Patient‐Specific Finite Element Modelling and Validation of Porcine Femora in Torsion

ABSTRACT The accuracy of biomechanical simulation has been improved by using high‐resolution computed tomography (CT) to define the geometry and material parameters. This technique has been used to assess numerous systems, including the mechanical properties of bone, fixation techniques post‐fracture and the performance of bone microarchitecture. In this study, a semi‐automated process for converting CT data into finite element (FE) models was used to model the mid‐shaft (diaphysis) of porcine femoral samples under sub‐maximal torsional and compressive load. Physical validation was undertaken to investigate if the all‐important geometry and material property mapping functioned correctly. Porcine femoral specimens were imaged using contiguous helical CT, which was converted to FE models using ScanIP from Simpleware, Exeter, UK. The heterogeneous material properties were estimated using density–elasticity relationships proposed in literature for human bone samples. Laboratory testing performed favourably, with a linear strain response validating the use of the array of linear material models used in simulation. The simulation procedure also performed well. Linear regression and mean error calculation demonstrated accurate correlation between predicted (from simulation) and observed (measured within the laboratory) results that offered improvement over the accuracy within comparative testing for human samples. Using FE modelling on a patient‐specific basis offers potential in a number of scenarios, including the determination of injury risk and design of protective equipment. The increased accessibility of animal samples allows large‐scale fracture testing of complex loading mechanisms and the potential to consider younger animal samples (to investigate the behaviour of developing bone). Spiral fractures of long bones have been demonstrated to be an indicator of non‐accidental injury in children. Combining the increased accuracy in torsional simulation in this study with younger sample testing may be employed to attempt to determine the causes of fracture from post fracture scans, aiding in the diagnosis of non‐accidental injury.