Osseointegration of a 3D Printed Stemmed Titanium Dental Implant: A Pilot Study

In this pilot study, a 3D printed Grade V titanium dental implant with a novel dual-stemmed design was investigated for its biocompatibility in vivo. Both dual-stemmed (n = 12) and conventional stainless steel conical (n = 4) implants were inserted into the tibial metaphysis of New Zealand white rabbits for 3 and 12 weeks and then retrieved with the surrounding bone, fixed, dehydrated, and embedded into epoxy resin. The implants were analyzed using correlative histology, microcomputed tomography, scanning electron microscopy (SEM), and transmission electron microscopy (TEM). The histological presence of multinucleated osteoclasts and cuboidal osteoblasts revealed active bone remodeling in the stemmed implant starting at 3 weeks and by 12 weeks in the conventional implant. Bone-implant contact values indicated that the stemmed implants supported bone growth along the implant from the coronal crest at both 3- and 12-week time periods and showed bone growth into microporosities of the 3D printed surface after 12 weeks. In some cases, new bone formation was noted in between the stems of the device. Conventional implants showed mechanical interlocking but did have indications of stress cracking and bone debris. This study demonstrates the comparable biocompatibility of these 3D printed stemmed implants in rabbits up to 12 weeks.


Introduction
Bone-anchored implants have been a standard treatment for edentulism since the mid-1980s, after Per-Ingvar Brånemark demonstrated the successful osseointegration of a titanium dental implant placed in human patients [1]. However, failure rates for clinical use dental implants range between 3 and 8% depending on the implant design and/or patients' health factors [2][3][4][5]. Although this appears to be a rather successful procedure, an epidemiological study reported 1200 emergency department visits due to dental implant failures from 2008 to 2010 in the US alone [6], signifying the continued burden of edentulism on the healthcare system. As such, methods to improve the clinical outcomes of dental implants are still actively pursued.
Due to its bone-bonding or osseointegrative ability, mechanical and chemical properties, and overall biocompatibility, titanium and titanium alloys have long been the dental implant material of choice [7][8][9]. Recently, considerable emphasis has been placed on surface treatment of implants, where surface roughness and texture modications have been shown to facilitate bone integration and cellular activity via microscale and nanoscale features [10][11][12][13]. In addition, a range of surface coatings, such as calcium phosphate, magnesium, and titania, have been explored with the intent of encouraging faster osseointegration [14][15][16]. While it is known that implant geometry can change the response of the bone-implant interface under loading [17], conventional machining processes have traditionally limited implant morphologies to conical-and screw-like designs.
However, with the technological advances in additive manufacturing, 3D printing of titanium and titanium alloys for new and innovative implant geometries are now possible. Additive manufacturing techniques, such as direct metal laser sintering (DMLS) and electron beam melting (EBM), are processes that can create three-dimensional metallic constructs by selectively melting metal powder in a layer-bylayer fashion. ese techniques are capable of creating complex porous features [18][19][20] and an inherent surface roughness, as the melted powder droplets solidify on the object surface. Current use of this technology for implant manufacturing has focused on creating open pore networks to mimic the trabecular bone, showing improved cellular activity and greater bone ingrowth in both rabbit and sheep models [21][22][23].
In this pilot study, the osseointegration of a dental implant with a novel dual-stemmed shape, enabled by DMLS, was assessed for the rst time. We present complementary histology, high-resolution X-ray, and electron microscopy analyses to investigate the bone-implant interface.

Implants.
Twelve dual-stemmed implants (herein referred to as the SIT implant) were produced via DMLS using an EOSINT M 280 (EOS GmbH, Krailling, Germany) with Ti6Al4V powder and were received from Stemmed Implant Technologies Inc. (Niagara Falls, Canada). Implants were 3 mm in diameter, with 1 mm diameter stems. In preparation for implantation, implants were cut into 6 mm in length, brie y sandblasted with 70 psi and a 90% glass bead/10% Al 2 O 3 media, and autoclaved. Final implants consisted of a body and stems, both 3 mm in length (Figure 1(b)). Four conical stainless steel mini-implant screws, with a tapered body and maximum of 2 mm diameter, were used as controls. e control implants were received from Stemmed Implant Technologies Inc., cut into a 6 mm length to match the length of the stemmed implants (Figure 1(a)), and autoclaved for sterilization prior to the implantation procedure. At higher magni cation, it is clear that the 3D printed surface (f) retains characteristic surface features representative of the powders used in its production and has a much higher roughness than conventionally machined implants (Figures 1(b), 1(c), 1(e), and 1(f)).  (e, f) SEM images of SIT implant surface. At higher magni cation, it is clear that the 3D printed surface (f) retains characteristic surface features representative of the powders used in its production and has a much higher roughness than conventionally machined implants (c).

Implant
from the McMaster Animal Research Ethics Board. e day prior to surgery and over the following four days, all rabbits received 10 mg/kg enro oxacin (Baytril ® , Bayer, Leverkusen, Germany) to prevent infection. During surgery, the rabbits were induced with xylazine, ketamine, and acepromazine, intubated, and placed on iso urane gas 2-3% inhalation with oxygen. Buprenorphine was administered at 0.5 mg/kg subcutaneously to prevent pain and readministered every 12 hours for 48 hours postoperatively. e surgical method was a cranial medial approach parallel to the tibial crest with a slow-rotating drill and irrigation with saline ( Figure 2). e implant bed therefore consisted of cylindrical holes through the cortical bone slightly smaller than the diameter of the implants. One implant of each type was inserted into the tibial metaphysis of each rabbit, where stemmed implants were pressed to t and conical implants were screwed in until ush with the bone crest. Incisions were closed with a layer of 4-0 Vicryl ® (Ethicon, Inc., Somerville, USA), and the skin was closed with stainless steel wound clips. Animals were then randomly split into two groups: one provided a 3-week healing period, and the other, 12 weeks. After the healing period, rabbits were euthanized by overdose of barbiturate. However, one animal at each time point was perfused with 2% glutaraldehyde in 0.1 M sodium cacodylate bu er solution, and bone-implant sections were removed for decalci cation with EDTA for histology.
Implants with surrounding bone tissue were collected and prepared into implant-bone blocks following the methodology for preparing undecalci ed bone outlined by Donath and Breuner [24]. Approximately 2 cm × 2 cm bone blocks containing the implants were xed in a solution of 1% glutaraldehyde and 1% paraformaldehyde in 0.1 M sodium cacodylate bu er for 7-10 days and subsequently dehydrated in a graded series of ethanol, followed by embedding in LR white acrylic resin for the 3-week specimens and Embed-812 epoxy resin for the 12-week specimens. Blocks were longitudinally sectioned using an Isomet ® low-speed saw (Buehler, Lake Blu , USA) and a diamond wafer blade to reveal the bone-implant interface.

Histology.
Rabbits were overdosed with 65 mg/kg body weight sodium pentobarbital I.V. and perfused via the left cardiac ventricle with 1 L lactated Ringer's solution followed by 1 L formalin (10% paraformaldehyde in phosphate bu er, pH 7.2). e fragment of the tibial bone with the metal implant was carefully removed, post xed in formalin for two days, and then placed in formalin supplemented with 4% EDTA (pH 7.2) for demineralization. Demineralizing solution was exchanged once per week over nine months. e metal implants were carefully removed from the soft bone, and the bone was sectioned for analysis at the site of the implants. e sections were dehydrated with a series of graded concentrations of ethyl alcohol (50-100%) and xylene, embedded in para n wax, cut into 5 µm thick, mounted onto glass slides, and stained with both hematoxylin and eosin. e histological analysis was performed under an Eclipse 50i light microscope (Nikon, Tokyo, Japan).

Microcomputed Tomography.
Visualization of whole implant-bone ingrowth was achieved using a SkyScan 1172 (Bruker, Billerica, USA) with a 100 kV X-ray beam, aluminiumcopper lter, 2.3 µm-2.6 µm pixel size, and 0.3-1°rotation step. NRecon and CTAn software (Bruker, Billerica, USA) were used to reconstruct and visualize the 3D volumes. e length of bone growth was measured using the image processing and analysis software ImageJ (National Institutes of Health, Bethesda, USA).

Scanning and Transmission Electron Microscopy.
Longitudinal implant-bone blocks were sputter coated with gold and imaged with a JSM-6610LV (JEOL Ltd., Tokyo, Japan) scanning electron microscope (SEM) at an accelerating voltage of 10 kV. Backscattered electron (BSE) images with compositional contrast enabled identi cation of regions of new bone growth along the implant length. Transmission electron microscopy (TEM) specimens were prepared using an in situ lift-out method ( Figure 3) on a NVision 40 (Carl Zeiss GmbH, Germany), a dualbeam instrument comprising a focused ion beam (FIB) milling instrument and a Schottky eld emission gun (FEG) lament SEM. Due to implant-bone separation caused during retrieval and sample preparation, an intact boneimplant specimen was not possible. However, the interface between old and new bone was successfully prepared for high-resolution analysis. TEM images were captured using a Titan 80-300 (FEI, Oregon, USA) operated at 300 kV with a high-angle annular dark-eld detector.

3.1.
Histology. Histological analysis of both SIT implants and control screw implants was completed after 3 and 12 weeks of implantation to determine the cellular activity and remodeling behaviour of the bone tissue with the implanted devices ( Figure 4). For both the SIT and control groups, implants resided primarily in the cortical bone. Following extraction of the control implant after 3 weeks, bone debris was present between the implant and the cortical bone (Figure 4(a)). is is in contrast to the cortical bone surrounding the SIT implant after 3 weeks, where there was no debris and the presence of multinucleated osteoclasts and hypertrophied osteoblasts suggested that the bone was being  (Figure 4(c)), meanwhile, the SIT implant was completely encased in the cortical bone, with a layer of bone forming over its surface. e remodeling of the bone was still active but showed the morphology of more mature bone (Figure 4(h)).

Microcomputed Tomography.
Prior to sectioning for SEM, the entire implant-bone blocks of the SIT and control implants were imaged by microcomputed tomography (µ-CT). Radiographs of both implant types revealed the top portion of the implants to be surrounded by the cortical bone with the remainder of the implant located in the medullary cavity ( Figures 5(a), 5(e), 6(a), and 6(e)). e SIT and control implants were shown to have new bone growing from the preexisting cortical bone, down and around each implant surface and into the medullary cavity, after 3 and 12 weeks (Figures 5 and 6). ree-dimensional renderings ( Figures 5(b), 5(f), 6(b), and 6(f)) of both implants provided a holistic perspective of the entire implant and surrounding bone volume. e new bone growth around the implants is simultaneously visualized with the growth down the implant length. e location of the reconstructions shown in e new bone accounted for 25% and 50% of the total bone length residing along the control and SIT implant surfaces, respectively, at 3 weeks (Figure 7). After 12 weeks, the new bone accounted for 35% and 55% of the total bone length residing along the control and SIT implant surfaces, respectively (Figure 7). While the bone growth down the SIT implant surface was greater than that down the control implant surface at both time points, a di erence in the extent of radial bone growth between the control and SIT implants was less evident. Qualitative results suggest that over the same time period, bone grows similarly around the SIT implant compared to controls.

Scanning Electron Microscopy.
Similar to the micro-CT results, SEM images did not show a trabecular bone transition underlying the cortical bone, which indicated some misplacement of the implant o the target anatomical position. As such, bone contact was only possible originating from the cortical bone crest. Imaging of the embedded sections with SEM enabled qualitative assessment of the bone-implant contact in this cortical region. ree weeks after implantation, the cortical bone was present within the threads of the control implant (Figure 8(a)).
is mechanically interlocked bone was in contact with the control implant, while the new bone further down the length of the implant was primarily not in direct contact. However, stress cracks were observed at the mechanically interlocked thread tips. In contrast, little to no bone was in contact with the SIT implant after 3 weeks (Figure 8(c)). e absence of threads also indicates a lack in mechanical interlocking. e bone structure around the SIT implant specimens appeared less developed with more porosity and randomly oriented osteocyte lacunae; however, in some cases, new bone formation was observed in between the stems of the SIT implant. Twelve weeks post implantation, the bone surrounding both implant types was more developed and in greater contact with the implant surfaces (Figures 8(b) and 8(d)). e arrow in Figure 8(d) points to bone growth within the microporosities of the SIT implant suggesting improved osteoconduction. As with the 3-week samples, stress cracks were also present within the bone from the 12-week control implant.

Transmission Electron Microscopy.
To fully assess the osseointegration between the bone and implant and the quality of bone tissue at the interface, higher spatial resolution than that achieved by SEM is required. Figure 9(a) shows a high-angle annular dark-eld (HAADF) image of the SIT implant-bone interface after 12 weeks of healing. A separation at the bone-implant interface, likely due to mechanical stresses during removal and resin in ltration,  was exaggerated by the FIB during TEM sample preparation. However, the matching contours of the bone and implant surface indicate that the implant and bone were likely in complete contact prior to retrieval. Preparation of a TEM specimen for the SIT implant at 3 weeks was not possible because of a lack of contact at the bone-implant interface. However, a specimen of the bone near the implant interface was removed for TEM. e di erence in bone quality near the implant surface at 12 and 3 weeks is shown in Figures 9(b) and 9(c), respectively. e collagen bers of the 12-week bone are more organized compared to the woven collagen bers and visible mineral clusters of the 3-week bone, highlighting the di erences in bone maturity.

Discussion
A novel implant design by Stemmed Implant Technologies Inc. is marked by a signi cant geometrical change that employs dual prong-like stems when compared to conventional implants that are generally conical threaded screws. e unique shape of this implant was achieved by the layer-by-layer, bottom-up approach of DMLS. It has been  International Journal of Dentistry proposed that the new SIT implant will better resist rotational forces experienced during mastication and bruxism, as well as reducing the amount of bone removed and damaged during surgery and insertion. However, it is important to note that the authors did not evaluate any of these claims in this study. is pilot study aimed to understand the bone-implant interactions of the SIT implant and to predict its potential success in clinical implant scenarios. is was conducted with histology, X-ray, and electron microscopies. While the nature of a pilot study limits this work to a small sample size, the reported ndings provide an initial assessment of the biocompatibility of the SIT implant and demonstrate its potential for further animal and clinical studies.
Histological analysis showed active bone growth and remodeling for both the SIT and control implants via osteoclast-and osteoblast-mediated bone matrix resorption and deposition. ese observations are similar to previous histomorphometric evaluations of other DMLS implants placed in both sheep and humans [25,26]. Early bone formation is evident by the presence of osteoblasts connected to the newly formed bone (Figure 4(e)).
e observation of bone debris for the control implant at 3 weeks and predominant osteoclast activity at 12 weeks compared to the SIT implant suggests a potential di erence in the rate of bone formation and remodeling between the implants. Bone debris in the peri-implant space at early healing time points has been observed previously with threaded implants [26,27] and may lead to delayed bone formation compared to implants devoid of threads [28]. Histological analysis indicates that the SIT implant shows comparable osseoconduction to conventional implants after 12 weeks, marked by complete bone encasement and active remodeling. e nondestructive basis of micro-CT has been demonstrated as a useful tool for visualizing the entire implant and bone volume in two and three dimensions. Contrast gradients enable di erentiation of new bone from old bone and identi cation of active remodeling sites, as well as sites lacking bone and osseointegration. ese micro-CT results indicate that the implants were only anchored in the cortical bone, despite the usually large amount of the trabecular bone present in the metaphysis of rabbits where implants were placed, indicating a potential misplacement. us, this study is limited to the evaluation of cortical bone only. A fairly small amount of bone growth was conducted from the cortical bone crest down the implant length for the control implant compared to a slightly larger amount on the SIT implant that was even clearly visible between the implant stems at 3 weeks. Ideally, placement in the trabecular bone would maintain bone trabeculae between the stems for added stability. e reason for the observed di erence in bone growth is somewhat unclear due to the potential interplay between di erences in both the osteoconductivity of titanium and steel and the radial bone growth required for the control implants due to the threaded design. By three weeks, the majority of the bone volume, which was to encapsulate the implants, had been deposited and was remodeled into more mature dense bone, but not a greater quantity, by 12 weeks.
While micro-CT is ideal for a general overview of bone growth, it lacks the resolution necessary to visualize submicron features at the bone-implant contact. e greater extent of bone growth from the cortical region down the implant surfaces for the SIT implant was con rmed by SEM.
e SIT implant conducted bone growth along its length and within the stems (Figure 8). e control implant initiated a limited amount of new bone formation, and large cracks were present within the cortical bone, perhaps caused by overtorquing during implant placement. Very tight integration, with no separation, was seen around the SIT implant, as the bone had grown into the micropores of the implant surface, which is an indicator of biocompatibility. Comparison of the bone-implant contact across studies remains challenging because of a lack of standardized methodologies employed to model bone growth and measure the bone-implant contact (BIC). Animal model selection, bone type, surgical procedure, heal time, sample preparation methodology, and selected implant length for BIC measurements varies across studies, all of which can in uence the BIC [16,26,29,30]. Nevertheless, previously reported BIC measurements of machined and analogous 3D printed implants after 2 weeks were 20% for both implant types [26]. In this study, we instead looked at the conduction of bone growth down the implant surface, since it was placed primarily in the cortical bone, and found as expected that the titanium SIT implant was a better conductor of bone growth at both early and late time points.
Due to bone-implant interface separation caused by FIB sample preparation, the exact integration between the bone and the SIT implant could not be analyzed; however, the maturity of the bone surrounding the implant could be evaluated to demonstrate the success of bone growth at the implantation site. TEM imaging of the lift-outs revealed di erences in the orientation of the collagen brils and the presence of mineral clusters after 3 and 12 weeks. is suggests that the mechanism of distant osteogenesis is occurring during healing after the insertion of the implants. In distant osteogenesis, mature bone acts as a substrate for osteogenic cells to form a matrix that gradually encroaches upon the implant surface [31].
is pilot study was limited to an investigation of the structural and biochemical interaction of the implant device in vivo via advanced imaging modalities. To further validate these results, future work should focus on determining the mechanical integrity of the bone-implant interface. Mechanical testing of the implanted devices would also be bene cial to improving the understanding of the overall system and the potential advantages to using additive manufacturing as a production method for dental implants.
is could potentially be completed in vivo through methods such as resonance frequency analysis to determine implant stability [32] and via pullout tests to con rm adequate mechanical strength [33]. Complementary information from in vitro testing, such as investigating cell viability [34], may provide additional insight into the biocompatibility of the device. Other works reporting 3D printed implant devices have shown promising cell viability and biocompatibility [35][36][37].
International Journal of Dentistry 9

Conclusions
Additive manufacturing provides a means for innovative dental implant designs with inherent surface features which facilitate bone integration. Initial observation of a dualstemmed 3D printed dental implant has shown successful bone growth and bone-implant contact similar to conventional and other 3D printed implants up to 12 weeks of healing in rabbits. In some cases, new bone formation was noted in between the stems of the device, although the stems were not within a trabecular bone region. Conventional implants showed mechanical interlocking but did have indications of stress cracking and bone debris. is pilot study demonstrates that this 3D printed implant design is biocompatible, as it allows for successful osseointegration in rabbits up to 12 weeks, and supports additional studies to obtain more statistical validation, including mechanical testing.

Conflicts of Interest
e authors declare no con icts of interest.