Spondyloarthritis & psoriatic arthritis
Performance of artificial intelligence tools in axial spondyloarthritis imaging assessment: A systematic literature review and meta-analysis
Authors' conclusion
Although AI showed good performance, human expert remains essential to reach diagnostic accuracy while maintaining clinical safety.
Study context
- Population
- Studies of Artificial Intelligence (AI) imaging assessment in axial spondyloarthritis (axSpA), PubMed and Scopus, 1st January 2010 to 7th June 202533 studies
- Comparison
- AI techniques vs human reader in interpreting imaging in axSpA
- Outcome
- Performance of AI (sensitivity, specificity, accuracy, ROC AUC) at subject level
Methodological detailsPRISMA, DerSimonian-Laird random effect model, subject and image levels, 1033 references identified, 46 full texts reviewed, Studies published 2020–2025, 58% in 2024/2025; 67% from Asia, 64% included MRI; 85% applied deep learning
- PRISMA
- DerSimonian-Laird random effect model
- Subject and image levels
- 1033 references identified, 46 full texts reviewed
- Studies published 2020–2025, 58% in 2024/2025; 67% from Asia
- 64% included MRI; 85% applied deep learning
Pooled result
As reported in the abstract
Pooled performance of AI tools in axSpA imaging assessment, subject level
Sensitivity
87%
95% CI 85; 90%
Specificity
80%
95% CI 75; 85%
Accuracy
84%
95% CI 81; 87%
ROC AUC
0.88
95% CI 0.86; 0.90
Further pooled results
Overall AI performance at image level
- On image level the results were similar
Original publication
Performance of artificial intelligence tools in axial spondyloarthritis imaging assessment: A systematic literature review and meta-analysis. Joint Bone Spine. Published 30 January 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →