TL;DR
This paper investigates how various parameters in a computational model affect bone remodeling in the human mandible under chewing and biting forces. They found that boundary conditions significantly influence bone density distribution, while initial density has a lesser impact. The study highlights the importance of these parameters for accurately simulating bone adaptation.
What Problem Does This Paper Address?
Bone is a dynamic tissue that constantly remodels itself in response to mechanical stimuli. This adaptive process is crucial for maintaining bone strength and integrity. While numerous computational models exist to simulate this bone remodeling, they often involve complex algorithms with multiple parameters. Accurately predicting how bone density changes, especially in critical areas like the mandible under normal physiological loads (chewing and biting), requires a thorough understanding of how these parameters influence the model's output. Discrepancies between simulation results and real-world observations (in vivo data) suggest a need to systematically explore the sensitivity of these models to various inputs. This paper aims to address this by conducting a parametric study on a specific bone remodeling algorithm applied to the mandible.
Methods
The researchers employed a finite element analysis (FEA) approach to simulate bone remodeling. The core of their method is a bone remodeling algorithm based on strain energy density per unit bone mass, which serves as the mechanical stimulus. This algorithm incorporates an "equilibrium zone" (or "lazy zone") where bone density doesn't change, and a "saturated density change rate" to ensure stability and prevent unrealistic "checkerboard" patterns in bone density distribution.
They developed 2D FE models of a section of the mandible, including teeth, surrounding tissues, cortical bone, and trabecular bone. The models were subjected to simulated chewing and biting forces. The study systematically varied several key parameters:
- Boundary Conditions: Two types were tested: constrained (vertical movement only) and free (allowing movement in all directions) at the model's edges.
- Initial Bone Density: Uniform initial densities were tested across a range, and also with small random perturbations.
- Reference Value of Mechanical Stimulus: This parameter, representing the target stimulus for bone adaptation, was varied.
- Width of the Equilibrium Zone: The size of the "lazy zone" was adjusted.
The simulations were run for 80 iterations, and convergence was defined by a minimal change in overall bone density.
Datasets
Not discussed in the paper. The study relies on a simulated biomechanical model and material properties derived from literature.
Key Insights and Ideas
- Boundary Conditions are King (for Distribution): The choice of boundary conditions at the model's edges dramatically alters the final bone density distribution, especially in areas away from the direct application of loads. This suggests that accurately defining how the mandible is supported and constrained is critical for realistic simulations.
- Initial Density is Less Influential (for Overall Density): While starting with very different initial bone densities can lead to different adaptation paths, the final average bone density at convergence tends to be similar across a wide range of initial values. This implies the remodeling process self-corrects to a degree.
- Higher Stimulus Threshold = Lower Density: A higher "reference value" for the mechanical stimulus (the target strain energy density) leads to a lower overall average bone density at convergence. This makes intuitive sense: if the bone "aims" for a higher stimulus, it needs to be less dense to achieve it.
- Equilibrium Zone Width: A Trade-off: The width of the equilibrium zone doesn't significantly impact the average bone density at convergence. However, increasing its width reduces the extent of both the highest and lowest bone density areas, leading to a more uniform distribution within those reduced extremes.
- Stability Mechanisms Work: The inclusion of an equilibrium zone and saturated density change rate successfully prevented the emergence of checkerboard patterns, a common artifact in simpler bone remodeling models.
Key Results
- Boundary Conditions: Constrained boundary conditions resulted in a more uniform bone density distribution in the lower center of the mandible compared to free boundary conditions, which showed more drastic changes.
- Initial Bone Density: For initial densities between 0.2 and 1.6 g/cm³, average bone density increased to converge around 1.6-1.7 g/cm³. For densities above 1.8 g/cm³, it decreased. At the 80th iteration, the differences in average bone density due to initial density were negligible.
- Reference Value of Mechanical Stimulus: Higher reference values (e.g., 0.014 J/g) resulted in lower average bone densities (around 1.32 g/cm³) compared to lower reference values (e.g., 0.002 J/g) which yielded higher densities (around 2.05 g/cm³).
- Equilibrium Zone Width: Average bone density at convergence was similar across tested widths (5% to 45%). However, increasing width reduced the spatial extent of extreme bone densities.
Limitations and Future Work
- Non-realistic Initial Conditions: The study acknowledges that assuming a uniform initial bone density doesn't reflect reality. For simulating specific scenarios like post-implant adaptation, more realistic initial conditions are needed.
- Simplified Boundary Conditions: While the study explored different boundary conditions, they were simplified. Future work could incorporate more complex boundary conditions, potentially including masticatory muscle forces, to better represent in vivo scenarios.
- Absence of Biological Signals: The current models focus solely on mechanical stimuli. Future research should aim to incorporate biological, chemical, and other factors that influence bone remodeling (e.g., aging, disease, hormonal influences).
- Quantitative Comparison: The authors highlight the need for more rigorous quantitative comparisons between simulation results and experimental (e.g., CBCT) data.
- Cortical Bone Remodeling: The current model primarily focuses on trabecular bone. Extending the algorithms to simulate cortical bone remodeling would be valuable.
Summary updated