Generation of a statistical model of the whole femur incorporating shape and material property distribution

Download

PDF - Pre print Download (264Kb)

Description/Abstract

Orthopaedic implant designs undergo extensive preclinical testing to ensure their reliability but in use there performance can vary significantly between patients. A possible reason for this is that vey few computational studies have incorporated interpatient variability. The vast majority use only a single bone model as constructing multiple models from souces such as CT scans is a time consuming and laborious task. This is despite evidence from recent studies indicating that bone geometry and bone quality have a significant effect on implant performance . The objective of this study is to build a statistical model of the whole femur which incorporates both the natural shape and material property variations which occur between patients. The model will be able to generate new femur instances and suggest the principal modes by which the femur varies in the population.