Methods for quantitative characterization of bone injury from computed-tomography images

Area of Science:

• Biomedical imaging
• Bone research
• Quantitative morphology

Background:

• Computed tomography (CT) imaging offers insights into bone structure for disease diagnosis.
• Current evaluation of bone structure and shape is often qualitative, limiting comparisons and early detection of bone quality changes.
• There is a need for quantitative methods to assess bone structure and shape for improved disease characterization.

Purpose of the Study:

• To develop and validate quantitative methods for bone morphometry, texture analysis, and shape analysis.
• To enable more precise characterization of bone diseases using CT images.
• To assess the framework's ability to detect the impact of hemophilia on bone structure.

Main Methods:

• Utilized bone morphometry, texture analysis, and shape analysis techniques on CT images.
• Validated the analysis methods using a mouse femur wound model to study hemophilia.
• Developed open-source software extensions for 3D Slicer to ensure accessibility and reproducibility.

Main Results:

• Shape analysis effectively localized and characterized spurious bone formation.
• Texture and bone morphometry analyses provided additional information on bone composition.
• One-dimensional (1D) textural features significantly differentiated injured femurs from healthy femurs, even in a small sample size.

Conclusions:

• The proposed quantitative analysis framework enhances the definition and understanding of bone disease phenotypes.
• The methods offer improved capabilities for inter-study comparisons and detecting subtle bone quality variations.
• The open-source software promotes robustness and reproducibility in medical image analysis for bone research.