2023-06-08 06:05:54
Artificial intelligence (AI) is playing an increasingly important role in medicine. At the European Congress of Rheumatology (Eular), which just ended in Milan, Dutch researchers demonstrated the power of this new tool, which could help treat rheumatic diseases earlier.
Early inflammatory arthritis is a clinical condition that is difficult to differentiate but has the potential to progress to rheumatoid arthritis (RA) or another form of arthropathy. In some cases, it may regress spontaneously or remain undifferentiated indefinitely. An important prognostic factor in this course is the presence of joint erosions, which can be detected with magnetic resonance imaging (MRI). MRI also allows direct evaluation of (teno-)synovitis and bone marrow edema.
Early prediction of rheumatoid arthritis based on MRI images of the hands and feet can be of great help to patients, who can ‘benefit’ from early treatment that can change the course of the disease. Typically, radiologists and rheumatologists use manual scoring systems to identify key features on MRI images.
A recent study by Li and colleagues at Leiden University Medical Center in the Netherlands investigated the use of artificial intelligence (AI) and deep learning to automatically analyze MRI scans and predict early rheumatoid arthritis in patients with symptoms of clinically suspicious arthralgia. The AI model was trained in several stages, first to understand the anatomy, then to distinguish rheumatoid arthritis patients from healthy subjects and finally to identify image features that are predictive of disease development.
The study involved the analysis of MRI scans of the hands and feet of 1,974 people with symptoms of onset arthritis or clinically suspected arthralgia, of which 651 went on to develop rheumatoid arthritis. The results of the studies showed that the AI model could predict rheumatoid arthritis with an accuracy comparable to that of human experts. The authors of the study conclude that automatic interpretation of MRI images by AI could enable automatic prediction of rheumatoid arthritis.
The researchers emphasize the importance of continuing to train the model using more MRI data from healthy subjects, which could further improve the accuracy of the predictions. Future research will also focus on the prediction of rheumatoid arthritis in the specific subgroup of undifferentiated arthritis.
In addition, this innovative approach not only confirms the importance of known inflammatory markers, such as synovial inflammation, but also paves the way for the identification of novel imaging biomarkers that may improve the understanding of the underlying pathological process in early rheumatoid arthritis.
Access to all functionalities is reserved for healthcare professionals.
If you are a healthcare professional, you must log in or register for free to get full access to this content.
If you are a journalist or if you wish to inform us, please write to us [email protected].
#indispensable #tool #modern #imaging