Researchers at Weill Cornell Medicine and Hospital for Special Surgery have developed a machine-learning tool that can identify different subtypes of rheumatoid arthritis. This could improve care for patients with this complex condition.
The study, published in Nature Communications, shows that AI can effectively analyze pathology samples from RA patients, potentially leading to better diagnosis and personalized treatment. Dr. Fei Wang noted that this tool could change how diseases are diagnosed and treated. While similar tools exist in oncology, Dr. Wang’s team is working to apply this technology to other medical fields.
Dr. Wang collaborated with Dr. Richard Bell and Dr. Lionel Ivashkiv to automate the process of identifying rheumatoid arthritis (RA) subtypes using tissue samples. This automation could help doctors choose the best therapy for each patient.
Pathologists manually classify arthritis subtypes by examining biopsy samples, which is slow, costly, and can lead to inconsistencies. Dr. Bell described it as a “bottleneck” in pathology research due to its time-consuming nature.
The team trained their algorithm on RA samples from one group of mice to learn how to identify different tissue and cell types and sort them by subtype. They tested the tool on another set of samples, discovering new insights like reduced cartilage damage after six weeks of RA treatment.
They then used the tool on human biopsy samples and found it could accurately type them. Now, they’re validating the tool with more patient samples and exploring how to use it in routine pathology work.
Dr. Bell said “this tool is the first step towards more personalized RA care. It helps identify a patient’s subtype for quicker treatment. The technology can spot tissue changes humans might miss, reducing costs and improving efficiency in clinical trials for RA treatments.”
Dr. Rainu Kaushal highlighted that this tool shows AI’s growing role in personalized medicine, offering new ways to detect and treat rheumatoid arthritis.
The team is developing similar tools to evaluate osteoarthritis, disc degeneration, and tendinopathy. Dr. Wang’s team is also exploring using broader data to define disease subtypes and has recently used machine learning to identify three subtypes of Parkinson’s disease.
Dr. Wang hopes this research inspires more computational studies for other diseases. Dr. Ivashkiv called this work a significant step in analyzing RA tissues to help patients.
In conclusion, the study shows that machine learning can effectively identify subtypes of rheumatoid arthritis, helping doctors choose the best treatments for patients. This approach can save time, reduce costs, and improve the accuracy of diagnoses, paving the way for more personalized care for RA and potentially other diseases.
Journal reference:
- Bell, R.D., Brendel, M., Konnaris, M.A. et al. Automated multi-scale computational pathotyping (AMSCP) of inflamed synovial tissue. Nature Communications. DOI: 10.1038/s41467-024-51012-6.