Original articleFinite element models predict in vitro vertebral body compressive strength better than quantitative computed tomography
Introduction
Noninvasive measurements of bone mineral density are central to the diagnosis and management of osteoporosis. However, quantitative computed tomography (QCT)-based estimates of bone mineral density have generally only shown modest correlations (r2 = 0.49–0.69) to in vitro measurements of vertebral strength [1], [2], [3]. Not surprisingly, vertebral bone mineral density alone is limited in differentiating between patients with and without vertebral fractures [4]. These limitations necessitate that clinical studies enroll hundreds and sometimes thousands of patients and follow them for many years to develop sufficient statistical power for analyses of disease and treatment effects [5], [6]. More accurate metrics should improve diagnosis and management of osteoporosis and may provide a more sensitive and therefore more timely measure of the effects of therapeutic interventions.
Although a wealth of literature exists on correlations between bone mineral density and vertebral strength [1], [2], [3], [7], [8], [9], [10], [11], [12], several theoretical considerations undermine the use of bone mineral density alone as a surrogate for vertebral strength. Bone mineral density is an integral value of material content and therefore cannot reflect the potentially important effects of subtle geometric and structural differences [9], [13] and densitometric inhomogeneities [14], [15] on vertebral strength. Studies that directly compared dual energy x-ray absorptiometry (DXA) and QCT found that both techniques were similarly correlated with the fracture strength of vertebral bodies (r2 = 0.49–0.74) [1], [2], [3]. However, other studies found that QCT is better than DXA in clinically assessing bone loss [16], [17] and fracture risk [16], [17], [18]. All these techniques possess limited generality because of their empirical nature.
Finite element models derived from QCT scans may improve predictions of vertebral strength because they mechanically integrate all the geometrical and material property data within the scans. Such QCT-based finite element models have found successful use in strength prediction of the proximal femur [19], [20], [21], and for the spine they have been used to investigate the theoretical effects of vertebroplasty [22], disc degeneration [23], and bone density [23], [24] on fracture risk. While one study found that geometry-based finite element models (r2 = 0.79) were a better predictor of vertebral strength than CT-derived estimates of density (r2 = 0.67) [25], an important consideration for clinical implementation is the extent to which the model generation can be automated from clinical QCT scans. The so-called voxel method [20], [23], [24], in which QCT voxels are converted directly into finite elements, is most attractive in this respect. Voxel-based finite element models of vertebrae [23], [24] have not been validated by any in vitro experiments and thus their advantage over QCT-derived bone mineral density is unclear.
The goal of this study was to establish that voxel-based finite element models predict vertebral compressive strength better than measures of QCT-derived bone mineral density and geometry. We used QCT scans to develop clinical measures of bone mineral density and patient-specific finite element models of a series of cadaveric thoracolumbar vertebrae. We then compared predictions of the strength of the vertebrae generated from the finite element models and those derived from analysis of the QCT scans alone against the in vitro measurements of the vertebral compressive strength.
Section snippets
QCT scanning
Data from 17 isolated vertebrae taken from 17 cadavers (T11–L4; age: 20–87 years; males = 8, females = 9) were used from a previous experiment [26]. In those experiments, the intervertebral disc material was removed from both endplates, and the posterior elements were removed by sectioning through the pedicles with an autopsy saw. Quantitative computed tomography scans were taken of the vertebral bodies using a clinical scanner (GE 9800; General Electric, Milwaukee, WI, USA: 140 kV, 70 mA, 1.5
Results
The finite element model-derived regression variables (FFE and KFE) showed better correlations with compressive strength than either the product of BMDQCT and Amin or BMDQCT alone (Table 1 and Fig. 4 ). All of the explanatory variables and fracture strength were positively correlated with one another (P < 0.005) to varying degrees (0.71 < r < 0.95; Table 2). The explanatory variables had a significant effect on the absolute value of the regression residual (P = 0.01). The mean of the absolute
Discussion
These results demonstrate that voxel-based finite element model-derived estimates of strength are better predictors of in vitro vertebral compressive strength than clinical measures of bone density derived from QCT with or without geometry. While BMDQCT is well correlated (r2 = 0.91) with the elastic modulus of machined specimens of vertebral trabecular bone [31], (volumetric) bone mineral density by itself does not reflect any structural information of the whole vertebra. Finite element models
Acknowledgements
Funding was provided by grants from the National Institutes of Health (AR41481) and the National Science Foundation (BES-9625030). The National Disease Research Interchange (NDRI) and the Willed Body Program of UCSF provided the human tissue used in these experiments. The authors would like to acknowledge Dr. Peter Bacchetti for his guidance on statistical issues, Dr. David Kopperdahl for his work in the QCT scanning and mechanical testing of the vertebrae, and Mr. Jesse Lanzon for his
References (40)
- et al.
Prediction of thoracic and lumbar vertebral body compressive strengthcorrelations with bone mineral density and vertebral region
Bone
(1995) - et al.
Randomised trial of effect of alendronate on risk of fracture in women with existing vertebral fractures
Lancet
(1996) - et al.
Lumbar vertebral body compressive strength evaluated by dual-energy x-ray absorptiometry, quantitative computed tomography, and ashing
Bone
(1999) - et al.
Mechanical strength of the thoracolumbar spine in the elderlyprediction from in situ dual-energy x-ray absorptiometry, quantitative computed tomography (QCT), upper and lower limb peripheral QCT, and quantitative ultrasound
Bone
(2002) - et al.
The predictive value of quantitative computed tomography for vertebral body compressive strength and ash density
Bone
(1989) - et al.
Inhomogeneity of human vertebral cancellous bonesystematic density and structure patterns inside the vertebral body
Bone
(2001) - et al.
Static and fatigue failure properties of thoracic and lumbar vertebral bodies and their relation to regional density
J Biomech
(1995) - et al.
Femoral strength is better predicted by finite element models than QCT and DXA
J Biomech
(1999) - et al.
Automated three-dimensional finite element modelling of bonea new method
J Biomed Eng
(1990) - et al.
Quantitative computed tomography for prediction of vertebral fracture risk
Bone
(1985)