Persistent homology analysis distinguishes pathological bone microstructure in non-linear microscopy images

Area of Science:

• Biomedical Engineering
• Materials Science
• Skeletal Biology

Background:

• Bone microstructure analysis is crucial for understanding skeletal health and disease.
• Non-linear microscopy techniques like SHG and TPaF offer detailed imaging of bone matrix and cellular components.
• Quantifying microstructural changes, particularly porosity, is challenging with traditional methods.

Purpose of the Study:

• To develop and validate a novel topological method for quantifying bone microstructure from non-linear microscopy images.
• To assess the method's ability to differentiate between healthy and pathologically altered bone tissue.
• To apply the method to a murine model of skeletal pathology.

Main Methods:

• Utilized persistent homology statistics with a signed Euclidean distance transform filtration on image patches.
• Analyzed second harmonic generation (SHG) and two-photon excited autofluorescence (TPaF) images of bone tissue.
• Applied the method to wild-type (WT) and osteocalcin-VEGF knockout (OcnVEGFKO) murine models.

Main Results:

• The topological method successfully quantified bone microstructure, including hole number, size, distribution, and crowding.
• Significant differences in topological statistics were observed between OcnVEGFKO and WT groups.
• High prediction accuracies (up to 98.7% for SHG images) were achieved in classifying bone samples based on genotype.

Conclusions:

• The developed topological method provides a robust and interpretable approach for bone microstructure analysis.
• This technique can effectively detect and quantify skeletal pathologies, offering high diagnostic accuracy.
• The method is capable of identifying abnormalities and features across various spatial scales in bone tissue.