Probabilistic Analysis of the Material and Shape Properties for Human Liver

Realistic assessments of liver injury risk for the entire occupant population require incorporating inter-subject variations into numerical human models. The main objective of this study was to quantify the variations in shape and material properties of the human liver. Statistical shape analysis was applied to analyze the geometrical variation using a surface set of 15 adult human livers recorded in an occupant posture. Principal component analysis was then utilized to obtain the modes of variation, the mean model, and a set of 95% statistical boundary shape models. Specimen-specific finite element (FE) models were employed to quantify material and failure properties of human liver parenchyma. The mean material model parameters were then determined, and a stochastic optimization approach was utilized to determine the standard deviations of the material model parameters. The distributions of the material parameters were used to develop probabilistic FE models of the liver implemented in THUMS human FE model to simulate oblique impact tests under three impact speeds. In addition, the influence of organ preservation on the biomechanical responses of animal livers was investigated using indentation and tensile tests.

Results showed that the first five modes of the human liver shape models accounted for more than 70% of the overall anatomical variations. The Ogden material model with two parameters showed a good fit to experimental tensile data before failure. Significant changes of the biomechanical responses of liver parenchyma were found after cooling or freezing storage. The force-deflection responses of THUMS model with probabilistic liver material models were within the test corridors obtained from cadaveric tests. Significant differences were observed in the maximum and minimum principal Green-Lagrangian strain values recorded in the THUMS liver model with the default and updated average material properties. The results from this study could help in the development of more biofidelic human models, which may provide a better understanding of injury mechanisms of the liver during automobile collisions.