The Hong Kong University of Science and Technology (HKUST) has led an international research team that has made a significant breakthrough in the field of Alzheimer’s disease (AD). They have successfully created an advanced model that uses artificial intelligence (AI) and genetic data to forecast an individual’s susceptibility to AD even before symptoms manifest.
This pioneering study opens up new possibilities for using deep learning techniques in predicting disease risks and unravelling the underlying molecular mechanisms. It has the potential to revolutionise the diagnosis, interventions, and clinical research related to AD and other prevalent conditions like cardiovascular diseases.
In a collaborative effort, the President of HKUST, and the Chair Professor and Director of HKUST’s Big Data Institute, along with their research team, delved into the potential of artificial intelligence (AI), particularly deep learning models, to predict the risk of Alzheimer’s disease (AD) using genetic information.
This study stands as one of the earliest instances of deep learning models being applied to assess AD polygenic risks in both European-descent and Chinese populations. The results demonstrated that these deep learning models outperformed other models in accurately identifying patients with AD and categorizing individuals into distinct groups based on their disease risks linked to various biological processes. This research showcases the promising role of AI in advancing the understanding and prediction of AD, benefiting both populations of European and Chinese descent.
Currently, Alzheimer’s disease (AD) diagnosis heavily relies on clinical assessments involving cognitive tests and brain imaging. However, by the time symptoms become evident, it is often too late for optimal intervention. Hence, early prediction of AD risk holds great potential for improving diagnosis and intervention strategies.
The integration of the advanced deep learning model with genetic testing allows for the estimation of an individual’s lifetime risk of developing AD with an impressive accuracy rate exceeding 70%. This approach presents a promising avenue for identifying individuals at high risk of AD at an earlier stage, enabling timely interventions and enhancing the development of effective strategies to combat the disease.
Alzheimer’s disease (AD) is a hereditary condition influenced by genomic variations. These genetic variants are present from birth and remain consistent throughout an individual’s life. Analysing an individual’s DNA information can provide valuable insights into their predisposition to AD, facilitating early intervention and timely management of the disease. While FDA-approved genetic testing for the APOE-ε4 genetic variant can provide an estimate of AD risk, it may not be sufficient to identify high-risk individuals due to the contribution of multiple genetic factors to the disease.
Therefore, it is crucial to develop tests that integrate information from multiple AD risk genes to accurately assess an individual’s relative risk of developing AD over their lifetime. This comprehensive approach enables a more precise determination of AD risk and enhances our ability to identify individuals who may require targeted interventions and monitoring.
The President of HKUST stated that the study showcases the effectiveness of deep learning techniques in genetic research and predicting the risk of Alzheimer’s disease. This significant breakthrough is expected to expedite large-scale screening and staging of AD risk within the population.
In addition to risk prediction, the approach enables the categorization of individuals based on their disease risk and offers valuable insights into the underlying mechanisms that contribute to the development and advancement of AD. The transformative potential of these findings will help advance the understanding and management of Alzheimer’s disease.
The Chair Professor and Director of HKUST’s Big Data Institute expressed how this study exemplifies the remarkable benefits of applying AI in the realm of biological sciences, particularly in biomedical and disease-related research. By employing a neural network, they successfully captured the complex relationships present in high-dimensional genomic data, resulting in enhanced accuracy in predicting Alzheimer’s disease risk.
Additionally, using AI-driven data analysis without human supervision, the research team successfully categorized individuals at risk into distinct subgroups, shedding light on the underlying mechanisms of the disease. This study highlights the elegant, efficient, and effective nature of AI in addressing interdisciplinary challenges. The Chair Professor firmly believes that AI will play a crucial role in various healthcare domains in the near future.
The study was a collaborative effort involving researchers from the Shenzhen Institute of Advanced Technology, University College London, and clinicians from local Hong Kong hospitals, including Prince of Wales Hospital and Queen Elizabeth Hospital.
The findings of the study have been recently published in Communications Medicine, highlighting their significance in the scientific community. The research team is currently working on further refining the developed model with the ultimate goal of integrating it into standard screening procedures.
