COST-BENEFIT ANALYSIS OF 3D FINITE ELEMENT MODELING OF THE TIBIA USING A HEALTHY RUNNING POPULATION

Abstract

Bone stress injuries (BSIs) are a common, painful injury in runners. Computational methods, like finite element analysis (FEA), are used to further understand BSI mechanisms. FEA can be performed on three-dimensional (3D) models, however, this method has several drawbacks such as being time consuming and computationally expensive. Alternatively, FEA can be performed on two-dimensional (2D) cross sections which is a more time effective method requiring less user involvement than the 3D models. In this study, magnetic resonance (MR) images of 18 subjects (9 males and 9 females) were used to create full 3D models and 2D cross-sections. The objectives of this study were to compare FEA results from 3D models and 2D cross-sections to identify if the differences are statistically and clinically significant, perform a cost-benefit analysis of the more complex 3D FEA (with ANSYS) to the simpler 2D FEA (with VA-BATTS), and perform a multivariate regression analysis to determine which variables (demographic or geometric) are the largest indicators for increased stresses in both sexes. The 3D tibia models were produced from the segmentation of MR images and patient specific running forces were applied in a 3D computational model using ANSYS. The distal third cross-sections were segmented from the same imaging set and running forces and moments were applied using VA-BATTS. The 33% location of the 3D model illustrated that the majority of the subjects experienced maximum principal and equivalent stresses at the outer medial edge and the outer lateral edge of the cross-sections respectively. When comparing the stresses and strains at the distal third, the 3D model resulted in lower average tensile, compressive, and shear stresses than the 2D cross-section. Based on the results, statistically significant differences in tensile, compressive, and shear stresses were seen between the two approaches. The likeliest contributors to those differences was the use of moments in the 2D cross-sections and the lack of moments in the 3D models. Furthermore, these results were clinically significant because of the difference in magnitude between the results as well as the differences in stress distribution. This study also determined the benefits and negatives of each method and identified that full 3D models were better for full bone visualization and 2D cross-sections were better for specific locations of the bone. Finally, this study found that age, sex, BMI, and cortical cross-sectional area (CSA) did not significantly affect the results of 3D FEA. This study could contribute to the understanding of mechanisms behind BSIs as well as help produce risk-mitigating strategies in the future.