Abstract
BACKGROUND: In patients with osteoarthritis, a detailed assessment of degenerative cartilage disease is important to recommend adequate treatment. Using a representative sample of patients, this study investigated whether MRI is reliable for a detailed cartilage assessment in patients with osteoarthritis of the knee. METHODS: In a cross sectional-study as a part of a retrospective case-control study, 36 patients (mean age 53.1 years) with clinically relevant osteoarthritis received standardized MRI (sag. T1-TSE, cor. STIR-TSE, trans. fat-suppressed PD-TSE, sag. fat-suppressed PD-TSE, Siemens Magnetom Avanto syngo MR B 15) on a 1.5 Tesla unit. Within a maximum of three months later, arthroscopic grading of the articular surfaces was performed. MRI grading by two blinded observers was compared to arthroscopic findings. Diagnostic values as well as intra- and inter-observer values were assessed. RESULTS: Inter-observer agreement between readers 1 and 2 was good (kappa = 0.65) within all compartments. Intra-observer agreement comparing MRI grading to arthroscopic grading showed moderate to good values for readers 1 and 2 (kappa = 0.50 and 0.62, respectively), the poorest being within the patellofemoral joint (kappa = 0.32 and 0.52). Sensitivities were relatively low at all grades, particularly for grade 3 cartilage lesions. A tendency to underestimate cartilage disorders on MR images was not noticed. CONCLUSIONS: According to our results, the use of MRI for precise grading of the cartilage in osteoarthritis is limited. Even if the practical benefit of MRI in pretreatment diagnostics is unequivocal, a diagnostic arthroscopy is of outstanding value when a grading of the cartilage is crucial for a definitive decision regarding therapeutic options in patients with osteoarthritis.