Interobserver Agreement of Post-operative Perth Computed Tomography Protocol Data in Total Knee Arthroplasty

Background: Total knee arthroplasty (TKA) has become the treatment of choice for advanced osteoarthritis. Identifying malalignment is central to improving TKA outcomes and providing optimal management of TKA patients with post-operative pain and dissatisfaction. Computed tomography (CT) imaging has become increasingly popular as a more precise way of analysing post-TKA component alignment and the Perth CT protocol remains the current predominant assessment tool. This study aimed to analyse and compare inter- and intra-observer agreement of a post-operative multi-parameter quantitative CT assessment (Perth CT protocol) in TKA patients. Methods: Post-operative CT images of 27 patients who underwent TKA were analysed retrospectively. Images were analysed by an experienced radiographer and a final-year medical student at least two weeks apart. Measurements for nine angles were collected: modified hip-knee-ankle (mHKA) angle, lateral distal femoral angle (LDFA) and medial proximal tibial angle (MPTA), femoral flexion and tibial slope, femoral rotation angle, femoral-tibial match rotational angle, tibial tubercle lateralisation distance, and Berger’s tibial rotation. Intra-observer and inter-observer intraclass correlation coefficients (ICCs) were calculated. Results: Inter-observer reliability for the measurements of all variables varied from poor to excellent (ICC: -0.003 to 0.981). Five out of the nine angles demonstrated good to excellent reliability. Inter-observer reliability was highest for mHKA in the coronal plane and the poorest for the tibial slope angle in the sagittal plane. The intra-observer reliability for both reviewers was excellent (0.999 vs. 0.989). Conclusion: This study demonstrates that the Perth CT protocol has excellent intra-observer reliability and good to excellent inter-observer reliability for five out of nine of the measured angles used to assess component alignment post-TKA, making it a useful tool for surgical outcome prediction and success.