auto scoRA : Deep Learning to Automate Sharp/van der Heijde Scoring of Radiographic Damage in Rheumatoid Arthritis
Study context
- Population
- Adult patients with RA; radiographs of the hands and feet155 in test set
- Participants
- 769 patients
- Comparison
- autoscoRA automated Sharp/van der Heijde (SvdH) scoring vs human scorer
- Outcome
- Agreement of the model with a human scorer for joint space narrowing, erosion and combined scores, joint level and summed total SvdH scores
Methodological detailsdeep learning-based, fully automated
- Deep learning-based
- Fully automated
Performance
As reported in the abstract
Intraclass correlation (excellent agreement with a human scorer)
0.9
Further findings
| Outcome | autoscoRA |
|---|---|
| Agreement with human reader for detecting longitudinal progression of joint damage across SvdH cutoffs | |
| Average agreement | 70% |
Agreement with first human reader, subset scored by a second human reader
- Model outperformed the second human reader in agreement with the first human reader
Original publication
auto scoRA : Deep Learning to Automate Sharp/van der Heijde Scoring of Radiographic Damage in Rheumatoid Arthritis. Arthritis Rheumatol. Published 21 April 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →