Reproducible assessment of radiolucent lines in total knee arthroplasty

The Knee Society Total Knee Arthroplasty Radiographic Evaluation and Scoring System was introduced to encourage uniform reporting of radiographic outcome. However, the method for evaluation of radiolucent lines has been shown to be unreliable. Because it has been shown that reducing the complexity of classification systems increases reliability and reproducibility, we questioned whether a simplification of the Radiographic Evaluation and Scoring System would improve reliability and reproducibility. A new system for assessment of radiolucent lines was introduced, and the interobserver reliability and intraobserver reproducibility were studied in 100 patients with 120 total knee replacements. For the new system the mean kappa intraobserver reproducibility coefficient was 0.71 (range, 0.62-0.85) for the femoral component, 0.86 (range, 0.80-0.96) for the tibial component, and 0.58 (range, 0.46-0.75) for the patella prosthesis. The mean interobserver reliability coefficient among three observers was 0.61 (range, 0.45-0.72) for the femoral component, 0.82 (range, 0.73-0.88) for the tibial component, and 0.58 (range, 0.43-0.72) for the patella prosthesis. The new system for assessment of radiolucent lines increased reliability and reproducibility and should supplement the Knee Society's Radiographic Evaluation and Scoring System.