Magnetic resonance imaging provides accurate measures of cartilage creep and biomechanical tissue properties: Ex vivo comparison to ground truth mechanical testing

Abstract

Cartilage mechanical properties have been suggested to be more effective biomarkers for early-stage osteoarthritis (OA) than conventional clinical pain and image feature detection, when compared with OA grading methods. However, limited research exists evaluating the feasibility of alternative methods, such as magnetic resonance imaging (MRI) techniques, to determine biomechanical properties. Therefore, this study aimed to evaluate the feasibility of clinical MRI for non-invasive evaluation of cartilage creep behaviour and biomechanical properties. Bovine cartilage samples (n = 12, diameter = 6 mm) were loaded at 0.25 MPa/s until reaching 1 MPa, then held under constant stress for 1 h using a counterbalanced study design with two different configurations. The first configuration used a custom-made, hydraulic-based MRI-compatible device to apply the load to the sample. During loading, 2D proton density-weighted fast spin echo MR images with fat suppression (CHESS method) were captured every minute. The second configuration used a universal testing machine as a ground truth (GT) reference. Time-dependent creep deformation was assessed in both configurations, and the instantaneous and equilibrium moduli were calculated at 1 min and at the end of the creep test, respectively. In addition, sample-specific fibril-reinforced poroelastic (FRPE) material parameters were estimated for both configurations using inverse finite element analysis of the measured creep data. The FRPE model successfully simulated experimental data, with mean R² values of 0.77 [95 % CI: 0.61, 0.92] for MRI and 0.98 [95 % CI: 0.95, 0.99] for GT. Results showed comparable deformation trajectories with no significant differences in the FRPE material properties between the configurations (i.e., E_f⁰,E_f^ε,E_nf,k₀,M). Only the mean instantaneous modulus at 1 min of creep was higher (p < 0.001) with MRI 4.5 [95 % CI: 2.9, 6.1] MPa compared to GT 2.9 [95 % CI: 2.3, 3.5] MPa. These findings demonstrate that MRI can capture cartilage creep deformation and estimate biomechanical properties with reasonable accuracy in an ex vivo setting. This advocates towards further development of the workflow for creep compression experiments in vivo. Yet, the workflow requires load-controlled relaxation and considerations of 3D contact mechanics of the human knee. While this work does not yet establish clear clinical applicability, it represents important evidence for non-invasive quantification of cartilage biomechanics. It is conceivable that our advancements may contribute to subject-specific estimation of inherent biomechanical tissue properties in the future.

Keywords: Bovine; Creep compression; Ex vivo; Fibril-reinforced poroelastic; Inverse finite element analysis; Mechanical properties.