Mount Sinai researchers have pioneered an AI tool, “HistoAge,” featured in Acta Neuropathologica, capable of estimating brain age, predicting age at death, and identifying age-related brain changes. By analyzing 700 hippocampal section images from aged brain donors, the AI model calculates age acceleration, offering crucial insights into neurodegenerative diseases. Its accuracy surpasses traditional methods, impacting fields from Alzheimer’s research to clinical diagnostics. Mount Sinai, alongside UCSD, received $8.5 million to lead the CFDE Data Resource Center, fostering biomedical research by enhancing data integration and utilization.
Researchers at Mount Sinai have achieved a groundbreaking feat by developing an artificial intelligence (AI) tool called “HistoAge,” as reported in the recent issue of Acta Neuropathologica. This innovative tool has the capability to estimate histopathological brain age, predict the age at death, and pinpoint areas of the brain susceptible to age-related changes.
As individuals age, their brains undergo intricate cellular and structural changes that often result in functional decline and increased vulnerability to neurodegenerative diseases, notably Alzheimer’s. Understanding the concept of age acceleration in the brain, which represents the variance between a person’s chronological and biological age, is crucial in explaining disease-related alterations and functional deterioration.
The researchers harnessed the power of AI and developed the “HistoAge” algorithm using multiple instance learning. To train the model effectively, the team curated nearly 700 digitized slide images of human hippocampal sections from aged brain donors. The hippocampus was specifically chosen due to its significant role in brain aging and neurodegenerative diseases.
The model was trained to estimate the age at death of each patient based on these images, a task considered insurmountable for human accuracy. The disparity between the actual age of the patient and the age predicted by the model was then used to calculate age acceleration.
Compared to existing methods like DNA methylation for estimating age acceleration, the estimations derived from HistoAge exhibited stronger associations with clinical and pathological outcomes linked to cerebrovascular disease, cognitive impairment, and abnormal levels of Alzheimer ’s-type degenerative protein aggregation.
This revolutionary HistoAge model holds immense potential for clinicians, offering a more precise means of determining brain age and comprehending the factors contributing to neurodegeneration in aging patients. According to Gabriel A. Marx, MD, MS, a resident in Neurology at Icahn Mount Sinai, “This model opens the floodgates for a slew of fascinating and essential analyses that bring us closer to finally understanding the aging brain and age-related brain diseases such as Alzheimer’s.”
The research team is now actively working on creating a multicenter dataset that can be utilized to train other AI models for further advancements in brain disease research. This achievement underscores Mount Sinai’s commitment to advancing healthcare through cutting-edge data analytics and AI applications.
In a related development, Mount Sinai, in collaboration with the University of California San Diego (UCSD), has been awarded $8.5 million to establish a data integration hub aimed at advancing National Institutes of Health (NIH) Common Fund initiatives. This initiative seeks to gather patient data to support biomedical research and the development of novel therapeutics for various diseases. Mount Sinai and UCSD will lead the CFDE Data Resource Center, a pivotal step toward enhancing the utilization of data produced by Common Fund initiatives and revolutionizing the landscape of medical research and patient care.
