A new study led by investigators from the Hospital of Special Surgery (HSS) in New York City finds that their computer vision tool effectively distinguishes rheumatoid arthritis (RA) from osteoarthritis (OA) in joint tissue taken from patients who have undergone total knee replacement (TKR). The results suggest that the machine learning model will help improve searches in the short term and improve patient care in the future. The results were presented today at the 2022 European Union of Societies for Rheumatology (EULAR) Conference.
TKR is often the only management option for patients with severe damage to the knee joint. Determining the disease that caused joint damage is essential to guide treatment plans, since RA is a systemic inflammatory disease that may also affect the eyes or the lining around the heart, while RA affects only the joints. “We know that there are many more immune cells present in the synovium, or joint tissue, of rheumatoid arthritis patients than in those with osteoarthritis,” said Bella Mehta, MBBS, MS, rheumatologist at HSS and lead author of the study. ‘But how many is not clear exactly.’
Dana Orange, MD, MS, rheumatologist at HSS, assistant professor at Rockefeller University and senior author of the study said. “However, these methods are imperfect.” For example, a recent study by HSS investigators found that evaluations by two highly experienced pathologists assessing the infiltration of a single type of immune cell known as lymphocytes on the same slides agreed only 67 percent of the time.1
Dr.. Orange, Mehta and their colleagues at HSS and collaborating institutions have developed and validated a computer vision tool that rapidly counts tens of thousands of cell nuclei in images of a whole slice of the synovium.2 In their current study, they measured 14 different pathologist-recorded traits in the synovium from 60 patients with rheumatoid arthritis and 147 patients with OA who underwent TKR, and used a computer vision tool to determine cell density.
The investigators identified significant differences between RA and OA features in the synovium. RA samples showed increased cell density; low number of mast cells, a type of white blood cell; and less evidence of fibrosis or scarring compared to OA samples. The probability of correctly distinguishing between RA and OA in the synovium was 85 percent when using 14 features scored by the pathologist alone, 88 percent when using the computer score for cell density alone and 91 percent when researchers combined the results of the pathologists and the computer cell density calculation. The researchers identified a cut-off point to distinguish RA from OA, and determined that the synovium contains more than 3,400 cells per mm.2 It should be classified as RA.
“Although our innovations are not yet ready for clinical use, they hold great promise for helping future pathologists,” said Dr. Orange. “Currently, we consider it a valuable tool for research purposes because it provides an accurate and 100% reproducible degree of inflammation and we look forward to its further development.”
In the future, Dr. Orange added, computer vision could be trained to gather other types of information from tissue samples, including what types of cells are present and whether they are close enough together that they are likely to communicate with each other. This more accurate assessment could enable doctors to more accurately know which cells are causing tissue damage and customize treatments accordingly.
Authors: Bella Mehta, MBBS, MSAnd the Susan M. Goodman, MDAnd the Edward F DiCarlo, MDAnd the Diana Jannatkhah, J Alex GibbonsAnd the Miguel Otero, Ph.D.And the Laura Donlin, Ph.D. (HSS), Tania Panellini, MD, PhD (Weill Cornell Medicine), William Robinson, MD, PhD (Stanford University), Peter K. Skolko, MDAnd the Mark P. Vigé, MDAnd the Jose A. Rodriguez, MD (HSS), Jessica Kirschman (Stanford University), James Thompson, David Slater, Damon Frieza (MITER Corporation), Zhenxing Xu, Fei Wang, Ph.D. (Weill Cornell Medicine), Dana Orange, MDAnd the Ms (HSS and Rockefeller University).
1. Orange D, Agios F, DiCarlo EF, et al. Identification of three subtypes of rheumatoid arthritis by integrating machine learning of synovial histological features and RNA sequencing data. rheumatoid arthritis. 2018; 70 (5): 690-701. doi: 10.1002/article 40428
2. Guan S, Mehta B, Slater D, et al. Quantitative rheumatoid arthritis using computer vision. ACR Open Rheumatol. 2022; 4 (4): 322-331. doi: 10.1002/acre 2.11381
HSS is the world’s leading academic medical center focused on musculoskeletal health. At its core it is the Hospital for Special Surgery, ranked No. 1 nationally in orthopedics (for the 12th consecutive year), No. 4 in rheumatology by U.S. News & World Report (2021-2022), and Best Children’s Orthopedic Hospital in New York, NJ and CT by US News & World Report list of “Best Children’s Hospitals” (2021-2022). In a survey of medical professionals in over 20 countries by Newsweek, HSS was ranked #1 globally in orthopedics for the second year in a row (2022). Founded in 1863, the hospital has the lowest complication and readmission rates in the state for orthopedic surgery, and among the lowest incidence rates. HSS was the first in New York State to receive a Magnet recognition for Excellence in Nursing Service from the American Nurses Credentialing Center five times in a row. Affiliated with Weill Cornell Medical College, HSS has major campuses in New York City and facilities in New Jersey, Connecticut and in the Long Island and Westchester County areas of New York State, as well as in Florida. In addition to patient care, HSS is a leader in research, innovation, and education. The HSS Research Institute has 20 laboratories and 300 employees focused on driving the advancement of musculoskeletal health through the prevention of degeneration, tissue repair, and tissue regeneration. The HSS Institute for Innovation is working to realize the potential of new drugs, treatments, and devices. HSS Education Institute is the trusted leader in advancing musculoskeletal knowledge and research for physicians, nurses, health professionals, academic trainees and consumers in more than 145 countries. The Foundation collaborates with medical centers and other organizations to enhance the quality and value of musculoskeletal care and to make world-class HSS care more widely available nationally and internationally. www.hss.edu.
Disclaimer: AAAS and EurekAlert! is not responsible for the accuracy of newsletters sent on EurekAlert! Through the contributing institutions or for the use of any information through the EurekAlert system.