Principal Investigator: Dr Narges Razavian
New York University School of Medicine, New York, USATags: 45449, Alzheimer's Disease, deep learning, early-detection, forecasting, mild-cognitive-impairment, vascular-dementia
Alzheimer’s disease is common, terminal and does not have a cure at the moment. By the time clinical symptoms emerge, it is years too late to be able to reverse the disease. 100% of clinical trials have failed so far, partly because they recruit patients when they already have clinical symptoms.
Our brain-changes start *decades* before any clinical symptom of AD emerges. Currently, there are no accurate imaging biomarkers from MRI that would allow us to detect the disease at its earlier stage. Existing methods of today are expensive, unpleasant to the patient, and involve injection of radioactive markers(PET), something we can not do at regular intervals. Therefore we are not very good at detecting AD before clinical symptoms start.
On the other hand, artificial intelligence has shown great success at sifting through thousands of images, and finding patterns that differentiate objects from each other. Our project aims to use AI to look into thousands of brain MRI scans, find patterns that are associated with early and late Alzheimer’s disease stage, compared to normal aging. If completed, this can impact clinical trials and development of treatments for a disease that has evaded cure up to now.