Semi-Automated Bone Tracking for Dynamic MRI Analysis of Knee Joint Kinematics

Motivation: The ability to analyze tibiofemoral kinematics during dynamic imaging is important for research and clinical assessment. Goal(s): To develop and validate a semi-automated pipeline to track tibiofemoral motion and extract kinematic parameters from sagittal plane CINE MRI. Approach: A combination of edge detection with connected-component labeling and frame-to-frame transformation optimization was used to track bone boundaries during knee flexion-extension cycles during dynamic MRI. The method was validated in five healthy volunteers and compared with manual segmentation. Results: Measured kinematics (flexion-extension range of motion: 24.1°±7.2°, anterior-posterior displacement: -18.8±3.2 mm, superior-inferior displacement: 0.8±1.3 mm) matched expected physiological ranges and showed higher accuracy than manual segmentation. Impact: Direct analysis of dynamic MRI frames can achieve accurate motion tracking of the tibia and femur without requiring high-resolution reference scans. This approach outperforms manual segmentation and streamlines the study of in vivo knee kinematics from imaging data.