Reliability of radiographic knee and hip osteoarthritis classification between radiologists and orthopaedic surgeons

Abstract Introduction Osteoarthritis (OA) is a common and disabling joint condition that requires accurate assessment for effective management. Variability in radiographic classification of OA severity can impact diagnostic accuracy and treatment decisions. This study evaluates the interobserver reliability of radiographic grading systems for hip and knee OA severity. Methods This retrospective study included 576 patients diagnosed with hip or knee OA between 2016 and 2023 who underwent radiographic evaluation. Radiographs were independently scored by three orthopaedic surgeons and three radiologists using established classification systems (Kellgren-Lawrence (K-L), Croft for hip; K-L, Ahlbäck, Brandt and Osteoarthritis Research Society International (OARSI) for knee OA). Interobserver agreement was assessed using kappa statistics. Results A total of 296 hip OA patients (mean age 69.5 ± 8.7, 62% female, 67% underwent surgery) and 280 knee OA patients (mean age 69.8 ± 8.4 years, 41% female, 53% underwent surgery) were assessed. For hip OA, interobserver agreement was moderate using K-L (κ = 0.552) and Croft (κ = 0.468). For knee OA, moderate agreement was found for the KL (κ = 0.598) and Ahlbäck (κ = 0.463), and substantial for Brandt (κ = 0.625), and OARSI (κ = 0.648) classifications. Overall, radiologists assigned higher OA severity scores than orthopaedic surgeons. Conclusion This study shows fair to moderate agreement between orthopaedic surgeons and radiologists in diagnosing hip and knee OA. Based on our data, we recommend the Kellgren-Lawrence as the grading system for lower extremity osteoarthritis. However, grading minute osteophytes and early-stage osteoarthritis remains challenging.