Deep learning-based quantification of knee effusion-synovitis volume on MRI – Technique development and validation
Authors' conclusion
A DL-based model for automated quantification of knee ESV on MRI was developed and validated.
Study context
- Population
- Baseline right-knee MRI from the Osteoarthritis Initiative101 knees manually segmented
- Participants
- 4,698 participants
- Comparison
- Effusion-synovitis volume (ESV) vs semiquantitative effusion-synovitis scores
- Outcome
- Segmentation performance of the DL-based model (Dice coefficient)
Methodological detailsdeep learning-based model, trained and tested using manual segmentations from 101 randomly selected knees
- Deep learning-based model
- Trained and tested using manual segmentations from 101 randomly selected knees
Performance
As reported in the abstract
Segmentation performance (Dice coefficient), test set
Mean Dice coefficient
0.79
95% CI 0.71–0.86
Further findings
Correlation of ESV with semiquantitative effusion-synovitis (sqES) scores (Spearman)
- ESV vs WORMSES, ρ
- 0.50 (95% CI 0.47-0.53)
- ESV vs MOAKSES, ρ
- 0.65 (95% CI 0.63-0.67)
- Moderate correlations
Associations with WORMS features and knee OA symptoms: ESV vs sqES scores
- WORMS features, minimum Δβstd
- 0.03 (95% CI 0.00-0.07)
- Knee OA symptoms, minimum Δβstd
- 0.03 (95% CI 0.00-0.07)
- ESV demonstrated larger effect sizes than sqES scores
Original publication
Deep learning-based quantification of knee effusion-synovitis volume on MRI – Technique development and validation. Osteoarthritis Cartilage. Published 22 June 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →