Micro-computed X-ray tomography (MicroCT) is one of the most powerful techniques available for the three-dimensional characterization of complex multi-phase or porous microarchitectures. The imaging and analysis of porous networks are of particular interest in tissue engineering due to the ability to predict various large-scale cellular phenomena through the micro-scale characterization of the structure. However, optimizing the parameters for MicroCT data capture and analyses requires a careful balance of feature resolution and computational constraints while ensuring that a structurally representative section is imaged and analysed.
In this work, artificial datasets were used to evaluate the validity of current analytical methods by considering the effect of noise and pixel size arising from the data capture, and intrinsic structural anisotropy and heterogeneity. A novel "segmented percolation method" was developed to exclude the effect of anomalous, non-representative features within the datasets, allowing for scale-invariant structural parameters to be obtained consistently and without manual intervention for the first time.
Finally, an in-depth assessment of the imaging and analytical procedures are presented by considering percolation events such as micro-particle filtration and cell sieving within the context of tissue engineering. Along with the novel guidelines established for general pixel size selection for Micro-CT, we also report our determination of 3 μm as the definitive pixel size for use in analysing connectivity for tissue engineering applications.