Spondyloarthritis & psoriatic arthritis
Quantitative prediction of radiographic progression in patients with axial spondyloarthritis using neural network model in a real-world setting
Authors' conclusion
Clinical and radiographic data-driven ANN allows precise mSASSS prediction in real-world settings.
Study context
- Population
- Patients with axial spondyloarthritis (axSpA) in a real-world setting
- Participants
- 555 patients
- Comparison
- Artificial neural network (ANN) models vs generalized linear model (GLM) for predicting modified Stoke Ankylosing Spondylitis Spine Score (mSASSS)
- Outcome
- Prediction of mSASSS of the radiographs at follow-up time points
Methodological detailsreal-world setting, training and testing datasets at a 3:1 ratio, explainable machine learning
- Real-world setting
- Training and testing datasets at a 3:1 ratio
- Explainable machine learning
Performance
As reported in the abstract
mSASSS prediction error (RMSE)
ANN (mSASSS as number of each score)
2.83
Root mean squared error (RMSE)
GLM or ANN (mSASSS as total score)
2.99-3.57
RMSE
- R2 of the fitted models
- 0.90-0.95
Further findings
- Structural changes constituting the mSASSS scoring systems were the most important contributing factors
- No detectable structural abnormalities at baseline were the most significant factors suppressing mSASSS change
Original publication
Quantitative prediction of radiographic progression in patients with axial spondyloarthritis using neural network model in a real-world setting. Arthritis Res Ther. Published 20 April 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →