Validation of an MRI Technique for the 6-DOF Knee Kinematics Measurement

Background: For total knee arthroplasty (TKA), the optimal rotational position of the femoral component is felt to be critically important. The current knee joint kinematics measurement technology is unable to identify the exact rotation axis of the knee joint, the main reasons being low measurement accuracy and insufficient three-dimensional data (2D-3D image matching technology). In order to improve the effect of TKA surgery, we proposed a knee joint kinematics measurement method, based on the MRI technology, and verified its measurement accuracy. We then employed this method to identify the personalized optimal rotation axis of the knee joint for TKA patients. Purposes: The purpose of the study was 1) to propose a method for measuring knee joint kinematics and verify its accuracy and 2) to propose a method for determining the optimal rotation axis of knee joint for TKA surgery, based on accurate kinematic measurement results. Materials and Methods: The experiment was divided into two parts: in vitro and in vivo. The purpose of the in vitro experiment was to verify the measurement accuracy of our method. We fixed two aquarium stones (approximately 10 cm * 10 cm * 10 cm in size, close to the size of the distal femur and proximal tibia) firmly on the fixed and moving arms of the goniometer/vernier caliper with glue and immersed the aquarium stones in the water to capture MRI images. The MRI images were then processed with MATLAB software, and the relative motion of the two aquarium stones was measured. The measurement accuracy of our method was verified via the scale reading of the moving arm on the goniometer/vernier caliper. In vivo, 36 healthy elderly participants (22 females, 14 males) were recruited from the local community; our method was then employed to measure the relative motion of the tibia and femur and to observe the rollback and screw home motion of the medial/lateral condyle of the femur, which was identified as specific kinematic features of the knee joint. Results: In vitro, all measurements were accurate to <1 mm and <1°. In vivo, all knee measurements showed rollback motion (the rollback distance of the medial femoral condyle was 18.1 ± 3.7 mm and that of the lateral condyle was 31.1 ± 7.3 mm) and screw home motion. Conclusion: In the application scenario of knee joint kinematics measurement, our method has an accuracy of <1° of rotation angle and <1 mm of translation for all reference points, and it can be employed to identify the most stable axis of the knee joint. Significance: Using our method to accumulate data on the knee rotation axis of more subjects to establish an average rotation axis of a given population may help in knee prosthesis design and reduce the patient dissatisfaction rate. Individually measuring the patient's rotation axis before TKA surgery and adjusting the prosthesis installation in TKA may further reduce the patient dissatisfaction rate, and automatic computer measurement may be realized in the future, but it is still time-consuming for now.