Purpose: Pre-radiographic knee osteoarthritis (OA) changes such as cartilage swelling and softening are indicative of cartilage degeneration. Magnetic Resonance (MR) T2 relaxometry imaging is sensitive to changes in cartilage composition in early OA. Increased knee cartilage T2 map values have been reported in individuals with anterior cruciate ligament (ACL) injury and/or OA. Tibiofemoral (TF) cartilage is known to adapt to activities of daily living such as walking. Kinematic changes with OA and/or ACL deficiency (ACLD) could cause repetitive loading of ill-adapted cartilage regions leading to collagen network breakdown and PG loss. Characterizing the T2 map value within the contact and non-contact TF cartilage regions could provide key information for detection of early OA. The purpose of this study was to develop 3D methodology for T2 map value computation within contact and noncontact TF cartilage regions during walking. Our hypothesis was that the average and range of T2 values would be different between the contact and non-contact regions for healthy and ACLD individuals. Methods: Three male participants volunteered for this study (age 55, 44, 34 years; 2 unilateral ACLD; 1 intact ACL with unilateral meniscal injury). Participants arrived at the lab in the morning and were fully non-weight bearing for at least 30min prior to imaging. MR imaging (3T GE Discovery 750) was performed in a supine position using morphological (3D high resolution steady-state fast precision) and T2 relaxometry (2D multislice multiecho Carr-Purcell Meiboom-Gill) sequences. Next, participants' knees were imaged using a dual fluoroscopy (DF) system at 120Hz while walking at 1.2m/s on a treadmill. 3D TF bone and cartilage models were generated from segmented morphological scans (Amira). Each TF cartilage model was volume meshed using 3D fournode tetrahedron elements with x0.01mm edge length resulting in x50,000 nodes and x200,000 elements. TF arthrokinematics were obtained using 2D-3D registration software (Autoscoper, Brown University) for x0.4s from 0.1s before to 0.3s (midstance) after heelstrike when the bones were visible within the calibrated DF field-of-view. TF cartilage T2 maps were computed using an algebraic algorithm applied to T2 relaxometry images. A custom Matlab (2014b) program applied DF derived bone alignments for each motion frame to respective cartilage volume mesh models to determine proximities between the femur and tibia cartilage volume mesh nodes. Femur cartilage node proximity was computed as the distance between each femur volume mesh node and the closest neighboring tibia cartilage volume mesh node, and vice versa for tibia cartilage node proximity. T2 map value for each femur and tibia cartilage volume mesh node was obtained from T2 map images. Femur and tibia cartilage contact and non-contact region for each motion frame was defined as nodes with 0<proximity<1mm and 1<=proximity<2mm, respectively. Paired t-tests were used for within-subject comparison of contact and non-contact regions' average and range of the T2 map values for the stance phase of walking. Results: T2 average and range values were significantly different between contact and non-contact regions for all participants (Table 1). T2 map range in the intact ACL contralateral limb was significantly lower in contact compared to non-contact region (Table 1). T2 map range in the ACLD limb's lateral femoral and tibial cartilage was significantly greater in contact compared to the non-contact region (Table 1). Conclusions: The developed methodology permitted subject-specific 3D analysis of T2 maps in TF cartilage contact and non-contact regions (Table presented) during stance phase of walking. T2 map average and range was different in TF cartilage contact and non-contact regions. Cartilage adapts to cyclic loading it experiences. In a healthy knee the regions of contact have lower average and range of T2 values compared to non-contact regions. The increase in average T2 map values for the ACLD knee is likely attributed to a perturbed normal kinematic path where the contact region may encompass cartilage ill-adapted to weight bearing. Although the number of participants is small, this study provides the tools to use subject-specific functional information to evaluate cartilage structure and composition. Work is ongoing to recruit more participants to develop characteristic T2 maps for aid in early OA detection.
Sharma, G. B., Kuntze, G., Beveridge, J. E., Bhatla, C., Shank, J., & Ronsky, J. L. (2015). Characterization of tibiofemoral cartilage T2 mapping in the contact and non-contact regions during walking using dual fluoroscopy and magnetic resonance imaging. Osteoarthritis and Cartilage, 23, A256–A257. https://doi.org/10.1016/j.joca.2015.02.466