COMBINING MUSCULOSKELETAL MODELING AND FEM IN DIABETIC FOOT PREVENTION

Recently the development of Patient-specific models (PSMs) tailored to patient-specific data, has gained more and more attention in clinical applications. PSMs could represent a solution to the growing awareness of personalized medicine which allow the realization of more effective rehabilitation treatments designed on the subject capabilities. PSMs have the potential of improving diagnosis and optimizing clinical treatments by predicting and comparing the outcomes of different approaches of intervention. Furthermore they can provide information that cannot be directly measured, such as muscle forces or internal stresses and strains of the bones. Given the considerable amount of diseases affecting motor ability, PSMs of the lower limbs have been broadly addressed in literature. Two techniques are mostly used in this area: musculoskeletal (MS) modeling and finite element (FE) analysis. (MS) models represent a valuable tool, as they can provide important information about the unique anatomical and functional characteristics of different subjects, through the computation of human internal variables, such as muscle activations and forces and joint contact forces. The flexibility and adaptability of FE analysis makes it a perfect solution to model biological geometries and materials and to simulate complicated boundary and loading conditions. Accurate and descriptive FE models would serve as an excellent tool for scientific and medical research. Furthermore they could be used in clinical settings if combined with medical imaging, in order to improve patient care. Several 3-dimensional (3D) foot FE models were recently developed to analyze the biomechanical behavior of the human foot and ankle complex that is commonly studied with experimental techniques like stereophotogrammetry, force and plantar pressure plates. In this context, many gait analysis protocols have been proposed to assess the 3D kinetics, kinematics and plantar pressure distribution. This evaluation has shown to be useful in characterizing the foot biomechanics in different pathologies like the diabetic foot. Diabetic foot is an invalidating complication of diabetes mellitus, a chronic disease frequently encountered in the aging population. It is characterize by the development of ulcers which can lead to amputation. Models for simulations of deformations and stresses in the diabetic plantar pad are required to predict high risk areas on the plantar surface and can be used to investigate the performance of different insoles design for optimal pressure relief. This work represents a first effort towards the definition of a more complete PSM which combining both a MS model and a FE model, can increase the understanding of the diabetic foot pathology. To achieve this objective, several limitations and issues have been addressed. As first, MS models of diabetic and control subjects were developed using OpenSim, to estimate muscle forces. The objective was to evaluate whether the diabetic population exhibit lower limb muscle strength deficits compared to the healthy one. Subjects routine gait analysis was performed and lower limb joints kinematics, kinetics, time and space parameters estimated by means of a modified version of the IORgait protocol. 3D lower limb joints kinematics and kinetics was also calculated with OpenSim. Both methodologies were able to highlight differences in joint kinematics and kinetics between the two populations. Furthermore MS models showed significant differences in healthy muscle forces with respect to the diabetic ones, in some of the muscles. This knowledge can help the planning of specific training in order to improve gait speed, balance, muscle strength and joint mobility. After the use of MS models proved to be applicable in the diabetic population, the next step was to combine them with foot FE models. This was done in two phases. At first the impact of applying the foot joints contact forces (JCFs) obtained from MS models as boundary condition on the foot FE models was verified. Subject specific geometries from MRI were used for the development of the foot FE models while the experimental plantar pressures acquired during gait were used in the validation process. A better agreement was found between experimentally measured and simulated plantar pressure obtained with JCFs than with the experimentally measured ground reaction forces as boundary conditions. Afterwards the use of muscles forces as boundary condition in the FE simulations was evaluated. Subject-specific integrated and synchronized kinematic-kinetic data acquired during gait analysis were used for the development of the MS models and for the computation of the muscle forces. Muscle insertions were then located in the MRI and correspondent connectors were created in the FE model. FE subject-specific simulations were subsequently run with Abaqus by conducting a quasi-static analysis on 4 gait cycle phases and adopting 2 conditions: one including the muscle forces and one without. Once again the validation of the FE simulations was done by means of a comparison between simulated and experimentally measured plantar pressures. Results showed a marked improvement in the estimation of the peak pressure for the model that included the muscles. Finally, an attempt towards the definition of a parametric foot finite element model was done. In fact, despite the recent developments, patient-specific models are not yet successfully applied in a clinical setting. One of the challenges is the time required for mesh creation, which is difficult to automate. The development of parametric models by means of the Principle Component Analysis (PCA) can represent an appealing solution. In this study PCA was applied to the feet of a small cohort of diabetic and healthy subjects in order to evaluate the possibility of developing parametric foot models and to use them to identify variations and similarities between the two populations. The limitations of the use of models have also been analyzed. Their adoption is indeed limited by the lack of verification and validation standards. Even using subjects’ MRI or CT data for the development of FEM together with experimentally acquired motion analysis data for the boundary and loading conditions, the subject specifity is still not reached for what regards all the material properties. Furthermore it should be considered that everything relies on algorithm and models that would never be perfectly representing the reality. Overall, the work presented in this thesis represents an extended evaluation of the possible uses of modeling techniques in the diabetic foot prevention, by considering all the limitations introduced as well as the potential benefits of their use in a clinical context. The research is organized in six chapters: Chapter 1 - provides a background on the modeling techniques, both FE modeling and MS modeling. Furthermore it also describes the gait analysis, its instrumentation and some of the protocols used in the evaluation of the biomechanics of the lower limbs; Chapter 2 - gives a detailed overview of the biomechanics of the foot. It particularly focuses on the diabetes and the diabetic foot; Chapter 3 - introduces the application of MSs for the diabetic foot prevention after a brief background on the techniques usually chosen for the evaluation of the motor impairments caused by the disease. Aim, material and methods, results and discussion are presented. The complete work flow is described, and the chapter ends with a discussion on new key findings and limitations. Chapter 4 – reports the work done to combine the use of musculoskeletal models with foot FEMs. At first the impact of applying the foot joints contact forces obtained from MS models as boundary condition on the foot FEMs is verified. Then the use of muscles forces (again obtained from MS models) as boundary condition in the FE simulations is evaluated. For both studies a brief background is presented together with the methods applied, the results obtained and a discussion of novelties and drawbacks. Chapter 5 – explores the possibility of defining a parametric foot FEM applying the Principle Component Analysis (PCA) on the feet of a small cohort of diabetic and healthy subjects. A background on the importance of patient specific models is presented followed by material and methods, results and discussion of what obtained with this study. Chapter 6 - summarizes the results and the novelty of the thesis, delineating the conclusions and the future research paths