*2.10. MicroCT*

Three-dimensional microcomputed tomography (μCT) was used to quantify bone regeneration 10 weeks after stage 2 of the Masquelet technique. The rats were killed, and the limb on which surgery was performed was collected, together with the surrounding soft tissues, and fixed by incubation in 10% phosphate-buffered formalin for two weeks. The area between the inner pins was scanned by microCT (Skyscan 1174, Bruker Micro-CT, Billerica, MA, USA) with a voltage source of 50 keV, a current of 745 mA and an isotropic resolution of 14.4 μm. Three-dimensional reconstruction was performed for all scans and analyzed with the same parameter setup (NRecon v.1.6 and CTAn v.1.11 software, SkyScan, Kontich, Belgium) to separate mineralized elements from the background, with the software histogram tool used to determine grayscale level threshold values. As a dedicated external fixator with a guide saw had been used to create the defect, it was possible to locate the 6 mm long defective region with precision (the distance between the two adjacent pins) and to identify it as the region of interest. The following data were collected within the region of interest: total defect volume (TV in mm3) and bone volume (BV in mm3) for the calculation of the bone volume ratio (BV/TV, as a %).

### *2.11. Statistical Analysis*

All results are reported as means ± standard error of the mean (SEM). The Shapiro-Wilk test was used to determine whether the data followed a normal distribution. An *F*-test was performed to verify the assumption of equal variances. Two-tailed Student's *t*-tests were used for comparisons if the data met both these requirements (normal distribution and

equal variances). If one or both the assumptions were not met, the PMMA and metakaolin groups were compared in non-parametric Mann–Whitney *U*-tests. Values of *p* < 0.05 were considered significant in all tests. Statistical analyses were performed with GraphPad Prism 5 statistical software (GraphPad Software Inc, La Jolla, CA, USA).
