Tibiofemoral Cartilage Deformation Determined in Cyclic Compression by Displacement-Encoded MRI

Introduction: Measurement of articular cartilage deformation is important to characterize the differences between healthy and pathological tissue and to compare native tissue with engineered constructs. Among these differences is the nonuniform distribution of extracellular matrix proteins through the tissue depth, which in turn gives rise to heterogeneous strains and anisotropic mechanical properties of cartilage [1-4]. Magnetic resonance imaging (MRI) is a noninvasive technique that can render tissue morphology and measure tissue deformation. One method of determining deformations is to map the change in tissue thickness using morphologic data and curve fitting routines [5]. Another method is the cartilage deformation by tag registration (CDTR) technique, which documented heterogeneous deformations through the volume of cartilage in unconfined compression [4]. This method relied on the registration of “tag” lines in tissue that were typically separated by 3-4 pixels. In this study, displacement-encoded spin echoes (DENSE [6]) with a fast spin echo (FSE) data acquisition were used to determine two-dimensional (2D) deformations in an intact porcine tibiofemoral joint. The technique is an improvement over previous methods since deformations are determined at each pixel location and throughout the volume of the deformed tissue. Materials and Methods: A single juvenile porcine leg was stored at –20°C until thawed at room temperature for testing. Skin and subcutaneous tissue were removed from the stifle (knee) joint without disruption of the joint capsule. The tibia and femur were fixed with polymethylmethacrylate in custom plastic cups and secured to an electro-pneumatic cyclic loading device, which constrained the joint to only compression/distraction. A preliminary study (n=11) found that 550 loading cycles at one times body weight (78 N) were required to ensure a steady-state load-deformation response before imaging. After steady-state was achieved, the Biospec 70/30 MR system (7.05 Tesla, Bruker Medical GMBH, Germany) acquired displacement-encoded phase information from a sagittal slice of the medial tibiofemoral joint. During each 10 s cycle, the displacement-encoding gradient was applied between two 90° RF pulses. Load was then applied, and FSE was used to acquire deformed cartilage data. The technique of cosine and sine modulation (+cos, –cos, +sin, –sin) to eliminate artifacts (CANSEL [7]) was utilized. DENSE imaging parameters were: encoding gradient strength=0 mT/m for a reference image and 20 mT/m for 2D (x and y) displacement encoding; gradient encoding time=3.0 ms, and mixing time=1000 ms. FSE imaging parameters were: TR=10000 ms; TE=37.42 ms; number of echoes per excitation=16; spatial resolution=250x250 μm2; slice thickness=2.0 mm; and number of averages=4. Displacement fields in the femoral and tibial cartilage were determined using MATLAB (v.7.0, The Mathworks, Natick, MA) software. Artifacts were eliminated from the acquired images by combining the phase-cycled data. A 2D inverse Fourier transform was then applied, and first-order phase corrections were performed. The differences in the phase information between a reference scan and a directional gradient scan were then computed, allowing for the direct calculation of the displacement within a pixel of interest in the direction of the gradient. A moving average algorithm smoothed the displacement data. In preliminary studies, deformations from repeated tests (n=9) of a silicone phantom determined the absolute precision of this method, which was defined as the RMS of deformations at standardized material locations, as 16.2 μm. Results: Nonuniform in-plane deformations were documented during simple compressive loading for articular cartilage in the femur and tibia of an intact porcine stifle joint (Fig. 1). The deformations in both the loading and transverse (y and x, respectively) directions were the largest at the contact between femoral and tibial articular cartilage. Deformation was negative in the loading direction, indicating compression. Deformation of the femoral articular cartilage in the transverse direction was close to zero, and that of the tibial articular cartilage indicated anterior movement. Discussion: Deformation of the articular cartilage of a cyclically compressed intact animal joint was documented using displacement-encoded MR imaging. Tibiofemoral contact in the medial condyle of the specimen occurred during loading as evidenced by compressive deformation patterns. Throughout the cartilage, the deformations were heterogeneous in response to the cyclic load. The femur was compressed with the maximal deformation occurring in the tibiofemoral contact region. Transverse and anterior displacement of the tibial articular cartilage relative to the femur also occurred. Anterior displacement was attributed to the geometry of opposing cartilage surfaces, characterized by a posterior slant of the tibial surface. Ligamentous and other soft tissue attachments and the possible shear deformation of the growth plate cartilage in the juvenile cartilage may have also contributed to the anterior displacement. This animal joint study expands on previous results which documented deformation in explants of articular cartilage [4]. In addition, the DENSE-FSE pulse sequence allowed for determination of displacements at every pixel in a region of interest, at a precision below the spatial resolution of the image. From this displacement data it is possible to compute the strain field in the tibiofemoral joint cartilage.