Étude des mécanismes de compensation des membres inférieurs qui peuvent influencer la balance sagittale après fusion lombaire grâce au système d’imagerie EOS

Abstract

Sagittal balance compensation mechanisms such as pelvic retroversion for an unbalanced spine have been well described in the literature. However, those implemented in the lower limbs such as knee flexion and their evolution after lumbar spine fusion has not been quantified. The adoption of whole body EOS imaging in routine practice has allowed radiographic analysis of the lower limbs, even in degenerative pathologies. This work aims to characterize and predict using machine learning the occurrence of knee flexion in relation to global and regional sagittal parameters of the spine in lumbar degenerative pathologies before and after lumbar fusion. This work is a single-center retrospective analysis including 108 patients who underwent lumbar spine fusion for degenerative pathology. All included patients had biplane EOS wholebody imaging before and after surgery with a minimum of 6 months of follow-up. Exclusion criteria were patients with prior spine surgery and total knee replacement before lumbar spine fusion. Twenty-seven parameters, including knee flexion and spinopelvic parameters, were measured using SterEOS software. The statistical analysis of the parameters, their evolution and their correlations were performed with XLSTAT-R and GraphPad Prism 9. The predictive methods of regressions and classifications are based on a database of 201 preoperative radiographs of patients with degenerative spine pathologies. They are developed under Python using the Scikit-Learn machine learning library. Postural parameters of 108 patients (65H; 43F) with an average age of 63.02 years were studied. Knee flexion was moderately correlated with ankle dorsiflexion angle (R=0.75) and weakly correlated with SVA C7(R=0.41), PI-LL (R=0.41) or PT (R=0.20). Patients with greater preoperative knee flexion (n=68) had significantly greater PI-LL mismatch (9.88 +/- 10.48° vs. 0.57 +/- 12.19° ) and VAS (5.7 +/- 4.2 cm vs. 2.1 +/- 3.9 cm) than the group with less knee flexion (n=44). Patients with exacerbated knee flexion postoperatively (n=57) had significantly greater PI-LL mismatch (p=0.01) as well as loss of lordosis (p=0.04) after surgery compared to the group with less flexion (n=51) without any impact on VAS. Prediction of flexion beyond a threshold of 10° of knee flexion from the parameters VAS, CAM, PI-LL and age is possible with an accuracy of 0.68, a sensitivity of 0.86 and a specificity of 0.44. The results of this work suggest that lower limb alignment in relation to global and regional sagittal spinal parameters can be quantified using whole body imaging. Knee flexion angle can help quantify lower limb compensation in relation to spinopelvic misalignment in patients operated on for degenerative pathologies. However, the relationship between spinopelvic parameters and knee alignment remains poorly correlated and there is considerable subjectdependent variability in the implementation of these compensation mechanisms. Further work is needed to fully understand the complex interplay of compensation mechanisms in the face of sagittal imbalance and spinopelvic misalignment.

Date	26 Sept 2023
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Carlos Vázquez (Supervisor) & Jacques A. de Guise (Co-supervisor)