Early Detection of Osteoarthritis in the Knee Using Echo Modulation Curve-based T2 Mapping: Analysis in the Osteoarthritis Initiative Dataset.

Motivation: Early detection of knee osteoarthritis relies on sensitive biomarkers for cartilage changes, yet conventional mono-exponential T2 mapping models are not accurate due to unaccounted physical effects. Goal(s): To implement and evaluate Echo-Modulation-Curves (EMC) T2 fitting to improve reproducibility and predictive power for early detection of osteoarthritis compared to standard models. Approach: T2 maps were estimated using mono-exponential (with and without the first echo), and EMC fitting. Reproducibility was assessed with Bland-Altman analysis, while predictive performance was evaluated using logistic regression and ROC analysis. Results: EMC fitting outperformed mono-exponential methods in both reproducibility and predictive accuracy, demonstrating stronger discrimination between controls and osteoarthritis incidence. Impact: EMC T2 fitting shows higher reproducibility and sensitivity in detecting early OA in the knee cartilage compared to mono-exponential models, suggesting that accurate T2 estimation can help identifying cartilage changes before radiographic evidence, potentially enabling earlier intervention strategies.