Segmentation and finite element analysis in orthopaedic trauma

Abstract

Background

Finite Element Analysis (FEA) has evolved into a crucial tool in orthopaedic trauma research and clinical practice. This review explores the broad applications of FEA in orthopedic surgery.

Main body

FEA involves several steps, including geometry representation, segmentation, 3D rendering, meshing, material property assignment, defining boundary conditions, and specifying contact conditions. The process utilizes patient-specific volumetric data—computed tomography (CT) scan, for example—and aims for a balance between computational efficiency and accuracy. FEA provides valuable outcome measures such as stress distribution, strain quantification, fracture gap motion, failure prediction, and implant stability. These measures aid in evaluating fracture fixation techniques, implant design, and the impact of different fixation strategies. FEA has found applications in femur and proximal humerus fracture fixation, distal femur fracture planning, tibial plateau fractures, and post-traumatic osteoarthritis. It plays a pivotal role in predicting fracture risk, assessing construct stability, and informing surgical decision-making. Additionally, FEA facilitates the development of custom surgical planning and personalized implants. To enhance accuracy, FEA is combined with cadaveric biomechanical analysis, providing a reference-standard representation of in vivo kinematics. Future research should focus on refining FEA models through increased validation using cadaveric models and clinical data.

Conclusion

FEA has revolutionized orthopaedic trauma research by offering insights into biomechanics, fracture fixation, and implant design. Integration with cadaveric biomechanical analysis enhances accuracy. Further validation efforts and integration into regular clinical practice are essential for realizing FEA’s full potential in individualized patient care. The combination of FEA and cadaveric analysis contributes to a comprehensive understanding of in vivo kinematics, ultimately improving patient outcomes.

Introduction

Cloud-based segmentation and computer aided design (CAD) generates medical surface mesh files and has the potential to provide efficient access to advanced diagnostic tools. One crucial aspect of this technology is segmentation, which involves delineating specific anatomical structures or regions of interest within medical images. By employing sophisticated algorithms and artificial intelligence, cloud-based platforms can automate and streamline the segmentation process. This may allow time savings and improvements in diagnostic accuracy when compared to traditional workflow that is performed in a Health Care Facilty (HCF) or that is provided by a company on site. Another potential added benefit of cloud-based medical segmentation is the ability to begin development of a medical surface mesh file with fewer dedicated resources (hardware, software, and training) in the HCF. The most recent developments in fracture surgery include obtaining a 3-D print that can be double bagged and taken sterilly into the operating room (Fig. 1).

Fig. 1.

Patient with a complex tibial plateau fracture. (A) Anatomic 3D model (Polyactic Acid (PLA) resin) of a patient’s complex bicondylar tibial plateau fracture, 3D printed (Formlabs 3D Printing; Somerville, MA) and utilized for surgical planning. (B) Intraoperative photograph (original material; Brigham and Women’s Hospital, Boston, MA, 02115) demonstrating the aforementioned tibial plateau fracture with the reduction held in place with surgical instruments prior to final fixation. (C) and (D) demonstrate a postoperative anterior-posterior knee radiograph and an coronal-reformatted image from a post-operative CT scan, respectively, showing medial and lateral plateau fixation with anatomic restoration of the tibial plateau articular surface, guided by the surgical plan devised with the model printed in (A)

Finite Element Analysis (FEA) is a computational simulation method that initially found a niche within orthopaedic implant development in the late 20th century [1, 2] and is now more widely utilized in orthopaedic research and clinical practice [3–7]. Applied to solid structures, FEA involves setting up and solving a series of partial differential equations to quantify stress and strain applied to a given structure or collection of elements. Described in simpler terms, FEA allows for a simulation analysis in which the user can virtually generate a bone or other anatomic structure, assign to it structure properties such as strength or elasticity varied throughout each discrete region of the structure, and toggle the forces applied to the structure to ultimately quantify the subsequent stress and strain response of the bone to manipulation. The intuitive application of this analytic method to orthopaedic trauma is determining the load needed for failure of a fixation construct based on various structural loading forces applied to the construct. As such, FEA has the potential to become an essential tool in orthopaedics and in particular orthopaedic trauma, providing valuable insights into biomechanical behavior and aiding in the development and optimization of effective surgical fixation strategies. This review aims to provide a comprehensive analysis of FEA in orthopaedic trauma, focusing on the analytic process of FEA, its applications, and future areas of research.

The analytic process of finite element analysis

FEA involves several steps for preparation and analysis. First, the geometry of the structure of interest, such as a bone, is computationally represented using elements and nodes. These elements and nodes have defined degrees of freedom, allowing for the calculation of displacements and forces at any given node. The geometry can be generated de novo using generic representations or generated from patient-specific data obtained through computed tomography (CT) scans that are imported into an FEA software.

For geometries to be generated through patient imaging such as CT scans, both segmentation and 3D rendering must be performed. Segmentation involves removing the aspects of a CT scan which are not pertinent to the geometry primarily analyzed. In other words, for a FEA focusing on mechanical stress placed on the femur, everything within the CT scan that is not within the borders of the femur bone (i.e. other bones, muscles, tendons) are “segmented” out, so that the resulting geometry retains solely the femur bone. The benefits of robust image segmentation include both user convenience for the remainder of the analysis as well as minimizing computational cost by only including the pieces of the structure which are necessary for the analysis.

Along with segmentation, image rendering is also necessary to view the medical volume on a 2D platform, such as a computer screen. One common way to volume render a medical volume is to map each voxel or sample value by opacity and color, after which a suitable file can be generated in a format that is amenable as the starting geometry for FEA.

The next step in FEA is termed “meshing,” which is the process of creating elements and nodes from a geometry. For the example, for the femur, this includes dividing the virtual bone model into small, interconnected elements. Each junction of these elements is a “node”, and this collection of elements and nodes is collectively termed a “mesh” of the original geometry. Meshing allows simulation of the stress and strain placed on discrete elements and at discrete nodes contained within the bone. The balance between solution time and representative accuracy is an important consideration when determining the number of nodes and elements. While a greater number of nodes and elements can provide better accuracy, it also increases computational cost. Element types and shapes (i.e. quadrilateral or tetrahedral) are then chosen to ensure appropriate representation of the structure. Likewise, quadratic elements (additional mid-edge node) can improve the accuracy of the simulations although they also increase the computational time.

Next, material properties including tensile strength and elasticity are assigned to the elements within the structure. For bone, it is important for these properties to account for bone heterogeneity across different populations [8]. In the earlier days of finite element modelling, different isotropic and linear elastic material properties were typically assigned uniformly to the trabecular and cortical bone structures [9]. More recently, relationships between bone mineral density and elastic modulus have been established as well as assignment criteria to the elements of FEA models [10]. CT Hounsfield Unit values can be used to estiamte the elastic modulus of each individual element in the FEA model. In FEA validation studies, these isotropic but inhomogeneous models have been shown to increase the accuracy compared with uniform material properties [11].

Boundary conditions are then defined to establish constraints for degrees of freedom at various nodes in the model. This step involves considering complex in vivo joint kinematics and the various load forces acting on the structure, including bone, muscle, tendon, and ligaments. All of these input factors—element types, material properties, boundary conditions—contribute to the ultimately quantification of strain within the FEA model.

Finally, contact conditions need to be defined between two interacting surfaces. A tie constraint assumes a fully bonded interface, which is typically used in the locking screw interactions between the screw and the bone and the screw and the plate [12]. Furthermore, surface-to-surface interactions are defined where two surfaces can move with a frictional behavior relative to each other. Thereby, the determination of a realistic friction coefficient is very important. In orthopaedic trauma, a friction coefficient of 0.3 or 0.4 is usually applied to the interface between different bone fragments, whereas 0.1 is used for the bone-plate interface [13].

Outcome measures in finite element analysis

FEA provides valuable outcome measures for assessing the biomechanical behavior of orthopaedic constructs, including stress distribution, quantification of strain, fracture gap motion, prediction of failure location following fixation, and implant stability [14–22]. By analyzing these parameters, surgeons can gain insights into the performance of fracture fixation techniques, implant design, and the impact of different fixation strategies.

Applications of finite element analysis in orthopaedic trauma

FEA has found numerous potential applications in orthopaedic trauma, allowing for a deeper understanding of fracture fixation, implant fatigue, and joint mechanics.

In femur [3, 23] and proximal humerus fracture fixation [3, 24–28]FEA has been instrumental in predicting fracture risk, assessing construct stability, evaluating the influence of different fixation strategies, and informing surgical decision-making. In studies utilizing magnetic resonance imaging (MRI) scans, FEA was found to accurately model femur strength and predict fracture based on comparison of results from direct mechanical testing of cadaveric femurs, suggesting that FEA can be utilized to evaluate femur strength and allow fracture risk stratification in a clinical setting [6]. Further, the use of FEA in assessing fracture fixation has involved determining the most effective use of plates, nails, screws; potential methods by which these implants can be improved; and comparison of different implants [29–31].

For proximal humerus fracture fixation, recent investigations have utilized FEA to improve designs of existing plates as well as to analyze the optimal use of locking plates, intramedullary nails, and screws fixation [32–34]. For distal femur fracture fixation, a recent study by He et al. showed that patients with distal femur fractures whose cases were preoperatively planned using FEA had less blood loss, fewer fluoroscopic images, and shorter operation time relative to the non-FEA group [35]. A 2023 study by Jitprapaikulsarn et al. investigated revision fixation strategies after failure of primary distal femur plating, providing FEA data supporting the stability of dual plating in the setting of revision distal femur fixation [36]. In a validated computational framework with 19 proximal humerus fractures, Schader et al. tested the orientation of six proximal locking screws with a parametric approach [24]. The evaluation of their large number of FEA simulations showed that subject-specific screw orientations significantly reduced the peri-screw bone strain, indicating a lower cut-out risk for patient specific configurations compared with the standard fixation.

Many studies have utilized FEA to evaluate optimal hardware designs and fixation strategies for tibial plateau fractures [37]. Samsami et al. utilized FEA to demonstrate that fracture patterns with a coronal split component often entail a destabilized medial tibia, altering the stress distributions placed on the hardware [38]. Forna et al. compared three methods of reduction, single medial, single lateral, and dual plating, for the treatment of a bicondylar tibial plateau fracture [39]. The authors found small differences in stress and strain around the fracture area between the dual plate and single lateral plating, suggesting that single lateral plating may be the optimal strategy for bicondylar tibial plateau fractures to prevent associated complications of dual plating while ensuring fracture stability [39]. Several other investigations have evaluated new types of plates and screws for tibial plateau fixation [37, 40–42]. With a validated FEA, Synek et al. investigated the axial stiffness and peri-implant strain of volar plate fixation in distal radius fractures [43]. Specifically, they simulated all possible configurations by using three to six distal screws. The deterioration of the biomechanical properties largely depended on the specific fracture case and the screw that was removed. They conclude that the optimal fracture fixation is patient-specific and that it is not only critical how many screws are used but also which ones.

In addition to optimizing acute fracture fixation, FEA is also valuable in studying contact stress in post-traumatic osteoarthritis, particularly in cases of intraarticular fractures in the ankle [44]. Historically, much of the research done with contact stresses in post-traumatic arthritis has been via cadaveric models, but FEA provides a low-cost option for simulation analysis which may further expand this area of research. Based on biomechanical optimization, considering the patient’s anatomy, weight, bone density, and case-specific fracture patterns, this process is unique in its consideration of the key role of plate positioning and specifically the number and spatial paths of the fixation screws, with respect to received real-life loading (such as walking, going up and downstairs) to a given fracture pattern in the knee joint. This solution may overcome many of the limitations of traditional subjective planning done by the surgeon by taking the guesswork out of the equation, providing him specific instructions of the most stable construct, minimizing interfragmentary movement, and allowing for surgeon’s higher confidence that the fracture fixation construct has the best possible stability. These enhancements can prevent fixation construct failure, especially in cases with inadequate bone quality.

Figure 2 illustrates the concept of optimized screw spatial paths using the Finite Element model. This example utilizes a parametric FEA subjected to different loading conditions to optimize applications in complex tibial plateau fractures considering cortical and trabecular bones based on CT data of a variety of patient demographics.

Fig. 2.

Retrospective FEA of hardware failure after three-plate fixation of a bicondylar tibial plateau fixation. In this case, the screws highlighted with an orange circle broke, with FEA above used to determine what caused this result. (A) demonstrates a digital 3D model of the original fracture, with each fragment segmented and higlighted in a distinct color. (B) demonstrates surgical fixation used, featuring medial, lateral, and posterior plating. The orange circles in this figure highlight the site at which the hardware failed. The highlighted fragments indicate continued nonunion of the original fracture. (C) demonstrates the post-fixation nonunion. The scale included in this figure demonstrate level of stress on various regions of bone, with red indicating the highest level of stress. As seen in the figure, there is considerable stress remaining at the proximal component of the fracture, contributing to increased likelihood of poor fracture healing and malunion or nonunion. (D) demonstrates level of stress on the hardware (screws) of fixation utilized in initial fixation. The regions highlighted with orange circles demonstrate areas of maximal strain (colored in red), indicating high likelihood of hardware failure, correlating with the findings shown in (B). These findings via FEA, if determined preoperatively, would inform surgeons to adjust their planned fixation construct to better distribute the level of stress and strain across the construct, minimizing likelihood of hardware failure (Reproduced, with permission, from: CustomSurg AG, FractureSolver)

Lastly, FEA has the potential to facilitate the development of custom surgical planning and personalized implants, improving patient-specific outcomes. For surgeons, the ability to not only preoperatively plan different fixation strategies but also to evaluate the stability of these fixation methods using a simulation approach may prove to be the next major advancement in orthopaedic trauma surgery. As the technology progress, this application of FEA is likely to have significant implications for patient care.

Interpreting FEA with cadaveric biomechanical analysis

FEA, when combined with cadaveric biomechanical analysis, accurately represents in vivo kinematics. Biomechanical outcomes obtained from applying force to cadaveric bone provide the reference-standard representation of how bone acts under stress in the human body. This allows accurate capture of the properties of the bone, soft tissue, and other structures, thus improving reliability of the model using that data. This allows validation of FEA models; by comparing data obtained from stress applied to a virtual model of bone to a cadaveric specimen, the accuracy of the former can be assessed appropriately.

Future research in FEA should focus on advancing validation techniques by incorporating more cadaveric models and clinical data. This will help refine and improve the accuracy of FEA models, ensuring their applicability in real-world scenarios. Efforts are currently underway for standardizing the validation and verification process regarding application of FEA into orthopaedic research and clinical practice [45]. Furthermore, the integration of FEA with other computational methods and emerging technologies, such as machine learning and artificial intelligence, holds promise for further enhancing the capabilities of FEA in orthopaedic trauma.

An important limitation to cadaveric studies is difficulty in truly mimicking an in vivo biomechanical environment. It is challenging to account for the complex, dynamic interaction between bone, soft tissue, ligaments, and tendons through a cadaveric sample; this, limits the reliability of cadaveric biomechanical analysis. FEA models have the advantage that it is relatively simple to apply actual patient-specific loads from musculoskeletal models, not only including joint reaction, but also ligament and muscle forces. Nonetheless, use of cadaveric specimens remains the most accurate option to date in validation of FEA in orthopaedic research.

Conclusion

FEA has already begun to revolutionize orthopaedic trauma research by providing valuable insights into biomechanical behavior, fracture fixation techniques, and implant design. The process of FEA, including geometry creation, meshing, material properties assignment, boundary conditions, and quantification of stress, strain, and displacement, allows for accurate analysis and optimization of orthopaedic constructs. While FEA has already found various applications in orthopaedic trauma research, further work is needed to validate models and integrate FEA into regular clinical practice–a worthwhile endeavor given its potential to provide individual-level patient care. By leveraging the strengths of FEA and cadaveric biomechanical analysis together, orthopaedic surgeons and research can achieve a comprehensive understanding of in vivo kinematics and, ultimately, improve patient outcomes.

Abbreviations

FEA

Finite element analysis

CT

Computed tomography

MRI

Magnetic resonance imaging

Author contributions

KY prepared the manuscript draft and organized the study outline. AF wrote the main manuscript text, incorporated revisions, and prepared figures. SC provided significant edits to the main manuscript draft and is an expert in the relevant field. AVK oversaw the study and offered edits to the manuscript.

Funding

Open access funding provided by Copenhagen University

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval

This study does not involve human participants.

Consent for publication

Consent has been provided appropriately. All patients sign a standardized form prior to surgery including permission for publication of deidentified material.

Competing interests

Dr. von Keudell is a Consultant for Stryker, Ossiform, Medical Insight and OrthoXel. He receives royalties from SpringerNature and is a cofounder of CustomSurg AG. Simon Comtesse has stocks in CustomSurg AG.