MRI after unicondylar knee arthroplasty: rotational alignment of components

This pilot study used magnet resonance imaging (MRI) to analyse the rotation of medial unicondylar knee arthroplasty (UKA) components and assessed how accurately the results could be reproduced. Knee MRI using a special protocol to reduce metal artefact was performed in ten patients who had undergone medial UKA. Two independent investigators measured the rotation angle of femoral (zirconium) and tibial (cemented full-poly or cemented modular metal-backed) components applying different reference lines for the latter. Statistical analysis comprised tests for reliability, variance between measurement techniques, standard deviations and limits of agreement. For all methods tested, there was sufficient inter- and intra-observer reliability. Lowest variances were, however, found for the femoral epicondyles, for both femoral and tibial components. A tangent to the dorsal epicondyles of the tibia also gave reproducible results with low variances for the tibial component. Almost all applied measurement techniques were reproducible by statistical definition, although some of them resulted in substantial differences between both, observations and observers. A variance test helps to distinguish better between clinically useful and less accurate references. MRI allows good reproducible rotation analysis via the femoral epicondyles for both femoral and tibial UKA implants. For the tibia, the tibial tuberosity, the eminentia and the tibial epicondyles in particular were shown to be less reliable. The dorsal epicondyles seem to be most suitable for the tibial component.