[Back]


Talks and Poster Presentations (without Proceedings-Entry):

E. DallŽAra, R. Schmidt, D. H. Pahr, P. Varga, J. Patsch, Y. Chevalier, F. Kainberger, P.K. Zysset:
"An Improved Technique for Inducing Compression Fractures of Vertebral Bodies in Vitro";
Poster: 36th European Symposium on Calcified Tissues, Wien; 05-23-2009 - 05-27-2009.



English abstract:
Vertebral compression fracture is a common medical problem in osteoporotic individuals. The computer tomography (CT)-based Finite Element (FE) method may be used to predict vertebral stiffness and validation with experimental tests in vitro. The aim of this study was to develop a novel technique for inducing compression fractures in human vertebral bodies in vitro and to make a preliminary comparison between the fracture load prediction of a previously published voxel-based FE method and the volumetric bone mineral density (vBMD).
First, the cortical endplates were removed from 37 vertebral bodies (T12-L5) extracted from ten donors (seven males and three females with age 44-82). The obtained slices were polished to obtain plane and parallel loading surfaces. Afterwards, each slice was scanned with a clinical CT and the vBMD was evaluated using a calibration phantom. The vertebral slices were carefully positioned in the testing system and loaded in compression beyond fracture up to a large deformation. Rotation of the upper loading plate was allowed by means of a ball joint. To circumvent testing device compliance, the displacement and the angles of rotation were measured directly on the loading plates with three sensors. A previously reported FE method based on calibrated CT images [1] was applied for computing both stiffness and failure load of the vertebral slices. Four samples with large calcifications such as osteophytes were excluded from the analysis. In agreement with clinical observation, most of the vertebrae underwent an anterior wedge fracture. The failure loads measured in this study (2.3-9.2 kN) were consistent with values found in the literature [2]. The high values of stiffness obtained in this study (16.8-54.8 kN/mm) are in line with the reduced sample thickness and the improved testing protocol. As expected, the FE method predicted both stiffness and failure load substantially better than vBMD as the correlation coefficients (R2) improve from 0.35 to 0.60 and from 0.36 to 0.84 respectively.
In conclusion, an improved technique for generating compression fractures was developed and successfully applied to a large set of human vertebrae. The obtained results will be exploited to identify the best FE modeling strategies to predict vertebral failure load in vivo.

[1] Chevalier et al, 2008, Spine.
[2] Crawford et al, 2003, Bone.