New AI model can identify neuroanatomical regions susceptible to age-related changes

Paper title: Histopathological brain age estimation using multiple instance learning

journal: acta neuropathologicaOctober 10, 2023

author: John F. Craley, MD, Professor of Pathology, Molecular and Cellular Medicine, Neuroscience, Artificial Intelligence and Human Health at the Icahn School of Medicine at Mount Sinai. Dr. Kurt W. Farrell, Assistant Professor of Pathology, Molecular and Cellular Medicine, Neuroscience, Artificial Intelligence and Human Health at Icahn Mount Sinai College. Gabriel A. Marx, MD, MSc, Icahn Mount Sinai Neurology Resident. and other co-authors.

conclusion: The aging brain undergoes structural and cellular changes that affect function and can increase susceptibility to neurodegenerative diseases such as Alzheimer’s disease. Age acceleration in the brain, the difference between biological age and chronological age, reveals insights into the mechanisms and normal functioning of one of the body’s most important organs. It can also explain age-related changes and functional decline, and identify early disease-related changes that signal the onset of brain damage.

For the first time ever, researchers at Mount Sinai used AI to develop an algorithm called HistoAge that predicts the age of death based on the cellular composition of human brain tissue specimens, with an average accuracy of within 5.45 years. announced. This powerful tool can also identify neuroanatomical regions vulnerable to age-related changes, which are indicators of potential cognitive disease.

how: Researchers examined a collection of approximately 700 digitized images of slides containing human hippocampal sections taken from elderly brain donors to develop a histological brain age estimation algorithm. The hippocampus is an ideal region for this analysis as it is known to be involved in both brain aging and age-dependent neurodegenerative diseases. The team then trained a machine learning model to estimate a person’s age at death based solely on the digitized parts. This is a task that is impossible for a human observer to perform with any degree of accuracy. They used the difference between the model’s predicted age and the actual age to derive the amount of accelerated aging in the brain.

result: When compared to current measures of accelerated aging (such as DNA methylation), HistoAge-based aging acceleration was found to be more strongly associated with cognitive impairment, cerebrovascular disease, and levels of Alzheimer-type abnormal protein aggregation. did. This study found that the HistoAge model is a reliable and independent metric for determining brain age and understanding the factors that drive neurodegeneration over time.

Why is the study interesting?: The researchers believe that the HistoAge model and subsequent similar algorithms represent an entirely new paradigm for assessing aging and neurodegeneration in human samples, and can be easily scaled up in clinical and translational research laboratories. He said that it could be introduced. Furthermore, this approach provides a more rigorous, unbiased, and robust indicator of the cellular changes underlying degenerative diseases. The team will then build a multicenter collaboration to develop large AI-enabled datasets that will be used to develop even more powerful AI models that have the potential to transform and enhance our understanding of brain diseases.

Dr. Craley, director of research at Mount Sinai, said:

“AI’s disruptive impact on brain research is a paradigm shift that will propel us towards the next generation of treatments. The HistoAge model will help us uncover critical causal relationships in debilitating brain diseases such as Alzheimer’s disease. It will be.”

Dr. Farrell, of Mount Sinai, said:

The use of modern computational approaches, such as AI, on human tissue samples taken from Mount Sinai’s vast and diverse collection will transform the way human diseases are evaluated. Our new HistoAge model is just one example of how AI is paving the way for further discoveries about the mechanisms of aging and neurodegeneration. Clinical scientists are increasingly using AI in research and diagnostic settings. This is a tool that is revolutionizing healthcare, and we are excited to optimize machine learning and become a leader in this space. This is not meant to replace the health system’s commitment to compassionate care, but rather to improve diagnosis and treatment for all patients.

Dr. Marx, director of research at Mount Sinai, said:

This model opens the floodgates to a number of fascinating and essential analyzes that will ultimately bring us closer to understanding brain aging and age-related brain diseases such as Alzheimer’s disease. This is the first time in pathology that we have been able to quantify the extent of brain aging. This approach not only allows us to discover genes that prevent or worsen brain aging, but also environmental risk factors that accelerate brain aging in individuals.

This research was supported by funding from the National Institutes of Health (R01AG054008, R01NS095252, R01AG060961, R01NS086736, P30AG066514, P50AG005138, R01AG062348, U24MH100931, and K01AG070326). Center for Research (P30 AG066514), the Winspear Family Center for Research Alzheimer’s Disease Neuropathology, the Rainwater Charitable Foundation/Tau Consortium, and a generous gift from Stuart Katz and Jane Martin. Researchers from the University of Pennsylvania and Boston University contributed to the study.