Abstract
Purpose: Scientific evidence pertaining to the evaluation of trueness of occlusal splints fabricated using different three-dimensional (3D) printers and build orientations compared to subtractive technologies is lacking.
Methods: Overall, one hundred and ten occlusal splints were manufactured using two different 3D printers and a dental mill. Five groups of ten were fabricated using the 3D printers at different build orientations (0, 30, 45, 60, and 90 degrees). In addition, a comparison group of ten occlusal splints was subtractively manufactured using a five-axis dental mill. All occlusal splints were scanned and exported as a standard tessellation language file. Analysis was conducted with metrology software with root mean square estimate average positive deviation and average negative deviation used as the measured outcome.
Results: The 0 degree printing orientation was the most accurate for printer one with the root mean square value of 0.05 ± 0.01 mm, and 60 degree printing orientation was most accurate for printer two with the RMS value of 0.11 ± 0.01 mm. Subtractively manufactured occlusal splint had significantly higher trueness with the lowest RMS value of 0.03 ± 0.05 mm.
Conclusion: Build orientations influence the trueness of additively manufactured occlusal splints while occlusal splints produced by subtractive manufacturing were statistically significantly more accurate.
Introduction
With the advent of digital technologies, the development of a digital workflow has revolutionised restorative and prosthetic dentistry [1]. Computer aided design (CAD) software can directly process digital models, and the laboratory workflow can be performed in a fully digital environment, allowing for the fabrication of different types of prostheses including occlusal splints (OS) [2]. Computer aided manufacturing (CAM) includes subtractive manufacturing (SM), and additive manufacturing (AM) often referred to as three-dimensional (3D) printing [3]. SM utilizes computer-assisted machine tools to mechanically mill a sintered or pre-sintered material in defined paths, to achieve the desired geometry in a wet or dry environment [4,5]. SM manipulates a block or puck of material through a series of digitally controlled cutting patterns reducing the block to desired shape. Whilst it has been shown that these processes can reduce the time involved compared to conventional methods, SM is inherently wasteful due to the reductive nature of the process and the use of cutting/grinding burs in the process [4].
AM is becoming an established technology in the dental field due to its wide range of applications [5]. AM allows for fabrication of an object, producing fine details, by adding materials layer by layer deposition, based on a computerized model [4,5]. AM is now used for various applications in dentistry including cast patterns, dental casts, OS, orthodontic appliances, surgical guides, and temporary restorations [5]. AM has the benefits of reducing material waste, lowering consumption of energy, minimizing the steps required to reach the final product and having predictable costs without compromising detail [4,6]. The most common AM processes used in prosthodontics currently are stereolithography (SLA) and digital light processing (DLP) which are forms of vat polymerization [7,8]. Vat photo polymerisation 3D printers fabricate the desired object by exposing a tank of liquid resin to ultraviolet light and forming layer by layer which join as a build plate rises [9]. Several biocompatible resin materials, such as KeySplint Soft (Keystone industries, Gibbstown, NJ, USA), have been developed for 3D printing of OS.
In comparisons of the trueness and clinical acceptance of OS for injection moulding, SM and AM demonstrated that SM was the most accurate and clinically acceptable group [10]. Superior accuracy of SM technologies compared to AM or conventional manufacturing methods has also been observed in denture bases [11,12]. Accuracy consists of both trueness and precision. Trueness describes the closeness of the arithmetic mean of test values to that of a true or accepted reference value, whilst precision describes the closeness of agreement between the test values [11]. No studies have investigated the trueness of OS materials printed at different orientations for AM technologies in comparison to SM technologies.
Scientific evidence on the influence that the printing orientation has on the accuracy of OS is limited. An in vivo investigation using 0, 30, and 90 degrees demonstrated that lower printing orientations were observed to display superior fit clinically [13]. Investigation of the effect that volumetric changes and orientation have on the accuracy of an occlusal splint has also been conducted where no effect on trueness was observed [14]. This study investigated the effect of printing orientation at 45 degrees with the posterior portions of the occlusal splint rotating away from the build plate as the variable in both groups. This study showed that the orientation the splints were manufactured at had a significant effect on accuracy, however no significant changes in accuracy were observed when the volume of the OS was different with the two samples used. This would suggest that further investigation with varying printing orientation is warranted but not different volumetric changes in OS. Investigations for other dental AM applications can provide an insight to the effect that printing orientation can have on accuracy and can help to inform experimental design. Recent studies have suggested that for interim restorations fabricated with AM, build directions and printing orientations, and build angles and support configurations may impact on both the compressive strengths and trueness of the material [15]. The evidence that OS trueness is dependent on printing orientations is limited, however, it has been observed in several other dental devices [16,17,18]. The effect of the printing orientations has been observed to have significant effects on accuracy for denture bases where angulations of 45 degrees are optimal [19]. This phenomenon has also been observed for surgical guides, provisional crowns, and dental models [16,20,21]. With the effect on accuracy that the change in printing orientation has on other dental applications, comprehensive investigations for OS are warranted. Literature on the trueness of OS fabricated by AM with different orientations and SM technologies may be lacking due to the recent introduction of biocompatible photopolymers approved for long-term clinical use by a variety of manufacturers.
The objective of this in vitro study was to assess the effect of build orientation on the trueness of the intaglio surface of an OS fabricated by AM technologies, using a biocompatible resin material, KeySplint Soft in two different commercial printers. Furthermore, the OS with highest levels of trueness from both the printers were compared with an OS fabricated by SM technology to enable a comparison with more established CAM technologies [9]. The Null hypothesis was that occlusal splints that are manufactured at different printing orientations have the same degree of trueness for multiple printers.
Materials and Methods
A maxillary full-coverage OS was designed using CAD software (Dental System; 3Shape, Copenhagen, Denmark) and exported as a standard tessellation language (STL) file. This OS design is defined as the reference model and was used to manufacture one hundred specimens from a biocompatible 3D printer photopolymer, KeySplint Soft, for occlusal splints using two DLP-AM printers: and Asiga Max UV (Asiga, Sydney, Australia) (P1) and Rapid Shape D30II (Rapid Shape, Heimsheim, Germany) (P2). Eleven groups were manufactured with a sample size of ten. G-power was used to calculate the required sample size. Assuming an effect size of 0.4, a sample size of 110 was considered adequate with a power of 80%, and an alpha error of 5% for comparison between five different orientations with a covariate of manufacturing type (Asiga/ Rapid Shape/ Milling). In addition, this sample size has been shown to be acceptable for in vitro experiments assessing trueness [8,10,12,17,19,22]. The OS was 3D printed at 5 build orientations; 0, 30, 45, 60, 90 degrees with 0 degrees being considered with the occlusal surface being parallel to the build plate on P1 and P2. The STL of the OS was imported into the relevant CAM software and the printing supports generated as per the manufacturer’s predetermined settings. Post processing was conducted according to the manufacturer’s instructions; comprising of cleaning in an isopropanol (>97%) bath for 5 minutes, followed by drying with compressed air, then finally post-curing (Otoflash G171; NK - Optik GmbH, Dietzenbach, Germany) at the 2,000 flashes setting per side in a nitrogen atmosphere. A further ten OS were subtractively manufactured in a five-axis dental mill machine (Programil7, Ivoclar. Schaan, Lichtenstein) as a control group. All OS were maintained in a controlled airconditioned room, then digitized by using a calibrated desktop laser scanner (E3; 3Shape A/S) within 48 hours. Scan spray powder with a particle size of 5 µm (Renfert-Scan spray, Renfert GmbH, Munich, Germany) was applied by an experienced user to aid in the digitization and counteract the translucent reflective nature of the printed and milled OS materials.
The digitized STL file of the scanned surface of the OS was then trimmed with a prototyping design tool (Meshmixer; Autodesk, San Francisco, CA, USA), at the periphery of the OS to include only the intaglio surface. The trimmed STL files of the intaglio surface were then compared with the reference model using metrology software (Geomagic Control; 3D Systems, Rock Hill, SC, USA). The following functions were performed in the software; initial alignment, best fit alignment, then a 3D comparison. Data set outputs were exported into SPSS Statistics (Version 26, IBM, New York, NY, USA), with mean positive deviations, negative mean deviations and root means square estimates (RMS) to interpret the outcomes of the study. These outputs were the determination of the trueness of each sample and thus the overall trueness of the individual groups. This technique has been used to compare the trueness and for the generation of color deviation maps of differing appliances [9,23,24,25,26]. The color deviation maps are used to determine areas of clinically significant change.
RMS deviation data from both P1 and P2 3D printers was tested for normality using Shapiro-Wilk test in relation to build orientation, and Levene’s test was used to test for homogeneity. Shapiro-Wilk test indicated that the data were normally distributed (P > 0.05) and Levene’s test demonstrated that the variances between the groups were equal. Subsequently, a two-way ANOVA was conducted to concomitantly evaluate the influence of the type of printer and build orientation on RMS trueness levels. A one-way ANOVA was also conducted to compare the RMS deviations between the five build orientations for Asiga and Rapid Shape printers separately. Games-Howell and Tukey Honestly Significant Difference tests were conducted for pair-wise comparisons between the build orientations among Asiga and Rapid Shape printers, respectively. A one-way ANOVA was then conducted to compare the RMS deviations between the best build orientations in P1 and P2 printers with the deviations observed from the milling technique. A P value of <0.05 was considered significant.
Results
A total of 110 OS were fabricated, fifty each with P1 and P2 printers by AM and ten using Programil 7 by SM. Results from the two-way ANOVA demonstrated that the type of printer (P < 0.0001), build orientation (P < 0.0001) and their combination (P < 0.0001) affected the trueness of the OS defined by their RMS. Levene’s test of homogeneity was found to be insignificant (P > 0.05). Representative color maps of the surface matching differences of each group are shown in Figs. 1,2,3. The P1 printer showed greater trueness at 0 degrees (Fig. 1A) with most inaccuracies at the outer regions of the OS. The P2 printer showed greatest trueness at a 60 degree printing orientation (Fig. 1B). OC manufactured with SM showed the greatest trueness (Fig. 1C). As the angulations increased, greater deviations were seen in the incisal edges and cusp tips for P2 and more deviations were also seen in the posterior regions in a bucco-lingual dimension (Fig. 2). At 90 degrees, large discrepancies were seen in the anterior region (Fig. 2D). The P2 printer showed increased deviations across all angulations (Fig. 3) with similar findings of increased deviations in the incisal edges and cusp tips as angulations increased, with 90 degrees (Fig. 3D) having the greatest level of deviations.
Table 1 demonstrates that the RMS scores increased as the angulation increased in P1 3D printer, 0 degree has the best trueness (Mean ± SD: 0.05 ± 0.01) while a 90 degree orientation has the highest level of deviation (Mean ± SD: 0.10 ± 0.03). However, the RMS levels improved with the increase in build orientation angulation from 0 degree (Mean ± SD: 0.13 ± 0.01) to 60 degrees (Mean ± SD: 0.11 ± 0.01) in OS printed with the Rapid Shape printer. The highest RMS scores occurred at 90 degrees in both printers. RMS trueness values differed between the build orientations (Fig. 4) that were printed using the P1 3D printer (F = 6.01, P < 0.01; Levene’s test, P = 0.054) as well as the P2 3D printer (F = 4.07, P < 0.007; Levene’s test, P = 0.639). On pair-wise comparison, deviation with 0 degrees was significantly lower than all other build orientations except 30 degree using the P1 3D printer. The P2 3D printer showed that 60 degrees (Mean ± SD: 0.11 ± 0.01) had less deviation compared to that of 90° (Mean ± SD: 0.13 ± 0.02).
SM had high levels of trueness except for the incisal edges and cusp tips (Fig. 1C), showing deviations up to 0.6 mm at these regions. Table 2 demonstrates that the trueness levels of P1 3D printer at 0 degrees (Mean ± SD: 0.05 ± 0.01) was significantly lower than the milling technique (Mean ± SD: 0.03 ± 0.005) but higher than the best trueness levels obtained using P2 3D printer at a build orientation of 60 degrees (Mean ± SD: 0.11 ± 0.01).
Table 1
Comparison of RMS accuracy levels between occlusal devices manufactured using Asiga Max and Rapid Shape D30II printers at different build orientations
| Build angle |
Asiga Mean (SD)* |
Post-hoc results‡ |
Rapid Shape Mean (SD)† |
Post-hoc results§ |
| 0 degree (A) |
0.05 (0.01) |
A < C, D, E |
0.13 (0.01) |
|
| 30 degree (B) |
0.07 (0.03) |
|
0.12 (0.01) |
|
| 45 degree (C) |
0.08 (0.02) |
C > A |
0.11 (0.01) |
|
| 60 degree (D) |
0.10 (0.03) |
D > A |
0.11 (0.01) |
D < E |
| 90 degree (E) |
0.10 (0.03) |
E > A |
0.13 (0.02) |
E > D |
*One-way ANOVA (F = 6.01, P < 0.01); †One-way ANOVA (F = 4.07, P = 0.007); ‡Games-Howell; §Tukey’s post hoc test
Table 2
RMS accuracy levels of the best build orientations in both the printers and those occlusal devices fabricated using milling technique
| Build angle |
Mean (SD)* |
Post-hoc results‡ |
| Asiga 0 degree (A) |
0.05 (0.01) |
A < B; A > C |
| Rapid Shape 60 degree (B) |
0.11 (0.01) |
B < A, C |
| Milling (C) |
0.03 (0.005) |
C > A, B |
One-way ANOVA (F = 146.81, P < 0.0001); ‡Tukey’s post hoc test; Levene’s test (P = 0.069)
Discussion
The objective of this study was to evaluate the effect of the build orientation on the trueness of the intaglio surface of OS fabricated by two validated 3D printers in comparison to subtractive technologies for an OS material. It was found that both the build orientation and the printer type influenced the trueness of AM OS. The Null hypothesis was thus rejected. The best build orientation for the P1 3D printer was found to be 0 degrees whilst for the P2 3D printer was 60 degrees. This contrasts with the manufacturer’s instructions for use, which recommended between 35 and 50 degrees.
This study was motivated by the lack of scientific evidence to guide OS printing orientation. OS are positioned on a long span of hard dentition and a high level of trueness is required. The results from P1 indicate that as the orientation of the print is increased from 0 degrees the level of trueness is reduced. This may be due to the increased number of layers required for higher printing orientations with errors having a cumulative effect on loss of trueness as the OC is 3D printed. The color deviations maps for P1 (Figs. 1A, 2) support this with larger areas of deviations being observed in the posterior regions and anterior incisal edges of the occlusal splint, an in vivo observational study supports these findings [13]. The P2 printer has less trueness compared to P2 overall with the best orientation being recorded at 60 degrees. There was no clear change in printing orientation across the groups (Figs. 1B, 3) however, the 45 and 60 degree orientations appear to be optimal which is supported by studies investigating the accuracy of other 3D printed dental devices [19,20]. The color deviation maps of P2 support the statistical analysis with fewer areas of large deviations in the 45 and 60 degree groups. The difference in results for the two 3D printers may be due to the different technologies present for separation detection technologies that govern when the build plate moves between layers.
The observed differences observed between the two 3D printers can be used by developers and end users. In vitro studies investigating the optimal printing orientation of denture bases show the effect printing orientation can have on trueness. Several recent studies report optimal printing orientations of 45, 60, and 75 degrees despite utilizing similar shapes and materials to 3D print [17,19]. These studies had similar methodologies but used different types of 3D printers as well as different types of resins, which may account for the differences. However, it highlights the fact that different 3D printers will manufacture optimally at different printing orientations. The different areas of deviation observed in the color deviation maps further highlights this with most distortions in trueness being observed in the anterior region for P1 (Figs. 1A, 2) while P2 (Figs. 1B, 3) showed more deviations in the palatal aspect of the OS. The results of this study show that for the same type of resin, different printers appear to influence the trueness of appliances fabricated. Therefore, caution should be exercised with claims of accuracy for photopolymers that utilize a range of 3D printers.
The comparison between the best AM groups and the SM group demonstrates the differences in the technology. SM was the truest group of all the groups analysed which is due to the highly automated process that involves fewer variables when compared to AM during manufacturing [4,12,26]. This is supported with the color deviation maps showing areas that lack trueness are the cusps of the posterior teeth and incisal edges of the anterior teeth. These areas would have been too narrow for the smallest milling bur (0.5 mm) to negotiate, therefore a milling compensation would have been built into the milling strategy, calculated by the mills CAM software [5]. This phenomenon was also observed by Wesemann et al. (2021) who reported that the accuracy of SM OS was the most accurate [9]. Therefore, it is postulated that OC manufactured via SM should be best practice, however, the deviations observed with the best groups of the AM OC remain clinically acceptable. Any further in vitro studies assessing the AM OC would benefit from including a SM group as a ‘gold standard’; for comparison.
There are several limitations to this study. The use of scan spray powder coating or extraoral scanning of the splint surface may affect the assessment of the trueness. However, this was necessary to scan the clear material with an optical scanner. The effect of scan spray powder has been documented to be in the magnitude of 13-16 μm and thus would not have a significant effect on the outcomes of the study [23,27]. A greater range of OC designs on different anatomical models would help to generate more clinically significant results, however, given the results of Reich et al. (2022) which showed different volumes had no significant difference, this may not be warranted [14]. A comparison group investigating conventional approaches to OS manufacture could be included, however, given that recent in vitro studies have shown that SM techniques are on par, this was excluded from the experimental design for this study [9]. The assessment of the OS in an in vivo context may strengthen the outcomes of this study. Future directions for the assessment of accuracy of OS should include the assessment of a greater range of printing orientations and a wider range of 3D printers to determine if the observed differences between the two 3D printers are not a specific phenomenon with the 3D printers included in this study. It may be recommended for material manufacturers to assess the accuracy for each combination of 3D printer and photopolymer resin to determine the best build orientation with regards to trueness as the result of this study indicates that 3D printers may perform differently for different 3D printers. The results of this study indicate that the build orientation is important for AM manufacture of OS and that SM gives the truest replication of OS.
Conflict of Interest
The authors declare no conflict of interest.
Acknowledgments
The authors thank Keystone Industries and Dentona for the supply of materials.
References
- 1. Ahmed KE (2018) We’re going digital: the current state of CAD/CAM dentistry in prosthodontics. Prim Dent J 7, 30-35.
- 2. Tallarico M (2020) Computerization and digital workflow in medicine: focus on digital dentistry. Materials (Basel) 13, 28-33.
- 3. Sulaiman TA (2018) Materials in digital dentistry - a review. J Esthet Restor Dent 32, 171-181.
- 4. Van Noort R (2012) The future of dental devices is digital. Dent Mater 28, 3-12.
- 5. Zimmermann M, Valcanaia A, Neiva G, Mehl A, Fasbinder D (2020) Influence of different CAM strategies on the fit of partial crown restorations: a digital three-dimensional evaluation. Oper Dent 43, 530-538.
- 6. Alharbi N, Wismeijer D, Osman RB (2017) Additive manufacturing techniques in prosthodontics: where do we currently stand? A critical review. Int J Prosthodont 30, 474-484.
- 7. You SG, You SM, Kang SY, Bae SY, Kim JH (2020) Evaluation of the adaptation of complete denture metal bases fabricated with dental CAD-CAM systems: an in vitro study. J Prosthet Dent 125, 479-485.
- 8. You SM, You SG, Kang SY, Bae SY, Kim JH (2021) Evaluation of the accuracy (trueness and precision) of a maxillary trial denture according to the layer thickness: an in vitro study. J Prosthet Dent 125, 139-145.
- 9. Wesemann C, Spies BC, Schaefer D, Adali U, Beuer F, Pieralli S (2021) Accuracy and its impact on fit of injection molded, milled and additively manufactured occlusal splints. J Mech Behav Biomed Mater 114, 104-109.
- 10. Einarsdottir ER, Geminiani A, Chochlidakis K, Feng C, Tsigarida A, Ercoli C (2019) Dimensional stability of double-processed complete denture bases fabricated with compression molding, injection molding, and CAD-CAM subtraction milling. J Prosthet Dent 124, 116-121.
- 11. International Organization for Standardization (1994) Accuracy (trueness and precision) of measurement methods and results—Part 1: Ggeneral principles and definitions. ISO 5725-1:1994, Geneve.
- 12. Goodacre BJ, Goodacre CJ, Baba NZ, Kattadiyil MT (2016) Comparison of denture base adaptation between CAD/CAM and conventional fabrication techniques. J Prosthet Dent 116, 249-256.
- 13. Vasques MT, Laganá DC (2018) Accuracy and internal fit of 3D printed occlusal splint, according to the printing position. Clin Lab Res Dent 1, 1-6.
- 14. Reich S, Berndt S, Kühne C, Herstell H (2022) Accuracy of 3D-printed occlusal devices of Different volumes using a digital light processing printer. Appl Sci 12, 1576-1581.
- 15. Alharbi N, Osman R, Wismeijer D (2016) Effects of build direction on the mechanical properties of 3D-printed complete coverage interim dental restorations. J Prosthet Dent 115, 760-767.
- 16. Yu BY, Son K, Lee KB (2021) Evaluation of intaglio surface trueness and margin quality of interim crowns in accordance with the build angle of stereolithography apparatus 3-dimensional printing. J Prosthet Dent 126, 231-237.
- 17. Yoshidome K, Torii M, Kawamura N, Shimpo H, Ohkubo C (2021) Trueness and fitting accuracy of maxillary 3D printed complete dentures. J Prosthodont Res 65, 559-564.
- 18. Chaiamornsup P, Iwasaki N, Tsuchida Y, Takahashi H (2021) Effects of build orientation on adaptation of casting patterns for three-unit partial fixed dental prostheses fabricated by using digital light projection. J Prosthet Dent 128, 1047-1054.
- 19. Cameron AB, Evans JL, Abuzar MA, Tadakamadla SK, Love RM (2022) Trueness assessment of additively manufactured maxillary complete denture bases produced at different orientations. J Prosthet Dent, Mar 5, doi: 10.1016/j.prosdent.2021.12.025.
- 20. Dalal N, Ammoun R, Abdulmajeed AA, Deeb GR, Bencharit S (2020) Intaglio surface dimension and guide tube deviations of implant surgical guides influenced by printing layer thickness and angulation setting. J Prosthodont 29, 161-165.
- 21. Yousef H, Harris BT, Elathamna EN, Morton D, Lin WS (2021) Effect of additive manufacturing process and storage condition on the dimensional accuracy and stability of 3D-printed dental casts. J Prosthet Dent 128, 1041-1046.
- 22. Chen L, Lin WS, Polido WD, Eckert GJ, Morton D (2019) Accuracy, reproducibility, and dimensional stability of additively manufactured surgical templates. J Prosthet Dent 122, 309-314.
- 23. Lehmann KM, Azar MS, Kämmerer PW, Wentaschek S, Hell EN, Scheller H (2011) The effect of optical conditioning of preparations with scan spray on preparation form. Acta Stomatol Croat 45, 20-26.
- 24. Alharbi N, Osman RB, Wismeijer D (2016) Factors influencing the dimensional accuracy of 3D-printed full-Coverage dental restorations using stereolithography technology. Int J Prosthodont 29, 503-510.
- 25. Hada T, Kanazawa M, Iwaki M, Arakida T, Soeda Y, Katheng A et al. (2020) Effect of printing direction on the accuracy of 3D-printed dentures using stereolithography technology. Materials (Basel) 13, 3405-3409.
- 26. Zafar MS (2020) Prosthodontic applications of polymethyl methacrylate (PMMA): an update. Polymers (Basel) 12, 2299-2304.
- 27. Burde AV, Dudea D, Cuc S, Moldovan M, Campian RS (2016) Three-dimensional evaluations of the coating thickness of two optical conditioning scanning sprays. Mat Plas 53, 65-67.
