Méthode d'évaluation fonctionnelle de la gonarthrose par étude de la pseudo-cinématique d'accroupissement

Abstract

In the literature, there is discordance about the use of arthritis radiological evidence for the evaluation of the pathology. Indeed, the studies that evaluated the correlation between the radiological grades and the function scores calculated from self-reported questionnaires obtained highly variable conclusions. When for some, severe knee arthritis correlates with an important function limitation, others did not find any relation between the two observations. From these results, it appears that radiological diagnostic alone can not explain the knee function. A quantitative measure of the articular function can complement the arthritis diagnostic. Several methods are accessible to analyse the 3D kinematics of the femoro-tibial joint with medical imaging. However, none allows the evaluation of the weight-bearing knee from a reproducible and accurate measurement, while minimising the prejudices the subject undergoes (exposition to ionizing radiations, invasive protocol). Hence, the main objective of this Ph.D. dissertation is to develop and validate a method to analyse the kinematic of the weight-bearing knee, with a solution to the problematic mentioned above. Then, the goal is to evaluate that the radiological evaluation of knee arthritis enriched with a kinematic measurement at the knee, correlates better with the function score. In the first part of the work, a method to compute the pseudo-kinematic of the femoro-tibial joint was developed. Five flexion angles of the knee were imaged with the low dose biplane EOS® system (0, 15, 30, 45 and 70°); the pair of radiographs of the knee in full extension allowed the 3D reconstruction of the femur, the tibia and the fibula. After the segmentation of the bony contours on every radiograph, the registration of the geometric models of the bony structures allowed to extract the position of the tibia relatively to the femur, from one flexion angle to the following one. The intrinsic 2D/3D registration was based on the iterative closest point algorithm. In addition, the introduction of both a weighted registration and the registration of the tibia/fibula as a rigid body considerably reduced the errors of the axial rotations computation. After the validation of the method on ten subjects, its measurement error in reproducibility and precision was assessed and lower than 0.4mm and 0.5° and the method sensitivity to the 3D reconstruction was lower than 1.7mm and 1.4°. With a study of the pseudo-kinematic adduction (ADD) by a clustering analysis, this parameter was found to be a functional radiological marker of the knee arthritis. The asymptomatic subjects were in mild ADD (<5°) or in abduction (ABD) whereas the arthritic subjects were is severe ADD >5° (14 patients with internal knee arthritis, 14 asymptomatic subjects). Considering knee ADD as a risk factor for the occurrence and progression the knee arthritis, it was possible to compute composite Kellgren-Lawrence grades for 28 subjects: the original Kellgren-Lawrence grade became worsened (higher) with severe knee ADD and decreased (lower) with knee ABD. The correlation of the KOOS (Knee Ostearthritis Outcome Score) function scores with these composite grades was lightly but not significantly higher than the correlation with the original Kellgren-Lawrence grades. In the last part of the work, the pseudo-kinematic screw-home mechanism was analysed (7 patients with internal knee arthritis, 5 asymptomatic subjects). The asymptomatic tibias were in external rotation (mean value 9°), the arthritic tibias were in a milder external rotation (0-5°). Moreover, for an asymptomatic subject, a low ratio of the sagittal slopes of the tibial plateaus and a low sagittal slope of the medial plateau were correlated with the internal tibial rotation. Also, results suggested the coronal slopes of the tibial plateaus had an influence on the external tibial rotation. This doctoral work provided a technical solution to evaluate the 3D kinematic of asymptomatic and arthritic knees with the EOS® system, with high reproducibility and accuracy. Also, knee adduction was found to be a functional radiological marker of knee arthritis and the necessity of complementing the radiological evaluation of the knee arthritis with such a measure was confirmed. Finally, the analysis of the knee morphology with the knee function revealed new hypotheses to a better comprehension of the joint, to a more complete evaluation of the pathology.

Date	24 Jul 2017
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Jacques A. de Guise (Supervisor) & Nicola Hagemeister (Co-supervisor)