A Novel Approach to Support Diagnosis and Treatment of Knee Osteoarthritis via Superimposition of Ground Reaction Force on Proximal Tibial Model

INTRODUCTION: Medial Knee Osteoarthritis (MKO) is generally associated with abnormal varus tibio-femoral joint alignment [1]. During walking, this may result in altered lower limb locomotion and unbalanced loading at the knee. In detail, while the contact forces are reduced on the lateral condyle in the varus alignment, the latter causes a medial shift of the weight-bearing axis, producing an abnormal moment arm at the knee that increases forces across the medial compartment, resulting in cartilage hypercompression and consequent wear within the medial condyle [2].
To delay or prevent end-stage MKO, the correction of this misalignment may be considered at an early disease stage to restore physiological tibio-femoral alignment and to provide medial compartment decompression [1,3]. This is generally achieved surgically via high tibial osteotomy (HTO). For the latter, the treatment efficacy may be provided by the analysis of the ground reaction forces (GRF), whose three-dimensional orientation reflects knee (mis)alignment if properly and reliably depicted upon real patient-specific knee anatomy (4).
Unfortunately, although multi-instrumental evaluations are executed frequently, these rarely involve reliable and robust combination of subject-specific medical imaging, i.e. computer tomography (CT), and gait analysis (GA), including GRF data. The aim of this study was to propose an original methodology merging CT imaging with GA and GFR data in order to depict a realistic patient-specific representation of the knee status during weight bearing activities before and after HTO.

METHODS: 25 patients (BMI<25; 40-65 years) affected with MKO and with a varus deformity < 20° were selected for HTO. All patients received pre- operative clinical scoring (via KSS, KOOS and EQ-5D), radiological examinations and instrumental evaluations; so far, these were also executed post- operatively at 6-month follow-up on 14 of these patients. In detail, standard GA [5] was performed during walking, squatting and stair climbing/descending using a 9-camera motion capture system (Vicon, Oxford, UK), combined with wireless EMG and force platforms (Kistler, Einterthur, CH) for GRF tracking, along with updated standard procedures. The latter implied motion data collection of 4 additional skin-based reflecting fiducial markers, positioned around the palpated tibial plateau bony rim. By still wearing the GA markers, patients received full lower-limb X-ray and CT in weight-bearing condition (Carestream, Rochester, NY-USA). In data processing, relevant DICOM files were segmented to extract computer-aided design (CAD) models for the tibia and the 4 additional markers; an anatomy based reference frame was defined on the proximal tibia. Afterwards, the trajectories of the 4 additional fiduciary markers as derived from GA during motion were registered on the corresponding landmarks from CT-based bone reconstruction. Resulting registration matrices were then used to superimpose GRF data on the reconstructed tibia morphology. Segmentation reproducibility and marker cluster deformation were assessed; intersections of GRF vectors with the tibial plateau were calculated and depicted throughout motion cycles, and reported in percentage of the medio-lateral tibial plateau width.

RESULTS SECTION: Pre-operative clinical and radiological scoring confirmed MKO and associated abnormal varism, this being up to 18° over the patients. The morphological characterization of GRF was successfully achieved pre- and post- HTO on patient-specific tibial plateau morphology. Pre- operative GFR patterns and peaks positions, along with those related to the three knee joint moments, were located medially to the tibial plateau, as expected. After HTO, these resulted lateralized and much closer to the tibial plateau spine, as desired (see figure for pre- and post-op intersection of GRF vector on tibial plateau plane from a well representative patient). In detail, when post-op is compared to pre-op, the difference of the centroids were, on average, 54.5±18.2 mm (min÷max: 35.7÷72.5 mm) more lateral during walking and 52.4±27.6 mm (24.7÷86.4 mm) during stair climbing. When reported in
% of the tibial plateau width, these values became 69.3±20.2 (45.9÷81.3) and 78.2±29.8 (43.2÷97.8), respectively. Post-op clinical scores and GA revealed a considerable overall improvement, especially in functional performances.

DISCUSSION: The reported novel approach allows a linkage between motion data, including GFR, and tibial-plateau morphology, depicting the unbalanced loading due to knee varism. Relevant pre- and post-op application provide a quantification of the effect of the original deformity and executed joint realignment. This approach may offer also a useful assistance for surgical planning in case of HTO as well as ideally in other orthopedic treatments.

SIGNIFICANCE/CLINICAL RELEVANCE: This study reports on an original methodology merging patient-specific knee morphology, reconstructed via medical imaging, and ground reaction forces. This is thought to be potentially useful for MKO diagnosis and knee surgical planning, e.g. in HTO.

REFERENCES:
[1] Zaffagnini S et al. (2013) KSSTA, 21:934-941.
[2] Sharma LS et al. (2010) Ann Rheum Dis, 69(11): 1940–1945.
[3] MacLeod AR et al. (2019). Bone Joint Res, 7: 639–649.
[4] Belvedere C et al. 28th Congress of the International Society Of Biomechanics (ISB). Digital Meeting 25-29 July, 2021. Abstract 5072952.
[5] Leardini A et al. (2007). Gait Posture, 26:560-71.