Variabilities in µQCT-based FEA of a tumoral bone mice model

Area of Science:

• Biomechanical engineering
• Medical imaging analysis
• Orthopedic research

Background:

• Finite element analysis (FEA) using micro-Quantitative Computed Tomography (µQCT) shows promise for fracture risk prediction.
• Current FEA model generation from µQCT images often requires manual segmentation and operator input, introducing potential variability.
• Understanding operator-induced variability is essential for reliable fracture prediction, especially in studies involving bone pathologies like tumors.

Purpose of the Study:

• To quantify operator-induced variability in µQCT-based FE models of mice tibias.
• To assess the sensitivity of predicted failure load to operator-dependent variables like model orientation and length.
• To evaluate the impact of different boundary conditions on variability in tumoral and sham bone models.

Main Methods:

• Two operators generated FE models from µCT scans of 8 mice tibias (tumoral vs. sham).
• Failure load was predicted using fixed support and spherical joints boundary conditions.
• Variability was assessed by comparing predicted loads between operators and across different model orientations and lengths.

Main Results:

• Operator differences in predicted failure load were substantial (-122% to 93%).
• Spherical joint boundary conditions showed less operator variability (9.8%) compared to fixed support (58.3%).
• Bone tibia orientation significantly affected variability, particularly with fixed support (44.7%), while length had minimal impact (<4%).

Conclusions:

• Operator influence is a significant factor in µQCT-based FEA of mice tibias, affecting predicted failure loads.
• The choice of boundary conditions and model orientation critically impacts the reproducibility of FEA results.
• Automation of the µQCT-based FE model generation process is necessary to enhance the reliability and consistency of fracture risk assessments.