Crowdsourcing is becoming an increasingly popular way to develop machine learning algorithms to address many clinical problems of a variety of diseases. Today, at the American College of Rheumatology (ACR) Annual Meeting, a multicenter team led by researchers at the Special Surgery Hospital (HSS) focused on developing better methods for quantifying joint injuries. We announced the results of the RA2-DREAM Challenge, which is an initiative of. People with rheumatoid arthritis (RA).
Joint damage in patients with rheumatoid arthritis is currently measured by visual inspection and detailed scoring of radiographic images of the small joints of the hands, wrists, and feet. This includes both narrowing of the joint space (indicating loss of cartilage) and bone erosion (indicating damage due to invasion of the lining of the inflamed joint). The scoring system requires specially trained professionals, which is time consuming and costly.Finding an automated way to measure joint damage is important for both Clinical research Lead authors of the study, S. Louis Bridges, Jr., MD, Ph.D. According to HSS’s doctor and director of medicine for the care of patients.
“Whether the machine learning approach can provide a fast and accurate quantitative score to estimate the next degree: Joint damage With limbs, it will be of great help to clinical research, “he said. Electronic health record From genetic and other research assays to find biomarkers associated with progressive injury. It’s tedious to visually inspect and score every image on our own, and outsourcing it can be exorbitant. “
“This approach may also assist rheumatologists by quickly assessing whether there is progression of injury over time, which will encourage changes in treatment to prevent further injury,” he said. Added. “This is very important in geographic areas where there is no specialized musculoskeletal radiologist.”
For the challenge, Dr. Bridges and his collaborators have partnered with Sage Bionetworks, a non-profit organization that helps researchers create DREAM (reverse engineering assessment and method dialogue) challenges. These contests focus on developing innovative artificial intelligence-based tools in life sciences. The investigator sent a call for submission, along with a grant to provide the winning team with a prize. Competitors came from a variety of disciplines, including computer scientists, computational biologists, and medical scientists. No radiologist had the expertise and training in reading radiographic images.
In the first part of the challenge, a set of images was provided to the team, along with a visually generated known score. These were used to train the algorithm. Next, an additional set of images was provided to allow you to test and improve tools developed by competitors. In the final round, a third set of images was provided without a score, and competitors estimated the amount of joint space narrowing and erosion. Submissions were judged according to the most exact reproduction of the gold standard visually generated score. There were 26 teams that submitted the algorithm and 16 final submissions. In total, competitors were provided with 674 sets of images from 562 different RA patients, all of whom participated in a previous National Institutes of Health-sponsored research study led by Dr. Bridges. In the end, four teams were selected as top performers.
It was important for the organizers of the DREAM Challenge that the scoring system developed through the project was free to use, not unique, for researchers and clinicians to use for free. “Part of the appeal of this collaboration was that it was all in the public domain,” said Dr. Bridges.
Dr. Bridges explained that additional research and development of computational methods is required before the tool can be widely used, but current research shows that this type of approach is feasible. “We still need to improve the algorithm, but we’re much closer to our goal than before the challenge,” he concludes.
Special Surgery Hospital
Quote: Global Efforts to Develop AI Tools for Automatic Assessment of Rheumatoid Arthritis Radiation Damage (November 6, 2021) https: //medicalxpress.com/news/2021-11-global-effort Obtained from -ai-tools-automated on November 6, 2021. .html
This document is subject to copyright. No part may be reproduced without written permission, except for fair transactions for personal investigation or research purposes. The content is provided for informational purposes only.
Global effort to develop AI tools for automated assessment of radiographic damage in rheumatoid arthritis Source link Global effort to develop AI tools for automated assessment of radiographic damage in rheumatoid arthritis
The post Global effort to develop AI tools for automated assessment of radiographic damage in rheumatoid arthritis appeared first on California News Times.