Final Report Summary - TISSSPECBONEFEM (Incorporation of multiscale tissue-specific properties into musculoskeletal finite element modelling)
Realistic physiological tissue-level strain fields in whole bones are not yet known as these cannot be measured experimentally and numerical simulations targeting this aim either do not include microstructural information or investigate a sub-volume with simplified loading. However, this knowledge would be essential for unraveling the relation of micro-structure and function, and for understanding the rules dictating processes of bone modeling, remodeling and healing.
The aims of this project were to (I) assess realistic physiological mechanical environment in cortical bone tissue with previously unmatched accuracy to understand the functional adaptation of the intact microstructure and (II) use this knowledge to investigate the variations observed in altered conditions, e.g. fracture healing.
In particular, this project targeted the analysis of the femur of the sheep, which is one of the most important large animal model used frequently for investigating regeneration of musculoskeletal tissues. Moreover, long bones of the sheep are built of plexiform tissue, which is replaced with osteonal (Haversian) bone continuously along the lifespan of the animal. The fingerprint of this process is an obvious tissue type distribution, which thus can be used for the analysis of the chronological and spatial course of remodeling.
Aim I was realized in two work packages. In the first work package, bone geometry, bone mineral density, tissue types (osteonal/plexiform) and micro-scale elastic properties of the ovine femur were characterized experimentally and described as the function of anatomical location and animal age by means of a site-matched database of all these properties. Further, homogenized finite element (FE) model of the ovine femur was developed which included i) tissue-specific, experimentally assessed, heterogeneous and anisotropic elastic properties via a homogenization approach and ii) physiologically relevant loading conditions as provided by an ovine-specific musculoskeletal model. These models were used to predict the physiological strain distribution in the entire bone during walk. An experimental testing setup was developed and applied to validate these models by measuring surface deformations fields during mechanical loading in the elastic regime and by site-matched comparison with the numerical results. Finally, a combined analysis of the collected site-matched multi-modal and multi-scale properties and the predicted physiological strains allowed for the investigation of structure-functions relationships, i.e. the relation between osteonal remodeling and tissue maturation versus the local mechanical stimuli, which were analyzed as the function of anatomical location and animal age. This analysis revealed that osteonal remodeling is triggered mainly by the local shear forces applied by muscles and not by the global compression and bending of the femur. An additional, originally not planned achievement was the assessment of cm-scale properties of cortical bone in in-vivo-like conditions using an ultrasound-based bidirectional axial transmission measurement approach, which provide in vivo perspectives for assessing case-specific material properties.
In the second work package, mm-scale cuboidal samples were analyzed both experimentally and numerically. In particular, mineralization and pore network morphology was quantified based on SR-µCT images and, as an originally not planned achievement, the complete stiffness tensor was measured using resonant ultrasound spectroscopy. All these properties were analyzed as the function of anatomical location, tissue type and animal age; and provided further insights into relations between structure and function as well as in the details of the remodeling process. SR-µCT-based µFE models of the same samples are currently being utilized to i) predict tissue-level deformations from physiological loading conditions imported from the results of the organ-scale FE models (work package 1); and ii) to analyze the effect of the anisotropy an heterogeneity of the micro-scale elastic properties as well as the influence of vascular and lacunar porosities on the numerically homogenized, mm-scale stiffness tensor of bone. Experimental equivalent of these homogenized stiffness tensors, as assessed by means of resonant ultrasound spectroscopy, are used to validate these µFE models.
An additional, originally not planned achievement is the FE-based analysis of the in situ deformation of osteocytes, which are the bone cells embedded in the bone matrix that sense mechanical signals and orchestrate the remodeling process accordingly. This case specific modeling approach could be accomplished due to recent advancements in imaging technology and a specific method, synchrotron X-ray phase nano-tomography, which is able to provide sufficient details on bone and its pore network geometry at the sub-micron scale.
With respect to aim II, a numerical framework has been developed, which is suitable also for investigation of altered conditions like fractures. This is planned to be applied on an established sheep osteotomy model, which is anticipated to elucidate the optimum strain levels required to promote the endogenous healing process, that is of direct clinical relevance. These are expected to contribute to the deeper understanding of the effect of the exact mechanical environment on the healing outcome, allow for improvements in the treatment of fracture and have impact on the everyday healthcare, These would ultimately increase patients' life quality and ease the hospitalization and financial burden of the increasing number of fractures on the medical health care of Europe.
In summary, the data collected and the tools developed in this project provide i) deeper understanding of structural-functional relationships and of the remodeling process in ovine femoral cortical bone, ii) a unique, site-matched multi-modal and multi-scale database of mineral density and elastic properties, serving as direct input for FE models, iii) physiologically relevant strain distribution in the intact ovine femur as predicted by experimentally validated FE models enriched with case-specific material properties, iv) the basis of analyzing altered conditions, e.g. fractures healing with fixations.
The aims of this project were to (I) assess realistic physiological mechanical environment in cortical bone tissue with previously unmatched accuracy to understand the functional adaptation of the intact microstructure and (II) use this knowledge to investigate the variations observed in altered conditions, e.g. fracture healing.
In particular, this project targeted the analysis of the femur of the sheep, which is one of the most important large animal model used frequently for investigating regeneration of musculoskeletal tissues. Moreover, long bones of the sheep are built of plexiform tissue, which is replaced with osteonal (Haversian) bone continuously along the lifespan of the animal. The fingerprint of this process is an obvious tissue type distribution, which thus can be used for the analysis of the chronological and spatial course of remodeling.
Aim I was realized in two work packages. In the first work package, bone geometry, bone mineral density, tissue types (osteonal/plexiform) and micro-scale elastic properties of the ovine femur were characterized experimentally and described as the function of anatomical location and animal age by means of a site-matched database of all these properties. Further, homogenized finite element (FE) model of the ovine femur was developed which included i) tissue-specific, experimentally assessed, heterogeneous and anisotropic elastic properties via a homogenization approach and ii) physiologically relevant loading conditions as provided by an ovine-specific musculoskeletal model. These models were used to predict the physiological strain distribution in the entire bone during walk. An experimental testing setup was developed and applied to validate these models by measuring surface deformations fields during mechanical loading in the elastic regime and by site-matched comparison with the numerical results. Finally, a combined analysis of the collected site-matched multi-modal and multi-scale properties and the predicted physiological strains allowed for the investigation of structure-functions relationships, i.e. the relation between osteonal remodeling and tissue maturation versus the local mechanical stimuli, which were analyzed as the function of anatomical location and animal age. This analysis revealed that osteonal remodeling is triggered mainly by the local shear forces applied by muscles and not by the global compression and bending of the femur. An additional, originally not planned achievement was the assessment of cm-scale properties of cortical bone in in-vivo-like conditions using an ultrasound-based bidirectional axial transmission measurement approach, which provide in vivo perspectives for assessing case-specific material properties.
In the second work package, mm-scale cuboidal samples were analyzed both experimentally and numerically. In particular, mineralization and pore network morphology was quantified based on SR-µCT images and, as an originally not planned achievement, the complete stiffness tensor was measured using resonant ultrasound spectroscopy. All these properties were analyzed as the function of anatomical location, tissue type and animal age; and provided further insights into relations between structure and function as well as in the details of the remodeling process. SR-µCT-based µFE models of the same samples are currently being utilized to i) predict tissue-level deformations from physiological loading conditions imported from the results of the organ-scale FE models (work package 1); and ii) to analyze the effect of the anisotropy an heterogeneity of the micro-scale elastic properties as well as the influence of vascular and lacunar porosities on the numerically homogenized, mm-scale stiffness tensor of bone. Experimental equivalent of these homogenized stiffness tensors, as assessed by means of resonant ultrasound spectroscopy, are used to validate these µFE models.
An additional, originally not planned achievement is the FE-based analysis of the in situ deformation of osteocytes, which are the bone cells embedded in the bone matrix that sense mechanical signals and orchestrate the remodeling process accordingly. This case specific modeling approach could be accomplished due to recent advancements in imaging technology and a specific method, synchrotron X-ray phase nano-tomography, which is able to provide sufficient details on bone and its pore network geometry at the sub-micron scale.
With respect to aim II, a numerical framework has been developed, which is suitable also for investigation of altered conditions like fractures. This is planned to be applied on an established sheep osteotomy model, which is anticipated to elucidate the optimum strain levels required to promote the endogenous healing process, that is of direct clinical relevance. These are expected to contribute to the deeper understanding of the effect of the exact mechanical environment on the healing outcome, allow for improvements in the treatment of fracture and have impact on the everyday healthcare, These would ultimately increase patients' life quality and ease the hospitalization and financial burden of the increasing number of fractures on the medical health care of Europe.
In summary, the data collected and the tools developed in this project provide i) deeper understanding of structural-functional relationships and of the remodeling process in ovine femoral cortical bone, ii) a unique, site-matched multi-modal and multi-scale database of mineral density and elastic properties, serving as direct input for FE models, iii) physiologically relevant strain distribution in the intact ovine femur as predicted by experimentally validated FE models enriched with case-specific material properties, iv) the basis of analyzing altered conditions, e.g. fractures healing with fixations.