Analyse biomécanique des contraintes et déformations exercées sur la moelle épinière et les nerfs rachidiens lors des manoeuvres de correction de la scoliose

Scoliosis is a complex spine deformation manifesting itself in the three anatomic planes (Asher et coll., 2006; Morcuende et coll., 2003) as well as on the spinal cord positioning inside the spinal canal (Smorgick et coll., 2012). When deformities reach a certain degree of severity (Cobb angle of 50°), a surgical procedure consisting on a posterior instrumentation and fusion is often required to correct and stop the progression of the deformities (Cavali, 2012). This instrumentation consists in applying complex forces and displacements on the spine through implants such as pedicle screws, hooks and rods in order to straighten the spine. The important forces generated during corrective maneuvers may cause spinal cord and spinal nerves injuries (Fujioka et coll., 2006; Miguel et coll., 2012), which occurs in 0.5% of the instrumented cases (Miguel et coll., 2012). The pathomechanisms and biomechanical impact of the stresses and strains exerted on the spinal cord and nerves during correction maneuvers is not yet well understood (Persson et coll., 2011). Few in vitro studies have addressed this type of injury, producing data on the neurological impact of tensile forces on the spinal cord (Jarzem et coll., 1992; Owen et coll., 1988) and compressive stresses on the spinal nerves (Rydevik et coll., 1991). However, biomechanical studies have not yet studied intra-operative spinal cord injuries. The objective of this research project was to develop a patient-specific biomechanical model of the spine and the spinal cord, capable of assessing the stresses and strains exerted on the spinal cord and nerves during a scoliosis correction maneuver.To achieve this objective, first, a simulation protocol based on a hybrid modeling, composed of a flexible multi-body model of the spine as well as a comprehensive finite element model (FEM) of the spine and spinal cord was developed and verified. For this first study, clinical data from a control case (case #1) as well as from a case with per-operative neurological complications (case #2) were utilized. The FEM geometry was personalized to patient-specific configuration by using 3D spine reconstructions from calibrated bi-planar radiographs and an elastic deformation method called the 3D geometric kriging. In order to have a more realistic representation of the FEM of the scoliotic spinal cord, the gray and white matter of the personalized model were displaced towards the concave side of the scoliotic curve, following the data from Smorgick et al. (2012). Mechanical properties of membranous components of the spinal cord were calibrated according to experimental data, mainly from dynamic set-ups. For other components of the model, the original mechanical properties from the FEM were utilized. Surgical instrumentation simulations were performed in a two level manner: first, a multi-body simulation was performed in order to obtain the displacement of vertebral landmarks located on the center of the top and bottom end-plates as well as at the top and bottom of the pedicles. Second, the displacement data obtained were imported to the personalized comprehensive FEM in order to reproduce the instrumentation maneuvers. For each one of the two surgical scenarios, tensile biomechanical parameters such as the internal medullar pressure (IMP) and strain were compared to the critical limits reported on the literature for the spinal cord (0.0067 MPa et 0.1 mm/mm, respectively), and compressive parameters, to the critical limits for spinal nerves (0.01 MPa). For the shear stress analysis, as there were no critical thresholds reported, the simulated results were compared between each simulation.