New insights on the proximal femur biomechanics using Digital Image Correlation

Abstract

Finite element analyses (FEAs) of human femurs are mostly validated by ex-vivo experimental observations. Such validations were largely performed by comparing local strains at a small subset of points to the gold standard strain gauge (SG) measurements. A comprehensive full field validation of femoral FEAs including both strains and displacements using digital image correlation (DIC) full field measurements, especially at medial and lateral surfaces of the neck that experience the highest strains, provide new insights on femurs' mechanical behavior. Five cadaver femurs were loaded in stance position and monitored at the shaft and neck using two DIC systems simultaneously. DIC strains were compared to SG measurements at a limited number of locations so to corroborate DIC measurements by the gold standard technique. These were used to quantitatively assess the validity of FEA strains prediction especially at the neck where fracture usually occurs. Strains measured by DIC correspond well to the SG observations. An excellent agreement was observed between DIC and FEA predicted strains excluding the superior neck surface: FE=1.02×DIC-17,r²=0.977. At the superior neck however, strains were not well predicted by FEA models: Although the FEA predicted high strains at the 'saddle region', these were not observed experimentally. On the other hand, strain concentrations were measured by DIC at numerous vessel holes which were not represented by FE models. Since fractures usually initiate at the subcapital region in stance position ex-vivo experiments, where numerus vessel holes exist, these vessel holes may be required to be accounted for in future FE models so to allow a better estimation of the fracture load. Full field measurements are mandatory to allow a better validation of fracture load and location predictions which are of high clinical importance.

Keywords: Digital image correlation; Femur; Finite element analysis.