Accounting for patient variability in finite element analysis of the intact and implanted hip and knee: A review

It is becoming increasingly difficult to differentiate the performance of new joint replacement designs using available preclinical test methods. Finite element analysis is commonly used and the majority of published studies are performed on representative anatomy, assuming optimal implant placement, subjected to idealised loading conditions. There are significant differences between patients and accounting for this variability will lead to better assessment of the risk of failure. This review paper provides a comprehensive overview of the techniques available to account for patient variability. There is a brief overview of patient-specific model generation techniques, followed by a review of multisubject patient-specific studies performed on the intact and implanted femur and tibia. In particular, the challenges and limitations of manually generating models for such studies are discussed. To efficiently account for patient variability, the application of statistical shape and intensity models (SSIM) are being developed. Such models have the potential to synthetically generate thousands of representative models generated from a much smaller training set. Combined with the automation of the prosthesis implantation process, SSIM provides a potentially powerful tool for assessing the next generation of implant designs. The potential application of SSIM are discussed along with their limitations.