Noninvasive prediction of compressive strength and fracture of thoracolumbar vertebrae: finite element analysis

A. Beltagi, S. Guitteny, F. Amirouche, I. Imran, S. Elhafez, G. Mohamed

Abstract: Prediction of vertebral properties and associated fracture risk is a serious challenge for patients who have osteoporosis. This study aimed to enhance the prediction capability of the computed tomography-based finite element model (CT/FEM) for the strength, stiffness, and failure pattern of the vertebrae under compression. Twelve thoracic and eight lumbar vertebrae (T12-14 & L1-2) were imaged using CT scanner with specific parameters for an accurate representation of the vertebral bone geometry to build the 3D models. Uniaxial compressive load testing of the vertebrae was conducted. Non-linear FE analysis was performed to simulate the experimental testing with the same boundary conditions. The average thoracic vertebral stiffness and strength was 9545 ±1880 N/mm (range 6736-13628 N/mm) and 10198 ± 1486 N (8094 - 12967 N) respectively. While for the lumbar part, the mean stiffness and strength were 9206 ±932 N/mm (8238 - 10693 N/mm) and 10125 ± 1445 N (8389 – 11931 N). There was no significant difference between thoracic and lumbar specimens in the stiffness (p = 0.637) or the strength (p = 0.919). The study showed a strong correlation between the predicted and measured results for the peak load (r = 0.90) and moderate correlation for the stiffness (r = 0.68). The experimental fracture cracks and their locations in the specimens agreed reasonably well with those of the CT/FE models. The non-linear FE analysis was successfully validated by the in-vitro experiment. This modeling approach could be adapted to assess in-vivo loading conditions, therapeutic effects, and surgical spinal implants.

Series on Biomechanics, Vol.35 No.3 (2021), 20-32

Keywords: compressive strength; Finite element analysis; osteoporosis; thoracolumbar spine; vertebral fracture

Date published: 2021-10-26