Image processing for femoral endosteal anatomy detection: description and testing of a computed tomography based...

Area of Science:

• Orthopedic surgery
• Medical imaging
• Biomechanical engineering

Background:

• Accurate three-dimensional (3D) modeling of the femoral endosteal cavity is crucial for personalized orthopedic implant selection and design.
• Current methods for assessing femoral anatomy may lack the precision required for optimal component fitting.

Purpose of the Study:

• To develop and validate a computed tomography (CT)-based image processing program for 3D femoral endosteal cavity shape modeling.
• To assess the accuracy of the developed program by comparing its measurements to manual cadaveric measurements.

Main Methods:

• Utilized CT imaging of 50 cadaver femora, acquiring 30 axial slices per femur.
• Applied various image analysis techniques, including thresholding and edge detection operators, for femoral cavity detection.
• Developed contour tracking algorithms with validity criteria for border pixel identification.
• Validated results by manually measuring dimensions after sawing femora into 10 slices, comparing CT-derived data with caliper measurements.

Main Results:

• The CT-based image processing program successfully generated 3D models and dimensional data (anteroposterior, mediolateral, oblique diameters, center points).
• A mean difference of 1.1 mm (±0.7 mm) was observed between CT-derived and manual measurements.
• Measurement differences varied based on femoral region and cortical thickness, being highest in proximal slices (1.3 mm ±0.8 mm) and lowest in distal slices (0.9 mm ±0.6 mm).

Conclusions:

• The developed CT-based image processing program provides acceptable accuracy for 3D femoral endosteal cavity modeling.
• The program's ability to accurately capture femoral dimensions supports its application in studying endosteal anatomy.
• This technology can serve as a basis for individualized femoral component selection and surgical design.