Finite element modeling of the human thoracolumbar spine.

STUDY
DESIGN: Biomechanical properties within cadaveric vertebral bodies were parametrically studied using finite element analysis after calibration to experimental data.
OBJECTIVES: To develop and validate three-dimensional finite element models of the human thoracolumbar spine based on quantitative computed tomography scans. Specifically, combine finite element modeling together with biomechanical testing circumventing problems associated with direct measurements of shell properties. SUMMARY OF BACKGROUND DATA: Finite element methods can help to understand injury mechanisms and stress distribution patterns within vertebral bodies as an important part in clinical evaluation of spinal injuries. Because of complications in modeling the vertebral shell, it is not clear if quantitative computed tomography-based finite element models of the spine could accurately predict biomechanical properties.
METHODS: We developed a novel finite element modeling technique based on quantitative computed tomography scans of 19 radiographically normal human vertebra bodies and mechanical property data from empirical studies on cylindrical trabecular bone specimens. Structural properties of the vertebral shell were recognized as parametric variables and were calibrated to provide agreement in whole vertebral body stiffness between model and experiment. The mean value of the shell properties thus obtained was used in all models to provide predictions of whole vertebral strength and stiffness.
RESULTS: Calibration of n = 19 computer models to experimental stiffness yielded a mean effective modulus of the vertebral shell of 457 +/- 931 MPa ranging from 9 to 3216 MPa. No significant correlation was found between vertebral shell effective modulus and either the experimentally measured stiffness or the average trabecular modulus. Using the effective vertebral shell modulus for all 19 models, the predicted vertebral body stiffness was an excellent predictor of experimental measurements of both stiffness (r2= 0.81) and strength (r2 = 0.79).
CONCLUSION: These findings indicate that modeling of the vertebral shell using a constant thickness of 0.35 mm and an effective modulus of 457 MPa, combined with quantitative computed tomography-based modeling of trabecular properties and vertebral geometry, can accurately predict whole vertebral biomechanical properties. Use of this modeling technique, therefore, should produce substantial insight into vertebral body biomechanical behavior and may ultimately improve clinical indications of fracture risk of this cohort.