Automatic identification of radius and ulna bone landmarks on 3D virtual models

Area of Science:

• Orthopedics
• Biomedical Engineering
• Anatomical Imaging

Background:

• Landmarks are crucial for analyzing bone morphology and biomechanics.
• Current methods for landmark identification on 3D radius and ulna models are manual.
• Automatic landmark identification is needed for objective forearm function measurement.

Purpose of the Study:

• To develop a knowledge-based method for automatically locating landmarks on 3D radius and ulna bone models.
• To establish coordinate systems for the radius, ulna, and forearm.
• To enable objective measurement of forearm function.

Main Methods:

• A knowledge-based algorithm was developed for landmark identification.
• The method does not require pre-labeled data.
• Coordinate systems for individual bones and the combined forearm were calculated.

Main Results:

• The algorithm's landmark placement was validated against expert-defined landmarks on cadavers.
• Median Euclidean distance differences ranged from 0.4 to 1.8 mm.
• Median angular differences for coordinate systems were within -1.4 to 2.0 degrees.

Conclusions:

• The developed automatic method is a validated first step for 3D forearm anatomy analysis.
• Applicability depends on specific use cases and accuracy requirements.
• This method supports the automatic analysis of three-dimensional forearm anatomy.