FROM SMALL ERRORS TO BIG IMPACTS: HOW SEGMENTATION INFLUENCES BIOMECHANICAL SIMULATION PRECISION

Accurate segmentation of medical imaging data is fundamental to the development of reliable biomechanical models, particularly those employing finite element analysis (FEA) to study articular cartilage mechanics. Segmentation errors, whether arising from manual processes or automated algorithms, can significantly influence model predictions, affecting clinical and research outcomes. The study explores the impact of segmentation accuracies on FEA models of articular cartilage and proposes a novel congruence method to moderate their effects. Using a simplified axisymmetric hip joint model and real anatomical geometries derived from MRI data, segmentation errors were simulated and their effects on biomechanical outputs, including contact pressure distribution and joint mechanics, were systematically analyzed. The simplified model was validated against analytical solutions, ensuring computational accuracy. Additionally, a congruence method was proposed to align cartilage surfaces and reduce the impact of segmentation errors. Simulations for real anatomical geometry as well as the congruence method were conducted and the results of each were reported. Preliminary results highlight the substantial influence of segmentation accuracy on model reliability and demonstrate the efficacy of the congruence method in improving biomechanical predictions. This work emphasizes the need for precise segmentation in advancing computational biomechanics and provides a foundation for refining modeling techniques to better inform clinical decision-making and orthopedic interventions.