Optimum parameters for each subject in bone remodeling models: A new methodology using surrogate and clinical data

• Published in: European journal of mechanics, A, Solids Vol. 91; p. 104409
• Main Authors: Dicati, Gabriela Wessling Oening, Gubaua, José Eduardo, Pereira, Jucélio Tomás
• Format: Journal Article
• Language: English
• Published: Elsevier Masson SAS 01.01.2022
• Subjects: Computerized tomography; Finite element method; Individualization of bone remodeling model; Kriging; Numerical uniqueness of solution; Finite element method; Numerical uniqueness of solution; Kriging; Computerized tomography; Individualization of bone remodeling model
• ISSN: 0997-7538; 1873-7285
• DOI: 10.1016/j.euromechsol.2021.104409

Bone remodeling models use experimental and theoretical parameters to simulate bone tissue behavior. The physical parameters are computed satisfactorily using numerical methods, e.g., the density distribution. Different subjects require specific parameters. However, traditional bone remodeling models do not consider distinct parameters for each subject. We aim to present a new methodology that accounts for specific parameters for each subject while reviewing bone remodeling, biological aspects, and coupling with the finite element method. We divide the new methodology into three steps: (a) obtaining the density distribution from the femur tomography and a finite element model; (b) implementing an algorithm for bone remodeling via Abaqus; and (c) implementing a Matlab code that combines the design of experiments, surrogate, bone remodeling model, and the finite element method to minimize the difference between the clinical and numerical density distributions. Furthermore, we considered subjects’ characteristics as the physical activity amount and body mass in the numerical simulations. The new methodology is valid whenever the remodeling model presents a tendency towards solution uniqueness. We applied the new methodology in 18 subjects and obtained a different set of parameters for each one. These parameters allowed characterizing a more accurate and realistic femoral density distribution for each subject. Furthermore, the new methodology decreased the density relative difference by 50%, compared with the traditional ones. •Optimum parameters for simulating bone remodeling of a specific subject.•Optimization process accounts for surrogate and clinical data.•Numerical and clinical bone density distributions became closer.•Detailed analysis of the bone remodeling needs specific parameters for each subject.