The KNee OsteoArthritis Prediction (KNOAP2020) challenge: An image analysis challenge to predict incident symptomatic radiographic knee osteoarthritis from magnetic resonance imaging (MRI) and X-ray images
Authors' conclusion
The KNOAP2020 challenge established a benchmark for predicting incident symptomatic radiographic knee osteoarthritis.
Study context
- Population
- Test set of knees from the PROOF study (overweight females) with baseline MRI, X-ray and clinical risk factors
- Participants
- 423 knees
- Comparison
- Submitted prediction models for incident symptomatic radiographic knee OA within 78 months
- Outcome
- Prediction of incident symptomatic radiographic knee OA (combined ACR criteria), assessed by ROCAUC and balanced accuracy (BACC)
- Follow-up
- 78 months
Methodological detailsimage analysis challenge, blinded ground truth, Most algorithms trained on Osteoarthritis Initiative data
- Image analysis challenge
- Blinded ground truth
- Most algorithms trained on Osteoarthritis Initiative data
Performance
As reported in the abstract
Model performance
| Test or model | ROCAUC | Balanced accuracy (BACC) |
|---|---|---|
| Highest ROCAUC model | 0.64 (95% CI: 0.57-0.70) | |
| Highest BACC model | 0.59 (95% CI: 0.52-0.65) |
ROCAUC
Highest ROCAUC model0.64 (95% CI: 0.57-0.70)
Highest BACC model— not stated in the abstract
Balanced accuracy (BACC)
Highest ROCAUC model— not stated in the abstract
Highest BACC model0.59 (95% CI: 0.52-0.65)
Original publication
The KNee OsteoArthritis Prediction (KNOAP2020) challenge: An image analysis challenge to predict incident symptomatic radiographic knee osteoarthritis from MRI and X-ray images. Osteoarthritis Cartilage. Published 12 October 2022. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →