Three-dimensional fiber patterning in the annulus fibrosus can be derived from vertebral endplate topography

Introduction

The annulus fibrosus (AF) of the Intervertebral disc (IVD) is composed of concentric lamellae of helically wound collagen fibers. Understanding the spatial variation of collagen fiber orientations in these lamellae, and the resulting material anisotropy, is crucial to predicting the mechanical behavior of the complete IVD.

Methods

This study builds on a prior model predicated on path-independent displacement of fiber endpoints during vertebral body growth to predict a complete, three-dimensional annulus fibrosus fiber network from a small number of subject-independent input parameters and vertebral endplate topographies obtained from clinical imaging. To evaluate the model, it was first fit to mid-plane fiber orientations obtained using polarized light microscopy in a population of bovine caudal discs for which computed tomography images vertebral endplates were also available. Additionally, the model was used to predict the trajectories based on human lumbar disc geometries and results were compared to previously reported data. Finally, the model was employed to investigate potential disc-related variations in fiber angle distributions.

Results

The model was able to accurately predict experimentally measured fiber distributions in both bovine and human discs using only endplate topography and three input parameters. Critically, the model recapitulated previously observed asymmetry between the inclinations of right- and left-handed fibers in the posterolateral aspect of the human AF. Level to level variation of disc height and aspect ratio in the human lumbar spine was predicted to affect absolute values of fiber inclination, but not this asymmetry.

Conclusion

Taken together these results suggest that patient-specific distributions of AF fiber orientation may be readily incorporated into computational models of the spine using only disc geometry and a small number of subject-independent parameters.