Non-invasive, in vivo measurement of the three-dimensional (3-D) motion of the tibiofemoral joint is essential for the study of the biomechanics and functional assessment of the knee. Real-time magnetic resonance imaging (MRI) techniques enable the measurement of dynamic motions of the knee with satisfactory image quality and free of radiation exposures but are limited to planar motions in selected slice(s). The aims of the current study were to propose a slice-to-volume registration (SVR) method in conjunction with dual-slice, real-time MRI for measuring 3-D tibiofemoral motion; and to evaluate its repeatability during passive knee flexion.

Eight healthy young adults participated in the current study, giving informed written consent as approved by the Institutional Research Board. A 3-T MRI system (Verio, Siemens, Erlangen, Germany) incorporated with a neck matrix coil was used to collect the MRI data. A 3-D scanning using the VIBE sequence was used to collect the volumetric data of the knee at fully extended position (TR = 4.64 ms, TE = 2.3 ms, flip angle = 15°, in-plane resolution = 0.39 × 0.39 mm² and slice thickness = 0.8 mm). A real-time MRI using the refocused radial FLASH sequence (TR = 4.3 ms, TE = 2.3 ms, flip angle = 20°, in-plane resolution = 1.0 × 1.0 mm², slice thickness = 6 mm) was used to acquire a pair of image slices of the knee at a frame rate of 3 fps during passive flexion.

The volumetric MRI data sets were segmented for the femur and tibia/fibula to isolate the sub-volumes containing bone segments. A slice-to-volume registration method was then performed to determine the 3-D poses of the bones based on the spatial matching between sub-volume of the bones and the real-time image slices. The bone poses for all frames were used to calculate the rigid-body kinematics of the tibiofemoral joint in terms of the flexion/extension (FE), internal/external rotation (IR/ER), abduction/adduction (Abd/Add) and joint center translations along three anatomical axis of the tibia. The procedures were carried out five times for repeatability analysis. The standard deviation (SD) of the rigid-body kinematics for each frame from the five trials were calculated and then averaged across all frames to give quantitative measures of the repeatability of the kinematic variables.

The repeatability analysis showed that the mean±SD of the averaged SD in FE, Abd/Add and IR/ER components across all subjects were 0.25±0.09, 0.46±0.13 and 0.77±0.16 degrees, respectively. The corresponding values for the joint translations in anterior/posterior, proximal/distal and medial/lateral directions were 0.21±0.04, 0.11±0.03 and 0.43±0.09 mm.

An SVR method in conjunction with dual-slice real-time MRI has been successfully developed and its repeatability in measuring 3-D motion of the tibiofemoral joint evaluated. The results show that the proposed method is capable of providing rigid-body kinematics with sub-millimeter and sub-degree precision (repeatability). The proposed SVR method using real-time MRI will be a valuable tool for non-invasive, functional assessment of the knee without involving ionizing radiation, and may be further developed for joint stability assessment.