Subject specific finite element mesh generation of the pelvis from biplanar x-ray images: application to 120 clinical cases

Several Finite Element (FE) models of the pelvis have been developed to comprehensively assess the onset of pathologies and for clinical and industrial applications. However, because of the difficulties associated with the creation of subject-specific FE mesh from CT scan and MR images, most of the existing models rely on the data of one given individual. Moreover, although several fast and robust methods have been developed for automatically generating tetrahedral meshes of arbitrary geometries, hexahedral meshes are still preferred today because of their distinct advantages but their generation remains an open challenge. Recently, approaches have been proposed for fast 3D reconstruction of bones based on X-ray imaging. In this study, we adapted such an approach for the fast and automatic generation of all-hexahedral subject-specific FE models of the pelvis based on the elastic registration of a generic mesh to the subject-specific target in conjunction with element regularity and quality correction. The technique was successfully tested on a database of 120 3D reconstructions of pelvises from biplanar X-ray images. For each patient, a full hexahedral subject-specific FE mesh was generated with an accurate surface representation.     

An eFTD-VP framework for efficiently generating patient-specific anatomically detailed facial soft tissue FE mesh for craniomaxillofacial surgery simulation

Accurate surgical planning and prediction of craniomaxillofacial surgery outcome requires simulation of soft tissue changes following osteotomy. This can only be achieved by using an anatomically detailed facial soft tissue model. The current state-of-the-art of model generation is not appropriate to clinical applications due to the time-intensive nature of manual segmentation and volumetric mesh generation. The conventional patient-specific finite element (FE) mesh generation methods are to deform a template FE mesh to match the shape of a patient based on registration. However, these methods commonly produce element distortion. Additionally, the mesh density for patients depends on that of the template model. It could not be adjusted to conduct mesh density sensitivity analysis. In this study, we propose a new framework of patient-specific facial soft tissue FE mesh generation. The goal of the developed method is to efficiently generate a high-quality patient-specific hexahedral FE mesh with adjustable mesh density while preserving the accuracy in anatomical structure correspondence. Our FE mesh is generated by eFace template deformation followed by volumetric parametrization. First, the patient-specific anatomically detailed facial soft tissue model (including skin, mucosa, and muscles) is generated by deforming an eFace template model. The adaptation of the eFace template model is achieved by using a hybrid landmark-based morphing and dense surface fitting approach followed by a thin-plate spline interpolation. Then, high-quality hexahedral mesh is constructed by using volumetric parameterization. The user can control the resolution of hexahedron mesh to best reflect clinicians’ need. Our approach was validated using 30 patient models and 4 visible human datasets. The generated patient-specific FE mesh showed high surface matching accuracy, element quality, and internal structure matching accuracy. They can be directly and effectively used for clinical simulation of facial soft tissue change.     

Anisotropic finite element models for brain injury prediction: the sensitivity of axonal strain to white matter tract inter-subject variability

Computational models incorporating anisotropic features of brain tissue have become a valuable tool for studying the occurrence of traumatic brain injury. The tissue deformation in the direction of white matter tracts (axonal strain) was repeatedly shown to be an appropriate mechanical parameter to predict injury. However, when assessing the reliability of axonal strain to predict injury in a population, it is important to consider the predictor sensitivity to the biological inter-subject variability of the human brain. The present study investigated the axonal strain response of 485 white matter subject-specific anisotropic finite element models of the head subjected to the same loading conditions. It was observed that the biological variability affected the orientation of the preferential directions (coefficient of variation of 39.41% for the elevation angle—coefficient of variation of 29.31% for the azimuth angle) and the determination of the mechanical fiber alignment parameter in the model (gray matter volume 55.55–70.75%). The magnitude of the maximum axonal strain showed coefficients of variation of 11.91%. On the contrary, the localization of the maximum axonal strain was consistent: the peak of strain was typically located in a 2 cm³ volume of the brain. For a sport concussive event, the predictor was capable of discerning between non-injurious and concussed populations in several areas of the brain. It was concluded that, despite its sensitivity to biological variability, axonal strain is an appropriate mechanical parameter to predict traumatic brain injury.     

Development of a Finite Element Model of Decompressive Craniectomy

Decompressive craniectomy (DC), an operation whereby part of the skull is removed, is used in the management of patients with brain swelling. While the aim of DC is to reduce intracranial pressure, there is the risk that brain deformation and mechanical strain associated with the operation could damage the brain tissue. The nature and extent of the resulting strain regime is poorly understood at present. Finite element (FE) models of DC can provide insight into this applied strain and hence assist in deciding on the best surgical procedures. However there is uncertainty about how well these models match experimental data, which are difficult to obtain clinically. Hence there is a need to validate any modelling approach outside the clinical setting. This paper develops an axisymmetric FE model of an idealised DC to assess the key features of such an FE model which are needed for an accurate simulation of DC. The FE models are compared with an experimental model using gelatin hydrogel, which has similar poro-viscoelastic material property characteristics to brain tissue. Strain on a central plane of the FE model and the front face of the experimental model, deformation and load relaxation curves are compared between experiment and FE. Results show good agreement between the FE and experimental models, providing confidence in applying the proposed FE modelling approach to DC. Such a model should use material properties appropriate for brain tissue and include a more realistic whole head geometry.     

Can sulci protect the brain from traumatic injury?

The influence of sulci in dynamic finite element simulations of the human head has been investigated. First, a detailed 3D FE model was constructed based on an MRI scan of a human head. A second model with a smoothed brain surface was created based on the same MRI scan as the first FE model. These models were validated against experimental data to confirm their human-like dynamic responses during impact. The validated FE models were subjected to several acceleration impulses and the maximum principle strain and strain rate in the brain were analyzed. The results suggested that the inclusion of sulci should be considered for future FE head models as it alters the strain and strain distribution in an FE model.     

Subject-specific finite element model of the pelvis: development, validation and sensitivity studies

A better understanding of the three-dimensional mechanics of the pelvis, at the patient-specific level, may lead to improved treatment modalities. Although finite element (FE) models of the pelvis have been developed, validation by direct comparison with subject-specific strains has not been performed, and previous models used simplifying assumptions regarding geometry and material properties. The objectives of this study were to develop and validate a realistic FE model of the pelvis using subject-specific estimates of bone geometry, location-dependent cortical thickness and trabecular bone elastic modulus, and to assess the sensitivity of FE strain predictions to assumptions regarding cortical bone thickness as well as bone and cartilage material properties. A FE model of a cadaveric pelvis was created using subject-specific computed tomography image data. Acetabular loading was applied to the same pelvis using a prosthetic femoral stem in a fashion that could be easily duplicated in the computational model. Cortical bone strains were monitored with rosette strain gauges in ten locations on the left hemipelvis. FE strain predictions were compared directly with experimental results for validation. Overall, baseline FE predictions were strongly correlated with experimental results (r2=0.824), with a best-fit line that was not statistically different than the line y=x (experimental strains = FE predicted strains). Changes to cortical bone thickness and elastic modulus had the largest effect on cortical bone strains. The FE model was less sensitive to changes in all other parameters. The methods developed and validated in this study will be useful for creating and analyzing patient-specific FE models to better understand the biomechanics of the pelvis.     

Growing multiblock structures: a semi-automated approach to block placement for multiblock hexahedral meshing

Finite element (FE) analysis is a cornerstone of orthopaedic biomechanics research. Three-dimensional medical imaging provides sufficient resolution for the subject-specific FE models to be generated from these data-sets. FE model development requires discretisation of a three-dimensional domain, which can be the most time-consuming component of a FE study. Hexahedral meshing tools based on the multiblock method currently rely on the manual placement of building blocks for mesh generation. We hypothesise that angular analysis of the geometric centreline for a three-dimensional surface could be used to automatically generate building block structures for the multiblock hexahedral mesh generation. Our algorithm uses a set of user-defined points and parameters to automatically generate a multiblock structure based on a surface's geometric centreline. This significantly reduces the time required for model development. We have applied this algorithm to 47 bones of varying geometries and successfully generated a FE mesh in all cases. This work represents significant advancement in automatically generating multiblock structures for a wide range of geometries.     

Morphing the feature-based multi-blocks of normative/healthy vertebral geometries to scoliosis vertebral geometries: development of personalized finite element models

Personalized Finite Element (FE) models and hexahedral elements are preferred for biomechanical investigations. Feature-based multi-block methods are used to develop anatomically accurate personalized FE models with hexahedral mesh. It is tedious to manually construct multi-blocks for large number of geometries on an individual basis to develop personalized FE models. Mesh-morphing method mitigates the aforementioned tediousness in meshing personalized geometries every time, but leads to element warping and loss of geometrical data. Such issues increase in magnitude when normative spine FE model is morphed to scoliosis-affected spinal geometry. The only way to bypass the issue of hex-mesh distortion or loss of geometry as a result of morphing is to rely on manually constructing the multi-blocks for scoliosis-affected spine geometry of each individual, which is time intensive. A method to semi-automate the construction of multi-blocks on the geometry of scoliosis vertebrae from the existing multi-blocks of normative vertebrae is demonstrated in this paper. High-quality hexahedral elements were generated on the scoliosis vertebrae from the morphed multi-blocks of normative vertebrae. Time taken was 3 months to construct the multi-blocks for normative spine and less than a day for scoliosis. Efforts taken to construct multi-blocks on personalized scoliosis spinal geometries are significantly reduced by morphing existing multi-blocks.     

Growing multiblock structures: a semi-automated approach to block placement for multiblock hexahedral meshing

Finite element (FE) analysis is a cornerstone of orthopaedic biomechanics research. Three-dimensional medical imaging provides sufficient resolution for the subject-specific FE models to be generated from these data-sets. FE model development requires discretisation of a three-dimensional domain, which can be the most time-consuming component of a FE study. Hexahedral meshing tools based on the multiblock method currently rely on the manual placement of building blocks for mesh generation. We hypothesise that angular analysis of the geometric centreline for a three-dimensional surface could be used to automatically generate building block structures for the multiblock hexahedral mesh generation. Our algorithm uses a set of user-defined points and parameters to automatically generate a multiblock structure based on a surface's geometric centreline. This significantly reduces the time required for model development. We have applied this algorithm to 47 bones of varying geometries and successfully generated a FE mesh in all cases. This work represents significant advancement in automatically generating multiblock structures for a wide range of geometries.     

Proposing a Radial Basis Function and CSDM Indices to Predict the Traumatic Brain Injury Risk

BackgroundMany head injury indices and finite element (FE) head models have been proposed to predict traumatic brain injury (TBI). Although FE head models are suitable methods with high accuracy, they are computationally intensive. Head motion-based brain injury criteria are usually fast tools with lower accuracy. So, the objective of this study is to propose new criteria along with an artificial neural network model to predict TBI risks, which can be fast and accurate.MethodsFor this purpose, 250 FE head simulations have been carried out at 5 magnitudes and 50 rotational impact directions using the SIMon model. The effects of directions and magnitudes of rotational impacts were assessed for cumulative strain damage measure (CSDM) values. Next, statistical analysis and neural network were applied to predict CSDM values.ResultsThe results of the present research showed that the direction of rotation in the sagittal and frontal planes had a considerable effect on the CSDM values. Furthermore, new brain injury indices and a radial basis function neural network have been proposed to predict CSDM values which having high correlation coefficients with SIMon responses.ConclusionsThe results of this research demonstrated that rotational impact directions should be used to develop new head injury criteria being able to predict CSDM values. However, findings of present research proved that head motion-based brain injury criteria and RBF network can be used to predict FE head model responses with high speed and accuracy.     

Design and numerical implementation of a 3-D non-linear viscoelastic constitutive model for brain tissue during impact

Finite Element (FE) head models are often used to understand mechanical response of the head and its contents during impact loading in the head. Current FE models do not account for non-linear viscoelastic material behavior of brain tissue. We developed a new non-linear viscoelastic material model for brain tissue and implemented it in an explicit FE code. To obtain sufficient numerical accuracy for modeling the nearly incompressible brain tissue, deviatoric and volumetric stress contributions are separated. Deviatoric stress is modeled in a non-linear viscoelastic differential form. Volumetric behavior is assumed linearly elastic. Linear viscoelastic material parameters were derived from published data on oscillatory experiments, and from ultrasonic experiments. Additionally, non-linear parameters were derived from stress relaxation (SR) experiments at shear strains up to 20%. The model was tested by simulating the transient phase in the SR experiments not used in parameter determination (strains up to 20%, strain rates up to 8s(-1)). Both time- and strain-dependent behavior were predicted accurately (R2>0.96) for strain and strain rates applied. However, the stress was overestimated systematically by approximately 31% independent of strain(rate) applied. This is probably caused by limitations of the experimental data at hand.     

Integration of Diffusion Magnetic Resonance Tractography into tomotherapy radiation treatment planning for high-grade gliomas

IntroductionFractionated radiotherapy in brain tumors is commonly associated with several detrimental effects, largely related to the higher radiosensitivity of the white matter (WM) with respect to gray matter. However, no dose constraints are applied to preserve WM structures at present. Magnetic Resonance (MR) Tractography is the only technique that allows to visualize in vivo the course of WM eloquent tracts in the brain. In this study, the feasibility of integrating MR Tractography in tomotherapy treatment planning has been investigated, with the aim to spare eloquent WM regions from the dose delivered during treatment.MethodsNineteen high grade glioma patients treated with fractionated radiotherapy were enrolled. All the patients underwent pre-treatment MR imaging protocol including Diffusion Tensor Imaging (DTI) acquisitions for MR Tractography analysis. Bilateral tracts involved in several motor, language, cognitive functions were reconstructed and these fiber bundles were integrated into the Tomotherapy Treatment planning system. The original plans without tracts were compared with the optimized plans incorporating the fibers, to evaluate doses to WM structures in the two differently optimized plans.ResultsNo significant differences were found between plans in terms of planning target volume (PTV) coverage between the original plans and the optimized plans incorporating fiber tracts. Comparing the mean as well as the maximal dose (Dmean and Dmax), a significant dose reduction was found for most of the tracts. The dose sparing was more relevant for contralateral tracts (P < 0.0001).ConclusionThe integration of MR Tractography into radiotherapy planning is feasible and beneficial to preserve important WM structures without reducing the clinical goal of radiation treatment.     

The peculiar properties of the falx and tentorium in brain injury biomechanics

The influence of the falx and tentorium on brain injury biomechanics during impact was studied with finite element (FE) analysis. Three detailed 3D FE head models were created based on the images of a healthy, normal size head. Two of the models contained the addition of falx and tentorium with material properties from previously published experiments. Impact loadings from a reconstructed concussive case in a sport accident were applied to the two players involved. The results suggested that the falx and tentorium could induce large strains to the surrounding brain tissues, especially to the corpus callosum and brainstem. The tentorium seemed to constrain the motion of the cerebellum while inducing large strain in the brainstem in both players involved in the accident (one player had mainly coronal head rotation and the other had both coronal and transversal rotations). Since changed strain levels were observed in the brainstem and corpus callosum, which are classical sites for diffuse axonal injuries (DAI), we confirmed the importance of using accurate material properties for falx and tentorium in a FE head model when studying traumatic brain injuries.     

In Vivo 7T MRI of the Non-Human Primate Brainstem

Structural brain imaging provides a critical framework for performing stereotactic and intraoperative MRI-guided surgical procedures, with procedural efficacy often dependent upon visualization of the target with which to operate. Here, we describe tools for in vivo, subject-specific visualization and demarcation of regions within the brainstem. High-field 7T susceptibility-weighted imaging and diffusion-weighted imaging of the brain were collected using a customized head coil from eight rhesus macaques. Fiber tracts including the superior cerebellar peduncle, medial lemniscus, and lateral lemniscus were identified using high-resolution probabilistic diffusion tractography, which resulted in three-dimensional fiber tract reconstructions that were comparable to those extracted from sequential application of a two-dimensional nonlinear brain atlas warping algorithm. In the susceptibility-weighted imaging, white matter tracts within the brainstem were also identified as hypointense regions, and the degree of hypointensity was age-dependent. This combination of imaging modalities also enabled identifying the location and extent of several brainstem nuclei, including the periaqueductal gray, pedunculopontine nucleus, and inferior colliculus. These clinically-relevant high-field imaging approaches have potential to enable more accurate and comprehensive subject-specific visualization of the brainstem and to ultimately improve patient-specific neurosurgical targeting procedures, including deep brain stimulation lead implantation.     

Effects of idealized joint geometry on finite element predictions of cartilage contact stresses in the hip

Computational models may have the ability to quantify the relationship between hip morphology, cartilage mechanics and osteoarthritis. Most models have assumed the hip joint to be a perfect ball and socket joint and have neglected deformation at the bone-cartilage interface. The objective of this study was to analyze finite element (FE) models of hip cartilage mechanics with varying degrees of simplified geometry and a model with a rigid bone material assumption to elucidate the effects on predictions of cartilage stress. A previously validated subject-specific FE model of a cadaveric hip joint was used as the basis for the models. Geometry for the bone-cartilage interface was either: (1) subject-specific (i.e. irregular), (2) spherical, or (3) a rotational conchoid. Cartilage was assigned either a varying (irregular) or constant thickness (smoothed). Loading conditions simulated walking, stair-climbing and descending stairs. FE predictions of contact stress for the simplified models were compared with predictions from the subject-specific model. Both spheres and conchoids provided a good approximation of native hip joint geometry (average fitting error ～0.5 mm). However, models with spherical/conchoid bone geometry and smoothed articulating cartilage surfaces grossly underestimated peak and average contact pressures (50% and 25% lower, respectively) and overestimated contact area when compared to the subject-specific FE model. Models incorporating subject-specific bone geometry with smoothed articulating cartilage also underestimated pressures and predicted evenly distributed patterns of contact. The model with rigid bones predicted much higher pressures than the subject-specific model with deformable bones. The results demonstrate that simplifications to the geometry of the bone-cartilage interface, cartilage surface and bone material properties can have a dramatic effect on the predicted magnitude and distribution of cartilage contact pressures in the hip joint.     

Computational neurotrauma—design, simulation, and analysis of controlled cortical impact model

The controlled cortical impact (CCI) model is widely used in many laboratories to study traumatic brain injury (TBI). Although external impact parameters during CCI tests could be clearly defined, little is known about the internal tissue-level mechanical responses of the rat brain. Furthermore, the external impact parameters tend to vary considerably among different labs making the comparison of research findings difficult if not impossible. In this study, a design of computer experiments was performed with typical external impact parameters commonly found in the literature. An anatomically detailed finite element (FE) rat brain model was used to simulate the CCI experiments to correlate external mechanical parameters (impact depth, impact velocity, impactor shape, impactor size, and craniotomy pattern) with rat brain internal responses, as predicted by the FE model. Systematic analysis of the results revealed that impact depth was the leading factor affecting the predicted brain internal responses. Interestingly, impactor shape ranked as the second most important factor, surpassing impactor diameter and velocity which were commonly reported in the literature as indicators of injury severity along with impact depth. The differences in whole brain response due to a unilateral or a bilateral craniotomy were small, but those of regional intracranial tissue stretches were large. The interaction effects of any two external parameters were not significant. This study demonstrates the potential of using numerical FE modeling to engineer better experimental TBI models in the future.     

Injury prediction and vulnerability assessment using strain and susceptibility measures of the deep white matter

Reliable prediction and diagnosis of concussion is important for its effective clinical management. Previous model-based studies largely employ peak responses from a single element in a pre-selected anatomical region of interest (ROI) and utilize a single training dataset for injury prediction. A more systematic and rigorous approach is necessary to scrutinize the entire white matter (WM) ROIs as well as ROI-constrained neural tracts. To this end, we evaluated injury prediction performances of the 50 deep WM regions using predictor variables based on strains obtained from simulating the 58 reconstructed American National Football League head impacts. To objectively evaluate performance, repeated random subsampling was employed to split the impacts into independent training and testing datasets (39 and 19 cases, respectively, with 100 trials). Univariate logistic regressions were conducted based on training datasets to compute the area under the receiver operating characteristic curve (AUC), while accuracy, sensitivity, and specificity were reported based on testing datasets. Two tract-wise injury susceptibilities were identified as the best overall via pair-wise permutation test. They had comparable AUC, accuracy, and sensitivity, with the highest values occurring in superior longitudinal fasciculus (SLF; 0.867–0.879, 84.4–85.2, and 84.1–84.6%, respectively). Using metrics based on WM fiber strain, the most vulnerable ROIs included genu of corpus callosum, cerebral peduncle, and uncinate fasciculus, while genu and main body of corpus callosum, and SLF were among the most vulnerable tracts. Even for one un-concussed athlete, injury susceptibility of the cingulum (hippocampus) right was elevated. These findings highlight the unique injury discriminatory potentials of computational models and may provide important insight into how best to incorporate WM structural anisotropy for investigation of brain injury.     

On the use of a patient-specific rapid-prototyped model to simulate the response of the human head to impact and comparison with analytical and finite element models

Every year, thousands of fatalities result from head injuries, the majority of which are sustained in automotive accidents. In this paper, an experimental study of the response of the human head to impact is presented. A rapid prototyped model of a human head was generated based on high-resolution magnetic resonance imaging (MRI) scan data. The physical model was subjected to low velocity impacts using a metallic pendulum and a sensitivity study was performed to explore the influence of various parameters, including mass and velocity of the impactor, on the response. The experimental response characteristics are compared with predictions from an analytical model as well as with numerical predictions from finite element (FE) models generated from the same MRI data set. The results from the experimental tests closely match those predicted by both the analytical and the FE models and thus provide us with substantive corroboration of all three approaches. The remarkable agreement obtained between the measured response characteristics of rapid-prototyped skulls and numerical (FE) models obtained from in vivo MRI data clearly demonstrates the potential use of rapid-prototyping to generate experimental models for head impact studies, and, more generally, for the study of the response of complex bio-structures to loading. In addition, the quantitative and qualitative accuracy of the predictions from the analytical model is clearly demonstrated by the FE and experimental corroboration. In particular, the analytical prediction that, as impact mass drops the impact duration becomes increasingly short, appears to be substantiated, which has important implications for the onset of high pressure and shear strain gradients in the brain with potentially deleterious effects.     

Comparison of strain measurement in the mouse forearm using subject-specific finite element models,strain gaging,and digital image correlation

Mechanical loading in bone leads to the activation of bone-forming pathways that are most likely associated with a minimum strain threshold being experienced by the osteocyte. To investigate the correlation between cellular response and mechanical stimuli, researchers must develop accurate ways to measure/compute strain both externally on the bone surface and internally at the osteocyte level. This study investigates the use of finite element (FE) models to compute bone surface strains on the mouse forearm. Strains from three FE models were compared to data collected experimentally through strain gaging and digital image correlation (DIC). Each FE model was assigned subject-specific bone properties and consisted of one-dimensional springs representing the interosseous membrane. After three-point bending was performed on the ulnae and radii, moment of inertia was determined from microCT analysis of the bone region between the supports and then used along with standard beam analyses to calculate the Young’s modulus. Non-contact strain measurements from DIC were determined to be more suitable for validating numerical results than experimental data obtained through conventional strain gaging. When comparing strain responses in the three ulnae, we observed a 3–14% difference between numerical and DIC strains while the strain gage values were 37–56% lower than numerical values. This study demonstrates a computational approach for capturing bone surface strains in the mouse forearm. Ultimately, strains from these macroscale models can be used as inputs for microscale and nanoscale FE models designed to analyze strains directly in the osteocyte lacunae.     

Development and validation of a subject-specific finite element model of a human clavicle

This study aimed to develop and validate a finite element (FE) model of a human clavicle which can predict the structural response and bone fractures under both axial compression and anterior–posterior three-point bending loads. Quasi-static non-injurious axial compression and three-point bending tests were first conducted on a male clavicle followed by a dynamic three-point bending test to fracture. Then, two types of FE models of the clavicle were developed using bone material properties which were set to vary with the computed tomography image density of the bone. A volumetric solid FE model comprised solely of hexahedral elements was first developed. A solid-shell FE model was then created which modelled the trabecular bone as hexahedral elements and the cortical bone as quadrilateral shell elements. Finally, simulations were carried out using these models to evaluate the influence of variations in cortical thickness, mesh density, bone material properties and modelling approach on the biomechanical responses of the clavicle, compared with experimental data. The FE results indicate that the inclusion of density-based bone material properties can provide a more accurate reproduction of the force–displacement response and bone fracture timing than a model with uniform bone material properties. Inclusion of a variable cortical thickness distribution also slightly improves the ability of the model to predict the experimental response. The methods developed in this study will be useful for creating subject-specific FE models to better understand the biomechanics and injury mechanism of the clavicle.