| The main objective of this thesis is to develop tools to estimate the flexibility of the scoliotic spine for the planning of surgical instrumentation. The hypotheses tested in this thesis are: that the segmental flexibility of the scoliotic spine affects the correction following the surgical instrumentation and that a biomechanical model of the surgical instrumentation incorporating patient-specific geometric and mechanical properties could estimate adequately the correction for the planning of surgical instrumentation.;A novel method was developed for the identification of patient-specific mechanical properties of the scoliotic spine using a multi-body model. Vertebrae were represented as rigid bodies and intervertebral elements were defined using a spherical joint and three torsion springs. The initial mechanical properties of motion segments were defined from in-vitro experimental data reported in the literature. They were adjusted using an optimization algorithm to reproduce the reducibility of scoliotic deformities measured on the lateral bending radiographs. The personalized model was then used to simulate the surgical instrumentation and to investigate on the relationships between the segmental flexibility of the scoliotic spine and the correction following the surgical instrumentation.;The adjustment of the mechanical parameters of 10 scoliotic patients allowed reducing of 50% the sum of the squared differences between simulated and experimentally measured Ferguson angles in lateral bending. The classification of the flexibility of spine segments based on the computed mechanical modulation parameters (alphai) allowed to discriminate flexible (alpha ≤ 0,8) and rigid (alpha ≥ 1,2) scoliotic curves. This study shows that the inter-individual variability of the scoliotic spine flexural rigidity is important (alphai = 2,5 +/- 2,0) and should be considered into biomechanical models.;The simulation of the surgical instrumentation maneuvers of 7 scoliotic patients adequately predicted the surgical correction. Differences between the simulated and measured Ferguson angles in the frontal and the sagittal planes respectively were 2,3° +/- 2,0° and 2,2° +/- 4,1° before the adjustment of mechanical properties. The personalization slightly improved the Ferguson angles predicted in the frontal plane (1° +/- 4°) and the sagittal plane (2,0° +/- 3,9°) but no significant change was observed for the plane of maximum curvature. The model also predicts plausible torque reactions (from 0,2 Nm to 28 Nm) and lateral forces (≤611 N except for one patient) between the rod and the implant during the rod rotation and translation maneuvers. (Abstract shortened by UMI.). |
