Conceptual foundations of a REFRAME-based approach to discriminate across total knee implant designs based on the positions of functional centres of rotation

In modern knee arthroplasty, surgeons increasingly aim for individualised implant selection based on data-driven decisions to improve patient satisfaction rates. The identification of an implant design that optimally fits to a patient's native kinematic patterns and functional requirements could provide a basis towards subject-specific phenotyping. The goal of this study was to achieve a first step towards identifying easily accessible and intuitive features that allow for discrimination between implant designs based on kinematic data. A squat-cycle was simulated on eight fresh frozen specimens mounted in a weight-bearing knee rig, each initially tested under native conditions, and then after implantation with four different implant types (CR/CS, MS, LS, and PS). The kinematic signals of these five configurations were compared to determine whether key differences between implants could be detected leveraging two methodological approaches: (1) statistical parametric mapping to directly compare waveforms and (2) simple paired t-tests to compare the three-dimensional coordinates of the functional centres of rotation determined using a previously published REference FRame Alignment Method (REFRAME). While statistical parametric mapping of the kinematic data revealed only small differences in certain comparisons (e.g. LS vs. PS, and MS vs. LS) under lenient statistical testing conditions, the application of REFRAME showed clear differences between implants (for all implant combinations except for CR/CS vs. LS), even under conservative statistical testing. Since for most implant combinations, significant differences in the centres of rotation were found using REFRAME, this approach could present a suitable tool for discriminating between the kinematics of different implant types. Preoperative assessment of joint kinematics, combined with this REFRAME application, could therefore provide a key approach for improved clinical selection of implant type.