Can Multi-Vertebral CT-Based Finite Element Models Accurately Predict Strains? An In Vitro Validation Study.

Area of Science:

• Biomechanics
• Computational modeling
• Spine research

Background:

• Current finite element (FE) models for vertebral fracture risk often analyze single vertebrae, overlooking intervertebral disc roles in load distribution.
• Incorporating intervertebral discs into multi-vertebrae FE models enables more realistic boundary conditions.
• Computed tomography (CT) scans provide vertebral material properties but lack disc information.

Purpose of the Study:

• To develop multi-level FE models of the spine using only CT data.
• To validate these models by comparing predicted displacements and strains against experimental measurements.
• To assess the impact of different intervertebral disc material properties on model accuracy.

Main Methods:

• A human T10-L1 spine segment was tested in flexion-compression.
• Digital image correlation measured surface displacements and strains during testing.
• A CT-based FE model was created, assigning HU-based properties to vertebrae and various hyperelastic properties to disc components (nucleus pulposus, anulus fibrosus).

Main Results:

• The FE model accurately predicted vertebral surface displacements (R² = 0.92-0.99).
• However, varying disc constitutive laws led to significant differences in predicted principal strain distributions compared to experimental data (average relative errors > 34%).

Conclusions:

• CT-based multi-level FE models can accurately predict vertebral displacements.
• Current approaches for modeling intervertebral disc material properties in these models are insufficient for accurate strain prediction.
• A revised modeling strategy for intervertebral discs is essential for enhancing the accuracy of CT-based multi-level FE models.