METU’s new AI model to detect Alzheimer’s at early stage
SCIENTISTS from Middle East Technical University have developed artificial intelligence-based technology that detects the risk of early-stage Alzheimer’s by interpreting a genetic test on a sample of saliva.
Within the scope of the Scientific and Technological Research Council of Türkiye’s (TÜBİTAK) “Priority Areas R&D Projects Support Program,” while the first prototype is being implemented, the next stages, including the clinical phase, will be carried out at the Neuroscience and Neurotechnology Joint Application and Research Center (NÖROM), where the functions of the brain are being investigated.
Dr. Yeşim Aydın Son, an NÖROM board member and METU Informatics Institute faculty member, told Anadolu Agency (AA) that studies are progressing in the field of health informatics to develop diagnostic systems based on computer modeling of radiological images.
“Preliminary research is being carried out in the clinic that can support doctors’ differential diagnosis of the disease,” Son said.
Explaining the model, Son said: “We analyzed three large data set infrastructures created by international working groups using artificial intelligence, and genetic profiles that may be risk indicators at an early stage were determined. In the technique we developed to verify these genetic profiles, we take saliva samples from the participants’ mouths and isolate their DNA. We detected the genetic changes that increase the risk of disease, and we also observed variations that may be protective. We have developed a molecular diagnostic kit prototype that can help ease the process. With these analyses, we aim to provide information to our doctors whether people are at risk or have a protective genetic structure against Alzheimer’s, based on this AI model.”
Stating that they verified the method via a group of 100 Alzheimer’s patients in partnership with Hacettepe University Geriatrics Department, Son said: “Our project has reached the prototype stage. We would now initiate the second phase of the project, most likely with TÜBİTAK’s clinical research project. Our main goal in all these studies is to develop a diagnostic kit or developing a test.
Stating that they will carry out a largescale clinical study at the NÖROM center to optimize the results they obtain, Son said: “We aim to support doctors’ diagnoses. When our clinical research begins, people over 65 who show symptoms such as ‘not being able to remember’ will be referred to us. We will do these genetic analyses and evaluate the patient based on factors that increase risk or provide protection. At the same time, we will follow some of these patients for annual follow-ups. This way, we will be able to better test the reliability of our model.”
Son pointed out the importance of early diagnosis as well as differential diagnosis in Alzheimer’s and said: “Even if we cannot prevent the degenerative events occurring in the brain, some methods have been featured in the literature detailing methods to slow them down. At this point, our research in early diagnosis is of great importance.”
A medical doctor who returned to Türkiye after completing her doctorate in bioinformatics in the U.S., Son explained that for the last 14 years, she has been contributing to this field at METU and that she has been doing research combining medicine and artificial intelligence for the last 10 years.