Development and initial evaluation of a finite element model of the pediatric craniocervical junction

Finite Element Modelling of a Synthetic Paediatric Spine for Biomechanical Investigation

Studies on paediatric spines commonly use human adult or immature porcine spines as specimens, because it is difficult to obtain actual paediatric specimens. There are quite obvious differences, such as geometry, size, bone morphology, and orientation of facet joint for these specimens, compared to paediatric spine. Hence, development of synthetic models that can behave similarly to actual paediatric spines, particularly in term of range of motion (ROM), could provide a significant contribution for paediatric spine research. This study aims to develop a synthetic paediatric spine using finite element modelling and evaluate the reliability of the model by comparing it with the experimental data under certain load conditions. The ROM of the paediatric spine was measured using a validated FE model at ±0.5 Nm moment in order to determine the moment required by the synthetic spine to achieve the same ROM. The results showed that the synthetic spine required two moments, ±2 Nm for lateral-bending and axial rotation, and ±3 Nm for flexion-extension, to obtain the paediatric ROM. The synthetic spine was shown to be stiffer in flexion-extension but more flexible in lateral bending than the paediatric FE model, possibly as a result of the intervertebral disc's simplified shape and the disc's weak bonding with the vertebrae. Nevertheless, the synthetic paediatric spine has promising potential in the future as an alternative paediatric spine model for biomechanical investigation of paediatric cases.

FEBio finite element model of a pediatric cervical spine

OBJECTIVE

The underlying biomechanical differences between the pediatric and adult cervical spine are incompletely understood. Computational spine modeling can address that knowledge gap. Using a computational method known as finite element modeling, the authors describe the creation and evaluation of a complete pediatric cervical spine model.

METHODS

Using a thin-slice CT scan of the cervical spine from a 5-year-old boy, a 3D model was created for finite element analysis. The material properties and boundary and loading conditions were created and model analysis performed using open-source software. Because the precise material properties of the pediatric cervical spine are not known, a published parametric approach of scaling adult properties by 50%, 25%, and 10% was used. Each scaled finite element model (FEM) underwent two types of simulations for pediatric cadaver testing (axial tension and cardinal ranges of motion [ROMs]) to assess axial stiffness, ROM, and facet joint force (FJF). The authors evaluated the axial stiffness and flexion-extension ROM predicted by the model using previously published experimental measurements obtained from pediatric cadaveric tissues.

RESULTS

In the axial tension simulation, the model with 50% adult ligamentous and annulus material properties predicted an axial stiffness of 49 N/mm, which corresponded with previously published data from similarly aged cadavers (46.1 ± 9.6 N/mm). In the flexion-extension simulation, the same 50% model predicted an ROM that was within the range of the similarly aged cohort of cadavers. The subaxial FJFs predicted by the model in extension, lateral bending, and axial rotation were in the range of 1-4 N and, as expected, tended to increase as the ligament and disc material properties decreased.

CONCLUSIONS

A pediatric cervical spine FEM was created that accurately predicts axial tension and flexion-extension ROM when ligamentous and annulus material properties are reduced to 50% of published adult properties. This model shows promise for use in surgical simulation procedures and as a normal comparison for disease-specific FEMs.

Experimental Analysis of Fabricated Synthetic Midthoracic Paediatric Spine as Compared to the Porcine Spine Based on Range of Motion (ROM)

The present study is aimed at investigating the mechanical behaviour of fabricated synthetic midthoracic paediatric spine based on range of motion (ROM) as compared to porcine spine as the biological specimen. The main interest was to ensure that the fabricated synthetic model could mimic the biological specimen behaviour. The synthetic paediatric spine was designed as a 200% scaled-up model to fit into the Bionix Servohydraulic spine simulator. Biomechanical tests were conducted to measure the ROM and nonlinearity of sigmoidal curves at six degrees of freedom (DOF) with moments at ±4 Nm before the specimens failed. Results were compared with the porcine spine (biological specimen). The differences found between the lateral bending and axial rotation of synthetic paediatric spine as compared to the porcine spine were 18% and 3%, respectively, but was still within the range. Flexion extension of the synthetic spine is a bit stiff in comparison of porcine spine with 45% different. The ROM curves of the synthetic paediatric spine exhibited nonlinearities for all motions as the measurements of neutral zone (NZ) and elastic zone (EZ) stiffness were below “1.” Therefore, it showed that the proposed synthetic paediatric spine behaved similarly to the biological specimen, particularly on ROM.

Finite element modeling to compare craniocervical motion in two age-matched pediatric patients without or with Down syndrome: implications for the role of bony geometry in craniocervical junction instability

OBJECTIVE

Instability of the craniocervical junction (CCJ) is a well-known finding in patients with Down syndrome (DS); however, the relative contributions of bony morphology versus ligamentous laxity responsible for abnormal CCJ motion are unknown. Using finite element modeling, the authors of this study attempted to quantify those relative differences.

METHODS

Two CCJ finite element models were created for age-matched pediatric patients, a patient with DS and a control without DS. Soft tissues and ligamentous structures were added based on bony landmarks from the CT scans. Ligament stiffness values were assigned using published adult ligament stiffness properties. Range of motion (ROM) testing determined that model behavior most closely matched pediatric cadaveric data when ligament stiffness values were scaled down to 25% of those found in adults. These values, along with those assigned to the other soft-tissue materials, were identical for each model to ensure that the only variable between the two was the bone morphology. The finite element models were then subjected to three types of simulations to assess ROM, anterior-posterior (AP) translation displacement, and axial tension.

RESULTS

The DS model exhibited more laxity than the normal model at all levels for all of the cardinal ROMs and AP translation. For the CCJ, the flexion-extension, lateral bending, axial rotation, and AP translation values predicted by the DS model were 40.7%, 52.1%, 26.1%, and 39.8% higher, respectively, than those for the normal model. When simulating axial tension, the soft-tissue structural stiffness values predicted by the DS and normal models were nearly identical.

CONCLUSIONS

The increased laxity exhibited by the DS model in the cardinal ROMs and AP translation, along with the nearly identical soft-tissue structural stiffness values exhibited in axial tension, calls into question the previously held notion that ligamentous laxity is the sole explanation for craniocervical instability in DS.

Asymptomatic rotational vertebral artery compression in a child due to head positioning for cranial surgery: illustrative case

BACKGROUND

The authors recently reported a series of children with vertebral artery (VA) compression during head turning who presented with recurrent posterior circulation stroke. Whether VA compression occurs during head positioning for cranial surgery is unknown.

OBSERVATIONS

The authors report a case of a child with incidental rotational occlusion of the VA observed during surgical head positioning for treatment of an intracranial arteriovenous fistula. Intraoperative angiography showed dynamic V₃ occlusion at the level of C2 with distal reconstitution via a muscular branch “jump” collateral, supplying reduced flow to the V₄ segment. She had no clinical history or imaging suggesting acute or prior stroke. Sequential postoperative magnetic resonance imaging scans demonstrated signal abnormality of the left rectus capitus muscle, suggesting ischemic edema.

LESSONS

This report demonstrates that rotational VA compression during neurosurgical head positioning can occur in children but may be asymptomatic due to the presence of muscular VA–VA “jump” collaterals and contralateral VA flow. Although unilateral VA compression may be tolerated by children with codominant VAs, diligence when rotating the head away from a dominant VA is prudent during patient positioning to avoid posterior circulation ischemia or thromboembolism.

Biomechanical effects of laminectomies in the human lumbar spine: a finite element study

BACKGROUND CONTEXT

Previous studies have analyzed the effect of laminectomy on intervertebral disc (IVD), facet-joint-forces (FJF), and range of motion (ROM), while only two have specifically reported stresses at the pars interarticularis (PI) with posterior element resection. These studies have been performed utilizing a single subject, questioning their applications to a broader population.

PURPOSE

We investigate the effect of graded PI resection in a three-dimensional manner on PI stress to provide surgical guidelines for avoidance of iatrogenic instability following lumbar laminectomy. Additionally, quantified FJF and IVD stresses can provide further insight into the development of adjacent segment disease.

STUDY DESIGN

Biomechanical finite element (FE) method investigation of the lumbar spine.

METHODS

FE models of the lumbar spine of three subjects were created using the open-source finite element software, FEBio. Single-level laminectomy, two-level laminectomy, and ventral-to-dorsal PI resection simulations were performed with varying degrees of PI resection from 0% to 75% of the native PI. These models were taken through cardinal ROM under standard loading conditions and PI stresses, FJF, IVD stresses, and ROM were analyzed.

RESULTS

The three types of laminectomy simulated in this study showed a nonlinear increase in PI stress with increased bone resection. Axial rotation generated the most stress at the PI followed by flexion, extension and lateral bending. At 75% bone resection all three types of laminectomy produced PI stresses that were near the ultimate strength of human cortical bone during axial rotation. FJF decreased with increased bone resection for the three laminectomies simulated. This was most notable in axial rotation followed by extension and lateral bending. IVD stresses varied greatly between the nonsurgical models and likewise the effect of laminectomy on IVD stresses varied between subjects. ROM was mostly unaffected by the laminectomies performed in this study.

CONCLUSIONS

Regarding the risk of iatrogenic spondylolisthesis, the combined results are sufficient evidence to suggest surgeons should be particularly cautious when PI resection exceeds 50% bone resection for all laminectomies included in this study. Lastly, the effects seen in FJF and IVD stresses indicate the degree to which the remainder of the spine must experience compensatory biomechanical changes as a result of the surgical intervention.

Febio finite element models of the human cervical spine

Finite element (FE) analysis has proven to be useful when studying the biomechanics of the cervical spine. Although many FE studies of the cervical spine have been published, they typically develop their models using commercial software, making the sharing of models between researchers difficult. They also often model only one part of the cervical spine. The goal of this study was to develop and evaluate three FE models of the adult cervical spine using open-source software and to freely provide these models to the scientific community. The models were created from computed tomography scans of 26-, 59-, and 64-year old female subjects. These models were evaluated against previously published experimental and FE data. Despite the fact that all three models were assigned identical material properties and boundary conditions, there was notable variation in their biomechanical behavior. It was therefore apparent that these differences were the result of morphological differences between the models.

The atlantoaxial capsular ligaments and transverse ligament are the primary stabilizers of the atlantoaxial joint in the craniocervical junction: a finite element analysis

OBJECTIVE

Prior studies have provided conflicting evidence regarding the contribution of key ligamentous structures to atlantoaxial (AA) joint stability. Many of these studies employed cadaveric techniques that are hampered by the inherent difficulties of testing isolated-injury scenarios. Analysis with validated finite element (FE) models can overcome some of these limitations. In a previous study, the authors completed an FE analysis of 5 subject-specific craniocervical junction (CCJ) models to investigate the biomechanics of the occipitoatlantal joint and identify the ligamentous structures essential for its stability. Here, the authors use these same CCJ FE models to investigate the biomechanics of the AA joint and to identify the ligamentous structures essential for its stability.

METHODS

Five validated CCJ FE models were used to simulate isolated- and combined ligamentous–injury scenarios of the transverse ligament (TL), tectorial membrane (TM), alar ligament (AL), occipitoatlantal capsular ligament, and AA capsular ligament (AACL). All models were tested with rotational moments (flexion-extension, axial rotation, and lateral bending) and anterior translational loads (C2 constrained with anterior load applied to the occiput) to simulate physiological loading and to assess changes in the atlantodental interval (ADI), a key radiographic indicator of instability.

RESULTS

Isolated AACL injury significantly increased range of motion (ROM) under rotational moment at the AA joint for flexion, lateral bending, and axial rotation, which increased by means of 28.0% ± 10.2%, 43.2% ± 15.4%, and 159.1% ± 35.1%, respectively (p ≤ 0.05 for all). TL removal simulated under translational loads resulted in a significant increase in displacement at the AA joint by 89.3% ± 36.6% (p < 0.001), increasing the ADI from 2.7 mm to 4.5 mm. An AACL injury combined with an injury to any other ligament resulted in significant increases in ROM at the AA joint, except when combined with injuries to both the TM and the ALs. Similarly, injury to the TL combined with injury to any other CCJ ligament resulted in a significant increase in displacement at the AA joint (significantly increasing ADI) under translational loads.

CONCLUSIONS

Using FE modeling techniques, the authors showed a significant reliance of isolated- and combined ligamentous–injury scenarios on the AACLs and TL to restrain motion at the AA joint. Isolated injuries to other structures alone, including the AL and TM, did not result in significant increases in either AA joint ROM or anterior displacement.

Morphometric changes at the craniocervical junction during childhood

OBJECTIVE

Current understanding of how the pediatric craniocervical junction develops remains incomplete. Measurements of anatomical relationships at the craniocervical junction can influence clinical and surgical decision-making. The purpose of this analysis was to quantitatively define clinically relevant craniocervical junction measurements in a population of children with CT scans that show normal anatomy.

METHODS

A total of 1458 eligible patients were identified from children between 1 and 18 years of age who underwent cervical spine CT scanning at a single institution. Patients were separated by both sex and age in years into 34 groups. Following this, patients within each group were randomly selected for inclusion until a target of 15 patients in each group had been reached. Each patient underwent measurement of the occipital condyle-C1 interval (CCI), pB-C2, atlantodental interval (ADI), basion-dens interval (BDI), basion-opisthion diameter (BOD), basion-axial interval (BAI), dens angulation, and canal diameter at C1. Mean values were calculated in each group. Each measurement was performed by two teams and compared for intraclass correlation coefficient (ICC).

RESULTS

The data showed that CCI, ADI, BDI, and dens angulation decrease in magnitude throughout childhood, while pB-C2, PADI, BAI, and BOD increase throughout childhood, with an ICC of fair to good (range 0.413-0.912). Notably, CCI decreases continuously on coronal CT scans, whereas on parasagittal CT scans, CCI does not decrease until after age 9, when it shows a continuous decline similar to measurements on coronal CT scans.

CONCLUSIONS

These morphometric analyses establish parameters for normal pediatric craniocervical spine growth for each year of life up to 18 years. The data should be considered when evaluating children for potential surgical intervention.

The occipitoatlantal capsular ligaments are the primary stabilizers of the occipitoatlantal joint in the craniocervical junction: a finite element analysis

OBJECTIVE

There is contradictory evidence regarding the relative contribution of the key stabilizing ligaments of the occipitoatlantal (OA) joint. Cadaveric studies are limited by the nature and the number of injury scenarios that can be tested to identify OA stabilizing ligaments. Finite element (FE) analysis can overcome these limitations and provide valuable data in this area. The authors completed an FE analysis of 5 subject-specific craniocervical junction (CCJ) models to investigate the biomechanics of the OA joint and identify the ligamentous structures essential for stability.

METHODS

Isolated and combined injury scenarios were simulated under physiological loads for 5 validated CCJ FE models to assess the relative role of key ligamentous structures on OA joint stability. Each model was tested in flexion-extension, axial rotation, and lateral bending in various injury scenarios. Isolated ligamentous injury scenarios consisted of either decreasing the stiffness of the OA capsular ligaments (OACLs) or completely removing the transverse ligament (TL), tectorial membrane (TM), or alar ligaments (ALs). Combination scenarios were also evaluated.

RESULTS

An isolated OACL injury resulted in the largest percentage increase in all ranges of motion (ROMs) at the OA joint compared with the other isolated injuries. Flexion, extension, lateral bending, and axial rotation significantly increased by 12.4% ± 7.4%, 11.1% ± 10.3%, 83.6% ± 14.4%, and 81.9% ± 9.4%, respectively (p ≤ 0.05 for all). Among combination injuries, OACL+TM+TL injury resulted in the most consistent significant increases in ROM for both the OA joint and the CCJ during all loading scenarios. OACL+AL injury caused the most significant percentage increase for OA joint axial rotation.

CONCLUSIONS

These results demonstrate that the OACLs are the key stabilizing ligamentous structures of the OA joint. Injury of these primary stabilizing ligaments is necessary to cause OA instability. Isolated injuries of TL, TM, or AL are unlikely to result in appreciable instability at the OA joint.

Chiari-related scoliosis: a single-center experience with long-term radiographic follow-up and relationship to deformity correction

OBJECTIVE Previous reports have addressed the short-term response of patients with Chiari-related scoliosis (CRS) to suboccipital decompression and duraplasty (SODD); however, the long-term behavior of the curve has not been well defined. The authors undertook a longitudinal study of a cohort of patients who underwent SODD for CRS to determine whether there are factors related to Chiari malformation (CM) that predict long-term scoliotic curve behavior and need for deformity correction. METHODS The authors retrospectively reviewed cases in which patients underwent SODD for CRS during a 14-year period at a single center. Clinical (age, sex, and associated disorders/syndromes) and radiographic (CM type, tonsillar descent, pBC2 line, clival-axial angle [CXA], syrinx length and level, and initial Cobb angle) information was evaluated to identify associations with the primary outcome: delayed thoracolumbar fusion for progressive scoliosis. RESULTS Twenty-eight patients were identified, but 4 were lost to follow-up and 1 underwent fusion within a year. Among the remaining 23 patients, 11 required fusion surgery at an average of 88.3 ± 15.4 months after SODD, including 7 (30%) who needed fusion more than 5 years after SODD. On univariate analysis, a lower CXA (131.5° ± 4.8° vs 146.5° ± 4.6°, p = 0.034), pBC2 > 9 mm (64% vs 25%, p = 0.06), and higher initial Cobb angle (35.1° ± 3.6° vs 22.8° ± 4.0°, p = 0.035) were associated with the need for thoracolumbar fusion. Multivariable modeling revealed that lower CXA was independently associated with a need for delayed thoracolumbar fusion (OR 1.12, p = 0.0128). CONCLUSIONS This investigation demonstrates the long-term outcome and natural history of CRS after SODD. The durability of the effect of SODD on CRS and curve behavior is poor, with late curve progression occurring in 30% of patients. Factors associated with CRS progression include an initial pBC2 > 9 mm, lower CXA, and higher Cobb angle. Lower CXA was an independent predictor of delayed thoracolumbar fusion. Further study is necessary on a larger cohort of patients to fully elucidate this relationship.