An Algorithm for Automated Separation of Trabecular Bone From Variably Thick Cortices in High-Resolution Computed...

Area of Science:

• Biomedical Engineering
• Orthopedics
• Medical Imaging

Background:

• Accurate structural measurements of cortical and trabecular bone are crucial in various scientific fields.
• Current automated techniques for separating these bone types are lacking in accuracy and efficiency.

Purpose of the Study:

• To develop and validate a structure-based algorithm for automated separation of cortical and trabecular bone in binarized images.
• To compare the algorithm's performance against manual segmentation methods.

Main Methods:

• A novel algorithm utilizes cortical thickness as a seed value to identify and separate bone regions.
• The method was validated on seven biological datasets from four species using micro-computed tomography (μ-CT) and high-resolution peripheral quantitative computed tomography (HR-pQCT).
• Image segmentation was performed using a spatially local threshold value.

Main Results:

• The algorithm demonstrated a significant speed improvement, being approximately 11 times faster than manual measurements.
• Median errors in cortical area were -4.47 ± 4.15%, and in cortical thickness were approximately 0.5 voxels (μ-CT) and <0.05 voxels (HR-pQCT).
• Overall difference in thickness measurements was -28.1 ± 71.1 μm.

Conclusions:

• A simple, implementable, repeatable, and efficient methodology for unbiased separation of cortical and trabecular bone has been developed.
• This automated approach facilitates the evaluation of large datasets and full-field analyses, overcoming previous limitations.