EO.3.S8

Introduction
The “study of fractures" is nearly synonymous with orthopaedic research, yet our discipline has long worked with a limited set of computational tools when studying the fracture mechanics of its most important material class: porous compressible solids. Conventional meshed-based simulation techniques such as finite element analysis (FEA) are limited in their ability to model the crushing, cracking, fragmentation, and compaction phenomena of both natural porous tissue and new biomimetic 3D printed implants with porous lattice structures.Accurate simulation of trabecular bone biomechanics is critical to the development of mass-produced and patient-specific devices that reduce rates of implant migration and improve long-term patient health, particularly for persons with osteoporosis.

Methods and results
We present a novel “mesh-free,” or particle-based, approach to the modelling of porous compressible solids in orthopaedics that is capable of accurately simulating complex phenomena such as lattice fracture and densification, useful in predicting the safety and efficacy of both mass manufactured and patient specific implants. Our discussion begins with a brief survey of the theoretical basis of mesh-free simulation and how they compare to FEA, followed by presentation of three original case studies that explore the practical application of this approach to problems in orthopaedics: 

1. Generation of microCT-derived bone tissue and validation versus physical cadaveric testing across multiple bone densities and trabecular morphologies (Figure 1); 
2. Simulated migration and cut-out testing of several commonly-used devices in trauma and spinal surgery in porous PU foam; and
3. Compression testing of a 3D printed titanium porous spinal cage used in fusion surgery (Figure 2).

Figure 1. MicroCT-derived simulated bone tissue used to evaluate migration and cut-out of a proximal femoral device under physiological loading.

Figure 2. Simulated loading to failure of a 3D printed titanium alloy spinal spacer.

Conclusion
We review the findings of the three case studies and discuss potential next steps for mesh-free computer simulation in orthopaedics research, clinical care, and by industry for the development of mass produced and patient specific devices, biomaterials, and surgical robotics. 

Bone is a remarkable material, capable of withstanding high mechanical loads while remaining lightweight and self-regenerating throughout most of our lifespan. However, with age and hormonal changes, bones weakened by osteoporosis become prone to fractures, posing significant clinical challenges. Managing patients at immediate fracture risk remains difficult, as current systemic anti-osteoporotic treatments act slowly and locally.

Local bone augmentation presents an alternative strategy to address these challenges. Over the past decades, techniques such as bone cements and metal implants have been explored to locally reinforce osteoporotic bone. In the spine, percutaneous vertebroplasty—a technique introduced in 1987—involves injecting polymethylmethacrylate (PMMA)-based cement into vertebral bodies. While effective, this procedure is limited by risks of cement leakage, necessitating precise control of injection volume and distribution. Similarly, the proximal femur has been a target for reinforcement using polymer-PMMA implants, calcium phosphate cements, or bioactive hydrogels. However, in both cases, determining the optimal regions for augmentation and the feasibility of material injection remains a significant challenge.

To address these issues, we adopted a multidisciplinary approach integrating advanced imaging, mechanical testing, and computational modeling. High-resolution 4D computed tomography (4D CT) was used to capture experimental data on cement flow in real bone structures. This data was then used in finite element models to evaluate mechanical strength improvements under both static loading and advanced side-fall simulations, reflecting real-world fracture scenarios. Furthermore, we used lattice Boltzmann simulations to predict cement flow dynamics at high resolution, allowing us to map precise material distributions based on individual bone microarchitecture.

While this approach demonstrates significant potential to optimize local bone reinforcement techniques, it also highlights the current gap between these advanced imaging and computational tools and their feasibility for clinical application. The future of this technology is promising, yet the need for accessible and efficient clinical solutions remains a critical hurdle to overcome.

Complex and unstable proximal humerus fractures (PHF) in the elderly are challenging to treat with considerable rates of failures even with state-of-the-art locking plates. A large part of these complications is related to mechanical factors, with loss of reduction via screw perforation and cut-out being the predominant failure modes.[1] Alleviating the risk of mechanical failures would contribute to improved outcomes of PHF. Computer simulations such as finite element (FE) analyses could help understand the reasons behind the multi-factorial problem of mechanical failures and aid in the improved use of currently existing fixation technologies or develop novel advanced designs ensuring higher stability. Towards this aim, we have developed an FE simulation framework for PHF that allowed incorporation of anatomical variance with nearly 50 virtual subjects, seven different fracture patterns, freely configurable locking plate fixations and various loading schemes.[2] We validated the underlying FE analysis technology to predict cyclic screw cut-out failure experimentally measured in instrumented cadaveric human PHF specimens (R²=90).[3] A series of in silico studies was then performed with the computational framework to systematically address clinically relevant questions related to the type and position of the plate, as well as the length, configuration and cement reinforcement of the locking screws, with each of these involving 24–42 virtual specimens and requiring 504–4608 simulations. The analyses provided clinically relevant findings suggesting maximizing screw length [4] and spread [5], prioritizing calcar screws for augmentation [6] and aiming towards a proximal plate position.[7] Beyond the use of existing implants, we utilized the in silico tool to optimize screw trajectories in a study involving 19 digital subjects and more than 5000 analyses.[8] The resulting design was confirmed in a subsequent biomechanical study to provide significantly improved stability compared to the standard implant.[9] A follow-up in silico study has shown only moderate but significant benefits for patient-specific implants versus standardized optimized ones.[10] These studies demonstrated the potential of validated in silico tools for improved care of PHFs, although the findings require clinical corroboration.

References

[1] Panagiotopoulou et al., Injury,50:2176-2195,2019

[2] Varga et al., Med.Eng.Phys.,57:29-39,2018

[3] Varga et al., J.Mech.Behav.Biomed.Mater.,75:68-74,2017

[4] Fletcher et al., Arch.Orthop.Trauma.Surg.,139:1069-1074,2019

[5] Fletcher et al., J.Shoulder.Elbow.Surg.,28:1816-1823,2019

[6] Varga et al., Bone.Joint.Res.,9:534-542,2020

[7] Fletcher et al., J.Orthop.Res.,37:957-964,2019

[8] Mischler et al., J.Orthop.Translat.,25:96-104,2020

[9] Mischler et al., Front.Bioeng.Biotechnol.,23:10:919721,2022

[10] Schader et al., J.Shoulder.Elbow.Surg.,31:192-200,2022

Abstract text: Loosening and migration of spinal pedicle screws, is a feared clinical complication of spinal fusion treatments in patients with poor bone quality. The authors have developed a novel removable expanding screw technology (1) that was adapted for a pedicle screw application, to reduce the risk of loosening. The implant features two wings deployed from the screw after implant insertion. After extensive voice of customer input from a spinal surgeon expert group, three concepts with unique design features were developed. The aim of this study is evaluate these concepts using a novel particle-based simulation to understand the implant design parameters that influence performance. This simulation models the failure of individual particles (2) which overcomes the limitations of Finite Element simulations with bone-implant failure analysis. The result of the simulation demonstrated 36% less bone damage on a single load cycle for one of the designs that had the most similarity to conventional surgical technique, compared to a standard screw of the same thread geometry. The surgeons were interested in investigating the optimal wing position and wing geometry for this lead design, as an extension to the initial particle-based simulation. The second simulation based the loading scenario on an established study where the screw-implant construct is fixed and the loading applied through the anterior vertebral body (3). To understand where S-REX fits against conventional implants a number of screw implant scenarios were evaluated: non-augmented, small wings, larger wings and PMMA cement augmented. In a 10 cycle simulation vs no augmentation the reduction in subsidence was 28%, 58% and 72% respectively. Based on these results the wings were moved closer to anterior cortex of the vertebral body and their expansion was maximised. Clinician feedback on this final version was positive and this version is currently being manufactured for biomechanical testing in the AO’s laboratories.

References: 1) Oldakowska et al. 2024 2) Kulper et al 2018 3) Leibsch et al 2018

Acknowledgements: the S-REX project is a collaboration between REX Ortho and the AO Foundation.

Keywords : Expanding Implant, 3D simulation, Biomechanics, Spine

Introduction: Orthopaedic device performance is often evaluated using standard methods, but musculoskeletal models can provide additional insights. This study investigates the implementation of biomechanical frameworks to enhance our understanding of orthopaedic device performance. One example presented is stress shielding in a shoulder implant construct. By incorporating realistic material properties and activities of daily living, models were developed to understand device performance under physiological conditions.

Method: A shoulder implant construct finite element model was created to evaluate stress shielding. A mean bone was developed from subject-specific data, which involved calculating the statistical bone shape and volumetric intensity data derived from CT scans. These volumetric data were converted to Young's modulus to represent bone material properties. A shoulder implant was added to the bone, and loads from six different activities of daily living were applied to the construct. Stress shielding was evaluated for individual activities, as well as for the aggregate of all activities.

Results: The analysis showed the effects of stress shielding under different musculoskeletal loads. Differences in stress shielding resulted from the different directions and magnitudes of the loading. The combined effect of all activities was also evaluated, and showed that the area with the highest stress shielding was on the proximal end of the construct.

Conclusion: This framework shows the potential for biomechanical models in evaluating orthopaedic devices. The findings highlight the importance of considering the sources of variability within the system, and the power of modelling to understand their effects.

This case study demonstrates the use of a particle-based bone model to explore the risk of fracture and high bone strains experienced during implantation by impaction of a cementless 3D printed knee implant. The objective was to determine the sensitivity to fracture strains of nine implant press-fit design factors, each at three levels of intensity, under realistic repeated impact conditions to simulate multiple hammer blows by a surgeon, until the implant was fully seated in a prepared medial tibial condyle cadaveric model with heterogeneous bone mineral density distribution derived from a CT scan. This study was designed using a Taguchi method orthogonal array to test twenty-seven design variants and was performed using Alfonso particle-based modeling and simulations. The simulation parameters allowed the model to be sensitive to crack formation such that the influence of the implant press-fit design factors could be analyzed. To quantify the differences in performance between designs, the percentage of failed bone particles in multiple regions of interest were recorded during each simulated test. Trends that were observed included a tendency for cortical bone cracks to form during impaction with greater keel length. Using an unpaired Student’s t-test, we found a statistically significant relationship between the bone failure ratio in the Anterolateral region and cortical bone cracking (p=0.04). This suggests that designs that reduce bone failure in the Anterolateral region may result in fewer cortical bone cracks during impaction.

Introduction: Lumbar interbody fusion (LIF) is the gold standard to treat severe lumbar interbody disk pathologies. Standard cages in titanium and peek exhibit certain limitations, including the potential for subsidence and pseudoarthrosis. This suggests an unmet need for the development of innovative implants, exploiting new materials, such as b.Bone™ (GreenBone Ortho, Italy), a biphasic ceramic obtained through a biomorphic transformation from rattan wood. A biodegradable interbody cage can both induce bone formation and be completely resorbed over time. However, the mechanical resistance of such brittle material becomes a critical aspect to be considered during the design process. Digital twins (DT) allow to save both time and costs compared to standard experimental try-and-error approaches, obtaining directly the optimized geometry to be manufactured and tested.  This work presents a DT-driven cage geometry optimization exploiting a fully-parametric lumbar spine model (J Biomech. 2024 Feb:164:111951) and LIF surgery simulations. The design process’ final goal is the optimized cage prototype manufacturing and test, according to standard ISO 23089-2:2021. 

Method: A lateral approach (LLIF) was selected and simulated through detailed finite element model considering 3 steps: (i) site preparation and vertebrae distraction, (ii) cage insertion and posterior fixation and (iii) application of complex loading conditions combining follower load with a bending movement in flexion, extension, axial rotation and lateral bending. Cage design was iteratively optimized to minimize failed volume and bone-cage interface strains (indicating subsidence). This involved adjusting  cage medio-lateral width and incorporating variable holes for instrumentation accommodation and graft placement. Optimized designs were digitally tested against ISO23089-2 standards (axial compression, torsion, shear and subsidence) before manufacturing.

Results: All cage designs led to a reduction up to 98% in the range of motion, ensuring an effective fixation. In terms of cage mechanical resistance, flexion was the most critical loading condition. Nevertheless, the optimized cage design presented a failed volume below 10% for every simulated task. As expected, the surrounding bone was highly strained, similar or lower than what it is typically observed for standard titanium and peek cages. 

Conclusions: DT allow deep investigation of the mechanical response of the ceramic cage once implanted in human lumbar spine and the interaction between cage and bone. The iterative process allowed a cage design optimization leading to rapid manufacturing of prototypes, which was preceded by a DT of the mechanical tests prescribed by the standard ISO23089-2, to further reduce the number of tested prototypes.