Related Experiment Video
Updated: Jun 28, 2025
06:31
Author Spotlight: Enhancing Rheumatoid Arthritis Research Through HR-pQCT Imaging Analysis
Published on: October 6, 2023
2.1K
Deep learning enables automatic detection of joint damage progression in rheumatoid arthritis-model development and
Mikko S Venäläinen1,2, Alexander Biehl1, Milja Holstila3
1Turku Bioscience Centre, University of Turku and Åbo Akademi University, Turku, Finland.
Rheumatology (Oxford, England)
|April 10, 2024
Summary
This study introduces AuRA, an automated algorithm for scoring rheumatoid arthritis (RA) joint damage. AuRA demonstrates strong performance in external validation and shows potential for monitoring radiographic progression in RA patients.
Area of Science:
- Radiology
- Artificial Intelligence
- Rheumatology
Background:
- Deep learning shows promise for quantifying joint damage in rheumatoid arthritis (RA).
- Evidence for detecting longitudinal changes at an individual patient level using AI is limited.
- Automated scoring algorithms can potentially reduce the burden of manual assessment.
Purpose of the Study:
- To introduce and externally validate an automated RA scoring algorithm (AuRA).
- To demonstrate AuRA's utility in monitoring radiographic progression in a real-world setting.
- To compare AuRA's performance against existing top-performing algorithms.
Main Methods:
- The AuRA algorithm was trained on expert-curated Sharp-van der Heijde scores from RA radiographs.
- External validation was performed using a separate cohort of RA patient radiographs.
- Longitudinal changes were assessed by comparing radiograph sets over an average 4.6-year interval.
Main Results:
- AuRA achieved a lower root mean square error (23.6) compared to top RA2-DREAM algorithms (35.0 and 35.6) in external validation.
- Performance improvements were most notable at higher expert-assessed damage scores.
- Predicted longitudinal changes by AuRA significantly correlated with expert assessments (Pearson's R = 0.74, P < 0.001).
Conclusions:
- AuRA exhibits superior external validation performance for RA radiographic assessment.
- The algorithm shows significant potential for detecting longitudinal changes in joint damage.
- AuRA is available for application in automatic radiographic progression detection, reducing manual scoring needs.
