Reliability and Convergent Construct Validity of Quantitative Ultrasound for Synovitis, Meniscal Extrusion, and Osteophyte in Knee Osteoarthritis With MRI

Aims To determine: 1) inter‐rater reliability of quantitative measurements of ultrasound‐detected synovitis, meniscal extrusion, and osteophytes; and 2) construct (convergent) validity via correlations and absolute agreements between ultrasound‐ and gold‐standard magnetic resonance imaging (MRI)‐outcomes in knee osteoarthritis. Methods Dynamic ultrasound images for supra‐patellar synovitis, meniscal extrusion, and osteophytes were acquired and quantified by a physician operator, musculoskeletal ultrasonographer, and medical student independently. On the same day, 3T MRI images were acquired. Effusion‐synovitis, meniscal extrusion, and osteophytes were quantified on sagittal or coronal proton‐density‐weighted fat‐suppressed noncontrast TSE sequences, respectively. Intra‐class correlation coefficients (ICCs), Pearson's correlations (r), and Bland–Altman plots were used to analyze inter‐rater reliability, and correlations, and agreements between the two imaging modalities. Results Eighty‐nine participants [48 females (53.9%)] with mean (standard deviation) age of 61.5 ± 6.9 years were included. The inter‐rater reliability was excellent for osteophytes (ICC range = 0.90–0.96), meniscal extrusion (ICC range = 0.90–0.93), and synovitis (ICC range = 0.86–0.88). The correlations between ultrasound pathologies and their MRI counterparts were very strong (ICC range = 0.85–0.98) except for lateral meniscal extrusion [0.66 (95% CI, 0.52–0.76)]. Bland–Altman plots showed 0.01, 0.05, 0.10, 0.53, and 0.60 mm larger size in ultrasound medial tibial and medial femoral osteophytes, medial meniscal extrusions, synovitis, and lateral meniscal extrusions with 95% limits of agreements [±0.39, ±0.44, ±0.85, ±0.70, and ±0.90 (SDs)] than MRI measures, respectively. The lines of equality were within 95% CI of the mean differences (bias) only for medial osteophytes and medial meniscal extrusion. Conclusion The quantitative assessment of synovitis, meniscal extrusion, and osteophytes generally showed excellent inter‐rater reliability and strong correlations with MRI‐based measurements. Absolute agreement was strong for medial tibiofemoral pathologies.