Accuracy assessment of 3D bone reconstructions using CT: an intro comparison

Area of Science:

• Orthopedic biomechanics
• Medical imaging
• Computational anatomy

Background:

• Computed tomography (CT) is vital for 3D joint reconstruction, aiding orthopedic implant design, computer-assisted surgery, and biomechanical analysis.
• Accurate 3D models of osseous and cartilage geometry are essential for these applications.
• Evaluating the precision of CT-based reconstructions against established methods is crucial.

Purpose of the Study:

• To assess the accuracy of 3D bone and cartilage surface models reconstructed from clinical CT scans.
• To compare reconstructed geometries with high-resolution bone digitizations serving as ground truth.
• To validate commercially available reconstruction techniques using standard CT protocols.

Main Methods:

• Utilized 8 fresh frozen cadaveric elbow specimens.
• Obtained ground truth bone geometry using an optical tracking system for digitization.
• Employed proximity mapping to calculate residual errors between CT-derived models and digitized bone surfaces.
• Assessed cartilage surface model accuracy using CT scans with air contrast.

Main Results:

• Mean error for bone surface reconstruction was <0.4 mm in cortical and <0.3 mm in subchondral regions.
• Mean error for 3D cartilage surface models derived from CT scans was <0.3 mm.
• Demonstrated high accuracy of reconstructed models compared to ground truth digitizations.

Conclusions:

• Clinical CT scanning protocols can generate accurate 3D bone and cartilage models.
• Commercially available reconstruction algorithms effectively represent true joint geometry.
• CT-based modeling provides a reliable tool for orthopedic applications requiring precise anatomical data.