Abstract
dc:description.abstractComputational modeling of knee joint biomechanics can provide valuable insights into joint function, injury mechanisms, and patient outcomes, and can support surgical planning and informed clinical decision-making. However, achieving consistent and repeatable results, or reproducibility, is a significant challenge that affects the credibility of these models. This thesis addresses reproducibility through two studies related to the KneeHub Project, which is a multi-institutional collaboration designed to investigate reproducibility in finite-element modeling of the knee. The first study involved model benchmarking, where predicted results of the calibrated models were compared with the experimental data to evaluate model reproducibility. While all models replicated key experimental trends in joint kinematics and kinetics, discrepancies arose from differences in ligament representation, coordinate system definitions, and calibration strategies. Errors of up to 6.6 ± 2.4 mm in anteriorposterior (AP) translation to 13.5 ± 12.9˚ in internal-external (IE) rotation were observed. Variability in the approaches used highlights the subjectivity inherent in modeling, even when using the same data and goals. The second study implemented a consensus workflow that was developed by the KneeHub community in an effort to improve reproducibility and credibility in the modeling process. This workflow was developed to study ACL injury, repair, and potential post-injury outcomes, highlighting the importance of having a clear context of use. The model developed by following the steps in the consensus workflow successfully captured physiological behaviors, like posterior femoral rollback during deep flexion. Simulations of anterior-posterior laxity and passive flexion showed numerical stability, confirming model robustness. A sensitivity analysis indicated that precise calibration of ligament slack-length and a high-quality geometric representation of the cartilage anatomy are more critical for accurate predictions on knee kinematics than variations in stiffness or mesh density. Overall, this work supports the notion that a community-based consensus workflow can enhance reproducibility and predictive capability in knee joint modeling, contributing to credible computational biomechanics and supporting future studies, model calibration, and clinical applications.
Degree
thesis:*- Name thesis:degree_name
- M.S. in Mechanical Engineering
- Level thesis:degree_level
- Masters Thesis
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Nazem, Maryam
- Contributors dc:contributor
-
- Peter J. Laz
- Casey Myers
- Martin Rhodes
Subjects
dc:subject × 9Rights
dc:rights- Statement dc:rights
-
- <p>Copyright is held by the author. User is responsible for all copyright compliance.</p>
- Maryam Nazem
- Language dc:language
- English (eng)