Diagnosing neurodegenerative diseases has long been one of the most demanding challenges in modern medicine. Conditions like Alzheimer’s disease and chronic traumatic encephalopathy share overlapping features at the cellular level, and accurate differentiation requires highly specialized expertise that is not uniformly available. A new artificial intelligence platform developed at the Icahn School of Medicine at Mount Sinai may be about to change that.
How AI Is Transforming the Detection of Neurodegenerative Diseases
Researchers at the Center for Computational and Systems Pathology at Mount Sinai developed the Precise Informatics Platform, a machine learning system designed to analyze digitized microscopic slides prepared from tissue samples of patients with a broad spectrum of neurodegenerative diseases. Using deep learning techniques, the platform was trained to identify neurofibrillary tangles, abnormal accumulations of tau protein in the brain that are a hallmark feature of Alzheimer’s disease and several other age related neurological conditions.
The system uses a convolutional neural network capable of detecting these tangles directly from digitized images with a high degree of accuracy. The study was published in Laboratory Investigation, a Nature medical journal.
Why Neurodegenerative Diseases Are So Difficult to Diagnose
The core challenge in diagnosing neurodegenerative diseases lies in the complexity and overlap of their pathological features. Tau protein tangles, for example, appear not only in Alzheimer’s disease but also in chronic traumatic encephalopathy and a range of additional conditions associated with aging and neurological decline. Distinguishing between these diagnoses at the tissue level traditionally requires a highly trained neuropathologist and is labor intensive, time consuming, and difficult to reproduce consistently across different institutions.
This variability in diagnostic quality has real consequences for patients. Inaccurate or delayed diagnosis limits access to appropriate treatment, complicates clinical trial enrollment, and slows the development of targeted biomarkers and therapeutics.
What the Precise Informatics Platform Offers
The Precise Informatics Platform is the first framework designed specifically for evaluating deep learning algorithms using large scale image data in neuropathology. Beyond detection, the platform supports data management, visual exploration, object outlining, multi user review, and the evaluation of algorithm results, making it a comprehensive tool for both clinical and research applications.
According to lead investigator John Crary, MD, PhD, Professor of Pathology and Neuroscience at the Icahn School of Medicine, the platform represents a major advance over existing approaches that are poorly reproducible and require intensive manual labor. The goal, he noted, is ultimately to achieve more efficient and accurate diagnosis of neurodegenerative diseases at scale.
Mount Sinai processes more than 80 million tests per year as the largest academic pathology department in the United States, giving researchers access to an exceptionally broad dataset for training and validating AI models.
The Broader Impact on Brain Disease Research
The implications of this technology extend well beyond diagnosis. By enabling more consistent and scalable identification of disease features in brain tissue, the platform accelerates the development of targeted biomarkers and opens new pathways for therapeutic research. As AI tools become more integrated into neuropathology workflows, the ability to detect and quantify neurodegenerative diseases earlier and more reliably could meaningfully improve patient outcomes across a wide range of conditions.
FOMAT conducts clinical research across multiple therapeutic areas including neurology. To learn more about active studies, visit FOMAT’s patient studies page.
For the full source, see the original article at R&D Magazine.