Computationally efficient prediction of bone-implant interface micromotion of a cementless tibial tray during gait

A proximal tibial bone with mapped material properties was virtually implanted with a tibial tray. A FE model, with six-degree-of-freedom loads sampled from telemetric patients during walking, was used to generate training data for the surrogate model. The linear response surrogate model was evaluated for six full gait cycles; the average and peak micromotion across the interface, and the percentage of bone-implant interface surface area experiencing micromotions less than 50 and greater than 150 µm were calculated both as a function of the activity cycle and as the composite peak micromotion throughout the cycle. Differences in root-mean-square (RMS) micromotion between FE and surrogate models were less than 14 µm. FE analysis time for a complete gait cycle was 15 h, compared to 30 s for the surrogate model. Surrogate models have significant potential to rapidly predict micromotion over the entire bone-implant interface, allowing greater range in loading conditions to be explored than is possible through conventional methods.