Fig. 4: Mechanobiological design optimization: Image-based FE analysis with mechanical testing validation, predicted fracture propagation using XFEM, and simulated bone ingrowth over 6 months using time-dependent mechanobiological numerical modelling.

A Schematic of the cutting guide, LS-PEK frame, βTCP lattice, and LS-PEK restraining crossbar customized to live and cadaveric sheep used for design, biomechanical in silico modelling, and mechanical testing (cadaveric only). The single gyroid TPMS scaffold is shown including parameters of the gyroid unit cell. B Image-based FE model of whole-mandible and artificial bone and in vitro mechanical testing set-up simulating incisal loading from action of masseter and medial pterygoid musculature. Force-displacement curve from mechanical testing showing implant failure at 1533 N. The LS-PEK implant fractured in the predicted locations during mechanical testing validating the XFEM-predicted crack propagation pattern (black dashed line). C von Mises and (D) maximum principal stress distributions from image-based FE simulations used to iteratively modify the design of the LS-PEK frame. E–H Bone growth mechanobiological simulations over 24 weeks for different contact boundary conditions at the interface between natural bone and the scaffold component of the LS-PEK frame, as well as at the interface between the LS-PEK scaffold and βTCP lattice. E Simulation assuming no contribution from the βTCP lattice. F Simulation assuming frictional contact conditions of the βTCP lattice. G Simulation assuming bonded contact conditions of the βTCP lattice. For E–G coloured regions represent newly formed bone with a density above 0.3 g/cm³. Dark blue indicates low-density bone (0.3 g/cm³), transitioning through green and yellow (0.5 g/cm³), to pink, which represents high-density bone (0.8 g/cm³). H Mean ( ± SD) bone mass from the three simulations (E, F) and G The estimated bone mass at week 24 was 3.89 g. Linear regression analysis established a significant relationship between bone growth and time after implantation, F(1,2) = 92.34, p = 0.0107, with a slope of 0.1097 and R-squared = 0.9788. Linear regression plot presents line of best fit (solid line) and 95% confidence interval (dotted line). Source data are provided as a Source Data file.