In vivo distribution of spinal intervertebral stiffness based on clinical flexibility tests.

Abstract

A numerical study was conducted to identify the intervertebral stiffness of scoliotic spines from spinal flexibility tests. To study the intervertebral 3-dimensional (3D) stiffness distribution along scoliotic spine. Few methods have been reported in literature to quantify the in vivo 3D intervertebral stiffness of the scoliotic spine. Based on the simulation of flexibility tests, these methods were operator-dependent and could yield to clinically irrelevant stiffnesses. This study included 30 patients surgically treated for severe idiopathic scoliosis. A previously validated trunk model, with patient-specific geometry, was used to simulate bending tests according to the in vivo displacements of T1 and L5 measured from bending test radiographs. Differences between in vivo and virtual spinal behaviors during bending tests (left and right) were computed in terms of vertebral rotations and translation. An automated method, driven by a priori knowledge, identified intervertebral stiffnesses in order to reproduce the in vivo spinal behavior. Because of the identification of intervertebral stiffnesses, differences between in vivo and virtual spinal displacements were drastically reduced (95% of the differences less than +/-3 mm for vertebral translation). Intervertebral stiffness distribution after identification was analyzed. On convex side test, the intervertebral stiffness of the compensatory curves increased in most cases, whereas the major curve became more flexible. Stiffness singularities were found in junctional zones: these specific levels were predominantly flexible, both in torsion and in lateral bending. The identification of in vivo intervertebral stiffness may improve our understanding of scoliotic spine and the relevance of patient-specific methods for surgical planning.