Effect of pedicle screw misplacement on the pull-out strength using personalized finite element modeling

Intraoperative misplacement of vertebral pedicle screws is prevalent. While a pedicle screw misplacement of up to 4 mm is often considered safe by clinical standards, this may reduce the pull-out strength thereby increasing the risk of postoperative screw loosening. This study aimed to compare the pull-out strength of ideally-placed and misplaced pedicle screws via personalized finite element (FE) modeling simulations. Three-dimensional FE models to simulate pull-out tests of pedicle screws were constructed. Vertebral geometries and material properties were both determined based on computed tomography images from lumbar spine (L1-L5) of a 29-year-old healthy male individual. Pedicle screws were bilaterally placed under four conditions: ideal placement (no cortex perforation) as well as minor medial, minor lateral, and severe lateral misplacements causing cortex perforations. Minor and severe misplacements corresponded to, respectively, grades C and D of the Gertzbein classification. After meshing and voxel-based vertebral material assignments based on two distinct mappings from literature, pull-out strengths were estimated by considering a strain-based damage model to accurately simulate bone fractures. Results indicated that the mean of FE-predicted pull-out forces for ideally-placed screws (1583 and 2412 N for material mappings 1 and 2, respectively) was within the range of previous experimental data (ranging from 1016 to 2443 N) thus confirming the model validation. The mean of all pull-out forces for each misplaced screw group was significantly smaller (p