Background: Identification of femoral component debonding in the work-up of painful total knee arthroplasty (TKA) often poses a diagnostic challenge. The purpose of this study was to compare the sensitivity and specificity of radiographs, computed tomography (CT), and magnetic resonance imaging (MRI) for identifying femoral component loosening with debonding at the time of revision of a primary TKA with a recalled polyethylene insert.
Methods: Using an institutional database, we identified all cases of revision TKA performed for this specific implant recall following a primary TKA between 2014 and 2022. Patients who had a preoperative radiograph, CT, and MRI were included (n = 77). Sensitivity, specificity, and positive and negative likelihood ratio (LR) for predicting loosening were compared among the imaging modalities, using the intraoperative evidence of implant loosening as the gold standard.
Results: At the time of revision surgery, the femoral component was noted to have aseptic loosening with debonding in 46 of the 77 (60%) of the TKAs. There were no significant differences in demographics in the cohort with femoral debonding compared to those with well-fixed implants. The CT demonstrated a sensitivity of 28% and a specificity of 97%, while the MRI demonstrated a sensitivity of 37% and a specificity of 94% for detecting femoral loosening due to debonding. Both CT and MRI demonstrated poor negative LRs for femoral loosening (LR 0.7).
Conclusions: In this series of revision TKAs for a specific recalled component, neither CT nor MRI accurately diagnosed femoral component debonding. For patients who have this implant, it is imperative to interrogate the implant-cement interface intraoperatively and prepare for full revision surgery as well as marked bone loss secondary to osteolysis.
Keywords: aseptic loosening; computed tomography; femoral debonding; magnetic resonance imaging; revision; total knee arthroplasty.
Copyright © 2024 Elsevier Inc. All rights reserved.