Detaillierte Beitrags-Information
| Beitragstitel | Preoperative Planning of Spinal Alignment by Statistical and Musculoskeletal Modeling |
|---|---|
| Beitragscode | P027 |
| Autoren | |
| Präsentationsform | Poster |
| Themengebiete |
|
| Abstract |
INTRODUCTION Statistical shape models (SSMs) have been widely used to create 3D model of anatomical structures such as bones and organs, yet their potential to predict abstract global musculoskeletal properties has not been investigated. As spinal sagittal alignment plays a crucial role in spine related pathologies and often the optimal or healthy alignment remains unknown, we aimed at adopting the SSM method to predict sagittal spinal alignment based on the configuration of sacrum and pelvis. The predicted alignment was then evaluated using computational models in order to assess biomechanical quality metrics. METHODS The 2D SSM to predict sagittal alignment of the spine was trained with 50 annotated lateral EOS images. The primary annotations localized the femoral heads (FH) (4 points on the right and left acetabula) and the sacrum (boundaries of the endplate). Additionally, for training of the SSM, each vertebra’s upper and lower endplates from C1 to L5 were annotated. From these the midpoints of each vertebra were derived, defining the spinal curvature. The trained SSM was then employed to predict a healthy sagittal alignment from positions of FH and sacrum. To compare the loads of the altered (predicted) alignment with original loading, two personalized musculoskeletal models were created with OpenSim (simtk.org) based on a validated template model, and a linear forward bending motion from upright standing to 30° lumbar flexion was simulated. RESULTS The alignment predicted by the SSM presents more pronounced lumbar lordosis and thoracic kyphosis. Sagittal balance determined by the sagittal vertical axis (SVA) was reestablished. The comparison of the joints loads between the different alignment models showed lower compression forces for the SSM predicted alignment at all levels throughout the whole motion. Shear forces at L3/4 and L4/5 were comparable. In flexed postures shear was reduced in upper levels, whereas L5/S1 shear forces were slightly higher in predicted alignment. CONCLUSION We showed the first application of a 2D SSM to predict sagittal spinal alignment. The ability of an SSM predicted alignment to reduce joints loads at lumbar segments was further demonstrated. The prediction of an optimal spinal alignment and its analysis using biomechanical models has the potential to change preoperative planning in spinal surgery. |