What does it mean to be diagnosed with dementia caused by Alzheimer’s? Will the disease advance rapidly over a few years or more moderately over decades? Researchers at Simon Fraser University (SFU) are delving into the data to help predict Alzheimer's development and progression.
Dementia of Alzheimer's Type (DAT) is a progressive, neurodegenerative disease affecting the psychiatric, cognitive and structural condition of the brain, and accounts for up to 80% of all dementia cases. Effective treatments for Alzheimer’s are limited, and DAT is fatal for one in three elderly patients—with a survival rate of four to 20 years. Early detection is critical for successful disease management.
SFU engineering science professor Mirza Faisal Beg and statistics and actuarial sciences professor Jiguo Cao analyzed the data of over 400 individuals in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database to predict time-to-conversion for DAT.
This is the first known study that performs a comprehensive survival analysis for subjects in various stages of the disease. The study found that a person’s genetic data provided the most reliable prediction of whether they would develop DAT, while cognitive testing was most helpful in determining survival rates at disease onset.
Advancing the capabilities of engineering and data science is a recurring theme in the two researchers’ work. Beg is the principal investigator at the Functional & Anatomical Imaging & Shape Analysis Lab (FAISAL), who applies engineering methods with medical school knowledge to build translational tools for clinical utility. Cao is the Canada Research Chair in Data Science and an award-winning statistician whose work has made important contributions to genetics, public health and medicine.
The study, Predicting time-to-conversion for dementia of Alzheimer's type using multi-modal deep survival analysis, was published in Neurobiology of Aging, and was a collaboration with researchers from SFU, the University of Victoria, the University of British Columbia, Wake Forest University, University of Nottingham, Memorial University of Newfoundland and Ohio State University.
We spoke with professors Beg and Cao about their research.
What is the Alzheimer’s Disease Neuroimaging Initiative database and what information is found there?
The Alzheimer’s Disease Neuroimaging Initiative (ADNI) is a longitudinal multicenter study designed to develop clinical, imaging, genetic and biochemical biomarkers for the early detection and tracking of Alzheimer’s disease. The landmark public-private partnership has made major contributions to research and has enabled the sharing of data between researchers around the world. The database contains clinical and genetic information, digital images (MRI and PET scans) and biological specimens of over 800 subjects.
What types of data did you use for your analysis? How did you go about analyzing the information?
We used features from MRI, genetic and “CDC”— which stands for cognitive tests, demographic, and cerebral spinal fluid biomarker data—found in the ADNI database. We started with the baseline data of 401 subjects and used a deep-learning model designed for survival analysis to predict subjects’ time-to-conversion to DAT using 63 features.
What new and significant findings did this research uncover?
Our study revealed that CDC data outperform genetic and MRI data in predicting DAT time-to-conversion for subjects with mild cognitive impairment (MCI). Conversely, genetic data provided the highest predictive power for subjects with normal cognition (NC) at the time of their visit. Additionally, combining MRI and genetic features improved time-to-event prediction compared to using either modality alone. Finally, adding CDC to any combination of features performed as well as using only the CDC features.
What is unique about this study that makes it the first of its kind?
Our study demonstrates the use of multi-modal survival analysis to predict the probability of a preclinical subject being diagnosed with DAT over time. This approach aims to enhance clinical utility and improve our understanding of how different types of data contribute to time-to-diagnosis predictions. To our knowledge, this is the first comprehensive survival analysis focusing on the prediction of the time-to-conversion to DAT for subjects at various stages of the disease. Additionally, we utilized multimodal data to explicitly compare the predictive power of each data modality and assess the impact of each modality on disease diagnosis and progression.
What does this research mean for the understanding of Alzheimer’s disease? Where do you go from here?
In the current clinical practice, a DAT diagnosis cannot be made until the patient exhibits clear signs of cognitive decline. Therefore, methods for predicting the probability of a patient developing Alzheimer's disease as a function of time, collectively known as survival analysis, are important tools in helping understanding the characteristics of DAT.
Our results are limited by the sample size and characteristics of the subjects selected from the ADNI database. One way to address this limitation would be to include additional Alzheimer’s disease-related databases such as the National Alzheimer’s Coordinating Center and the National Institute on Aging Genetics of Alzheimer's Disease Data Storage Site to perform independent evaluation, which can lead to a more robust model.
Given that treatments for the disease are limited and the burden of Alzheimer's and dementia is significant for individuals, loved ones, caregivers and the healthcare system overall, the more we know about the condition, the better. We hope our research can contribute to a better understanding of the disease and point in promising directions for further research.
For more visit the Functional & Anatomical Imaging & Shape Analysis Lab (FAISAL), and Jiguo Cao’s research page.
SFU's Scholarly Impact of the Week series does not reflect the opinions or viewpoints of the university, but those of the scholars. The timing of articles in the series is chosen weeks or months in advance, based on a published set of criteria. Any correspondence with university or world events at the time of publication is purely coincidental.
For more information, please see SFU's Code of Faculty Ethics and Responsibilities and the statement on academic freedom.