Semi-automated bone tracking in dynamic CINE MRI during controlled knee motion

Area of Science:

• Biomedical Engineering
• Radiology
• Orthopedics

Background:

• Dynamic magnetic resonance imaging (MRI) allows in vivo visualization of bone movement during knee motion.
• Quantifying joint kinematics from dynamic MRI is challenging due to image quality limitations.
• Accurate assessment of knee joint motion is crucial for understanding biomechanics and diagnosing conditions.

Purpose of the Study:

• To develop a semi-automated pipeline for tracking femoral and tibial motion from CINE MRI during knee flexion and extension.
• To evaluate the accuracy, processing time, and consistency of the developed method against manual segmentation.

Main Methods:

• A semi-automated algorithm combining Canny edge detection, connected-component labeling, and frame-to-frame transformation optimization was developed.
• The method was validated on five healthy volunteers using a dedicated MRI-compatible device for controlled knee movements.
• Bone displacements were tracked in the 2D image coordinate system, comparing semi-automated results to manual segmentation.

Main Results:

• The semi-automated method achieved an average bone boundary alignment error of 0.40 ± 0.02 mm.
• Processing time was reduced from ~15 minutes (manual) to <5 minutes (semi-automated) per dataset.
• The semi-automated method demonstrated improved precision with smaller standard deviations in displacement measurements compared to manual segmentation.

Conclusions:

• The developed semi-automated method offers a reliable and time-efficient approach for quantifying relative bone positions in dynamic MRI.
• This technique facilitates the analysis of volitional knee motion, supporting clinical and research applications.
• The improved efficiency and reliability make it suitable for analyzing dynamic MRI data.