Bone quality depends on tissue properties (modulus of elasticity, mineral density, bone matrix quality and cellular behavior) 1 and structural properties (geometry, macro and micro-architecture) 2.
To assess bone tissue properties under various systemic 3 and local conditions 4, bone defects created surgically in animal models are used to provide a better understanding of the repair process 5,6. For that purpose, a wide range of imaging modalities are used, including micro-computed tomography (micro-CT), which is a nondestructive and high-resolution method 7, currently considered the gold standard for quantifying the bone micro-architecture in animal models 8. In this method, images are collected from multiple viewing angles around the sample and a three-dimensional (3D) reconstruction on a micrometric scale is performed 9,10 in a relatively short turnaround time with a high throughput 11, sharing the same physical principles as that of a clinical computed tomography scanner, but with higher resolution imaging as result of its smaller voxel sizes and a longer exposure 10,12.
Some of the numerous advantages of micro-CT, such as the narrow, monochromatic X-ray with high flux, which allows an in-depth exploration of the X-ray imaging modality 12; the possibility of analyzing the internal structure without cutting the sample 13, thus allowing further histological or mechanical tests using the same sample 14 or even repeat the micro-CT scanning of the same sample as many times as necessary 12,15, can be mentioned. It also provides a simultaneous precise quantification of different parameters, including mass, mineral density, geometry, and trabecular and cortical parameters 16.
The evaluation of bone microarchitecture by micro-CT includes steps such as acquisition of X-ray projection images, 3D reconstruction of the obtained images, 3D analysis and reporting the results, using specific software 6,8,17. There are variables associated with those steps that can affect the morphological results 18 and, standardize the micro-CT methodology, a guideline was published aimed at studies that assess bone microstructure through micro-CT in animal models 8.
The segmentation of the analyzed structures is the first step in structural analysis and is critical for the accurate quantification of architectural parameters 19. This is dependent on two steps. First, the definition of a region-of-interest (ROI) and then, its thresholding (or binarization) to separate the tissues that must be quantified.
Regarding the ROI definition, a more precise and reliable contour can be obtained using the manual method 8, however, this method depends on the operator's experience, and it seems to be more time-consuming in comparison to automatic method based on predefined shapes (circular, rectangular, triangular, and others) 6. One of the disadvantages of using predefined formats is that they cannot fully involve the ROI when the defect has irregular borders 19. However, in lesions with regular shapes, such as circular lesions created by drills, the ROI definition method with predefined formats can represent an easy and quick alternative 8,20.
Despite the existence of studies addressing the micro-CT methodology 6,8,19 there is a lack of information about the influence of the ROI definition on morphometric results in surgically created bone defects. In this context, this study aims to assess whether there is an agreement between the manual and predefined ROI definition methods during the 3D analysis of bone defects by micro-CT. Our hypothesis was that there is an agreement between both ROI definition methods.