Enhancing the Efficiency of Subject-Specific Knee Joint Biomechanical Simulations With Applications to Osteoarthritis
Date
2024-08-14
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
There are three challenges in conventional subject-specific modelling techniques. First, having an accurate material model is essential for studying biomechanical response of musculoskeletal systems. For instance, the stresses and strains in knee joint articular cartilage are influenced by site-specific variations in collagen fibril orientations that vary with aging, which is ignored in finite element (FE) analyses using a generic knee geometry. The other challenge is related to the manual geometry reconstruction from biomedical images, which is a time-consuming process and not practical for clinical applications. Finally, estimating joint forces and moments with conventional methods requires marker-based motion capture facilities to study the kinematics of human body motion and convoluted human body modeling to determine the kinetics of motions. Although marker-based motion capture offers the precise measurement of marker positions on the body and enables the calculation of kinematics and kinetics in a controlled setting, its data collection is labor-intensive and requires staff with expensive equipment and specialized technical experience. Subject-specific FE modeling provides a viable approach for the study of cartilage mechanics in normal and pathomorphological knee, thus providing insight into the mechanics of knee articular cartilage. Therefore, in this project we aimed to facilitate the development of subject-specific FE models of the knee. We developed and validated: 1) a 3D remodeling algorithm of collagen fibrils within knee joint cartilage under simulated gait, 2) a semi-automatic segmentation routine of knee joint geometry from magnetic resonance images (MRIs), and 3) a markerless motion capture to perform kinematics and kinetics analyses. To develop the cartilage fibril remodeling algorithm, a fibril-reinforced, biphasic cartilage model was integrated with 3D human knee joint geometry. For the MRI segmentation, we used 3D Swin UNETR, a statistical shape model (SSM) and automated filtering techniques to extract the distal femur, proximal tibia, femoral and tibial cartilages. To facilitate the estimation of joint forces and moments, OpenCap, a markerless motion capture software, was used during a cycling task. This technique is intended to expedite kinematics and kinetics analysis. The ultimate goal of this project is to develop an efficient pipeline for subject-specific FE modeling.
Description
Keywords
Citation
Kakavand, R. (2024). Enhancing the efficiency of subject-specific knee joint biomechanical simulations with applications to osteoarthritis (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.