- Email: [email protected]
Finite element simulation for the prediction of mechanical failure in the lumbar spine surgery
$420
Journal o f Biomechanics 2006, Vol. 39 (Suppl 1)
applied displacement boundary conditions on the brace are removed allowing the brace to position itself freely on the patient. Results: The simulation results are quite similar to the documented immediate effect of a brace. With a strap tension of 60 N, correction of frontal curvatures can be up to 50%. The lordosis and the sacral slope decrease, there's no significant correction of the axial rotation and rib hump. When removing the trochanter pad, the brace tilts laterally because of the removal of the lever arm and of the reaction force on the right thoracic region. The computational time is 10 minutes with a 3 GHz processor. The residual disequilibrium force is inferior to 1 N and the residual penetration in the contact interface is inferior to 1 mm. Discussion: The new simulation process is considered an improvement compared to previous models where external forces are applied directly on the trunk model. Because of the explicit representation of the brace and its design parameters, it could be a promising tool to study brace biomechanics and improve its effectiveness. 4527 Tu, 11:00-11:15 (P20) Optimum vertebral body shape and density for sustaining load with minimum mass M. Kasra. Department of Mechanical Engineering, McMaster University,
Hamilton, Ontario, Canada According to Wolff's law bone remodeling can be considered as an optimization process during which bone tissue adapts its structure and density according to loading environment having minimal mass needed to bear load. The objective of this study was to perform an optimization analysis of a parametric finite element model of a lumbar disc-vertebra-disc unit in order to find the geometry and bone density of a vertebra which gives the minimum vertebral weight for an optimum strength. Using ANSYS program, a three dimensional parametric finite element model of a lumbar disc-vertebra-disc unit has been developed. Different mathematical functions represented different geometries of different parts of a motion segment. The analyses were performed for intervertebral discs of normal material and geometry. The model was analyzed under a failure load of 6000N in axial compression simulating the loading and constraint condition of a L5 vertebral failure test. In the optimisation analysis, anterior and posterior cortical thicknesses (CTA and CTP) and cortical wall concavity (EEA) defining vertebral cortical wall geometry plus cancellous bone apparent density (DCAN) were the design variables (DV). The design variables were optimized for the minimum bone weight (Objective), considering the limits of state variables (SV) which were maximum von Mises stresses in the cortical shell and cancellous bone. The optimization process indicated a concave shape for the vertebral cortical wall. The optimized cancellous bone apparent density and cortical thicknesses also compared very well with average values of those reported for lumbar vertebral bodies. This study indicated that a design optimization process would lead to a concaved vertebral wall with a thin cortical shell and a low cancellous bone density suggesting that vertebral body is formed in such a manner that provides optimum vertebral strength for the minimum weight. This parametric modelling and optimisation technique may introduce an important research tool for studying bone remodelling process in response to applied loads. 7422 Tu, 11:15-11:30 (P20) Finite element simulation for the prediction o f mechanical failure in the lumbar spine surgery T. Mosnier 1, V. Lafage 1, L. Rillardon 2, J. Dubousset 1, J. Pratt 3, W. Skalli 1.
1Laboratoire de BioM6canique, Paris, France, 2Hopital Beaujon, Clichy, France, 31nstitut de Biomecanica de Valencia, Spain Introduction: Mechanical complications of lumbar spine surgery are sometimes related to the difficulty of adequate surgery planning. The aim of this study is to provide a patient specific finite element model of the lumbar spine with simulation of the surgical gesture and to use it as a predictive tool of mechanical failure for lumbar spine surgery. Materials and Method: A 3D finite element model of the lumbar spine with geometrical personalization was considered. Posterior implants and main characteristics of degenerative pathologies were modeled. After validation with regard to in vitro data (24 specimens and 4 instrumentations), the model was used to simulate real cases. Applied loads depended on patient characteristics (weight, imbalance), and simulation results were post-processed to assess stresses in the discs and the implants. For purpose of validation, pre and postoperative data were collected for 66 patients instrumented with rigid screw-rod systems, either at the L4-Sacrum (24 patients) or at L5-Sacrum levels (42). Two subsets were considered: "non complicated cases" (53), and "complicated cases" with an implant mechanical failure (13, i.e. 11 screw breakage and 2 screw loosening). Blind comparison was performed between simulations results and clinical outcome. Results and Discussion: Among the 66 patients, results highlighted the specific behaviors of 9 patients (3 at L4/S and 6 at L5-S levels) for which mechanical loads on implants were markedly higher. The analysis of the
Oral Presentations clinical outcome indicated that all of them were "complicated case". None of the "non complicated cases" demonstrated numerical results with particularly high stresses. Conclusion: Finite element simulations allowed to predict 9 on 13 failure cases among a total of 66 patients. This is a promising step towards the possibility to use finite element modeling as a clinically relevant simulation tool for surgery planning. 7066 Tu, 11:30-11:45 (P20) Development and validation o f a three dimensional finite element model o f w h o l e cervical spine S. Basa, V. Balasubramanian. Rehabilitation Bioengineering Group,
Department of Biotechnology, liT Madras, Chennai, India Understanding complex mechanisms involved in neck trauma is of interest in mitigating effects of injury and preventing them. A geometrically accurate, detailed and validated finite element (FE) model can provide insight into the underlying mechanisms of neck injuries and surrogate invitro studies. Previous FE models predicted various biomechanical parameters under complex loads but consisted of two or three spinal segments only. In order to get realistic response of multi level component of the whole cervical spine, a three dimensional anatomically accurate FE model (C1-7) was developed. Geometry was acquired from computer tomography images of an adult free from spinal disorder. Important anatomical features viz., cortical bone, cancellous bone, transverse process, spinous process, laminae, intervertebral disc, spinal canal were clearly defined for each spinal segment. The model consisted of 52,381 solid elements with 13,713 nodes. Material properties were obtained from literature [3]. This model has been validated against invitro experiments [1,2] under flexion, extension, torsion and axial compression. Loads and boundary conditions were simulated similar to invitro experiment against which the current model was validated. Load~Jisplacement response of FE model agreed well with invitro experimental results. Displacements were highest at C1-C2 when compared to other levels for flexion (8 mm), extension (8 mm) and axial torsion (12.5mm). Comparison of FE model response with invitro studies substantiates that this model can effectively reflect the behavior of human cervical spine. This model can be used to study applications of interest related to cervical spine trauma and dysfunction. References [1] Panjabi M.M., Crisco J.J., Vasavada A., Oda T., Cholewicki J., Nibu K., Shin E. Spine 2001; 26(24): 2692-2700. [2] Shea M., Edwards W.T., White A.A., Hayes W.C. J Biomech 1991; 24(2): 95107. [3] Yoganandan N., Kumaresan S., Voo L., Pintar E, Larson S. Med Eng Phys 1996; 18(7): 569-574.
6097 Tu, 11:45-12:00 (P20) Influence o f b o u n d a r y conditions on failure mechanisms of the human vertebral body S.K. Eswaran, T.M. Keaveny. Orthopaedic Biomechanics Laboratory,
University of California, Berkeley, CA, USA Knowledge of the failure mechanisms of the vertebral body may lead to improved diagnosis and treatment of osteoporosis. This study addresses whether the overall micro-mechanics of the vertebral body depends on the loading conditions - v i a an intervertebral disc vs. a PMMA layer, which is typically used in biomechanical experiments [1,2]. Towards this end, 60-micron-resolution micro-CT based finite element models of 13 elderly vertebrae (75±9 years), having up to 220 million degrees of freedom, were analyzed using state-ofthe-art supercomputer techniques. Two linear elastic analyses were run using custom code [3] for each vertebra in compressive loading conditions - via a PMMA layer or a degenerated intervertebral disc. All bone tissue having principal strains above the 90 th percentile was identified as "highly-strained" tissue, at most risk of failure within each vertebra. Significant differences (p<0.005) in the relative amount of highly-strained tissue were found in all material groups (cortical shell, endplate and trabecular bone). In particular, there were minimal amounts of highly-strained tissue in the endplate when loading was through the PMMA layer (0.3±0.2% by volume) versus through the intervertebral disc (19.5±2.3% by volume). This protection of the endplate for PMMA loading was a result of reduced bending and tensile stretching of the endplate, which otherwise occurred in the presence of the disc. While the general variation of the amount of highly-strained tissue as well as cortical load sharing across transverse slices was similar between the two loading modes, loading through a PMMA layer lead to a slightly greater load bearing role for the cortical shell (10% increase on average) and as a result, a greater amount of highly-strained tissue in the cortical shell. These differences in the distribution of the highly-strained tissue and specifically, the reduced role of the endplate when loading was through the PMMA layer suggest that the failure mechanisms may change based on the loading conditions of the vertebral body.
Journal o f Biomechanics 2006, Vol. 39 (Suppl 1)
applied displacement boundary conditions on the brace are removed allowing the brace to position itself freely on the patient. Results: The simulation results are quite similar to the documented immediate effect of a brace. With a strap tension of 60 N, correction of frontal curvatures can be up to 50%. The lordosis and the sacral slope decrease, there's no significant correction of the axial rotation and rib hump. When removing the trochanter pad, the brace tilts laterally because of the removal of the lever arm and of the reaction force on the right thoracic region. The computational time is 10 minutes with a 3 GHz processor. The residual disequilibrium force is inferior to 1 N and the residual penetration in the contact interface is inferior to 1 mm. Discussion: The new simulation process is considered an improvement compared to previous models where external forces are applied directly on the trunk model. Because of the explicit representation of the brace and its design parameters, it could be a promising tool to study brace biomechanics and improve its effectiveness. 4527 Tu, 11:00-11:15 (P20) Optimum vertebral body shape and density for sustaining load with minimum mass M. Kasra. Department of Mechanical Engineering, McMaster University,
Hamilton, Ontario, Canada According to Wolff's law bone remodeling can be considered as an optimization process during which bone tissue adapts its structure and density according to loading environment having minimal mass needed to bear load. The objective of this study was to perform an optimization analysis of a parametric finite element model of a lumbar disc-vertebra-disc unit in order to find the geometry and bone density of a vertebra which gives the minimum vertebral weight for an optimum strength. Using ANSYS program, a three dimensional parametric finite element model of a lumbar disc-vertebra-disc unit has been developed. Different mathematical functions represented different geometries of different parts of a motion segment. The analyses were performed for intervertebral discs of normal material and geometry. The model was analyzed under a failure load of 6000N in axial compression simulating the loading and constraint condition of a L5 vertebral failure test. In the optimisation analysis, anterior and posterior cortical thicknesses (CTA and CTP) and cortical wall concavity (EEA) defining vertebral cortical wall geometry plus cancellous bone apparent density (DCAN) were the design variables (DV). The design variables were optimized for the minimum bone weight (Objective), considering the limits of state variables (SV) which were maximum von Mises stresses in the cortical shell and cancellous bone. The optimization process indicated a concave shape for the vertebral cortical wall. The optimized cancellous bone apparent density and cortical thicknesses also compared very well with average values of those reported for lumbar vertebral bodies. This study indicated that a design optimization process would lead to a concaved vertebral wall with a thin cortical shell and a low cancellous bone density suggesting that vertebral body is formed in such a manner that provides optimum vertebral strength for the minimum weight. This parametric modelling and optimisation technique may introduce an important research tool for studying bone remodelling process in response to applied loads. 7422 Tu, 11:15-11:30 (P20) Finite element simulation for the prediction o f mechanical failure in the lumbar spine surgery T. Mosnier 1, V. Lafage 1, L. Rillardon 2, J. Dubousset 1, J. Pratt 3, W. Skalli 1.
1Laboratoire de BioM6canique, Paris, France, 2Hopital Beaujon, Clichy, France, 31nstitut de Biomecanica de Valencia, Spain Introduction: Mechanical complications of lumbar spine surgery are sometimes related to the difficulty of adequate surgery planning. The aim of this study is to provide a patient specific finite element model of the lumbar spine with simulation of the surgical gesture and to use it as a predictive tool of mechanical failure for lumbar spine surgery. Materials and Method: A 3D finite element model of the lumbar spine with geometrical personalization was considered. Posterior implants and main characteristics of degenerative pathologies were modeled. After validation with regard to in vitro data (24 specimens and 4 instrumentations), the model was used to simulate real cases. Applied loads depended on patient characteristics (weight, imbalance), and simulation results were post-processed to assess stresses in the discs and the implants. For purpose of validation, pre and postoperative data were collected for 66 patients instrumented with rigid screw-rod systems, either at the L4-Sacrum (24 patients) or at L5-Sacrum levels (42). Two subsets were considered: "non complicated cases" (53), and "complicated cases" with an implant mechanical failure (13, i.e. 11 screw breakage and 2 screw loosening). Blind comparison was performed between simulations results and clinical outcome. Results and Discussion: Among the 66 patients, results highlighted the specific behaviors of 9 patients (3 at L4/S and 6 at L5-S levels) for which mechanical loads on implants were markedly higher. The analysis of the
Oral Presentations clinical outcome indicated that all of them were "complicated case". None of the "non complicated cases" demonstrated numerical results with particularly high stresses. Conclusion: Finite element simulations allowed to predict 9 on 13 failure cases among a total of 66 patients. This is a promising step towards the possibility to use finite element modeling as a clinically relevant simulation tool for surgery planning. 7066 Tu, 11:30-11:45 (P20) Development and validation o f a three dimensional finite element model o f w h o l e cervical spine S. Basa, V. Balasubramanian. Rehabilitation Bioengineering Group,
Department of Biotechnology, liT Madras, Chennai, India Understanding complex mechanisms involved in neck trauma is of interest in mitigating effects of injury and preventing them. A geometrically accurate, detailed and validated finite element (FE) model can provide insight into the underlying mechanisms of neck injuries and surrogate invitro studies. Previous FE models predicted various biomechanical parameters under complex loads but consisted of two or three spinal segments only. In order to get realistic response of multi level component of the whole cervical spine, a three dimensional anatomically accurate FE model (C1-7) was developed. Geometry was acquired from computer tomography images of an adult free from spinal disorder. Important anatomical features viz., cortical bone, cancellous bone, transverse process, spinous process, laminae, intervertebral disc, spinal canal were clearly defined for each spinal segment. The model consisted of 52,381 solid elements with 13,713 nodes. Material properties were obtained from literature [3]. This model has been validated against invitro experiments [1,2] under flexion, extension, torsion and axial compression. Loads and boundary conditions were simulated similar to invitro experiment against which the current model was validated. Load~Jisplacement response of FE model agreed well with invitro experimental results. Displacements were highest at C1-C2 when compared to other levels for flexion (8 mm), extension (8 mm) and axial torsion (12.5mm). Comparison of FE model response with invitro studies substantiates that this model can effectively reflect the behavior of human cervical spine. This model can be used to study applications of interest related to cervical spine trauma and dysfunction. References [1] Panjabi M.M., Crisco J.J., Vasavada A., Oda T., Cholewicki J., Nibu K., Shin E. Spine 2001; 26(24): 2692-2700. [2] Shea M., Edwards W.T., White A.A., Hayes W.C. J Biomech 1991; 24(2): 95107. [3] Yoganandan N., Kumaresan S., Voo L., Pintar E, Larson S. Med Eng Phys 1996; 18(7): 569-574.
6097 Tu, 11:45-12:00 (P20) Influence o f b o u n d a r y conditions on failure mechanisms of the human vertebral body S.K. Eswaran, T.M. Keaveny. Orthopaedic Biomechanics Laboratory,
University of California, Berkeley, CA, USA Knowledge of the failure mechanisms of the vertebral body may lead to improved diagnosis and treatment of osteoporosis. This study addresses whether the overall micro-mechanics of the vertebral body depends on the loading conditions - v i a an intervertebral disc vs. a PMMA layer, which is typically used in biomechanical experiments [1,2]. Towards this end, 60-micron-resolution micro-CT based finite element models of 13 elderly vertebrae (75±9 years), having up to 220 million degrees of freedom, were analyzed using state-ofthe-art supercomputer techniques. Two linear elastic analyses were run using custom code [3] for each vertebra in compressive loading conditions - via a PMMA layer or a degenerated intervertebral disc. All bone tissue having principal strains above the 90 th percentile was identified as "highly-strained" tissue, at most risk of failure within each vertebra. Significant differences (p<0.005) in the relative amount of highly-strained tissue were found in all material groups (cortical shell, endplate and trabecular bone). In particular, there were minimal amounts of highly-strained tissue in the endplate when loading was through the PMMA layer (0.3±0.2% by volume) versus through the intervertebral disc (19.5±2.3% by volume). This protection of the endplate for PMMA loading was a result of reduced bending and tensile stretching of the endplate, which otherwise occurred in the presence of the disc. While the general variation of the amount of highly-strained tissue as well as cortical load sharing across transverse slices was similar between the two loading modes, loading through a PMMA layer lead to a slightly greater load bearing role for the cortical shell (10% increase on average) and as a result, a greater amount of highly-strained tissue in the cortical shell. These differences in the distribution of the highly-strained tissue and specifically, the reduced role of the endplate when loading was through the PMMA layer suggest that the failure mechanisms may change based on the loading conditions of the vertebral body.