Sunday , May 9 2021

AI technology helps predict future Alzheimer's disease diagnosis



A study published in Radiology showed that a deep learning model predicted Alzheimer's disease with 82% specificity and 100% sensitivity about 6 years prior to diagnosis using fluoride 18 fluorodecluglucose PET imaging studies of the brain.

"Differences in patterns of glucose uptake in the brain are very subtle and diffuse" Jae Ho Sohn, MD, from the Department of Radiology and Biomedical Imaging at the University of California, San Francisco, said in a press release. "People are good at finding specific biomarkers of disease, but metabolic changes represent a more global and subtle process."

Researchers investigated whether a deep learning algorithm could predict the final diagnosis of Alzheimer's disease among patients undergoing fluoride 18 fluorodeoxyglucose PET in the brain.

Investigators gathered prospecting fluoride 18-fluorodecluglucose PET brain images from the Alzheimer's disease drug imaging initiative dataset, containing 2,109 study studies from 1,002 patients and a retrospective independent test set that contained 40 study studies from 40 patients. They trained the 90% of the data set deep learning algorithm, tested it for the remaining 10% and the independent test set. The model was evaluated with sensitivity, specificity and recipient's operating characteristics.

brain

A deep learning model predicted Alzheimer's disease with 82% specificity and 100% sensitivity about 6 years prior to diagnosis using fluorine oxide glucose PET formation studies of the brain, according to study evidence.

Source: Shutterstock.com

The deep learning algorithm learned the metabolic patterns associated with Alzheimer's disease, according to the press release. The model achieved the area under the recipient's operative character curve of 0.98 (95% CI, 0.94-1) when tested for the ability to predict the final Alzheimer's disease diagnosis in the independent test set, Sohn and colleagues reported. The algorithm achieved 82% specificity at 100% sensitivity in detection of the disease on average 75.8 months prior to final diagnosis.

After comparing the performance of the algorithm with that of radiologic readers, the investigators also found that the algorithm surpassed readers (57% sensitivity and 91% specificity; P <.05).

"We were very pleased with the performance of the algorithm. It could predict each case as advanced to Alzheimer's disease," says Sohn in the release. "If we diagnose Alzheimer's disease when all symptoms have revealed, brain volume loss is so significant that it's too late to intervene. If we can detect it earlier, it's an option for investigators to find better ways to slow down or to and stop the disease process. " – by Savannah Demko

Disclosures: Sohn reports contributions from UCSF. See the entire study for all other writers relevant financial information.


Source link