Comprehensive guideline for X-ray microtomography-guided bone analysis and histology applied to the assessment of defect repair in long and flat bones

Bone regeneration studies require reliable analytical tools capable of assessing both microstructural and cellular aspects of tissue repair. This article presents a comprehensive methodological guide that integrates microcomputed tomography (micro-CT) and histological analyses for standardized evaluation of bone defect repair in rodent models, focusing on the tibia (long bone) and calvaria (flat bone). We describe detailed surgical protocols for defect creation, imaging workflows for micro-CT acquisition, reconstruction, and morphometric quantification, alongside histological processing and semi-quantitative scoring of tissue parameters. Illustrative results demonstrate how micro-CT enables three-dimensional, non-destructive quantification of mineralized tissue, while histology provides complementary insights into non-mineralized structures, vascularization, and cellular organization. The combined use of these techniques enhances reproducibility, ensures methodological consistency, and allows for robust comparisons across studies. By establishing a standardized workflow, this protocol supports translational research in bone repair and offers a valuable resource for evaluating regenerative strategies and biomaterials.

Over the years, various research groups around the world have been studying bone tissue both in physiological conditions and under diverse pathological circumstances. Currently, research on bone injuries in various contexts is receiving significant attention due to advances in tissue engineering. This area of study is critical for public health, as bone injuries can severely impact an individual's quality of life and productivity, while also contributing to increased public spending (Pivonka and Dunstan, 2012). Various methodologies that simulate different types of bone injuries have been developed for translational research, demonstrating high reproducibility. These include methods such as creating fractures followed by immobilization or the production of circumferential defects. Lesion models in rodents, particularly in rats and mice, are commonly used for analyzing bone repair. This process typically involves the use of a surgical drill with a variable diameter that is connected to a micromotor, allowing for well-defined osteotomies (Kaur et al., 2020; Uusitalo et al., 2005; Fan et al., 2023). This technique provides advantages over bone fractures, such as clear delimitation and standardization of lesion size, yielding statistically representative and reproducible results without requiring the immobilization of the animal.

Most studies focus on creating bone defects in specific bones, such as the femur and tibia for long bone injuries, or the calvaria for flat bones (Fan et al., 2023; Uusitalo et al., 2005). However, the analysis parameters differ between these two types of bones due to their distinct anatomical structures. The repair process in flat bones has unique characteristics; it does not involve endochondral ossification and relies on cells migrating from the dura mater to facilitate healing. This process can take longer depending on the size of the injury (Gao et al., 2023). Therefore, it is crucial to select appropriate analysis times to accurately evaluate the primary features of the repair. For such analyses, imaging techniques like microcomputed tomography (micro-CT) and histomorphological evaluation should be utilized.

Micro-CT is regarded as the gold standard for analyzing calcified tissues due to its advanced capabilities for non-destructive structural characterization. This technique allows samples to be preserved for further analyses and enables both 2D and 3D evaluations, with significant applications in bone defect analysis. It can perform both qualitative and quantitative analyses.

Qualitatively, micro-CT provides informative three-dimensional images of the trabecular and cortical architecture of calcified tissues. Quantitatively, it facilitates the calculation of various morphometric parameters, including volume, thickness, and mineral density. Additionally, it quantifies three-dimensional morphometric measures such as the bone volume to total volume ratio (BV/TV), trabecular thickness (Tb.Th), the number of trabeculae per unit length (Tb.n), as well as bone porosity and the interconnectivity index (ICI) (Waarsing et al., 2004). The technique utilizes X-rays, generated by the braking of electrons in the equipment source. These X-rays possess a short wavelength, which grants them high penetrating power. The imaging of bone microstructure relies on the differential absorption of radiation by various materials. However, one limitation of this technique is that if the samples do not exhibit sufficient contrast in absorption throughout their entire volume. This is the case of immature bone, characterized by the coexistence of tissues with differing levels of mineralization. It can lead to images that appear uniform, obscuring the desired microstructural differences. Conversely, when materials do not absorb radiation at detectable levels, image formation can be compromised. This issue can arise in demineralized or extremely porous bone samples, where the density is inadequate for the necessary differential absorption to take place (Keklikoglou et al., 2019).

The operation of X-ray micro-CT is based on the image contrast created by variations in beam attenuation, which involves absorption and scattering processes. There are three main ways that X-rays interact with matter: the photoelectric effect, Compton scattering, and pair production. The dominance of each interaction mode depends on the energy of the photons and the atomic number (Z) of the absorbing material. Together, these phenomena contribute to the reduction of X-ray beam intensity.

The X-rays used in this technique are generated by the deceleration of electrons at the equipment's source. These rays have a short wavelength, which gives them a high penetration power. Among the mechanisms of X-ray interaction with matter, the photoelectric effect is the most significant for low-energy photons and high-Z elements. This effect is characterized by the complete absorption of the incident photon by an inner-shell electron, resulting in the “clean” removal of the photon from the beam. This total transfer of energy makes the photoelectric effect particularly important in imaging techniques such as micro-CT.

As the X-ray beam passes through a sample, attenuation occurs, which is directly influenced by differences in the density and composition of the material. Therefore, the level of X-ray transmission is determined by both the mass and the absorption coefficient of the sample (Schoeman et al., 2016).

The micro-CT analysis process generally consists of three stages. The first stage is physical scanning. During this stage, the sample is placed in the microtomography using a sample holder that is appropriate for the sample's shape and stability. This setup helps prevent any movement during the scanning process. Next, a collimated X-ray beam is directed through the sample, and a detector measures the attenuated radiation that passes through, capturing two-dimensional projection images from multiple angles as the sample rotates 180–360 degrees (Cnudde and Boone, 2013).

In the second stage, the 2D projection images undergo mathematical processing through reconstruction algorithms. This process transforms the projections into sectional slices of the sample. These slices are then stacked to create a three-dimensional volume of the analyzed sample. The third stage involves extracting quantitative data from specific regions within the set of reconstructed images and analyzing this data.

Histological processing enables analysis at the cellular and tissue levels, allowing for the identification of structures that micro-CT does not visualize. This includes the formation of various types of connective tissue, such as fibrotic tissue, cartilage, and newly formed non-mineralized bone matrix. Additionally, it helps distinguish between primary and mature bone (Trejo-Iriarte et al., 2019).

The integration of histological analyses with micro-CT in bone repair studies has provided complementary information and yielded excellent results (Trejo-Iriarte et al., 2019; He et al., 2017). However, there is a lack of studies that support researchers in guided micro-CT and histological analysis. To address this, we provide a novel, comprehensive protocol featuring comparative results for mouse calvaria and rat tibiae, bridging the gap between digital imaging and traditional histology. Our aim is to establish a methodological guide that facilitates the evaluation of defect repair in these bone models. We have described the parameters and metrics used to ensure reproducibility of results and standardization of analyses, allowing for consistent comparisons across different studies. The methods described in this study are applicable to the analysis of any flat or long bone models.

Surgery process

Trefine drill Maximus −2 mm (TRE2C30)

Micromotor Marathon Talmax Handy 702 110v

Sterile scalpel

Suture thread (Nylon 0.4)

Forceps, scissors, retractors

Sterile saline solution 0.9%

Anesthetics: xylazine and ketamine

Antiseptic: povidone-iodine 10%.

Histology process

Paraformaldehyde (PFA) or 10% buffered formaldehyde

Phosphate buffered solution (PBS)

Ethylenediaminetetraacetic acid (EDTA)

Alcohol in increasing concentrations (70%, 80%, 90%, 100%)

Xylene

Paraffin

Harris' haematoxylin solution

Distilled water

Eosin solution

Glass slides

Glass coverslips

Slide mounting medium (Entellan ®)

Rotary microtome (Leica RM2255, Leica Biosystems, Germany)

Floating bath for histology

Light microscope (Olympus BX51, Olympus Corporation, Japan) equipped with a color CMOS wireless digital camera for digital image acquisition.

Computer with image analysis software (https://imagej.net/ij/).

X-ray microtomography and image analysis

SkyScan 1174 (® Bruker micro-CT—Belgium)

Image Analysis software (Skyscan™ CT- analyser software, NRecon, CTVox, CTan)

XRM-Versa 500 (Zeiss, Oberkochen, Germany/XRdadia)

Fiji/ImageJ (Eliceiri/LOCI group, Madison; the Jug group, Milan; Tomancak lab, Dresden) open source software for image analysis, processing and visualization (Schindelin et al., 2012)

Avizo 3D 2023.2 software (Thermo Fisher Scientific, Waltham, MA, USA): 2D and 3D visualization, VOI (volume of interest) selection and crop.

Methods

Ethical approval

The procedures involving animals were conducted according to the guidelines set forth by the National Council for the Control of Animal Experimentation (CONCEA). This study received approval from the Animal Experimentation Ethics Committee of UFSJ under protocol numbers 9802240321 and 6822081122. The animals had unrestricted access to food (Nuvilab CR-1, Nuvital Nutrientes S/A, São Paulo, Brazil) and water at all times.

Surgical method

Bone defect in mice's skull

To perform the calvaria defect, 8-week-old male Swiss mice were obtained from the animal breeding center at the Federal University of São João del-Rei (UFSJ) in Brazil.

The animals were anesthetized intraperitoneally using a combination of xylazine (12.5 mg/kg) and ketamine (12.5 mg/kg). The area of interest was then shaved and disinfected with 10% povidone-iodine solution (Figures 1A–F). A longitudinal incision approximately 1.5 cm long was made with a sterile scalpel, extending from just in front of the eyes to the midline of the skull (Figure 1C). The skin was held back manually with the thumb and index finger, allowing access to the calvaria. Two circular bone defects, each about 2.3 mm in diameter, were created on the right and left margins of the parietal bone using a trephine drill (PM 699, Jet ® Carbide Burs, Brazil) connected to a micromotor (MicroMotor 210/105L, Strong ®, Korea; Figure 1D). Continuous irrigation with sterile saline solution was used throughout the procedure to prevent overheating of the bone.

Tibial bone defect

Defects in long bones necessitate the use of larger animals, as the thickness of the lower limbs (tibia and femur) makes it challenging to create a 2 mm defect without a high risk of fracture. Therefore, we chose to use 12-week-old male Wistar rats for this study.

A bone defect with a diameter of 2 mm was created in the middle third of the proximal epiphysis of the right tibia. This procedure utilized a low-speed micromotor (MicroMotor 210/105L, Strong ®) coupled with a 0.2 mm diameter cylindrical cutting burr (PM 699, Jet ® Carbide Burs). The animals were anesthetized intraperitoneally using a combination of xylazine (12.5 mg/kg) and ketamine (0.75 mg/kg). Following this, trichotomy and aseptic preparation of the surgical area were performed using 10% povidone-iodine.

For the procedure, the animals were placed in a dorsal decubitus (supine) position, with the right hind limb in external hip rotation and triple flexion (Figures 1G–N). A longitudinal skin incision approximately 1 cm long was made, and the skin was retracted using retractors to expose the bone tissue. The osteotomy was carried out using a 2.0 mm diameter trephine drill, reaching a depth of about 2.0 mm, while continuously irrigating with sterile 0.9% sodium chloride solution to prevent local temperature elevation and minimize thermal damage to the bone tissue. After completing the procedure, the skin was sutured, and the area was cleaned again with antiseptics, followed by final antisepsis using 10% povidone-iodine (Costa et al., 2023).

Periods of analysis

Bone repair in long bones is evaluated at various intervals after surgery. Assessments occur between 3 and 7 days to analyze the inflammatory phase, at 14 days to evaluate callus formation, and approximately between 28 and 30 days to examine the late phase of repair and remodeling. For skull defects, samples are collected at later time points. In this protocol, bone repair in skull defects is assessed at 15, 28, and 56 days following surgery (Figure 2).

Bone samples

To obtain the bones used in this study, rats were subjected to a 2 mm defect in the tibial bone and were subsequently euthanized within 30 days. Similarly, mice underwent a 2.3 mm defect in the skull and were euthanized after 28 days. After euthanasia, the injured bones were collected and fixed in 10% formalin.

micro-CT: acquisition and Analysis

The following sections present the various stages of the micro-computed tomography procedure, accompanied by illustrative diagrams and representative images obtained from flat and long bones. First, the image acquisition process is described in detail, followed by the image processing and three-dimensional reconstruction stages, performed using specialized software. Finally, the main morphometric parameters evaluated are presented, highlighting their contribution to the quantitative analysis of bone structure.

Image acquisition

Two micro-CT scanners were employed for this work: a SkyScan 1174 X-ray microtomography (Bruker MicroCT, Belgium) and an XRM-Versa 500 (Zeiss/XRadia). The tibia and calvaria samples were positioned within the scanners in specific supports to ensure stability during imaging procedures. The acquisition parameters were carefully adjusted to maximize projection quality, resolution, and contrast, enabling the accurate detection of neo bone tissue.

Image acquisition settings

XRM-Versa 500: the acquisition settings for the tibia sample imaging were 30 kV and 2 W applied on the X-ray source tube, 300 ms exposure time per projection, 360° rotation with angular increments of 0.5°. The CCD camera was binned 2 × 2 with 1,024 × 1,024 pixels. The resulting spatial resolution of the reconstructed image is 7 μm/voxel.

Bruker SkyScan 1174: the calvaria sample was imaged with 50 kV and 800 μA applied on the X-ray source tube, the projection is an average of three frames with a 2,000 ms exposure time each, 360° rotation with angular increments of 0.7°. The resulting spatial resolution of the reconstructed image is 10.11 μm/voxel with a 1,304 × 1,024 × 14 bit CCD camera coupled to a P43 scintillator screen with automatic lenses and no filters.

Comprehensive step-by-step guide to micro-CT software

XRM-Versa 500 Acquisitions: the sample must be placed on the stage of the micro-CT, ensuring good fixation during acquisition. As can be seen in Figure 3, a polymer tube (part of an Eppendorf) was used as support.

A critical requirement of the micro-CT imaging process is that the sample remains stationary relative to the holder during rotation. Any movement shifts the rotational axis, leading to motion artifacts and inaccurate reconstructions. It is not unusual for biological samples to have some moisture content. Wet samples can be a problem during micro-CT acquisitions. The moisture content is highly sensitive to the X-rays, absorbing their energy. This can alter fluid properties, causing dehydration, resulting in either sample displacement or in shrinkage. While this effect is more common in soft tissues like cartilage, it can also occur in newly formed bones, and in bones preserved in fluids such as alcohol, distilled water, or formaldehyde. The preservation fluids reach its inner structure and can react with X-Rays during the imaging process. It might result in a blurred image (du Plessis et al., 2017).

To avoid these effects, the sample can be placed in the holder surrounded by a soft foam saturated with a fluid, such as water or the same solution used to preserve the sample. The fluid in the foam replaces any moisture the sample loses to evaporation and also absorbs some of the X-ray energy. This helps to mitigate or even eliminate the negative effects mentioned above.

Settings: properly setting the acquisition parameters is crucial for accurate analysis of micro-CT images of flat and long bones. Each micro-CT system is unique, with specific features like filters, X-ray source power, and X-ray beam energy. These differences mean that the optimal settings for a scan may vary from one type of micro-CT to another.

Five main parameters must be set before acquisition, which are alignment, spatial resolution, beam hardening filter, X-ray source energy, and projection exposure time.

Once a sample is placed on the micro-CT stage, it must be perfectly aligned. This alignment is crucial because the sample needs to rotate around its center, which must be precisely located at the midpoint between the X-ray source and the detector. Proper alignment prevents shifting during the scan. After aligning the sample, the spatial resolution is determined by how the sample appears in the projection image. The XRM-Versa 500 utilizes two types of magnification to achieve this: optical and geometrical.

Optical magnification: this is controlled by one of the five available objective lenses: 0.4×, 4×, 10×, 20×, and 40×. The choice of lens depends on the sample's size and the required spatial resolution.

Geometrical magnification: this involves moving the X-ray source and/or the CCD camera backwards and forward to achieve the desired frame. Because the X-ray beam is conical, these movements alter the size of the projection on the detector, effectively changing the magnification.

Spatial resolution refers to an imaging system's capability to differentiate between two closely spaced objects or features in an image. It can be calculated based on data determined by the system and the image, however, in this work, to simplify the understanding, it is defined as the size of each voxel/pixel in the 3D/2D image.

X-ray tubes are polychromatic, meaning they produce radiation across a range of energies. For example, if the system is set to 160 kV, it generates a spectrum of photon energies ranging from 0 keV up to a maximum of around 160 keV. When this beam passes through a sample, the lower-energy photons are more likely to be absorbed at the edges. Since the detector simply counts the total number of photons and does not differentiate between energy levels, it interprets this absorption as a layer of denser material at the sample's border. This results in the beam hardening artifact, an irregularity in the image's gray levels that makes the edges of the sample appear brighter than they actually are (Brooks and Di Chiro, 1976).

To mitigate beam hardening, a filter is placed directly at the X-ray source window. The XRM-Versa 500 system has a set of 12 filters made from a special Zeiss-manufactured glass, each one designed for a specific energy range between 30 and 160 kV. For optimal results, the transmission rate (the ratio of the original beam to the attenuated beam after passing through the sample and filter) should be between 22 and 36%. To adjust this rate, the voltage must be increased or decreased.

For a high-quality image with a good signal-to-noise ratio, an average of 5,000 photons should reach the CCD camera (photons that passed through the sample + filter). To adjust this number, one should also adjust the exposure time, which is the duration for which a single projection is acquired. A longer exposure time allows more photons to reach the detector.

Other parameters such as rotation, step angle, and camera binning are also important for a micro-CT acquisition. The rotation of the sample affects the quality of the final image. A full 360° rotation increases the signal-to-noise ratio and improves the overall image quality. A 180° rotation is the minimum required for an acquisition. The angular step, the degree of rotation between each captured projection, is determined by the desired spatial resolution and the size and complexity of the sample. A smaller angular step provides more projections, which is necessary to capture fine details and achieve a higher resolution. Camera binning is a setting that improves the signal-to-noise ratio by combining data from multiple adjacent pixels into a single, larger pixel. It results in less noisy images with lower spatial resolution. No binning images increases the size of the image file and the amount of data to be processed. More details about 180° and 360 rotation reconstruction parameters can be found in Sasov (2002).

A significant limitation of micro-CT is the inverse relationship between sample size and spatial resolution. Achieving high spatial resolution typically requires small samples; for example, a resolution of 1 μm/voxel is generally only attainable for samples approximately 1 mm in diameter. Furthermore, the Field of View (FOV) is constrained by the detector size; if the bone model is too large, the system must be equipped with a detector capable of covering the entire volume.

Image reconstruction: each commercial micro-CT scanner has its own software for both image acquisition and reconstruction. The acquisition step, as described previously, involves the system capturing numerous 2D projections, or radiograms, of the sample as it rotates. The real meaning of the micro-CT technique lies in the reconstruction step. This process takes all those 2D projections and uses a mathematical algorithm to turn them into a single, detailed 3D image of the sample. The most common algorithm used for this purpose is the Filtered Back projection (FBP). The 3D image can be digitally sliced into 2D reconstructed images of the sample.

Reconstruction software, in addition to applying the reconstruction algorithm, also provides tools to correct common micro-CT artifacts, such as; beam hardening artifact, as previously discussed it is primarily mitigated using physical filters. However, for fine-tuning, a digital correction is also necessary to fully remove the artifact; and ring artifacts, that are generated by malfunctioning pixels in the CCD camera, often called “dead pixels,” which stop working correctly. As the sample rotates during the scan, the failure of a single pixel is projected as a perfectly round artifact in the reconstructed 2D image slices.

Zeiss equipment features a function called Dynamic Ring Removal (DRR). This feature is applied during the acquisition process and works by randomly shifting the sample a few pixels off-center for each projection. By doing this, any faulty pixels in the detector do not form a distinct ring artifact in the final reconstructed image. The system automatically corrects these random shifts before the reconstruction process begins, ensuring a high-quality, artifact-free image. The main parameters to set up at the Zeiss reconstruction software are center shift and beam hardening correction.

Even if a sample is properly positioned on the micro-CT stage, as mentioned in the previous session, slight shifts from its rotational center can still occur. These movements can be caused by minor temperature fluctuations that happen when the X-ray tube is active or when the micro-CT chamber is open for sample replacement. Even minimal changes in temperature is enough to shift the sample by a few micrometers. The reconstruction software corrects this issue by adjusting the virtual rotation center of the sample image. The software shifts this center within a specific range of pixels until a perfectly aligned image is achieved. This process can be performed either automatically by the system or with assistance from the operator.

The software provides three different beam hardening correctors for the operator to choose from. Detailed information about the optical/geometrical magnification system of the Zeiss scanners can be found in the manuals; ZEISS Xradia 520 Versa-Submicron X-ray Imaging: Extending the Limits of Your Exploration; and 40×-Prime Objective from ZEISS-Enhance Resolution and Image Quality on ZEISS VersaXRM 730, accessible via Zeiss webpage (https://www.zeiss.com/microscopy/en/home.html).

Each corrector can be applied with a different intensity level, which is also selected by the operator. The operator can then compare the results of various correctors and intensity settings directly within the software to find the optimal combination. Ultimately, achieving a high-quality image is highly dependent on the operator's experience with the software and their ability to choose the right parameters.

Image analysis: among the software packages utilized in this work, only Fiji/ImageJ is open-source and freely accessible to anyone. Although each program has its own particularities, most of the processing workflows described in this paper can be replicated using it. Image analysis involves determining structural parameters of a sample based on quantitative data extracted directly from its digital image. For this process to be valid, we assume that the image is a faithful digital representation of the actual sample microstructure. While challenges regarding sufficient spatial resolution must typically be addressed, particularly when analyzing complex microstructures, we consider the spatial resolution used in this work to be adequate for this specific type of sample.

A significant limitation in image analysis stems from computational constraints. Micro-CT 3D datasets are substantially larger than conventional 2D microscopy images, as they typically consist of stacks exceeding 1,000 slices. Consequently, implementing this methodology requires a high-performance workstation with sufficient processing power and memory.

The entire process chain for image analysis typically consists of three main steps; -preprocessing: preparing the image to make it suitable for segmentation; -segmentation: identification of the specific phases or structures that need to be analyzed; and; -quantification: determining the relevant structural parameters and metrics based on the segmented regions.

Following, these steps are described for the analysis of the tibia sample, performed using the ImageJ/Fiji software tools; this process also applies to calvaria samples.

Image preprocessing: ImageJ/Fiji offers a variety of filters to preprocess micro-CT images, and the operator must select the filter that provides the most suitable result for their dataset. An essential practice is applying the filter to all 2D slices at once. The images should be imported into the software using the following steps; = >File = >Import = >Image Sequence… Once imported, the available filters can be found under; = >Process = >Filters. Each filter has an adjustable intensity level that must be set by the operator. The improper or excessive application of filters can have a negative effect on the image, caution is required.

Given the high quality of the original tibia images, no smoothing filter was necessary. However, a brightness adjustment was performed via; = >Image = >Adjust = >Brightness/Contras. In Figure 4A the brightness adjustment application can be observed. In addition to eliminating low gray-level pixels, this step homogenizes the image and effectively highlights the object of study.

The ROI was selected before the segmentation. The delineation of the ROI can be executed even prior to preprocessing. This makes the dataset smaller, accelerating the subsequent processes.

Region of interest: the key point for a robust image analysis lies in the proper delimitation of the Region of Interest (ROI) in the sample's image. ROI is the portion of an image that contains the information most relevant to the analysis. Usually, the ROI is determined as the largest cuboid that can be extracted from the 3D image; however, it is essential that this volume remains representative of the sample.

ROI selection is a critical step in the image characterization pipeline. To ensure accuracy, the analysis must encompass only the defect area designated for new bone growth. In this case study, because the defect was a drilled circular hole, the ROI was defined as a circular area matching the diameter of the original injury. This ensures that only newly formed bone is assessed; including any area outside this diameter would inadvertently measure pre-existing bone tissue.

When the analysis focuses on specific irregular parts of a sample, such as the trabecular or cortical bone regions, an irregular and more inclusive ROI should be used. This ensures that the measured volume is more representative and includes a sufficient amount of the sample's complex microarchitecture.

For ROI setting, ImageJ/Fiji provides several tools that improve the overall visualization for the operator. In addition to scrolling through the 2D slices directly at the bottom of the image window, the three orthogonal planes (Figure 4B) and the 3D image rendering (Figure 4C) are also accessible. For 3D image access; = >Plugins = >3D Viewer, and for the views; = >Image = > Stacks = >Orthogonal Views.

If the operator opts for working with another orthogonal plane as shown in Figure 4D, the image data set can be resliced via; = >Image = >Stacks = >Re-slice [l]. Figure 4E shows the result of a re-slice. Besides that, for additional adjustment of the image, it can be flipped, rotated or moved, for that use; = >Image = >Transform, and choose an option.

The ImageJ/Fiji's contains some specific tools for rectangles, ellipse, polygon, and freehand selection. They can be found on the toolbar, at the left-hand low level of the software interface. If the operator needs to change the ROI throughout the slices to adapt it to an irregular form, many of the mentioned section tools can be used in different slices. Each selection can be saved in: Analyse = >Tools = >ROI Manager, it is just to add each selection; = >Add [t], and then use = >More = >Interpolate ROIs. All the selected ROIs will be interpolated to generate one VOI, or Volume of interest.

Since this analysis focuses on the neo-bone grown within a circular defect drilled in a tibia, the ROI was selected using the ellipse selector tool (as can be seen in Figure 4E). Figure 4 illustrates the process of defining the ROI. in a long bone sample. For this specific figure Avizo software was used for better visualization. The image on the left shows the 3D reconstruction of the long bone, followed by its version displaying the delineated Region of Interest (ROI). The highlighted orange portion is the ROI itself; it represents the bone volume in which quantitative analyses are performed, focusing specifically on the area related to bone growth. Figures 4C–E show the ROI selection and segmentation processes.

Some procedures in ImageJ/Fiji cannot be undone. A key tip is to always duplicate images before applying any modifications. To do this, simply use a right mouse button click followed by = >Duplicate. After the selection of the ROI being established, the rest of the image can be erased via; = >Edit = >Clear Outside. To manage the size of the resulting dataset, duplicate the image, but restrict this operation solely to the range of slices that contain the primary ROI for analysis.

Segmentation: segmentation is a key step in the image analysis workflow. This process generates the object of analysis, the digital representation of the sample. ImageJ/Fiji offers many tools to convert original grayscale images (and colored ones) into a segmented result, ranging from automatic conversion to operator-assisted methods and the application of automated segmentation algorithms. The final choice of the threshold that best fits the analyzed image ultimately depends on the user, whether they rely on algorithmic guidance or personal judgment. While algorithms serve as valuable guides during the segmentation process, any fine-tuning remains the responsibility of the operator (Moreira et al., 2012).

In this work, the image was segmented via a thresholding method assisted by the operator, which consists of choosing a threshold in the image's gray level histogram to convert it into an image with just two levels, white (bone phase) and black (voids or pore phase). The result can be seen in Figures 4D, E. For this option access; = >Image = >Adjust = >Threshold.

Automatic conversion is applicable only when the image clearly displays two distinct phases. For automatic conversion; = >Process = >Binary = >Make Binary.

The function of automated segmentation algorithms varies based on the image's characteristics. These algorithms generally examine the grayscale histogram data to find an optimal threshold. An operator can preview the results of multiple algorithms to ensure they choose the most effective option for their segmentation task. For this access; = >Image = >Adjust = >Auto Threshold. In the Auto Threshold window, the operator can find the options in Method space. To test them out, select Try All. Figure 5 shows the result for the image in Figure 4E. If the algorithm is not applicable, the image remains the same. There are two options to apply the algorithms, either; = >Image = >Adjust = >Threshold, then replace Default by the chosen algorithm, and; = >Image = >Adjust = >Auto Threshold, then define the chosen one.

Characterization: for the long bone with ImageJ/Fiji BV/TV, Tb.n, Tb.Th, and Tb.Sp were determined, which according to Bouxsein et al. (2010), form the minimum set recommended for trabecular bone morphological characterization. The bone volume fraction (BV/TV) is a study of bone morphometry, particularly used to analyze trabecular bone. It is the ratio of the volume occupied by the solid bone structure (trabeculae) to the total volume of the tissue sample. Trabecular number (Tb.n) is used to quantify the density of the trabecular bone. It is the number of trabeculae per unit length within a given volume of tissue. All these parameters, including the thickness of the trabeculae (Tb.Th) and the space between them (Tb.Sp) are measured via 3D image analysis.

Since the segmented image (Figures 4D, E) is irregular (not a cuboid), it requires prior processing before analyzing Tb.Sp parameter. Since the algorithm analyzes the space between the trabeculae (the black phase; Figure 5A), this region must be delineated, separating the spaces within the sample from the surrounding void. Although there's no tool for this process, this problem can be solved by applying a specific filter combined with mathematical operations. The Top-Hat filter can be used to digitally delineate (or wrap) the irregular image. This filter performs a pair of non-linear morphological operations to extract small elements and details based on their brightness relative to the surrounding area. First, duplicate the segmented image, then access; = >Process = >Filters = >Top Hat, then choose the radius of pixels that will fill the gaps (Figure 5C), and select both options including Preview (Figure 5D). If the gaps are not filled yet, increase the radius (If the chosen radius is too large, the operator can enlarge the image via; = >Image = >Adjust = >Canvas Size. Opt by left the original image in the Center). To finalize, access; =Process = >Image Calculator… and subtract (Figure 5E) the Segmented original image (Figure 5B) from that one resulting from Top Hat filter application (Figure 5D). The resulting image is the inverse (Figure 5F) of the original (Figure 5B), however, segmented from the rest of the void. Both processes can be applied to the whole data set.

Bone volume fraction: typically, BV/TV is reported as a single average value representing the entire region of interest (ROI), which is expected to be locally representative of the analyzed bone tissue. This work, however, presents a new methodology to evaluate the BV/TV histogram as a function of the depth within the established ROI, offering a more detailed analysis than a simple average. The analysis of the BV/TV distribution (histogram) enables a quantitative assessment of the spatial homogeneity of the newly formed bone tissue. The following process must be applied to the original segmented image (Figure 5B) and its inverse (Figure 5F); = >Analyze = >Analyze Particles… and just select Summarize. The analysis yields a table containing several determinations, including the slice number and its corresponding %Area. This %Area represents the volume fraction for each slice: it is the percentage of trabecular bone in the segmented image, and the percentage of voids when analyzing the inverse image.

The following Equation 1 enables one to determine BV/TV for each slice:

The Figure 6A shows the BV/TV distribution for the analyzed ROI.

Trabecular thickness and trabecular space: before performing any length measurements, the image must be calibrated to set the proper scale. The resolution of the long bone image is 7 μm/voxel, which means that every voxel (or pixel) has edges 7 μm long. Access; = >Analyze = >Set Scale… Set the Distance in pixels to 1, the Known distance to the resolution of the image (seven in this case), and the Unit of length to “um,” the system will understand it as “micrometer”. With the scale set, a scale bar can be inserted in the image; = >Analyze = >Tools = >Scale bar (Figure 6A).

Measurements of trabecular distances (Tb.Th and Tb.Sp) are based on the size of inscribed spheres that can fit within the trabeculae and the spaces between them. This three-dimensional measurement method not only yields the distance values but also provides a colored 3D map of the sphere sizes. To perform this measurement, the BoneJ plugin must be installed; = >Help = >Update… in the Update Manager window, set BoneJ in the Manage Update Sites; = > Apply and Close = > Apply Changes, and restart ImageJ/Fiji.

For regular shapes like cuboids, the BoneJ software calculates both Tb.Th and Tb.Sp simultaneously; however, the process differs for irregular regions of interest. For irregular ones access; = >Plugins = >BoneJ = >Thickness and set just Trabecular thickness + Show thickness maps and apply this for both data sets of images, original segmented and its inverse (Figure 6E). In this case, the inverse image application will result in Tb.Sp, whilst for the original segmented the Tb.Th.

Processing and reconstruction of defect in flat bone

The images obtained from the micro-CT initially consist of two-dimensional projections created during the sample's rotation throughout the acquisition process. However, these projections cannot be directly interpreted for structural analysis, necessitating several processing steps. The first step involves using NRecon software (a Bruker-microCT program version 1.7.1.0), which converts the two-dimensional projections into stacked slices, forming a complete three-dimensional volume of the sample. At this stage, adjustments to the histogram can be made, along with artifact corrections and the removal of imperfections that arise from the acquisition process. This ensures enhanced fidelity in the reconstruction.

Next, the generated volume is explored using CTvox software (version 3.3.1), which is designed for three-dimensional visualization. This allows for volumetric rendering, rotation on different axes, and qualitative inspection of the sample's morphology. This step is crucial for assessing the quality of the reconstruction and identifying any inconsistencies before proceeding with quantitative analysis. Finally, the images are processed in CTan software (a Bruker-microCT program version 1.18.8.0), specifically designed for morphometric and quantitative analysis of the reconstructed volume. During this step, segmentation and extraction procedures are applied to structural parameters such as volume, porosity, thickness, and mineral density, providing robust quantitative data for characterizing the tissue or biomaterial under study.

The methodological flow illustrated in Figure 7A outlines the necessary steps to transform the raw images generated by the micro-CT into three-dimensional and quantitative information. This integrated process, encompassing projection, reconstruction, visualization, and analysis, ensures that the data obtained are converted into reliable and scientifically relevant information, enabling detailed structural analyses of the samples.

Reconstruction with NRecon software

NRecon software (version 1.7.1.0) can reconstruct any bone morphology. Here, we demonstrate the process using a flat bone model as an example. To begin the process, two-dimensional projection images need to be reconstructed into three-dimensional volumes using NRecon software. This software converts raw data into a series of cross-sectional images for further analysis.

To open the dataset, navigate to the File menu and select “Open dataset.” Choose the folder that contains the projection images in TIFF format (Dimensions: 2,452 × 1,640). When you open any of the hundreds of obtained projections, the software will automatically recognize the sequence of projections. Next, define the reconstruction limits. The main screen will display a sample projection with three boundary lines: two red lines that indicate the upper and lower limits of the reconstruction region, and one green line serving as a central reference. The reconstruction will be processed exclusively in the area between the red lines (Figure 7B). It is crucial to select this region precisely, as the extent of the reconstructed area directly affects both the volume of data generated and the processing time. To enhance the efficiency of the procedure, it is recommended to clearly delimit the region of interest. This selection can also be made using the Top and Bottom commands on the panel. The green line between the red lines will indicate a slice for preview. Finally, click Preview to generate a preliminary reconstruction of an intermediate slice.

To adjust reconstruction artifacts, navigate to the Settings tab. Here, select “Smoothing” to apply slight smoothing, which will help reduce high-frequency noise. Adjust the smoothing slider as needed for your image. Use the “Misalignment Compensation” feature to correct any potential misalignments of the rotation axis. The “Ring Artifact Correction” option can be employed to eliminate characteristic circular artifacts; adjust the intensity until these rings disappear without compromising the bone texture. Additionally, the “Beam Hardening Correction” helps reduce halo-shaped artifacts around the edges of the sample. In the Output tab, adjust the reconstruction intensity histogram bar to remove background or oversaturated areas, which will enhance the contrast of the bone tissue (Figure 7B). Make sure to adjust the parameters gradually, observing how each change affects the image. This way, you can minimize artifacts while preserving important structural details.

In the Output panel, specify the folder where you want to save the reconstruction files and select your preferred file format. The 16-bit TIFF format is recommended, as it preserves the full dynamic range of intensity and is compatible with quantitative analysis software like CTAn and ImageJ. After returning to the Start menu, click on “Add to Batch.” Once you have reviewed all your settings, click “Start Batch” to generate the file containing all the slices. A loading screen will appear shortly afterward (Figure 7B).

3D reconstruction in CTvox

The software CTvox (version 3.3.1) is designed for volumetric rendering of samples, allowing users to transform a set of reconstructed tomographic slices into realistic 3D objects. Its interface supports intuitive navigation and dynamic manipulation of both objects and cameras. Users can also utilize flexible planar cutting tools to create sectional views of the microstructure. Additional features include options for custom color selection for different material phases and background adjustments, which facilitate both qualitative and quantitative analyses of internal characteristics.

Open CTvox and click File > Open. Navigate to the folder containing the images reconstructed by NRecon and select the first image (Dimensions: 2,452 × 2,452). After loading the volume, the 3D view will appear automatically (Figure 8A).

After loading the reconstructed volume, the software displays an intensity histogram (gray values) in the side menu. Often, low-intensity contributions, such as background noise, dominate the distribution, obscuring relevant structures (Figure 8B). To minimize this artifact, follow these steps to apply a selective opacity adjustment directly to the histogram:

a) Select the opacity channel: activate the Opacity channel from the top menu of the histogram. Initially, a pink linear curve, representing uniform opacity, will be displayed.

b) Suppress low-intensity noise: add a control node in the lower left region of the histogram (where intensity values are close to zero), and adjust the curve to assign an opacity of approximately 0 (total transparency) to these voxels. This step visually eliminates peripheral noise without altering the original data.

c) Optimize the opacity curve: smoothly modify the curve to ensure a gradual transition between transparent (noise) and opaque (sample) regions, refining the sharpness of the rendered volume.

d) Adjust color and lighting: use the Red, Green, Blue, and Light (L) channels to assign specific colors to density ranges and simulate lighting effects, enhancing microstructural contrasts. Test predefined profiles (Load) to ensure reproducibility.

e) Control reversibility: use Reset active curve (per channel) or Reset all curves (global) to reset adjustments, ensuring methodological flexibility.

Use the rotation and zoom tools to find the viewing angles that best show the defect and the progression of bone repair. When the image is as you want it, click on the camera icon (Capture image) or go to File > Save Image to save your 3D image.

Image analysis in CT-analyser (CTan)

The CT-analyser CTan software (version 1.18.8.0) from Bruker is widely utilized for the quantitative analysis of bone structures using micro-CT. It enables accurate characterization of critical morphological parameters such as porosity, mineral density, trabecular thickness, trabecular separation, and 3D connectivity. With advanced segmentation and image processing tools, CTan allows for the differentiation between cortical and trabecular bone, assessment of microdamage, and quantification of structural changes in studies related to bone diseases, treatment efficacy, and biomaterials. This software provides robust and reproducible data for preclinical research and tissue bioengineering. For the microstructural analysis of bone samples in CTan software, users should follow the sequence of operations indicated by icons on the software's home page:

Open project (“Folder” Icon): import the Reconstructed Volume. The set of 3D images in TIFF 16-bit format must be loaded into the software, ensuring that previous artifact corrections (e.g., ring correction and beam hardening) have been applied.

Raw images preview and slice selection: after loading the reconstructed volume, the interface displays a numbered cross-sectional view (slices) of the material, along with a side window that previews the volume. To define the 3D region of interest (VOI), follow these steps:

a) Navigate to the desired initial slice using the side scroll bar (indicated by a red line over the image) or directly select the corresponding number from the list of slices (Figure 9A);

b) Right-click on the initial slice and select “Set as Top of Selection” to exclude all upper slices;

c) Repeat the procedure on the final slice by selecting “Set as Bottom of Selection” to exclude lower slices.

This selection is crucial for focusing the analysis on the anatomically relevant region (e.g., trabecular bone), thereby excluding non-analytical structures like cartilage or edge artifacts that could distort morphometric parameters, such as overestimating porosity. Accurate delimitation ensures that the results reflect only the microstructure of interest.

It is important to differentiate between the terms: Region of Interest (ROI) refers to an area selected in a single cross-sectional image for 2D analysis, while Volume of Interest (VOI) encompasses a contiguous set of ROIs across multiple images, forming a 3D volume for three-dimensional morphometric analysis. Therefore, accurately delimiting these areas is vital to ensure data accuracy and relevance. During 3D volumetric rendering, an interpolation algorithm acts as a form of antialiasing to generate intermediate voxels between selected slices, ensuring geometric continuity and smoothing of the microstructure to minimize discretization artifacts and preserve the morphological integrity of trabecular or porous structures. Although an automatic interpolation tool is available, it is not recommended for abrupt transitions in density or geometry.

3. Saving the ROI: after delineating the ROI, click on the floppy disk icon (Save Regions of Interest) to store the selection made. This step is essential for ensuring that the configuration can be reloaded and replicated in subsequent stages of data processing and analysis. If the ROI is not saved, it cannot be recovered, requiring it to be manually redefined.

4. Region of interest (ROI) selection: delimit the ROI to exclude edge artifacts and focus the analysis on anatomically representative areas (e.g., metaphysis of long bones). When you begin selecting the ROI, the software automatically selects the entire lower area of the slice, highlighting it in red (Figure 9B). For manual delineation, use the Lasso tool (indicated by the lasso icon or a free cursor), activated with the left mouse button, allowing you to draw irregular contours that correspond to the sample anatomy (Figure 9C). In homogeneous or symmetrical samples, a single slice selection can be interpolated across the entire volume using geometric extrapolation algorithms, which automatically replicate the ROI in subsequent slices. For irregular structures (e.g., bones, fractures, porous interfaces), the ROI should be defined slice by slice or every few slices, manually adjusting for geometry variation. This approach ensures that non-representative regions (e.g., cartilage, noise) are excluded, thus maintaining the integrity of metrics such as porosity or connectivity. If the ROI selection is incorrect, use the “Scissors” icon (or Clear Selection) to remove inappropriate selections. Calculations are limited to the selected region or volume of interest.

5. Binary selection preview: this step provides a real-time preview of how the threshold applied to the image affects the binarization process. It is a vital tool for making precise adjustments and validating parameters. Users can instantly observe how voxels are classified as solid phase (represented in white) or background/pores (represented in black) based on the threshold set in the histogram. This allows for the identification of artifacts or residual noise captured during segmentation (Figure 9D). Additionally, this tool helps validate the structural continuity of interconnected elements, such as bone trabecular structures or porous networks, and it can detect issues related to over- or under-segmentation, including the loss of fine details or the unintended inclusion of noise. Once the binarization parameters are established, the minimum and maximum threshold values must be meticulously documented to ensure standardization across all study samples. These parameters should be consistently replicated when setting up the analysis Task List to guarantee comparability and reproducibility of results.

6. After that, it is important to check that the parameters for bone material are selected. To do this, go to File → Preferences → Nomenclature and change the setting from “General” to “Bone.” This adjustment automatically sets the morphometric parameters for specific bone tissue metrics. In addition, in the Histogram tab, enter the linear attenuation values (Hounsfield units – HU) obtained from calibration phantoms. A calibration phantom is a standardized reference object containing specific, known concentrations of calcium hydroxyapatite (HAp). Because micro-CT images represent X-ray attenuation as varying levels of gray, quantifying these gray levels allows researchers to measure the degree of tissue mineralization. To achieve this, it is crucial that the phantoms are scanned either simultaneously with the bone samples or independently under the exact same acquisition parameters (e.g., voltage, current, exposure time, and filter). For studies with rodents, HAp phantoms with known mineral densities (e.g., 250, 500, 750 mg/cm 3) should be used (Figure 9E). By plotting the known HAp densities of the phantom against their measured average gray levels, the software establishes a linear calibration curve. Establish a calibration curve that relates the gray values of the images to the actual bone mineral density. This step is essential for converting the arbitrary gray values of the sample images into biologically relevant quantitative data (e.g., bone mineral density – BMD), enabling accurate assessment of the hard callus maturation and the quality of the newly formed bone.

7. Custom processing preview: this step is essential as it allows real-time visualization of the effects of custom processing (e.g., morphological operations, smoothing filters, artifact removal) applied to binary images before final execution. The following tabs will be available for this purpose:

a) Collection: selection and organization of samples to be processed in batches (Figure 9F).

b) Task list: definition of processing operations and metrics to be calculated (e.g., porosity, trabecular thickness). This list is customizable and lists the sequential actions that will be performed automatically during the analysis (Figure 9F).

c) Internal: configuration of internal software parameters (e.g., segmentation algorithms). External: Integration with external tools or scripts for complementary analysis (Figure 9F).

d) Output: defining the format and destination of the results (e.g., tables in.CSV format, binary images; Figure 9F).

After configuring the Task List and processing parameters, the next step is to consolidate multiple scans and their corresponding ROIs to perform the analysis in an automated and standardized manner. Here is the step-by-step guide:

8. Ensure that all scan files (e.g., 16-bit TIFF) and their saved ROI files (e.g.,.roi format) are in specific directories and have consistent naming conventions (e.g., Sample01_ROI.roi).

9. Using the BATch MANager Module: above the standard tabs in the custom processing window, you will find the BATch MANager icon (represented by the Batman symbol), which is responsible for managing and executing batch processing analyses. Click on the BATch MANager icon to open the batch processing interface. In this interface, you will see the Task List selected in the previous steps immediately above (Figure 9G).

10. In the main panel, click “Add” (or Add Samples) and select the reconstructed volume file from the first sample (e.g., Sample01; Figure 9G).

11. For each sample loaded, click on the arrow next to the “Add” button and select “Load ROI.” Associate the corresponding ROI file (e.g., Sample01_ROI.roi), ensuring that the predefined region of interest is applied. Repeat the process for all samples in the study (Figure 9G).

12. After loading all samples and ROIs, click “Start” to begin the automated analysis. The software will process each sample sequentially, applying the Task List and specified ROIs.

At the conclusion of the processing, an Excel file (.xlsx) will be automatically generated. This file will include all the parameters selected in the Task List for each sample, as well as the data organized into tabs based on metrics or samples, depending on your configuration. Additionally, it will contain processing metadata, such as threshold values and the software version. This step finalizes the quantitative analysis flow, providing data that is ready for statistical analysis and graph generation.

Parameters for evaluation by micro-CT analysis

According to Bouxsein et al. (2010), the main quantitative parameters used in morphometric analysis by microcomputed tomography (micro-CT) for characterizing trabecular microarchitecture are presented in Table 1.

Histological processing

The samples were fixed in buffered Millong formalin for 48 h. They were then decalcified in EDTA (Synth ®) diluted in PBS (pH 7.4) at 200 mg/L. The tibiae were kept in the decalcifying solution for 52 days, while the calvaria were kept for only 10 days, with the solution being changed every 3 days. After that, the center of the lesion of the samples was sectioned, with the tibiae being sectioned in the transverse plane and the calvaria being sectioned in the coronal plane. After that, the material was washed in running water for 1 h. In the next step, dehydration in ethanol, paraffin embedding, and histological sectioning into 5 μm thick transverse slices were performed, followed by staining with Hematoxylin and Eosin (H&E; Figure 10).

Histological analysis parameters

Hematoxylin and eosin staining is an effective tool for micrometric visualization of tissue, enabling a better understanding of tissue composition and structure, as well as visualization of cells present in the tissue (Fischer et al., 2008). Several studies have demonstrated the efficiency of this technique for evaluating bone regeneration in different contexts, such as surgical lesions or fractures (Costa et al., 2023; Medrado et al., 2025; Li et al., 2017).

To obtain representative results, the analysis of bone sections stained with hematoxylin and eosin (H.E.) should include images at lower magnification (4 ×) (Figure 11A and 11B) to illustrate the general appearance of the lesion, as well as higher magnifications (10 × or 40 ×) (Figure 11A1–A5 and 11B1–B3) to visualize the presence of bone trabeculae in the medullary region of long bones (Figure 11A1) microscopic structures, such as cells and connective tissue (Figure 11A2–A5 and 11B1–B3). The areas chosen for zooming should be representative and provide a clear view of the structures within the lesion.

The macroscopic structure gives a comprehensive view of the lesion, highlighting aspects of mineralization depicted by the pink-stained matrix. This matrix can be observed around 14 days, which is when mineralization of the lesion's cartilage matrix begins. It may appear as a continuous layer closing the lesion or discontinuously at various points within interspersed connective tissue. Considering the kinetics of bone regeneration, a high amount of mineralized matrix in this area indicates efficient repair. The presence of mineralization in medullary regions (bone trabeculae) signifies areas of hard callus that have not yet been resorbed, while reduced amounts of bone trabeculae may suggest more effective repair.

Osteocytes are clearly visible within the mineralized matrix. Their arrangement may reflect the stage of bone maturation. In primary bone, osteocytes are more numerous and distributed in a disorganized manner, whereas in mature bone, they are arranged concentrically, indicating the formation of lamellae. Notably, variations in the color of the primary bone matrix can be observed, with a darker shade than that of mature bone, ranging from light purple to reddish pink.

Osteoblasts can be seen attached to the upper and lower ends of the mineralized matrix or on the inner surface of the medullary bone matrix. Their morphology may be cuboidal or compact in certain regions, depending on their metabolic activity.

Vascularization is crucial for the bone regeneration process. Numerous treatments have demonstrated effectiveness in enhancing bone repair by promoting vascularization. In bone sections stained with H.E., neovascularization appears as red structures within spaces in the matrix. At 40 × magnification, it is possible to observe red blood cells surrounded by endothelial cells.

In long bones, over periods of up to 14 days, cartilage is often seen as a purple-stained connective tissue matrix. Within this cartilage matrix, chondrocytes can be visualized with their characteristic morphology at various stages of proliferation. The presence of cartilage for extended periods may indicate delayed repair, whereas its absence during shorter periods may suggest improvement in the regeneration process.

In flat bones, a connective tissue matrix connecting the two ends of the lesion is observed. Repair in the calvaria is more limited compared to long bones. Mineralization initially occurs from the ends but stops within a few weeks, leaving the center of the lesion filled with fibrous connective tissue, resembling pseudoarthrosis. In lesions in the calvaria, mineralization may occur at scattered points within the connective tissue at the center of the lesion. Studies evaluating different treatments for bone repair have noted increased mineralization in the connective tissue, making this parameter an excellent indicator of improvement in repair.

The parameters evaluated can be semi-quantified using a score. A score from 1 to 5 should be assigned to each parameter of each section evaluated by optical microscopy. The data should be properly plotted in a statistical program of the researcher's choice, followed by comparative statistical analysis between different groups. The parameters to be evaluated can be customized according to the study hypothesis; however, a suggested analysis is attached below for studies of long bone and flat bone (Table 2) injuries.

Quantifying osteocyte density in the lesion region provides valuable data for analysis and can be performed simply.

At 40 × magnification, capture photos in five random fields distributed throughout the lesion region (Figure 12A).

Open the captured images in ImageJ software and select the “Multi-point tool.” With a double click, you can configure the marker parameters such as color and size (Figures 12B, C).

Click on the desired cell, and a marker will appear in the click region accompanied by a number indicating the total number of cells marked so far. The total number of cells marked in the image can be viewed in the lower region of the menu opened by double-clicking on the “Multi-point tool” (Figure 12C).

Calculate the average number of osteocytes per field. Record the numbers obtained in an organized Excel spreadsheet.

Calculate the average number of osteocytes obtained for the analyzed field area to obtain the osteocyte density/μm 2 result.

Plot the data in a statistical program of the researcher's choice and perform the necessary statistical analysis.

To demonstrate the benefits of using micro-CT and histological analysis together in assessing bone lesions, samples collected from previous studies were processed using both techniques. The results obtained from the proposed methodology were illustrated using images that facilitated a detailed visualization of the structures and allowed for cross-referencing information across different techniques. This approach enabled a more comprehensive analysis of bone lesions (Figures 13A–D).

The three-dimensional reconstructions produced by micro-CT of the tibiae (Figure 13A), and of the calvaria (Figure 13C) provided precise delimitation of the region of interest and enabled the quantification of essential microstructural parameters. These included bone volume fraction (BV/TV; Figure 13E), trabecular thickness (Tb.Th; Figure 13F), trabecular number (Tb.N; Figure 13I), and trabecular separation (Tb.Sp; Figure 13J). These indices offered an objective and volumetric assessment of the quality and organization of mineralized tissue, allowing for consistent comparisons between experimental groups.

While similar mineralized structures could also be visualized using hematoxylin and eosin staining, the clarity of these images was inferior to that obtained from micro-CT. Moreover, histology is limited to two-dimensional sections that represent the lesion, while micro-CT allows for a comprehensive three-dimensional analysis of the entire region of interest.

However, micro-CT has limitations in detecting non-mineralized tissues, resulting in the loss of structural and cellular information. In this regard, histological analyses (Figures 13B and 13D) complemented the findings by allowing the assessment of parameters such as inflammatory infiltrate (Figure 13G), new bone formation (Figure 13H), fibrous connective tissue (Figure 13K), and osteocyte density (Figure 13L), which is fundamental to understanding the regeneration process.

The evaluation of bone regeneration is a complex and sequential biological process that necessitates analytical tools capable of addressing both microstructural dimensions and cellular responses. This study employs a combined analysis of computed microtomography (micro-CT) and histology, aligning with the scientific consensus that this methodological combination represents the standard of excellence for comprehensively characterizing bone defect repair. Notably, there is an increasing frequency of studies evaluating the osseointegration of biomaterials and treatment efficacy that utilize these two techniques together, reflecting their recognized complementarity and robustness (Trejo-Iriarte et al., 2019; He et al., 2017).

This study demonstrates several potential applications for modeling bone injury using standardized models in rodents, such as well-defined circular defects created in the tibias (long bones) and calvaria (flat bones) using a surgical drill. The use of precise osteotomy, as opposed to fracture models, standardizes lesion sizes, yielding statistically more representative results. Both calvarial and tibial defect models are commonly referenced in the literature (Fan et al., 2023; Uusitalo et al., 2005). The synergy between micro-CT and histology is crucial and widely supported by existing research. Two different micro-CT scanners were utilized in this study. Although both instruments offer nearly identical spatial resolutions, they were selected based on the institutional workflows of the collaborating research groups, with each system dedicated to a specific type of bone specimen. For example, studies aimed at evaluating the osseointegration of biomaterials and implants often use this combination to correlate structural integration metrics (such as bone volume fraction (BV/TV) and density) quantified by micro-CT with cellular responses and tissue maturation observed histologically.

He et al. (2017) and Trejo-Iriarte et al. (2019) demonstrate the effectiveness of combining these analyses in evaluating bone repair. They emphasize that when assessing parameters like trabecular thickness or bone volume (via micro-CT), it is essential to balance these findings with the characterization of tissue type and maturation stage. Our results confirm that, as noted in the studies by Trejo-Iriarte et al. (2019) and He et al. (2017), the integration of computed microtomography and histology is critical for evaluating bone regeneration. This integration is pivotal because each technique provides information that the other cannot achieve independently.

Computed microtomography excels due to its non-destructive nature and its capability to offer a three-dimensional assessment of microarchitecture. It enables accurate quantification of essential morphometric parameters for characterizing bone volume and structure, including BV/TV, trabecular thickness (Tb.Th), trabecular number (Tb.n), and connectivity density (Conn.D). Standardizing these parameters, according to established guidelines (Bouxsein et al., 2010), is crucial for ensuring result comparability. Thus, computed microtomography serves as the primary tool for answering the question: how much mineralized bone has been formed?

However, as noted in studies like that of Keklikoglou et al. (2019), microtomography has limitations. Its physical principle relies on the differential absorption of X-rays, which depends on the mineral density of the sample. Consequently, samples without sufficient contrast, such as demineralized bone tissue or areas with extreme porosity lacking density, may yield compromised images. Histology functions as an indispensable biological validation tool in such cases. Although destructive, histological processing offers high-resolution analysis at both cellular and tissue levels. It allows for distinguishing between primary bone and mature bone (Trejo-Iriarte et al., 2019), which is crucial since computed microtomography primarily quantifies mineral density without assessing tissue maturation.

Moreover, histology is the only technique that can visualize and classify non-mineralized structures, including various types of connective tissue (such as fibrous tissue, often found at the center of unrepaired calvarial lesions) and vascularization. In long bone models, histology facilitates monitoring cartilaginous callus formation (soft callus) and its subsequent replacement, characterizing endochondral ossification—a process absent in flat bone repair.

Given the advantages and limitations of both techniques, one significant benefit of the methodologies outlined in this guide is the ability to integrate micro-CT and histological analyses within the same experimental workflow. Micro-CT is a non-destructive technique, allowing samples to be scanned before decalcification and histological processing. This preserves the tissue for subsequent microscopic evaluation. This approach enables a direct correlation between three-dimensional morphometric parameters and cellular and tissue-level observations from histology. Additionally, anatomical landmarks, cortical boundaries, and residual biomaterial particles identified in the reconstructed micro-CT volume can aid in locating and matching corresponding histological sections, thereby enhancing the interpretation of bone repair events.

Additionally, longitudinal in vivo micro-CT imaging can be performed during the healing period, allowing for repeated monitoring of the animals throughout the entire process. This technique provides valuable temporal information on tissue mineralization during defect closure. Following this, high-resolution ex vivo micro-CT and histological analyses can be conducted after euthanasia. This multimodal approach offers complementary structural, temporal, and biological insights, enhancing the robustness and reliability of bone regeneration studies as well as the translational relevance of the experimental results.

In conclusion, although histology and micro-CT are both image-based techniques used to analyze bone tissue, their purposes are a bit different. Histology provides a biological overview, aiming to assess cells, the extracellular matrix, and tissue physiology, validating the biological quality and maturation stage of the newly formed bone, distinguishing it from surrounding fibrous tissue (Trejo-Iriarte et al., 2019). In contrast, micro-CT focuses on the architectural structure of mineralized bone, enabling the quantification of bone volume and void spaces within the tissue (such as BV/TV, Tb.Th). A key distinction between the two imaging methods is that micro-CT is non-destructive, whereas histology requires a proper sample preparation, including the sectioning of the tissue into thin slices. However, this is not an impediment to analyzing the same specimen with both methods, as long as micro-CT is conducted first. The increasing prevalence of studies utilizing this methodological duality in the literature reinforces its necessity for a complete assessment of bone repair, considering both structural integrity and biological validation of newly formed tissue. In summary, histology and micro-CT serve as complementary techniques that enable a comprehensive evaluation of both physiological and structural properties of newly formed bone tissues.

The software packages and programming tools utilized in this work represent only a small selection of the numerous open-source and commercial options available. Although individual software packages employ unique workflows, the overall image analysis pipeline for flat and long bones presented here can be successfully replicated using alternative platforms with minor methodological adjustments.

This study presents a comprehensive methodological guide that combines microcomputed tomography (micro-CT) and histology for the standardized analysis of bone defect repair in both tibia (a long bone) and calvaria (a flat bone) models. We demonstrated that micro-CT is crucial for the three-dimensional quantification of mineralized bone microarchitecture, while histology provides valuable insights into cellular and tissue characteristics. The proposed standardization enhances reproducibility and comparability across studies, establishing this combined approach as the gold standard for robust evaluation in bone regeneration research.