Evaluating Mineral Deposition in Bone‐Regenerative Collagen‐Based Scaffolds Using High‐Resolution microCT and 3D Computational Segmentation

Abstract

Bone regeneration depends on coordinated interactions among cells, extracellular matrices, and the dynamic microenvironment within engineered constructs. Mineralized collagen‐glycosaminoglycan (MC‐GAG) scaffolds are designed to replicate key features of native bone matrix, yet their mineral accumulation and mechanical performance remain lower than those of mature bone, making precise evaluation of mineral deposition essential. High‐resolution microCT imaging provides nondestructive, three‐dimensional (3D) visualization of mineral distribution, density, and microarchitecture with substantially improved fidelity. Through dynamic, angle‐dependent calibration and scan‐specific recalibration, our microCT scanner HiCT minimizes motion‐ and temperature‐related artifacts and enables consistent, high‐resolution imaging suitable for small scaffold systems. This protocol outlines an integrated workflow combining high‐resolution microCT scanning with advanced image segmentation to assess mineralization within MC‐GAG scaffolds. The method includes scaffold fabrication, chemical cross‐linking, and human mesenchymal stem cell culture, followed by stable positioning in custom 3D‐printed racks to ensure precise imaging geometry. Reconstruction and threshold‐based segmentation in a DICOM viewer such as ORS Dragonfly allow separation of scaffold structure from higher‐density mineral phases, providing detailed visualization and quantitative analysis of mineral deposition. This reproducible framework enhances the characterization of bone‐regenerative biomaterials and supports efforts to engineer scaffolds that more closely mimic the structural and functional properties of native bone. © 2026 The Author(s). Current Protocols published by Wiley Periodicals LLC.

Basic Protocol 1: Fabrication of bone regenerative mineralized collagen scaffolds

Basic Protocol 2: High‐resolution microCT scanning and microstructural analysis of the MC‐GAG scaffolds

INTRODUCTION

Over the past decade, our group has developed and extensively characterized a nanoparticulate mineralized collagen‐glycosaminoglycan (MC‐GAG) scaffold designed to promote skeletal regeneration in the absence of exogenous progenitor cells or growth factors (Lee et al., 2015; Ren et al., 2015; Ren, Tu, et al., 2016). This biomaterial facilitates bone formation through enhanced mineral deposition and osteogenic differentiation. Despite its regenerative capacity, the in vivo efficacy of MC‐GAG scaffolds remains limited to ∼50% of the mineralization and mechanical strength observed in native bone (Ren et al., 2019; Ren, Tu, et al., 2016; Ren, Weisgerber, et al., 2016; Weisgerber et al., 2014, 2015; Zhou et al., 2017).

Micro‐computed tomography (microCT) has emerged as an advanced, nondestructive imaging modality for quantifying three‐dimensional (3D) mineral distribution, density, and microarchitecture within bone and bone‐regenerative constructs (Boerckel et al., 2014; Guldberg et al., 2003; Müller, 2009). Unlike traditional two‐dimensional (2D) imaging methods such as X‐ray radiography, microCT enables high‐resolution, volumetric visualization and quantitative analysis of bone morphology, mineralization patterns, and structural integrity in situ (Peister et al., 2011; Thimm et al., 2013). Its rapid image acquisition, minimal sample preparation, and high precision reduce experimental variability and human error while providing reproducible and quantitative assessment of mineralized tissue.

High‐resolution microCT systems (<50‐µm spatial resolution) further enhance these capabilities by offering significantly improved spatial resolution and image fidelity (Taschereau et al., 2025). Through dynamic, angle‐dependent calibration models and scan‐specific recalibration, our high‐resolution microCT system HiCT corrects for gantry motion, mechanical drift, and temperature‐dependent expansion, which are common limitations of conventional microCT (Khoury et al., 2009; Li et al., 2019; Hu et al., 2011; Zhao et al., 2015). This approach ensures consistent image quality, minimizes artifacts such as ghosting or doubling, and enables reliable detection of fine mineral features within small scaffolds.

In these protocols, we employ high‐resolution microCT imaging with enhanced spatial resolution and contrast to enable precise evaluation of mineral distribution within MC‐GAG scaffolds. This integrated imaging and its analysis approach provide a comprehensive and reproducible framework for assessing mineralization and structural organization in advanced bone‐regenerative scaffold systems.

Basic Protocol 1. FABRICATION OF BONE REGENERATIVE MINERALIZED COLLAGEN SCAFFOLDS

This protocol describes the fabrication and chemical cross‐linking of MC‐GAG scaffolds, followed by the seeding and culture of human mesenchymal stem cells (hMSCs) on the scaffolds. It further details the fixation of the constructs for HiCT to evaluate mineralization and scaffold structure.

Materials

• Type 1 collagen (isolated from bovine tendon; cat. no. C9791, Sigma‐Aldrich, St. Louis, MO)

• Chondroitin‐6‐sulfate (isolated from shark cartilage; cat. no. C1610, Spectrum Chemicals, Gardena, CA)

• Calcium nitrate hydrate (Ca(NO₃)_2·H₂O; cat. no. 202967, Sigma‐Aldrich, St. Louis, MO)

• Calcium hydroxide (Ca(OH)₂; cat. no. 239232, Sigma‐Aldrich, St. Louis, MO)

• Phosphoric acid (ACS reagent, ≥85 wt.% in H₂O; cat. no. 6950017, Sigma‐Aldrich, St. Louis, MO)

• Dulbecco's phosphate‐buffered saline (DPBS), pH 7.4, sterile (cat. no. D8537, Sigma‐Aldrich, St. Louis, MO)

• 100% ethanol (cat. no. BP28118100, Thermo Fisher Scientific, Waltham, MA)

• 1‐Ethyl‐3‐(3‐dimethylaminoproyl) carbodiimide (EDAC; cat. no. 341006, Sigma‐Aldrich, St. Louis, MO)

• N‐Hydroxysuccinimide (NHS; cat. no. 130672, Sigma Aldrich, St. Louis, MO)

• Dulbecco's modified Eagle's medium (DMEM; cat. no. 11995065, Thermo Fisher Scientific, Waltham, MA)

• Fetal bovine serum (FBS; cat. no. S11150, Bio‐Techne Corporation, Minneapolis, MN)

• l‐Glutamine (cat. no. 25030081, Thermo Fisher Scientific, Waltham, MA)

• Penicillin/streptomycin (cat. no. 15140122, Thermo Fisher Scientific, Waltham, MA)

• Human mesenchymal stem cells (hMSCs; cat. no. PT‐2501, Lonza, Inc., Allendale, NJ)

• Formalin (cat. no. SF100‐4, Thermo Fisher Scientific, Waltham, MA)

• Lyophilizer

• Biopsy punch

• Laminar flow hood

• 1Prepare a collagen‐glycosaminoglycan‐calcium phosphate suspension by mixing microfibrillar type 1 collagen (1.9 wt%), chondroitin‐6‐sulfate (0.17 wt%), calcium nitrate (Ca(NO₃)₂•H₂O; 0.4 wt%), and calcium hydroxide (Ca(OH)₂; 0.9 wt%) in a solution of phosphoric acid.

  For example, for a 100‐ml batch: add type 1 collagen (1.9 wt%; 1.90 g) and chondroitin‐6‐sulfate (0.17 wt%; 0.17 g) (assuming the final suspension is 100 g total); then add the calcium salts at 0.4% (w/v) Ca(NO₃)₂•H₂O (0.40 g per 100 ml; or 0.31 g per 100 ml if using Ca(NO₃)₂•H₂O for equivalent Ca²⁺) and 0.9% (w/v) Ca(OH)₂ (0.90 g per 100 ml); and finally bring to 100 ml with the phosphoric acid solution.

• 2Freeze the suspension at a constant cooling rate of 1°C/min, decreasing from room temperature to a final temperature of –10°C, and then lyophilize under vacuum (<200 mTorr, 0°C), resulting in 4‐mm‐thick sheets.
• 3Using a biopsy punch, cut the scaffolds into disks 8 mm in diameter and 3.5 mm thick (density 0.0548 ± 0.0045 g/cm³). Sterilize the disks by submerging them in 100% ethanol and then rehydrate them in DPBS overnight at room temperature.

• 4Weigh the MC‐GAG scaffolds.
• 5Immerse the scaffolds in 100% ethanol under a laminar flow hood and incubate overnight at room temperature to sterilize and dehydrate them.
• 6Working in the laminar flow hood, gradually replace the ethanol with DPBS by first replacing the 100% ethanol with 70% (v/v) ethanol for 1 hr and then gently washing the scaffolds with DPBS one to five times (adding ~30 ml DPBS each time and waiting 1 hr between washes). Allow the scaffolds to hydrate for 24 hr (or at minimum overnight) in DPBS to ensure complete hydration.
• 7Cross‐link the hydrated scaffolds by immersing them in an aqueous solution containing EDAC and NHS at concentrations corresponding to a 5:2:1 molar ratio of EDAC/NHS/COOH (where COOH represents collagen carboxyl groups). Incubate the scaffolds in the solution for 2 hr at room temperature.

  Using the dry collagen weight per scaffold determined in step 4, calculate the amounts of EDC and NHS. For mineralized scaffolds, we use 0.0111 g/scaffold EDC and 0.0026 g/scaffold NHS.

• 8Place the scaffolds in fresh DPBS on a shaker with gentle shaking for 2 hr to remove any residual reagents.

• 9Transfer each sample to an individual well of a 24‐well plate and add DMEM supplemented with 10% FBS, 2 × 10⁻³ M l‐glutamine, 100 IU/ml penicillin, and 100 µg/ml streptomycin (hMSC growth medium). Incubate overnight in a 37°C, 5% CO₂ incubator.
• 10Add 10 µl of cell suspension to each MC‐GAG scaffold and incubate them for 15 min at 37°C. Flip the scaffolds over and add another 10 µl of cell suspension to each, for a total of 3 × 10⁵ hMSCs seeded on each. Allow to attach for 1 hr and then incubate overnight in hMSC growth medium.
• 11After 24 hr, replace the culture medium. Continue to culture the cells, changing the medium every 2‐3 days.
• 12After 8 weeks, aspirate the medium and fix the hMSC‐seeded MC‐GAG scaffolds in 10% formalin for 24 hr. Store in 70% ethanol at 4°C (Fig. 1).

Figure 1.

Representative images of MC‐GAG scaffold before crosslinking and after culture with hMSCs. Photographs show the scaffold before crosslinking (MC‐GAG), immediately after seeding with hMSCs (0 weeks), and after 8 weeks of culture, at which point the sample was fixed in formalin for subsequent microCT analysis.

Basic Protocol 2. HIGH‐RESOLUTION microCT SCANNING AND MICROSTRUCTURAL ANALYSIS OF THE MC‐GAG SCAFFOLDS

This protocol outlines the preparation, high‐resolution microCT imaging, and computational processing steps required for the evaluation of MC‐GAG scaffolds. It describes the design and use of custom designed and 3D‐printed scaffold racks to ensure stable sample positioning during scanning, followed by optimized high‐resolution microCT acquisition parameters that provide consistent, high‐fidelity imaging across samples. The protocol further details reconstruction of raw projection data, application of dynamic thresholding strategies, and segmentation workflows in the DICOM viewer ORS Dragonfly to distinguish scaffold architecture from mineralized regions. These procedures enable precise 3D visualization and quantitative analysis of mineral distribution, density, and microstructural organization within MC‐GAG scaffolds.

Materials

• Polylactic Acid‐Light Weight (ePLA‐LW) filament (eSUN 3D, Shenzhen Esun Industrial Co., Ltd.)

• Prusa MK4 3D printer

• PrusaSlicer software (©2011‐2019 Alessandro Ranellucci; © 2016‐2023 Prusa Research)

• Dragonfly image‐processing software or other DICOM viewer

• U‐Net model, pre‐trained (Object Research Systems)

• 1Design an appropriate dome‐shaped scaffold rack; we use a rack containing three vertically stacked compartments (each ∼1.4 cm × 1.4 cm × 1.5 cm in height) designed to conform to the fixed circular sample holder of our microCT scanner HiCT (Fig. 2A; Taschereau et al., 2025).

Figure 2.

Design and fabrication of a 3D‐printed scaffold rack for microCT imaging. (A) Design of the dome‐shaped scaffold rack in PrusaSlicer (isometric and top views). (B) 3D‐printed scaffold rack made of PLA‐LW, featuring three vertically stacked compartments open on one side for scaffold placement (overall dimensions: 1.4 cm × 1.4 cm × 1.5 cm). (C) The scaffold rack containing MC‐GAG scaffolds is placed inside a plastic container, which is used to securely position the rack on the sample bed of HiCT. (D) Placement of the custom 3D‐printed scaffold rack in the microCT scanner; the inset (red dashed box) highlights the scanning bed, and the blue arrow indicates the plastic holder containing the rack. (E) User interface of the HiCT operation software showing the imaging protocols used in this study. The red circle indicates the “Normal” preset and “CUP SIZED (∼25 mm)” image dimension setting.

• 2Import the STL file of the rack model into PrusaSlicer for slicing, using the generic PLA protocol with modified nozzle temperature 240°C and extrusion multiplier 0.53, to generate a 3D‐printer‐specific G‐code file.
• 3Fabricate the racks using the Prusa MK4 3D printer (Fig. 2B), printing all components using ePLA‐LW filament.

  ePLA‐LW filament was selected for its low X‐ray attenuation property and stability during microCT imaging.

  The dome‐shaped geometry was specifically engineered to fit within our microCT scanner HiCT's cylindrical sample chamber while securely holding three scaffolds in a reproducible orientation during imaging. Modification may be needed for a different high‐resolution microCT scanner.

• 4Place up to three scaffolds in the 3D‐printed rack, and then secure the rack in a small plastic container to stabilize the samples during scanning (Fig. 2C).
• 5Open the HiCT sample chamber and position the container on the sample holder (Fig. 2D).
• 6In the Study section of the HiCT operation software, enter the project name. Select “CUP SIZED (∼25 mm)” for the field of view (FOV; Fig. 2E).
• 7Initiate the HiCT scan by clicking the “Normal (35 min)” button on the software interface.
• 8Reconstruct the acquired projection data by turning on the HiCT's reconstruction program.
• 9Remove the scanned scaffolds from the HiCT chamber.

• 10Open the Dragonfly software and select “Import DICOM Images”.
• 11Load the complete DICOM series from the HiCT scanner.
• 12In the Visualization panel, choose “grayscale” as the LUT (look‐up table) to enhance contrast between mineralized regions and backgrounds noise.

  The grayscale LUT provides a neutral, intensity‐based visualization that accurately reflects voxel density, making it suitable for setting threshold boundaries.

• 13
  Under the Segmentation or Thresholding module, apply the Auto‐Thresholding function:

  • Define range = Upper Otsu
    All reconstructed volumes were calibrated to Hounsfield unit (HU) scaling.
• 14
  After applying Upper Otsu, ORS Dragonfly automatically generates an initial HU threshold range (Fig. 3A):

  • Lower limit: –49.37 HU
  • Upper limit: 3026.71 HU
    These values represent an AI‐derived estimate for separating scaffold material from background based on histogram‐driven segmentation.

Figure 3.

Segmentation and virtual sectioning of 3D‐printed scaffold racks using ORS Dragonfly. (A) Threshold selection interface showing the automatically generated HU range (–49.37 to 3,026.71) obtained using the Upper Otsu method. (B) Representative orthogonal HiCT slices (frontal, sagittal, coronal, and axial), demonstrating separation of the scaffold compartments from the surrounding dome‐shaped rack with enhanced contrast for visualization.

• 15Inspect the auto‐thresholded volume in axial, sagittal, and coronal views (Fig. 3B).
• 16Select “Create ROI” and click “Generate Surface”.

  Regions of interest (ROI) are generated to isolate each scaffold set and the corresponding 3D‐printed scaffold stand. The resulting surface model provides a complete 3D representation of the thresholded scaffold, including the dome‐shaped outer frame and internal compartments. Window and level settings are adjusted to optimize visualization and downstream segmentation.

• 17
  Open the segmentation panel and load the HU range previously determined for scaffold identification (Fig. 4A):

  • Scaffold reference HU range: 250‐13,590.62 HU

Figure 4.

HU‐based segmentation of MC‐GAG for MC‐GAG scaffold and mineralization. (A) Scaffold segmentation using an HU range of 250‐13,590.62. (B) Mineralization segmentation using an HU range of 350‐13,590.62. Orthogonal HiCT views demonstrate the separation of scaffold structure and mineralized regions prior to 3D reconstruction.

• 18
  Adjust the lower HU boundary upward to isolate mineralized regions (Fig. 4B):

  • Mineralization HU range: 350‐13,590.62.
    In our work, reconstructed image volumes first underwent manual denoising to ensure data fidelity. Isolated voxel islands smaller than a 2 × 2 × 2 voxel (20 µm per voxel) were removed prior to segmentation. A refined mineral ROI was then generated by thresholding all voxels above 350 HU, a cutoff specific to our HiCT scanner and reconstruction settings. This threshold was determined by identifying the histogram inflection point, separating soft‐tissue‐equivalent material from mineralized tissue, and was further validated using calibration scans of phantoms with known mineral densities. Users working with other microCT systems should determine their own system‐appropriate threshold through the same general workflow comprising phantom‐based calibration, histogram analysis, and validation against known standards rather than relying on the absolute HU value reported here.
• 19Apply the updated HU range to generate the mineralization mask, confirming that only high‐density mineralized voxels remain.
• 20Overlay the scaffold and mineralization masks to verify accurate separation of mineralized and non‐mineralized regions.
• 21Convert the mineralization mask into a 3D surface model to visualize and quantify mineral distribution within each scaffold compartment.
• 22Denoise the scaffold‐containing volumes using the pre‐trained U‐Net model to enhance scaffold boundary clarity and reduce reconstruction noise.
• 23Perform global threshold segmentation of the scaffold outline using the Otsu method to identify scaffolds.
• 24Apply light morphological cleanup, removing isolated 1 × 1 × 1 voxel artifacts to ensure a clean binary scaffold mask.
• 25Convert the binary scaffold mask to a 3D mesh representation.
• 26Refine the mesh using Catmull‐Clark subdivision surface smoothing to improve surface continuity and enhance visualization of scaffold geometry.

• 27Open the ROI Painter or 3D visualization module and load the previously generated scaffold mask and mineralization mask.
• 28
  Assign color values to each mask for clear differentiation:

  • a.
    Set the mineralization mask to a blue color map (Fig. 5A).
  • b.
    Set the scaffold mask to a gray or light grayscale color map (Fig. 5B).

Figure 5.

3D renderings of scaffold architecture and mineral deposition. (A) Isolated mineralization network rendered as a stand‐alone volume to illustrate the spatial distribution of high‐density mineral voxels. (B) Composite structural rendering showing the mineral network (blue) overlaid onto the scaffold outline (gray). (C) Quantification of scaffold, mineral, and combined volumes.

• 29Adjust opacity settings for each mask to improve visibility of overlapping structures.
• 30Enable volume rendering or surface rendering for both masks, ensuring that the mineralized regions appear prominently within the scaffold volume.
• 31Inspect the reconstructed 3D model by rotating, zooming, and slicing through the volume to confirm that mineralized regions are accurately localized within the scaffold boundaries.
• 32Export the composite 3D rendering as an image using ORS Dragonfly's export functions.
• 33Save the segmentation workspace to preserve the scaffold and mineralization masks for future quantitative analysis (e.g., mineral volume, distribution mapping).
• 34Generate the final composite structural rendering by overlaying the scaffold outline mesh with the internal mineralization network.
• 35Adjust mesh smoothing, transparency, and lighting parameters to ensure that the mineralized regions remain clearly visible while preserving the overall scaffold geometry.
• 36Confirm that the composite rendering accurately represents both scaffold macro‐architecture and internal mineral deposition patterns.
• 37Export the final composite structural rendering for publication‐quality visualization and archival.

COMMENTARY

Critical Parameters

Consistent scaffold positioning and rack fit

The placement of the sample within the 3D‐printed rack is crucial. Accurate microCT analysis depends on stable and reproducible scaffold placement within the rack. Even minor shifts or tilting during scanning introduce motion artifacts, alter HU distributions, and compromise segmentation. It is absolutely necessary to ensure that all scaffolds are seated in identical orientations and firmly secured within the scanner holder.

Voxel size and scan resolution

The detection of mineralized voxels at 20‐µm scale requires high‐resolution acquisition. Any deviation in voxel size, scan duration, or reconstruction settings directly affects mineral segmentation accuracy. One must maintain identical acquisition parameters across all samples and avoid downsampling.

microCT acquisition parameters (projections, rotation step, voxel size, scan duration)

To support reproducibility, typical imaging conditions used in this protocol are summarized in the microCT acquisition subsection. Each scan is acquired with 3600 projections over 360° (0.1° rotation step) and reconstructed with an isotropic voxel size of ∼20 µm, yielding a typical scan time of ∼18 min per sample. We have also clarified our dynamic calibration schedule: calibration is performed at the start of each imaging session, and repeated when a new batch of samples is introduced or when routine quality checks indicate potential drift. These additions allow users to more accurately reproduce the imaging conditions described in this protocol.

HU threshold determination for mineral segmentation (≥350 HU)

The lower HU cutoff is one of the most sensitive parameters. Small adjustments (±25‐50 HU) substantially change the measured mineral volume. The 350‐HU threshold used in this protocol is specific to our HiCT scanner and reconstruction settings; therefore, users must determine an appropriate system‐specific threshold using histogram analysis and calibration phantoms. This ensures consistent identification of mineralized voxels across different scanners and imaging environments.

Consistency in mesh smoothing and composite rendering parameters

Scaffold outline segmentation, mesh smoothing (Catmull‐Clark), and rendering settings (opacity, color mapping, lighting) must remain constant across samples. Variations in visualization parameters can artificially exaggerate or obscure mineral deposition patterns, affecting both qualitative interpretation and downstream quantitative analyses.

Troubleshooting

An overview of troubleshooting is provided in Table 1.

Table 1.

Troubleshooting for microCT

Problem	Possible cause	Solution
Blurry or streaky image	Scaffold moved during scan	Ensure scaffold is firmly seated; add a small spacer; confirm rack fits tightly in the holder.
Scaffold and mineral look similar	Low contrast or incorrect scan settings	Increase resolution or exposure; recalibrate with phantoms.
Scaffold edges look soft or distorted	Denoising applied too heavily	Reduce denoising strength; confirm correct model; compare denoised versus raw data.
Not enough mineral visible	HU threshold too high	Lower threshold in small steps (∼25 HU); check histogram inflection point.
Too much of the non‐mineralized regions included	HU threshold too low or noisy image	Raise threshold; denoise with U‐Net first; apply consistent artifact removal.
Scaffold boundary looks rough or has holes	Noise affecting Otsu threshold	Apply U‐Net before Otsu thresholding; adjust morphological cleanup; avoid over‐smoothing.
Mineral is hard to see inside scaffold	Opacity settings not optimized	Lower scaffold opacity; increase mineral opacity; use contrasting colors.
Mineral appears outside scaffold	Misalignment of masks	Ensure both masks come from same dataset; reapply alignment; check for cropping.
Exported image looks low quality	Low export resolution settings	Increase export resolution; save as PNG or TIFF.

Understanding Results

When performed correctly, this protocol should produce clear microCT reconstructions that accurately capture both the scaffold architecture and the internal mineral network. Users should expect (1) a smooth, continuous scaffold outline after thresholding and mesh generation, (2) a mineralization mask containing distinct high‐density voxel clusters segmented at ≥350 HU, and (3) composite 3D renderings in which mineral deposits (blue) appear appropriately embedded within the scaffold structure (gray), as shown in the benchmark figures.

Correct mineral segmentation will show clustered, spatially coherent mineral voxels with minimal background noise. Over‐segmentation appears as diffuse speckling or mineral outside the scaffold boundaries, typically caused by too low an HU threshold or insufficient denoising. Under‐segmentation, often due to overly aggressive cleanup or a threshold set too high, results in sparse or fragmented mineral clusters. Similarly, irregular scaffold edges or gaps in the scaffold surface indicate that denoising or Otsu thresholding was not applied uniformly.

Unexpected findings, such as mineral localized only in certain scaffold compartments or mineral appearing outside the scaffold, may reflect biological variability, scaffold movement during scanning, or segmentation errors. As an additional benchmark, users can compare their quantitative outputs with representative values from a typical dataset: for example, a mineral volume of ∼0.26 mm³ within a scaffold volume of ∼196 mm³, corresponding to a mineral volume fraction of ∼0.13% and ∼47,000‐50,000 mineralized voxels. These values are illustrative and will vary by biological condition, but they provide a useful reference range for confirming that segmentation and reconstruction are performing as expected. Users should compare both their visual and quantitative results with the figures provided to determine whether the protocol has been successfully reproduced and whether the dataset is suitable for downstream analysis.

Time Considerations

Preparing and printing the custom dome‐shaped scaffold rack requires ∼1 hr, and mounting scaffolds into the rack and positioning them in the scanner takes additional 10 min. High‐resolution microCT data acquisition generally requires 60 min, followed by 2 hr for data reconstruction. Initial denoising, including manual removal of 2 × 2 × 2 voxel artifacts, adds ∼10 min per dataset. Segmentation steps including U‐Nest denoising, threshold‐based mineral isolation, scaffold outline identification, and morphological cleanup typically require 30 min, and 3D mesh generation and composite rendering add an additional 15 min. However, users may adjust the resolution settings according to their specific needs, with the understanding that these choices will directly influence both the acquisition time and the resulting data size.

Author Contributions

Youngnam Kang: Conceptualization; investigation; writing—original draft; writing—review and editing; data curation; supervision; methodology; project administration. Kaavian Shariati: Writing—original draft (portions of the manuscript); data curation; data interpretation; writing—review and editing. Catherine T. Cascavita: Writing—original draft (portions of the manuscript); data curation; writing—review and editing. Xiaoyan Ren: Critical review and revision of manuscript. Brendan A. C. Harley: Critical review and revision of the manuscript. Shili Xu: Conceptualization; investigation; writing—original draft; writing—review and editing; critical review and revision of the manuscript. Justine C. Lee: Conceptualization; investigation; writing—original draft; writing—review and editing; critical review and revision of the manuscript.

Conflict of Interest

The authors declare no conflict of interest.

Kang, Y. , Shariati, K. , Cascavita, C. T. , Ren, X. , Harley, B. A. C. , Xu, S. , & Lee, J. C. (2026). Evaluating mineral deposition in bone‐regenerative collagen‐based scaffolds using high‐resolution microct and 3D computational segmentation. Current Protocols, 6, e70306. doi: 10.1002/cpz1.70306

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.