Development and Validation of a Computationally Efficient Finite Element Model of the Human Lumbar Spine: Application to Disc Degeneration.

INTRODUCTION

This study develops and validates an accurate, computationally efficient, 3-dimensional finite element model (FEM) of the human lumbar spine. Advantages of this simplified model are shown by its application to a disc degeneration study that we demonstrate is completed in one-sixth the time required when using more complicated computed tomography (CT) scan-based models.

METHODS

An osseoligamentous FEM of the L1-L5 spine is developed using simple shapes based on average anatomical dimensions of key features of the spine rather than CT scan images. Pure moments of 7.5 Nm and a compressive follower load of 1000 N are individually applied to the L1 vertebra. Validation is achieved by comparing rotations and intradiscal pressures to other widely accepted FEMs and in vitro studies. Then degenerative disc properties are modeled and rotations calculated. Required computation times are compared between the model presented in this paper and other models developed using CT scans.

RESULTS

For the validation study, parameter values for a healthy spine were used with the loading conditions described above. Total L1-L5 rotations for flexion, extension, lateral bending, and axial rotation under pure moment loading were calculated as 20.3°, 10.7°, 19.7°, and 10.3°, respectively, and under a compressive follower load, maximum intradiscal pressures were calculated as 0.68 MPa. These values compare favorably with the data used for validation. When studying the effects of disc degeneration, the affected segment is shown to experience decreases in rotations during flexion, extension, and lateral bending (24%-56%), while rotations are shown to increase during axial rotation (14%-40%). Adjacent levels realize relatively minor changes in rotation (1%-6%). This parametric study required 17.5 hours of computation time compared to more than 4 days required if utilizing typical published CT scan-based models, illustrating one of the primary advantages of the model presented in this article.

CONCLUSIONS

The FEM presented in this article produces a biomechanical response comparable to widely accepted, complex, CT scan-based models and in vitro studies while requiring much shorter computation times. This makes the model ideal for conducting parametric studies of spinal pathologies and spinal correction techniques.