X-Ray Radiomics Predicts Knee Osteoarthritis Pain Progression
Researchers used FNIH Osteoarthritis Biomarkers Project x-ray data (follow-up 24–48 months) to build and validate nomogram models combining subchondral bone radiomics and clinical features to predict knee osteoarthritis (KOA) pain progression in 450 participants. Using LASSO-selected features and SHAP for interpretability, the nomograms achieved AUCs of 0.766 and 0.753, showed good calibration (mean absolute errors 0.012 and 0.008), and offered positive net benefit across thresholds.
Key Points
- 1Identify Wavelet-HH_gldm_HighGrayLevelEmphasis as the primary radiomics predictor of KOA pain progression
- 2Demonstrate nomograms achieve AUCs 0.766 and 0.753 with strong calibration and decision-curve benefit
- 3Enable clinicians to stratify KOA patients for early intervention using cost-effective x-ray based models
Scoring Rationale
Solid predictive performance and clinical applicability, balanced by moderate novelty and need for external validation.
Sources
Public references used for this report.
Practice with real Ad Tech data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Ad Tech problems