," explains Sohn. "But metabolic changes are more complicated."
Researchers trained their artificial intelligence using data from the Alzheimer's Disease Non-Violence Initiative (ADNI). Among other things, this data collection contains thousands of PET images for Alzheimer's at a very early stage of the disease. 90 percent of these entries, researchers used algorithm training, the remaining 10 percent to control success.
For the final examination, the AI had to finally analyze 40 images that had not been submitted before. The result describes the son in the following way: "The algorithm was able to reliably determine any event that occurred at the beginning of Alzheimer's disease."
In addition to the 100 percent rate, the doctors surprised the case very early. On average, the system recognizes the symptoms more than six years before the actual diagnosis of the disease. "We are glad about this result," says the Son. However, the doctor also knows that the series of tests was still relatively small and the result of the subsequent tests should be confirmed.
In his algorithm, however, he considers it an important tool in Alzheimer's therapy: "If we can detect a disease faster, it will give researchers the opportunity to find better ways to slowly or even stop the disease process."