Association between the cartilage area ratio derived from 3D MRI and patient-reported symptoms in medial knee osteoarthritis

BACKGROUND: Knee osteoarthritis (OA) is characterized by progressive cartilage deterioration in the medial compartment. Three-dimensional (3D) MRI enables quantitative assessment of cartilage morphology, and the cartilage area ratio is a sensitive metric that integrates focal defects. This study investigated associations between medial compartment cartilage area ratios and numerical rating scale (NRS) scores in patients undergoing high tibial osteotomy (HTO). METHODS: We analyzed 176 patients who underwent HTO for medial knee OA. All had preoperative 3.0 T MRI and completed a 10-item NRS questionnaire. Total NRS was the sum of Q1-Q4 and Q6-Q10 (Q5 excluded; maximum 90). Cartilage area ratios for the medial femur and medial tibia were calculated using automated 3D MRI analysis, and the medial meniscus coverage ratio was also measured. Item-level associations were assessed using Spearman correlation with Holm correction. Multivariable linear regression of total NRS adjusted for age, sex, body mass index (BMI), and Kellgren-Lawrence (KL) grade was performed. Subgroup analysis in KL grade 3-4 patients and sensitivity analyses progressively excluding high cartilage area ratio values were conducted. RESULTS: Total NRS scores showed significant negative correlations with both medial femoral (rs = - 0.31) and medial tibial (rs = - 0.26) cartilage area ratios. Item-level analyses showed the strongest correlations for instability during running (Q8) and grinding/noise (Q9). In multivariable regression, both medial femoral (standardized β = - 0.245, P = 0.007) and medial tibial (standardized β = - 0.226, P = 0.012) cartilage area ratios remained independently associated with total NRS, together with BMI, while age, sex, and KL grade showed no significant associations. The KL grade 3-4 subgroup retained significant correlations (n = 95; femoral rs = - 0.26; tibial rs = - 0.26). Sensitivity analyses confirmed that the inverse correlations were not driven by ceiling effects. Meniscus coverage ratios correlated positively with cartilage area ratios (medial femoral rs = 0.41; medial tibial rs = 0.47) but not with total NRS. CONCLUSIONS: Lower cartilage area ratios in the medial compartment are independently associated with greater symptom severity, particularly instability during running and mechanical symptoms. This supports cartilage area ratios as clinically relevant structural biomarkers in knee OA.