AI Automates Analysis of Pathology Slides for Distinguishing Rheumatoid Arthritis Subtypes

Recent advancements in understanding the histopathological diversity of RA synovial tissue have identified three distinct phenotypes based on cellular composition, highlighting the need for targeted therapeutic approaches. Currently, pathologists manually categorize arthritis subtypes by analyzing cell and tissue characteristics in human biopsy samples, a process that is both time-consuming and costly, potentially leading to inconsistencies in diagnosis. To address these challenges, a new machine-learning tool has been developed to enhance the accuracy and efficiency of RA phenotyping in both pre-clinical and clinical settings.

In their study published August 29 in Nature Communications, investigators at Weill Cornell Medicine (New York, NY, USA) and Hospital for Special Surgery (HSS, New York, NY, USA) demonstrated the capability of artificial intelligence (AI) and machine learning technologies to effectively subtype pathology samples from RA patients. This differentiation among the RA subtypes can guide clinicians in selecting the most appropriate therapy for individual patients. Initially, the team trained the algorithm using samples from a specific mouse model, optimizing its ability to identify and categorize tissue and cell types into subtypes. This algorithm was then validated with another set of samples, revealing its potential to track treatment impacts, such as reduced cartilage degradation after six weeks of standard RA treatments.

Subsequently, the tool was applied to human biopsy samples, where it proved to be both effective and efficient in classifying clinical samples. The researchers are continuing to validate this tool with additional patient samples and exploring optimal ways to integrate it into existing pathological workflows. This technology not only promises to streamline the subtyping process, thereby reducing research costs and enhancing the efficacy of clinical trials, but also offers novel insights into RA by identifying tissue changes that might be overlooked by human observers. Ongoing development efforts by the researchers aim to create similar diagnostic tools for other conditions like osteoarthritis, disc degeneration, and tendinopathy, and extend machine learning applications to identify subtypes of other diseases, such as Parkinson’s disease, based on broader biomedical data sets.

“Our tool automates the analysis of pathology slides, which may one day lead to more precise and efficient disease diagnosis and personalized treatment for RA,” said Dr. Fei Wang, a professor of population health sciences and the founding director of the Institute of AI for Digital Health (AIDH) in the Department of Population Health Sciences at Weill Cornell Medicine. “It shows that machine learning can potentially transform pathological assessment of many diseases.”

"By integrating pathology slides with clinical information, this tool demonstrates AI's growing impact in advancing personalized medicine," added Dr. Rainu Kaushal, senior associate dean for clinical research and chair of the Department of Population Health Sciences at Weill Cornell Medicine. "This research is particularly exciting as it opens new pathways for detection and treatment, making significant strides in how we understand and care for people with rheumatoid arthritis."

Related Links:
Weill Cornell Medicine
HSS