Pictured above is the lead author, Ghazal Mirabnahrazam, a research assistant currently completing a master's degree in engineering science at SFU

Combining genetics and brain MRI can aid in predicting chances of Alzheimer’s disease

June 28, 2022

In a recent study, SFU engineering researchers in the Functional and Anatomical Image & Shape Analysis Lab (FAISAL) investigated how the use of brain MRI and genetics can aid in the prediction of Dementia of Alzheimer’s Type (DAT). They examined patients at different stages of the disease to develop a biomarker to indicate the future probability of developing DAT.

“Our findings reveal that while genetic features have lower predictive power than MRI features, combining both modalities can improve the performance in predicting the future conversion to DAT,” says study lead author Ghazal Mirabnahrazam, a research assistant currently completing a master’s degree in engineering science at SFU.

They found that dementia scores deriving from genetic information have a better predictability of future DAT progression in currently normal patients who will develop DAT at a later time, as opposed to MRI, which have a better predictability of future DAT in patients with mild cognitive impairment. 

“In a clinical setting, clinicians can use our model to predict a quantitative score indicating the similarity between a subject's observed patterns based on MRI and genetic data at the time of clinical visit and DAT patterns,” says senior author Mirza Faisal Beg, a professor in the School of Engineering Science

“This is extremely useful, specifically at the MCI (mild cognitively impaired) stage in identifying those who will progress to DAT in the future.”

Having the ability to predict the future chances of DAT from only using the baseline information is important as it gives clinicians the ability to have awareness of the disease and time to plan out a suitable, personalized care plan for patients according to their onset of developing DAT.

Read the full news story from SFU News here >>