Subject-specific finite element (FE) modeling of the mandible bone has recently gained attention for its higher accuracy. A critical modeling factor is including personalized material properties from medical images especially when bone quality has to be respected. However, there is no consensus on the material model for the mandible that realistically estimates the Young's modulus of the bone. Therefore, the present study aims to review FE studies employing heterogeneous material modeling of the human mandible bone, synthesizing these models, investigating their origins, and assessing their risk of bias.

This computational study aimed to evaluate the bone-implant interaction of dental implants made from materials with varying Young's moduli under various conditions of bone quality. A subject-specific numerical workflow was developed by integrating boundary conditions obtained from a musculoskeletal multibody simulation (MMBS) into a finite element (FE) analysis of the mandible bone. Implants made from commercially pure titanium (cp-Ti), zirconia ceramic (ZrO₂), low-stiffness β-titanium alloy (β-Ti), and poly-ether-ether-ketone (PEEK) were evaluated during a clenching scenario. A systematic analysis was performed using statistical modeling to examine ten variations in bone quality, including cortical thickness and homogeneous bone stiffness. Additionally, two CT-based subject-specific comparisons were carried out using mandibles with distinctly different bone qualities. Implants made from materials with lower stiffness resulted in increased peri-implant strain and stress levels. In the statistical analysis, these effects were not significant when accounting for inter-individual variability of the bone qualities (p > 0.05). Cortical bone stiffness strongly correlated with peri-implant bone stress (r = 0.96 ± 0.01), while trabecular bone stiffness correlated with maximum (r = 0.71 ± 0.01) and minimum (r = -0.83 ± 0.02) principal strain in the bone. In the subject-specific analysis, stress and strain in the peri-implant bone increased for the low-quality bone and were significant for a PEEK-based implant (p < 0.001). Within the restrictions of the simplified numerical models and limited generalizability of the present findings, materials with lower stiffness may reduce peri-implant stress shielding but simultaneously increase stress at the bone-implant interface. However, their overall effect was not statistically significant when inter-individual variation in bone quality were considered.

Introduction: Finite element (FE) models of the mandible are widely used for biomechanical analysis and preoperative planning in craniomaxillofacial surgery. However, their predictive reliability depends on a realistic representation of geometry, material properties, boundary conditions, and muscular loading, for which no clear consensus currently exists. Therefore, we developed a sequentially coupled computational framework that integrates boundary conditions from a validated inverse-dynamic, scaled-generic musculoskeletal multibody simulation (MMBS) of the human masticatory system under unilateral clenching into a subject-specific, quasi-static FE model of the mandible. Methods: The MMBS model was scaled to a mandibular geometry reconstructed from computed tomography (CT) images. Time-dependent muscle forces and mandibular landmark displacements were transformed from the MMBS to the FE coordinate system and implemented as loading and boundary conditions, respectively. Heterogeneous bone material properties were assigned from CT data. Additionally, a baseline FE model, using common literature assumptions, including static generic muscle forces, fixed boundary conditions, and homogeneous bone material properties, was defined. Starting from this baseline model, sensitivity analyses of the calculated strain and stress distributions in clinically relevant regions of interest were conducted to quantify the influence of bone material properties, applied muscle forces, and the muscle attachment area, thereby isolating the effect of each modeling advancement and demonstrating the value of the proposed MMBS-FE framework. Results: The sequentially coupled MMBS-FE framework predicted realistic mandibular biomechanics with the highest von Mises stresses and logarithmic strains occurring in the balancing-side condyle and the working-side molar region. Bone material properties were the dominant source of variation, causing changes of peak stress up to 161% and peak strain up to 723%, followed by more modest effects from applied muscle forces (up to 31%) and a negligible influence of muscle attachment area definition. Discussion: These findings highlight the critical importance of realistic modeling of heterogeneous bone material properties and personalized, time-dependent coordinated muscle forces that serve as loading in subject-specific mandibular FE models. This proposed MMBS-FE model serves as a proof-of-concept framework and offers recommendations for biomechanical modeling of the human masticatory system under unilateral clenching conditions. By incorporating these recommendations, researchers can improve pre-clinical evaluation of craniomaxillofacial implants and surgical strategies.

Finite element (FE) modeling is widely used in biomechanical research, enabling computational investigations of anatomical structures. To develop accurate FE models, high-resolution medical imaging such as computed tomography (CT) is essential. However, CT exposes patients to ionizing radiation. Recently, radiation-free three-dimensional zero-echo-time magnetic resonance imaging (3D ZTE MRI) has shown promising results for bone visualization. Therefore, this feasibility study investigated whether 3D ZTE MRI is reliable for creating subject-specific FE models. Human femora from four female subjects were imaged using 3D ZTE MRI. Static FE analyses of the femora were performed in Abaqus/CAE, assuming a biphasic homogenous linear-elastic material. A sensitivity analysis was conducted to evaluate varying bone material properties. After loading with 2000 N, the mean displacement of the femoral head amounted to −1.30 ± 0.26 mm (vertical) and −9.03 ± 2.21 mm (horizontal), while the femoral neck exhibited compressive (inferior: −1366.03 ± 182.70 µm/m) and tensile (superior: 1038.69 ± 135.82 µm/m) strains. Overall, the results demonstrate displacement and strain predictions similar to previous experimental and computational studies based on CT data. Therefore, despite several simplifications, these findings confirm the feasibility of radiation-free 3D ZTE MRI for subject-specific FE modeling, enabling future simultaneous assessment of bone morphology and surrounding soft tissues.