Error introduced by common reorientation algorithms in the assessment of rodent trabecular morphometry using...

Area of Science:

• Orthopaedic Research
• Biomedical Imaging
• Preclinical Research

Background:

• Quantitative bone analysis using micro-computed tomography (μCT) is standard in preclinical research.
• Virtual image reorientation is a common processing step to standardize μCT data.
• Understanding reorientation-induced errors is crucial for accurate bone characterization.

Purpose of the Study:

• To quantify the error introduced by common reorientation algorithms in μCT-based bone analysis.
• To evaluate the impact of different resolutions, rotation types, and interpolation methods on bone parameters.
• To identify strategies for minimizing reorientation error in μCT studies.

Main Methods:

• Mouse and rat tibial metaphyses were scanned using μCT at resolutions from 6 to 30 μm.
• Trabecular regions of interest (VOIs) were manually selected and analyzed with and without rotation (In-Plane, Triplanar).
• Image interpolation was performed using Nearest-Neighbor, Linear, and Cubic methods; densitometric and morphometric variables were computed.

Main Results:

• Significant reorientation errors were observed across all bone parameters, increasing with voxel size.
• Nearest-Neighbor interpolation showed minimal error for densitometric parameters, unlike Linear and Cubic methods.
• In-Plane rotation introduced greater error than Triplanar rotation; morphometric error depended intricately on rotation and interpolation combinations.

Conclusions:

• Reorientation in μCT bone analysis can introduce substantial error, particularly at lower resolutions and with specific interpolation methods.
• Error can be eliminated by avoiding reorientation or by de-rotating VOIs.
• Minimizing error involves high-resolution scanning, appropriate interpolation selection, and consistent processing.