Intra- and inter-observer reliability of implant positioning evaluation on a CT-based three-dimensional postoperative matching system for total knee arthroplasty.

Background The evaluation of postoperative total knee arthroplasty (TKA) alignment mainly relies on measurement data obtained from plain radiographs. The aim of this retrospective observational study was to document the intra- and inter-observer reliability in assessment of TKA component positioning after surgery using a three-dimensional (3D) computed tomography (CT) image matching system. Methods Fourteen knees from 14 patients who received primary TKA were included, and images were analyzed by blinded readers not associated with the surgeries. The examiner digitized the reference points according to defined landmarks, and the designated size component was superimposed to the 3D reconstructed CT model for measurement. In addition to the evaluation of implant position against the coronal and sagittal lower limb mechanical axes that were defined based on bony landmarks, implant position against axes connecting implant-based reference points that are easier to indicate was evaluated. Results The overall intra- and inter-observer reliabilities determined by the intraclass correlation coefficients (ICC) of the implant alignment measurement for both femoral and tibial components were good (ICC > 0.60), except in the direction of femoral flexion and extension, for both mechanical and implant-based axes. The difference between implant alignment measurements according to the traditional mechanical axis and the implant-based axis ranged between means of 0.08o and 1.70o and were statistically significantly different. Conclusions The postoperative evaluation of implant position in the coronal and sagittal planes using 3D-CT image matching is reliable and has good reproducibility except for the sagittal alignment assessment of the femoral component. The measured implant position according to the traditional mechanical axis and the implant-based axis were slightly but significantly different.