Deep learning enables automatic detection of joint damage progression in rheumatoid arthritis—model development and external validation
Authors' conclusion
AuRA had the best external validation performance and demonstrated potential for detecting longitudinal changes in joint damage.
Study context
- Population
- Hand and foot radiographs of patients with RA; training data from two previous clinical studies, external validation data from Turku University Hospitalexternal validation cohort
- Participants
- 205 participants
- Comparison
- AuRA vs two top-performing RA2-DREAM algorithms
- Outcome
- Agreement with expert-assessed Sharp-van der Heijde total scores (RMSE) in the external validation cohort
Methodological detailsmodel development, external validation, Training data n = 367 from two previous clinical studies
- Model development
- External validation
- Training data n = 367 from two previous clinical studies
Performance
As reported in the abstract
RMSE vs expert-assessed scores, external validation cohort
| Test or model | Value | Root mean square error (RMSE) |
|---|---|---|
| AuRA | 23.6 | 23.6 |
| RA2-DREAM solutions | 35.0 |
Value
AuRA23.6
RA2-DREAM solutions— not stated in the abstract
Root mean square error (RMSE)
AuRA23.6
RA2-DREAM solutions35.0
Further findings
Correlation of AuRA-predicted longitudinal changes with changes in expert-assessed scores (54 patients, average interval 4.6 years)
- Pearson's R
- 0.74
- P
- P < 0.001
Original publication
Deep learning enables automatic detection of joint damage progression in rheumatoid arthritis—model development and external validation. Rheumatology (Oxford). Published 10 April 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →