A Laplace-Hamming Binarization Approach for Second-Generation HR-pQCT Rescues Fine Feature Segmentation

Area of Science:

• Biomedical Engineering
• Orthopedics
• Medical Imaging

Background:

• Second-generation high-resolution peripheral quantitative computed tomography (XCTII) offers high-resolution in vivo bone microstructure assessment.
• The standard XCTII image processing protocol may omit fine trabecular and cortical bone features, impacting analysis accuracy.
• Accurate segmentation of bone microstructure is vital for diagnosing and monitoring bone diseases.

Purpose of the Study:

• To implement and evaluate a novel Laplace-Hamming (LH) binarization approach for optimizing fine structure segmentation in XCTII images.
• To compare the reproducibility and accuracy of the proposed LH segmentation with the manufacturer's standard Gaussian-based binarization.
• To assess the impact of LH segmentation on key bone structural parameters in both trabecular and cortical compartments.

Main Methods:

• A Laplace-Hamming (LH) binarization approach was developed for XCTII image segmentation.
• Reproducibility was assessed using repeat XCTII scans of radii and tibias from 20 healthy volunteers.
• Accuracy was evaluated by comparing XCTII scans (standard and LH) with micro-CT (μCT) data from cadaveric bone phantoms.

Main Results:

• The LH approach successfully segmented fine bone features missed or distorted by the standard method.
• LH segmentation significantly reduced errors in trabecular volume fraction (BV/TV) and thickness (Tb.Th), but increased error in trabecular separation (Tb.Sp).
• LH improved cortical porosity (Ct.Po) correlation with μCT and reduced error in cortical pore diameter (Ct.Po.Dm), enhancing precision for multiple parameters.

Conclusions:

• The proposed Laplace-Hamming (LH) segmentation approach yields improved binary masks for XCTII images.
• LH segmentation reduces bias and enhances accuracy and reproducibility of key bone microstructure outcome metrics.
• This method offers a more precise assessment of both trabecular and cortical bone compartments, advancing in vivo bone analysis.