Spondyloarthritis & psoriatic arthritis
Deep learning algorithm for semiquantification of spinal inflammation in axial spondyloarthritis
Authors' conclusion
The high consistency of the scoring pipeline with human readers suggested that the deep learning-based algorithm has the potential to provide semiquantitative assessment of spinal inflammation based on SPARCC in axial SpA.
Study context
- Population
- Patients with axSpA undergoing whole spine MRI with STIR at 3T
- Participants
- 330 patients
- Comparison
- Deep learning-based Spondyloarthritis Research Consortium of Canada (SPARCC) scoring pipeline vs human readers
- Outcome
- Agreement (ICC) and correlation (Pearson) of SPARCC scores between human readers and deep learning pipeline
Methodological detailsdeep learning (attention Unet), whole spine MRI, STIR, 3T
- Deep learning (attention Unet)
- Whole spine MRI, STIR, 3T
Performance
As reported in the abstract
SPARCC scores, human readers vs deep learning pipeline
Intraclass correlation coefficient (ICC)
0.80
Pearson coefficient
0.82
Further findings
| Outcome | Deep learning pipeline |
|---|---|
| Spinal inflammation identification (BMO model) | |
| Sensitivity | 0.90 |
| Specificity | 0.84 |
Localisation in images with spinal inflammation (VB-IVD model), Dice coefficient
- Dice coefficient, vertebral body (VB)
- 0.81
- Dice coefficient, intervertebral disc (IVD)
- 0.80
Original publication
Deep learning algorithm for semiquantification of spinal inflammation in axial spondyloarthritis. RMD Open. Published 16 January 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →