Figure 5

Overview of the methodological workflow. (a) Healthy volunteers were MRI scanned in lateral decubitus with the knee positioned in different degrees of knee flexion, approximated to be 0°, 30°, 60° and 90° of flexion using a goniometer. (b) Semi-automated segmentation was used to derive the osseous meshes in every bony configuration. The cartilage layer was predicted based on the methodology as described by Van Oevelen et al.¹⁶. (c1) The medial and lateral menisci were modeled as elastic, volume preserving meshes that delineate the femoral condyles for the 4 bony configurations. (c2) For each case and in every degree of flexion, the medial and lateral meniscus was additionally manually segmented. A comparison between the segmented and predicted meniscus was performed for validation purposes. (d) To continuously describe meniscal dynamics, the 4 bony configurations and the according meniscal positions served as input for subject-specific PPSA modeling. The bony and meniscal kinematics were extracted for the 4 goniometer-based flexion angles at the time of scanning to compare segmented and PPSA-predicted meniscal anatomy for validation purposes.