Three-Dimensional Printing of Clinical Scale and Personalized Calcium Phosphate Scaffolds for Alveolar Bone Reconstruction

Abstract

Objective.

Alveolar bone defects can be highly variable in their morphology and, as the defect size increases, they become more challenging to treat with currently available therapeutics and biomaterials. This investigation sought to devise a protocol for fabricating customized clinical scale and patient-specific, bioceramic scaffolds for reconstruction of large alveolar bone defects.

Methods.

Two types of calcium phosphate (CaP)-based bioceramic scaffolds (alginate/β-TCP and hydroxyapatite/α-TCP, hereafter referred to as hybrid CaP and Osteoink™, respectively) were designed, 3D printed, and their biocompatibility with alveolar bone marrow stem cells and mechanical properties were determined. Following scaffold optimization, a workflow was developed to use cone beam computed tomographic (CBCT) imaging to design and 3D print, defect-specific bioceramic scaffolds for clinical-scale bone defects.

Results.

Osteoink^™ scaffolds had the highest compressive strength when compared to hybrid CaP with different infill orientation. In cell culture medium, hybrid CaP degradation resulted in decreased pH (6.3) and toxicity to stem cells; however, OsteoInk^™ scaffolds maintained a stable pH (7.2) in culture and passed the ISO standard for cytotoxicity. Finally, a clinically feasible laboratory workflow was developed and evaluated using CBCT imaging to engineer customized and defect-specific CaP scaffolds using OsteoInk^™. It was determined that printed scaffolds had a high degree of accuracy to fit the respective clinical defects for which they were designed (0.27 mm morphological deviation of printed scaffolds from digital design).

Significance.

From patient to patient, large alveolar bone defects are difficult to treat due to high variability in their complex morphologies and architecture. Our findings shows that Osteoink^™ is a biocompatible material for 3D printing of clinically acceptable, patient-specific scaffolds with precision-fit for use in alveolar bone reconstructive procedures. Collectively, emerging digital technologies including CBCT imaging, 3D surgical planning, and (bio)printing can be integrated to address this unmet clinical challenge.

1. Introduction

Alveolar bone defects – whether they result from congenital malformations, disease, trauma, or surgery – are prevalent, costly to treat, and have significant effects on quality of life [1–9]. Variable etiologies of these defects lead to highly distinct clinical presentations, creating a difficult challenge when attempting to regenerate the normal physiological shape, architecture, and function of the missing bone and soft tissue. Despite recent advances in bone grafting modalities to address clinically challenging situations (e.g., orofacial clefting, alveolar trauma, oral cancer, multiple tooth loss, and severe periodontitis), it continues to be difficult to deliver predictable treatment outcomes for severe alveolar defects due to their complex and highly variable morphologies [10–16]. Given the significant functional and esthetic impairments that result from alveolar defects, there is high demand for personalized regenerative strategies for their reconstruction.

A range of inert alloplastic materials have been employed, including polycaprolactone, polyether ether ketone, and titanium [17–19]. However, these materials have drawbacks such as poor tissue integration and increasing deterioration of soft tissue, which can lead to scaffold exposure, infection, and ultimately removal [20]. The basic concept underlying regenerative approaches to treat these defects involves surgical placement of biomaterials to create a microenvironment conducive to enabling the appropriate cells to populate the site of interest, interact, and differentiate to yield tissue regeneration [21–23]. Currently, in clinical practice the common formulations of calcium phosphate materials have shown promise due to their mineral content, density, solubility, and three-dimensional microarchitecture [24–26]. In a series of cell therapy clinical trials, we have investigated the use of β-tricalcium phosphate (β-TCP) as a scaffold matrix to deliver stem cells to alveolar bone defects to regenerate alveolar bone [27–29]. Though these studies demonstrated varying degrees of regenerative success in using this material as a delivery scaffold, bone regeneration was limited to the treatment of larger, more complex defects with variable morphologies. This is because their applications largely use a “one-size fits-all approach” where calcium phosphate material serve as an osteoconductive space maintenance material to facilitate cell ingrowth. As such, it was determined that for these defects, a more personalized, defect-specific approach is needed.

Additive manufacturing, also known as three-dimensional (3D) printing is an emerging technology that could have great utility in the treatment of these defects [30, 31]. This technology enables the design and engineering of personalized, defect-specific scaffolds that can be printed to meet very specific shapes, sizes, and morphologies. Readily scalable and reproducible, this approach has been used to generate scaffolds for predictable regeneration of complex tissues throughout the body [32–34]. Nonetheless, their investigation for alveolar bone regenerative applications has been limited.

In the present study, we hypothesized that through 3D imaging, design, and bioprinting, a digital workflow could be established to yield customized, clinical-scale calcium phosphate (CaP) scaffolds engineered in dimensions that could be precision fit into clinical alveolar defects. The aims of the study were to first determine the ideal CaP-based scaffold architecture that would yield optimal mechanical properties. Next, we aimed to evaluate the biocompatibility of these scaffolds with human mesenchymal stem cell populations (derived from alveolar bone). Finally, we sought to develop a standardized digital workflow, whereby clinical alveolar defects could be imaged to aid in the design and fabrication of clinical-scale CaP scaffolds with dimensions that would enable a precision-fit into the original defects.

2. Materials and methods

2.1. Fabrication and mechanical testing of CaP-based scaffolds

For the fabrication of alginate/β-TCP scaffolds (hybrid CaP; 35% β-TCP by weight with 65% alginate [4% alginic acid in water]), a Bioplotter (Envision TEC GBMH, Gladbeck, Germany) 3D printer was used. Alginate was used as a binder for the β-TCP. The components were mixed 4× times with a blender whisk for 1 min, respectively. A metal spatula was used to place the product in the Bioplotter cartridge. Cylindrical constructs (5 mm in diameter and 8 mm thick) were printed into 60-mm petri dishes in triplicate with varying infill orientations. All constructs were printed at 25 +/− 5°C with a regular, open lattice internal design. Each petri dish was large enough to hold seven scaffolds, all which were soaked in 4% calcium chloride (CaCl₂, Millipore-Sigma, St. Louis, MO, USA) in water for 10 min. The CaCl₂ was then discarded, and the scaffolds were soaked with fresh CaCl₂ for an additional 5 min. The scaffolds were stored at 4°C in the presence of a low-lint tissue paper (Kimwipes, Kimberly-Clark Professional, Roswell, GA, USA) soaked in CaCl₂.

Hydroxyapatite/α-TCP (OsteoInk^™, regenHU, Villaz-St.-Pierre, Switzerland) scaffolds were printed using the 3D Discovery Evolution bioprinter (regenHU). Scaffold parameters remained the same and were imported into the accompanying BioCAD (regenHU) design program. Human Machine Interface (HMI) software was initiated. OsteoInk^™ cartridges were assembled on the printhead using a 5 cc cartridge adapter and a 22G needle was used to print the scaffolds. The OsteoInk^™ scaffold was stored in 37°C with 100% humidity for 4 days according to the manufacturer’s instructions for maximum setting.

The mechanical properties of hybrid CaP and Osteoink^™ with varying infill orientation (i.e., 40°, 45°, 60°, and 90°) were evaluated using compressive strength test. Three replicates of hybrid CaP with dimensions of ~ 8 × 3 mm were subjected to unconfined compression at a strain rate of 1 mm/min using a screw-driven universal testing machine (TestResources Inc., Minnetonka, MN, USA). The compressive strength of OsteoInk^™ was determined using the MTESTQuattro universal testing machine (ADMET, Inc., Norwood, MA, USA), with the same test parameters. During testing, force-displacement responses were monitored, and compressive strength of the specimen was calculated by dividing the maximum load attained during the test by the cross-sectional area of the specimen.

2.2. Measurement of pH

To determine the effect of degradation on the pH around the hybrid CaP and OsteoInk^™ scaffolds, printed scaffolds were incubated in minimum essential alpha medium (α-MEM, ThermoFisher Scientific, Inc., Waltham, MA, USA) at 37°C. Briefly, the scaffolds were placed in a 24-well plates and fully covered with 300 μL α-MEM supplemented with 15% fetal bovine serum (FBS, ThermoFisher Scientific, Inc.). At baseline, 5 min, 1 h, and 24 h, 15 μL of the medium was removed for pH testing using Litmus strips (ThermoFisher Scientific, Inc.).

2.3. Alveolar bone derived mesenchymal stem cells (aBMSCs) isolation and expansion

The use of aBMSCs for these experiments was approved by the Institutional Review Board at the University of Michigan (IRB#HUM00034368). aBMSCs were derived/isolated from adult alveolar bone tissue with informed consent from patients undergoing distinct surgical procedures in the Periodontics and Oral Medicine Department (University of Michigan School of Dentistry), as previously detailed [35]. Briefly, aBMSCs were isolated from alveolar bone marrow tissue samples (~ 0.5 cc) using a bone scraper and resuspended in cold α-MEM. Centrifugation was carried out at 600 g for 10 min at room temperature. The supernatant was then removed, and the cell pellet resuspended in 5 mL α-MEM with 15% FBS. Next, the cell suspensions were transferred to T-25 tissue culture flasks and allowed to sit undisturbed without any change in media for 5 days in a 37°C humidified tissue culture incubator at 5% CO₂. After 5 days in culture, the non-adherent cells were removed, and the media was then changed to α-MEM supplemented with 10% FBS and changed every 2–3 days. Once the adherent cells reached 80–90% confluence, the aBMSCs were collected and subcultured to passage 8.

2.3.1. Stem cell seeding efficiency and viability on scaffolds

To confirm the cell-seeding efficiency and viability of the 3D printed CaP scaffolds, aBMSCs were seeded onto them. After 24 h, the scaffolds were removed from medium and placed into an empty 24-well plate; cell seeding efficiency and viability were then assessed, as previously described [36]. Briefly, the medium was then collected in a 1.5 mL tube for each scaffold. 0.3 mL DPBS was added to rinse the well and then collected in the same 1.5 mL tube. Next, 100 μL TrypleE reagent (ThermoFisher Scientific, Inc.) was added, and the solution was allowed to incubate for 5 min. 200 μL medium was used to neutralize the TrypleE and it was mixed via repeated pipetting. This solution was transferred to the 1.5 mL tube. Another 200 μL of medium was used to pick up any residue from the wells. The tubes were spun in the centrifuge at 400 g for 5 min at room temperature. The supernatant liquid was aspirated with caution so as not to disturb the pellet. The cells were resuspended in 200 μL TrypleE and again neutralized with 200 μL medium, pipetting to mix. 18 μL of the cell suspension was mixed with 2 μL of trypan blue dye and placed on a hemocytometer for cell counting. Cells on the upper and left border were included in the counts, while cells on the lower and right border were excluded. Cells that appeared to have taken up the blue dye were assumed to be apoptotic and, therefore, not viable. The percentage of live cells was calculated:

$$Viability(\%)=\frac{N_{live}}{N_{live}+N_{dead}}$$

where N_live and N_dead are the number of live and dead cells, respectively.

$$\text{Cell-Seeding}\text{Efficiency}(\%)=\frac{N_{\text{seeded}}-N_{\text{floating}}}{N_{\text{seeded}}}$$

where, N_seeded is the cell number seeded and N_floating is the cell found in media after 24 h of incubation.

2.3.2. Cytotoxicity of CaP-based scaffolds

Cytotoxicity testing of the scaffolds was adopted from ISO Standard 10993–5 – Biologic Evaluation of Medical Devices. Hybrid CaP (n=4) and OsteoInk^™ scaffolds (n=5/group) (1.32 mm high, 8 mm diameter, volume = 66.35 mm³) were disinfected twice using a 70% EtOH wash, three times with PBS, and once using basal medium without FBS. Sterile culture medium was prepared based on the surface area (3 cm²/mL), as defined in ISO 10993–12. For a disk of h mm thickness and an r mm radius, surface area = 2πrh + 2πr² = 2πr(h+r). This came out to 0.445 mL per scaffold. pH of the medium was maintained between 7.2–7.4. The scaffolds were incubated for 48 h at 37°C under a 5% CO₂ humidified atmosphere.

After thawing from stock, aBMSCs were passaged 2–3 times before use. Cells were removed from culture flasks by enzymatic digestion (trypsin/EDTA) and the cell suspension was centrifuged (200 g, 3 min). The cells were then resuspended in culture medium and the cell suspension was adjusted to a density of 100,000 aBMSCs/mL. A multichannel pipette was used to dispense 100 μL of culture medium only (blank) into the peripheral wells of a 96-well tissue culture microtiter plate. In the remaining wells, 100 μL of the cell suspension was dispensed (100,000 aBMSCs/mL = 10,000 aBMSCs/plate), avoiding the peripheral columns and rows. The cells were incubated at 37°C with 5% CO₂ for 24 h (~ 1 doubling period) to form a semi-confluent monolayer. This incubation period ensured cell recovery, adherence, and progression to the exponential growth phase. After 24 h of incubation, the cells were checked for even distribution. The supernatant liquid was collected, centrifuged at 11,500 rpm for 5 min at room temperature, and transferred to new tubes. The cells were then exposed to one of the two CaP compounds over a range of concentrations. The culture medium was removed, and the cells were treated with 4 distinct (100%, 50%, 25%, and 12.5%) concentrations of test sample extract in 100 μL of treatment medium. Serial dilutions were prepared: 180 μL of scaffold-conditioned medium was diluted with 180 μL of the medium incubated alone for the 50% dilution, and the same for the 25% and 12.5% medium, respectively. An untreated well of culture medium alone served as a blank. For the negative control, only the 100% concentration was tested. The samples were again incubated at 37°C with 5% CO₂ for 24 h.

After 24 h, the aBMSCs were examined under a phase-contrast microscope for morphological alterations or systematic errors in seeding. Culture medium was removed. The plates were washed once with 150 μL of PBS and then 20 μL of MTS solution (CellTiter 96 Aqueous One Solution Cell Proliferation Assay, Promega Corporation, Madison, WI, USA) was added and incubated for 2 h under the same conditions. After 2 h, the MTS solution was discarded. The plates were shaken and transferred to a microplate reader equipped to read the absorbance (reference wavelength 650 nm). MTS absorption was detected at 490 nm to determine cell viability. A decrease in the number of living cells results in a decrease in metabolic activity of the cells. This directly correlates to the amount of blue-violet formazan formed, as monitored by the optical density at 490 nm. To calculate the reduction in viability compared to the blank, the following equation was used:

$$Viability\%=\frac{{\text{OD}}_{490\text{e}}}{{\text{OD}}_{490\text{b}}}\times 100$$

where, OD490e is the mean value of the measured optical density of the 100% extracts of the test sample and OD490b is the mean value of the measured optical density of the blanks. The lower the viability percentage calculated, the higher the cytotoxicity of the material.

2.4. Construction of 3D volumetric label maps and segmentations

Cone-beam computed tomography (CBCT) scans of the maxilla and mandible were obtained from patients (n=6) who had been treated with alveolar bone grafting procedures at the Periodontics and Oral Medicine Department (University of Michigan). These scans were converted from DICOM files to GIPL files for de-identification purposes. Using a combination of the open source software programs 3DSlicer (3D Slicer Software, www.slicer.org) and ITK- SNAP (ITK-SNAP software, www.itksnap.com), a 3D label map of the hard tissue was created and cleaned to produce a segmentation model of the bone (including the defect site of interest) for each patient. The label map of the bone was created by adjusting the range file or density of tissue selected (Intensity Segmenter, 3D Slicer Software, www.slicer.org). The segmentation was cleaned and cropped in ITK-SNAP to remove artifacts and unnecessary anatomy in order to improve visualization and decrease file size.

2.4.1. Scaffold prototype and evaluation of fit

After the six 3D segmentations were completed, one case was utilized to develop a customized graft prototype using a combination of ITK-SNAP and MeshMixer open-source softwares. The selected case had a significant horizontal and vertical alveolar defect of the left mandible in the area of teeth #18–20 and required both horizontal and vertical bone augmentation. The scaffold was initially constructed in ITK-SNAP and then adjusted in MeshMixer. The region of the mandible containing the defect and the scaffold were 3D printed from a clear resin material using a Formlabs Form 2 printer. The fit was evaluated qualitatively. The printed mandibular defect region, scaffold, and scaffold placed on the defect region were scanned using a TRIOS (3Shape A/S, Copenhagen, Denmark) intraoral scanner. The scaffold was virtually aligned onto the defect and the void detected and quantified in ITK-SNAP.

2.4.2. Fabrication of customized OsteoInk^™ scaffolds and evaluation of fit

To assess scaffold printing accuracy, three customized scaffolds were designed and printed (3DDiscovery, regenHU) with OsteoInk^™. A series of 2D cross-sections were sliced from the STL model, and tool paths were generated using computer-aided manufacturing software (BioCAM; regenHU). OsteoInk^™ was loaded into 5 mL syringes and extruded through a 22G needle. The scaffolds were printed on a glass microscope slide at a feed rate of 5 mm/s and an extrusion pressure of 350 kPa with a rectilinear filling pattern of 50% density and then set at 37°C with 100% humidity for 4 days. The printed scaffolds were scanned with the intraoral scanner and exported as STL files. The scans and design files were imported into a 3D analysis software (Geomagic Control X; 3D Systems, Inc., Rock Hill, SC, USA), and were superimposed with the best matching algorithm. The deviation between the design and printed scaffold was measured at 30 evenly distributed sites on each scaffold.

2.5. Statistical analysis

Statistical analysis was performed using Prism software (GraphPad Software, San Diego, CA, USA). Data were presented as mean ± standard deviation, unless otherwise noted. Statistically significant differences were determined by unpaired Student t tests or one-way ANOVA tests, and statistical significance was defined as p<0.05.

3. Results

3.1. Mechanical characterization and optimization of scaffolds

Scaffolds with varying infill orientations (40°, 45°, 60°, and 90°) were printed and their orientations viewed macroscopically and microscopically with microCT and scanning electron microscopy (SEM, Tescan, Warrendale, PA, USA) (Fig. 1A). Following verification of infill orientation, the mechanical performance of the scaffolds and the effects of infill on compressive strength were evaluated. The load-displacement curve (Fig. 1B) showed the linear region at low displacement value, suggesting initial stiff mechanical response of the printed scaffold. Compared to the hybrid CaP scaffolds, Osteoink^™ scaffolds could not be displaced to 50% of their length, and they required a 5× greater force just to achieve 25% displacement (2 mm). When the compressive strength of the hybrid CaP and Osteoink^™ scaffolds were compared, the Osteoink^™ scaffolds had the highest compressive strength for all four infill orientations tested (Fig. 1C). The compressive strength of the hybrid CaP (3.02 ± 0.42 MPa) and OsteoInk^™ (4.67 ± 0.43 MPa) scaffolds was highest with a 60° infill orientation.

Fig. 1 – Mechanical properties of calcium phosphate scaffolds.

(A) MicroCT and SEM images of various interlayer angles were captured and tested mechanically, including 40° 45°, 60°, and 90°. (B) Scaffolds were placed on a screw-driven universal testing machine for compressive strength testing, where they were displaced to 50% of their size and the required force was measured and plotted on a force-displacement graph. (C) Compressive strength comparison between hybrid CaP scaffolds and Osteoink™.

3.2. pH testing of alginate and β-TCP components in media

Hybrid CaP and OsteoInk^™ scaffolds were placed in media and the pH of the media was tested over 24 h. Both scaffolds showed a decrease in pH of the media over time, with the hybrid CaP causing a more rapid decrease in pH (pH=6.3 ± 0.4) over 24 h as compared to OsteoInk^™ (pH=7.2 ± 0.0) (Fig. 2A). Due to the rapid decrease in media pH caused by the hybrid CaP scaffolds, the acidity of the two components of the hybrid CaP scaffolds (β-TCP and alginate) were tested independently to determine which of these components was responsible for the acidity. Both the β-TCP and alginate samples were tested over six time points. During the first 4 h, the pH dropped for both the alginate and β-TCP samples at a similar rate (Fig. 2B). For the alginate samples, the pH started at 7.1 ± 0.1, decreased to 6.4 ± 0.1 at 4 h, and at the 24-h time point it had increased to 6.9, where it remained over the next 7 days. In contrast, the pH of media subjected to β-TCP demonstrated a continual decrease in pH from 6.9 at the onset to 6.3 at 4 h; and by 24 h, it had decreased to 6.2, where it remained for 7 days. These data demonstrated that early in culture, both components contributed some level of pH to the media, but by 24 h and up to 7 days, β-TCP was causing a more significant decrease in pH than alginate (p<0.001).

Fig. 2 – Acidity and biocompatibility of calcium phosphate scaffolds.

(A) Hybrid CaP scaffolds exhibited a greater decrease in pH over a period of 24 h relative to Osteoink™. (B) The acidity of β-TCP and alginate samples were tested independently over six time points in culture medium and by 24 h; β-TCP caused significantly more acidity in the medium than alginate (p<0.001). (C) Photomicrographs of aBMSCs in culture after loading of hybrid CaP scaffolds. (D) Osteoink™ scaffolds retained significantly more stem cells than the hybrid CaP scaffolds. (E) Cells cultured in the presence of Osteoink™ scaffolds yielded significantly higher cell viability than those cultured in medium with hybrid CaP scaffolds. (F) According to ISO standards, cell viability is shown relative to varying concentrations of beta-TCP and OsteoInk compared to a control and Triton X-100 (at 100% concentration, < 70% viability indicates a cytotoxic effect; at concentrations below 50%, < 100% viability indicates cytotoxicity). At 100% concentration, 81.7% ± 10.7% aBMSCs survived once exposed to Osteoink™ scaffolds and 60.1% ± 6.3% survived after exposure to hybrid CaP scaffolds, indicating that Osteoink™ passed the standard, but that hybrid CaP material exhibited cytotoxicity.

3.3. Scaffold biocompatibility with stem cells

Qualitatively, when attempting to seed cells on hybrid scaffolds and observing the cells in culture after 24 h, it was clear that many of the cells were rounded and unhealthy (Fig. 2C) and their seeding efficiency on this scaffold was significantly lower than on Osteoink^™ (39% on hybrid CaP vs. 90% on Osteoink^™ scaffolds) (Fig. 2D). This observation suggested that the hybrid CaP scaffolds exhibited some level of toxicity towards the aBMSCs when in culture. The quantitative cell viability of aBMSCs seeded on fabricated scaffolds was assessed using the trypan blue dye exclusion test. The hybrid CaP scaffolds exhibited significantly lower cell viability (76.9% ± 6.9%) compared to viability of cells cultured in the presence of Osteoink^™ (90.2% ± 8.8%) and cells cultured in media alone (98.4% ± 1.4%) (Fig. 2E). The difference in viability suggested that aBMSCs were more compatible with OsteoInk^™ than hybrid CaP, so an additional assay was used to compare the cytocompatibility of the two materials. Hybrid CaP and OsteoInk^™ materials were subjected to ISO Standard 10993–5 for biologic evaluation of medical devices. The undiluted OsteoInk^™ sample showed higher cell viability (mean=81.7% ± 10.7%; pass) than the hybrid CaP scaffolds (mean=60.1% ± 6.3%; fail), indicating that OsteoInk^™ was more cell-friendly (p=0.008) and less toxic to the cells. At 50% dilution, OsteoInk^™ had 99.9% ± 9.8% viability (pass), while hybrid CaP was at 96.3% ± 2.8% (fail). For dilutions of 25% and lower, both materials passed the test with 100% viability (Fig. 2F). According to the ISO standards (> 70% for undiluted and ≥ 100% for 50% or further diluted samples), the OsteoInk^™ scaffolds passed and the hybrid CaP scaffolds failed the cytotoxicity test for material, suggesting OsteoInk^™ was the more “cell-friendly” of the two bioceramic compositions.

3.4. Personalized, clinically suitable scaffolds for large, complex alveolar defects

To establish a standardized digital workflow for engineering personalized scaffolds, a clinical case requiring a large autogenous block graft was used as a “model” scenario to represent an indication for which this digital approach would be ideally suited. First, a cone-beam computed tomography (CBCT) scan taken of a patient to receive a large autogenous bone block graft was used to design and create a customized scaffold prototype. The location of interest was the left side of the posterior aspect of the mandible (teeth #18–20) (Fig. 3A). The patient underwent the bone grafting procedure (autologous ramus block graft), whereby the block required sectioning and chairside adaptation to the defect (Fig. 3B). The procedure took > 3 h to complete and was performed in preparation for future placement of oral implants and eventual prosthetic restorations.

Fig. 3 – Standardized digital workflow for engineering defect-specific scaffolds.

(A) 3D digital reconstruction of (B) alveolar bone defect clinically treated with a 25 mm block of autologous bone harvested from the left ramus. (C) The block of bone was shaped and fixated with screws to the defect site on the left mandible in the area of teeth #18–20. (D) CBCT data from the same patient was utilized to determine the feasibility of creating a custom-fit graft modeled using the exact dimensions of the defect site. Models of the mandible and graft were subsequently 3D-printed with resin-based material and evaluated qualitatively for fit. (E) The model and graft were scanned separately, then together; next, they were aligned using specialized 3D software to evaluate fit. The graft is shown in translucent blue with the void colored yellow. The two STL files were fused, as represented by the red graft (in red). The void was quantified, with the greatest distance measuring 1.10 mm from the external surface of the bone to the internal surface of the graft.

The CBCT image of the defect was used to generate a 3D segmentation that accurately represented the defect of interest to create a patient-specific scaffold that could fit into the defect. This prototype scaffold and a replica of the mandible were 3D printed in a resin material using a Formlabs Form 2 printer (Fig. 3C). The mandibular resin model and scaffold were scanned separately, then together, using an intraoral scanner. They were aligned using MeshMixer software, and the void between the external surface of the bone and the internal surface of the graft was quantified via a label map. The void was generally < 1.0 mm measured from the surface of the bone to the internal surface of the scaffold, with the maximum distance measuring 1.10 mm (Fig. 3D). It was determined that fusion of the scaffold and the mapped “void” could create a virtual graft with no void. The small void distance yielded with this methodology provided the evidence needed to test reproducibility of this digital workflow to engineer clinically ready, defect-specific, calcium phosphate (CaP) scaffolds for other large alveolar and craniofacial defects. To perform this analysis, CBCT images from patients with large, variably sized alveolar defects were used to engineer clinically ready, customized Osteoink^™ scaffolds designed based on specific morphologies of the defects. Following generation of digital 3D reconstructions of these defects, the scaffolds were designed according to the methodology described above and printed with Osteoink^™. Following printing, the scaffolds were scanned and their digital designs compared to their actual printed morphology to evaluate how closely their actual dimensions adapted to the defect (Fig. 4). In evaluating the morphological fit of 30 sites among each of the scaffolds, the average fit deviation was only 0.27 mm (std dev. 0.52) (Table 1). These data demonstrated that regardless of the defect morphology or size, this marginal discrepancy was well within the range of what would be clinically acceptable for adaptation of a customized, digitally designed and printed CaP scaffold in a large alveolar bone defect.

Fig. 4 – Fit of printed scaffolds to variably sized clinical defects.

(A) Defect and digital design of defect-specific scaffolds from patient 1. (B) The dimensions of the printed Osteoink scaffolds were compared to the 3D design using superimposed color mapping to determine accuracy of the precision-fit of the scaffold with the alveolar defect. Models were superimposed by the best matching algorithm and the deviation of fit between the two model surfaces averaged among 30 different points.

Table 1.

Fit data for defect-specific scaffolds.

Patient 1	1.73	0	0.16	0.3
Patient 2	1.65	0	0.02	0.7
Patient 3	1.73	0	0.64	0.5
Overall average	1.70	0	0.27	0.5

4. Discussion

The reconstruction of large, alveolar bone defects remains a significant clinical challenge, and there is an urgent need to engineer personalized scaffolds to repair defects. The extrusion-based 3D printing is versatile fabrication method for bone scaffolds due to its capability in printing wide range of biomaterials to engineer geometrically complex scaffolds [30]. With the tremendous increase in demand of 3D printing for precise and repeatable high-throughput scaffold fabrication in medicine and dentistry, a variety of 3D printers are now available. In this investigation, we evaluated the mechanical and biological characteristics of two types of calcium phosphate utilizing two separate high-end extrusion-based printers with comparable XYZ resolution [37]. Furthermore, we assessed ideal calcium phosphate scaffold materials for use in a digital workflow to generate clinical scale, personalized, precision-fit scaffolds that can be used to regenerate large alveolar defects.

First, we evaluated the mechanical property of the OsteoInk^™ and β-TCP/alginate scaffolds (hybrid CaP) as they are intended for eventual replacement of hard tissue in a load bearing area. The mechanical competence of 3D printed scaffolds is governed by porosity, infill orientation, and reinforced material [38, 39]. In this study, compressive strength of the Osteoink^™ was higher than the hybrid CaP scaffolds. This finding is in accordance with a previous report whereby the engineering of a hybrid tri-calcium phosphate/chitosan material resulted in a reduction in the compressive strength and modulus of the tri-calcium phosphate material [40]. Mechanical properties are reduced because of structural instability caused by the presence of two materials with contrasting mechanical properties. As a result, our findings suggest that the OsteoInk^™ scaffold will be more beneficial for structural stability to the hybrid CaP scaffold for alveolar bone regeneration.

The dissolution kinetics of β-TCP have been well-documented and are desirable for tissue regeneration because the material can be completely replaced with vital bone in a relatively short period of time. An important clinical consideration of β-TCP degradation is its effects on stem and progenitor cells in the wound healing microenvironment [41, 42].To quantify the effects of material degradation on the pH of the local microenvironment in culture, we conducted studies evaluating the acidity of the β-TCP degradation byproducts. These studies showed that for the β-TCP/alginate scaffolds, both the β-TCP and alginate components caused some degree of acidity in the media; however, β-TCP caused a more marked and long-term decrease, while alginate caused an initial pH decrease with eventual rebound to neutral pH after 7 days (Fig. 2). β-TCP/alginate caused more acidity (pH 6.3 ± 0.42 at 24 h) than OsteoInk^™ scaffolds (pH=7.2 ± 0.00 at 24 h). Since tricalcium phosphate (β-TCP) is commercially manufactured by treating hydroxyapatite (a calcium salt) with phosphoric acid and slaked lime, it should be expected that some acidity resulted from placing β-TCP in solution. These findings are of particular relevance in determining the suitability of a material for its use as a delivery vehicle and scaffold for cell transplantation. In an in vivo environment, transient acidity of a material may be diluted and neutralized by the dynamic metabolic microenvironment, while in an in vitro setting where cells are placed directly on the material, acidity is cytotoxic and non-desirable. Some reports have shown no biocompatibility issues with β-TCP ceramics when tested in vitro with murine cells [42, 43]. However, cellular behavior can be greatly influenced by chemical and physical properties of the scaffold, including, pH, degradation products, hydrophilicity, roughness, among others. It is known that low pH inhibits cell proliferation and decreases cell viability due to apoptosis and cell necrosis [44, 45]. Notably, the Osteoink^™ showed higher pH compared to hybrid CaP and was suitable for cell proliferation and migration.

In addition to having appropriate mechanical strength and cytocompatibility, scaffold materials must also be customizable and defect-specific in order to address large alveolar bone defects reconstruction. Regardless of the clinical grafting technique, precise defect adaptation is key to facilitating graft stability, perfusion, and integration [46]. As such, the central aim of this study was to develop a standardized digital workflow for utilizing CBCT data to design and engineer customized scaffolds with precision adaptation to clinical defects of variable morphologies. Creating 3D segmentations from CBCT data is a well-established, yet time-consuming process. We determined that acquiring a high-quality CBCT scan with high resolution and a field of view restricted to the region of interest (defect site) is the first step towards a standardized protocol and overall digital workflow for patient-specific bone reconstruction. Removing artifacts and scatter from metallic restorations, fixed appliances, and implants takes a significant amount of time when using a large field of view. High resolution images and more localized fields of view in the areas of interest are key parameters in the image acquisition as a first step in the digital workflow to accurately and consistently distinguish the hard tissue boundaries within the defect. We were able to successfully design and print a scaffold prototype that qualitatively fit precisely into a printed mandibular defect with minimal voids between the outer surface of the defect site and the inner surface of the scaffold. Having these scaffolds printed and prepared for surgical placement prior to surgery would significantly reduce the amount of time required for grafting, because the graft harvesting step would be eliminated, as would the time needed for shaping and molding the graft during the procedure.

Once an efficient digital workflow for generating customized, precision-fit scaffolds was established, we wanted to determine whether the clinically ready material (Osteoink^™) could be printed with predefined morphologies to fit large bone defects in patients with alveolar defects of varying morphologies. Using a completely digital workflow, defects were imaged with CT scans, customized scaffolds were designed virtually based on the defect morphology, and Osteoink was used to 3D print different shaped scaffolds. When comparing the morphology of the printed scaffolds with the printing parameters and actual defect morphologies from which they were designed, there was a very high degree of precision with which the actual scaffolds matched their designs and respective defects. Clinically, if trying to implant these scaffolds, these void spaces (average = 0.27 mm) would be negligible and filled with the natural fibrin clot that forms in the initial wound healing phase of graft integration. Next, we aim to explore this approach in large animal models and ultimately clinical studies.

5. Conclusion

To the best of our knowledge, this is the first report outlining a completely digital workflow for fabricating customized, clinically-ready, CaP scaffolds for implantation into alveolar defects of varying morphologies. Collectively, our findings have established a strong foundation for future studies to evaluate how to streamline this personalized dental regenerative medicine approach among clinicians and laboratory personnel to treat large alveolar and craniofacial bone defects. Worth noting, as CaP scaffolds are shown to be conducive to cell attachment, they could be used as carriers for the delivery of stem cells in cell-based therapeutics for regeneration of severe alveolar defects.