- Email: [email protected]
Accuracy of 3-dimensional computer-aided manufactured single-tooth implant definitive casts
RESEARCH AND EDUCATION
Accuracy of 3-dimensional computer-aided manufactured single-tooth implant definitive casts Michele Buda, DDS, MSD,a Manuel Bratos, DDS, MSD, CDT,b and John A. Sorensen, DMD, PhD, FACPc
ABSTRACT Statement of problem. The integration of the digital workflow into routine prosthodontic practice for single-tooth implant surgery and fixed prosthesis fabrication has occurred at a remarkable pace in the last 5 years. With the greater demands of esthetics and precision, the definitive implant analog cast must ensure accurate implant positioning as well as an accurate relationship to adjacent teeth. Purpose. The purpose of this in vitro study was to evaluate the accuracy of the 3-dimensional (3D) implant position of definitive casts produced by 3D printing and analog technology. Material and methods. A master patient model was created from a dentate typodont. The maxillary left central incisor was removed, and a Straumann RC implant was positioned for a screw-retained prosthesis. A laboratory scanner with an accuracy of 5 mm was used for all scanning. A scanbody was connected to the master model implant and scanned to create a master patient file, which served as the control master patient for all comparisons. The two 3D printing systems used for this study were the Statasys Objet500 (group OBJ), an industrial Polyjet production system, and the Formlabs Formlab 2 (group FORM), a budget SLA Vat system. In addition, a conventional gypsum cast (group GYP) with an implant analog was made with elastomeric impression material. With a sample size of 10 per group, each gypsum cast and 2 printed group casts were scanned with the D2000 laboratory scanner 5 times per cast. Convince software (3Shape) was used for 3D analysis to calculate accuracy. The following variables were measured: implant analog vertical displacement, horizontal displacement of implant platform and apex, degree of tilting in the vertical axis, and rotational position change around the vertical axis. Means and standard deviations were calculated for trueness. One-way ANOVA and the post hoc t test with Bonferroni correction were used to investigate any significant differences among the experimental groups (a=.05). Results. For vertical displacement of the implant body, group OBJ had the lowest value of e30 ±24 mm. The values obtained for OBJ and FORM were significantly different from that obtained for GYP (P<.05). For horizontal displacement of the implant shoulder, Group OBJ had the lowest value, 85 ±12 mm, and the difference among these groups was significantly different (P<.05). The value for horizontal displacement of the implant apex was 123 ±25 mm for group OBJ and not significantly different from that obtained for group GYP (136 ±40 mm) but significantly different from that obtained for group FORM (326 ±54 mm). Also, the analysis of implant body tilting in the vertical axis showed significant differences between the values obtained for groups GYP and OBJ and between the values obtained for groups OBJ and FORM. With regard to implant rotational position change around the vertical axis, the values obtained for the gypsum cast and group FORM were not statistically different from those obtained for the master patient control model (P>.05). However, the implant orientation of group OBJ was significantly different from the orientation of groups GYP and FORM (P<.05). The actual clinical relevance of these printing system discrepancies is yet to be determined because the level of clinical acceptable discrepancy in the x, y, and z vectors is still undefined. Conclusions. This study showed statistically significant differences in accuracy among the implant analog cast fabrication systems; however; the level of clinical acceptable discrepancy is still undefined. Although further research is needed, this study supports the conclusion that the Polyjet industrial printing system was more accurate than the conventional implant analog gypsum cast. (J Prosthet Dent 2018;-:---)
Supported by Tylman Research Grant 2016-2017, American Academy of Fixed Prosthodontics. a Former postgraduate student and graduate, Department of Restorative Dentistry, School of Dentistry, University of Washington, Seattle, Wash; and Private practice, Treviso, Italy. b Affiliate Lecturer, Department of Restorative Dentistry, School of Dentistry, University of Washington, Seattle, Wash. c Professor, Department of Restorative Dentistry, Associate Dean for Graduate Studies, and Director, Biomimetics Biomaterials Biophotonics Biomechanics & Technology Laboratory, School of Dentistry, University of Washington, Seattle, Wash.
THE JOURNAL OF PROSTHETIC DENTISTRY
1
2
Volume
Clinical Implications The industrial 3-dimensional printing system proved more accurate than conventional elastomeric impressions and gypsum casts. Currently the cost of these high-quality printing systems is excessive except for large commercial laboratories. As with all technology, capital costs will decrease, making them affordable and sufficiently precise.
Applications of computer-aided design and computeraided manufacturing (CAD-CAM) to implant prosthodontics have been increasingly used to fabricate implant frameworks and monolithic restorations.1,2 Without an intraoral scanner, the gold-standard method used to record implant location and orientation requires an intraoral impression, from which a definitive cast is obtained.3 The definitive casts are usually scanned with an optical laboratory scanner, and a framework or restoration is then digitally designed and milled by a multi-axis milling machine with different restorative materials.4 Recently, with the rapid development of optical scanning technology, digital impressions have become more popular for implant-supported dental prostheses as well as conventional tooth-supported prostheses.5 A digital impression can be created with 2 different methods6: scanning a conventional definitive cast (indirect technique) or with an intraoral scanner (direct technique). The workflow for fabricating an implantsupported prosthesis could be entirely digital or combined with traditional laboratory procedures.7 When obtained with an intraoral scanner, the position of the implants is transferred to CAD-CAM machinery with a digital standard tessellation language (STL) file to fabricate the prosthesis. In this situation, a physical cast is lacking. Nevertheless, in most situations, physical casts are still needed so that they can be mounted on articulators to evaluate occlusion and adjust interproximal contacts. To allow technicians and dentists to perform these essential steps, 3-dimensional (3D) printed or milled casts can be obtained from the initial STL file. Accuracy describes closeness to the real dimensions of the object. Few studies have investigated the accuracy of 3D-printed or milled casts. Nonetheless, accuracy might depend on both the printing resolution and the 3D positioning of the implant analog in the cast relative to the teeth and soft tissues.8 A consensus is lacking regarding the accuracy of printing alone or printing combined with intraoral scanning for STL data acquisition. The accuracy of digital impressions has been compared with conventional impressions involving dental implants.9,10 Lee et al11
THE JOURNAL OF PROSTHETIC DENTISTRY
-
Issue
-
evaluated the accuracy of milled casts by using a laboratory scanner instead of a coordinate-measuring machine (CMM) to determine the true measurements as the control. Milled casts and gypsum casts were scanned with a laboratory scanner, and STL files were compared with reference file with Geomagic superimposition software. The 3D coordinate axes were defined, and the STL data sets from the gypsum and the digitally milled casts were individually aligned to the reference data set by a repeated best-fit algorithm based on the selected surfaces of the maxillary left first premolar and molar. The authors found a statistically significant difference between the gypsum and milled cast in the vertical axis of the implant position. In previous studies, both linear measurements between casts and CMMs have been used. The limitation of linear measurements is the uncertain replication of references points, whereas CMMs are unable to measure anatomic shapes. Therefore, superimposition of STL files obtained from a high-accuracy scanner offers a more accurate and consistent way to measure differences. The term ‘3D printing’ refers to the process of joining materials together to form objects from 3D model data. It is used to describe a wide variety of technologies known as additive manufacturing. According to the American Society for Testing and Materials, there are 7 categories of additive manufacturing: vat photopolymerization, material jetting, binder jetting, material extrusion, power bed fusion, sheet lamination, and direct energy deposition. Vat photopolymerization and material jetting are used to manufacture polymeric objects used in dentistry. In the vat photopolymerization system, an ultraviolet laser polymerizes a liquid photopolymer. The build platform is lowered into the vat as each layer polymerizes. In material jetting technology, inkjet print heads are used to jet liquid photopolymers onto a build platform. The material is immediately polymerized by ultraviolet lamps and solidified, allowing layers to be built up. In this study, 3D-printed casts were included because they are most often used by dental laboratories and are less expensive to produce than their milled counterparts. The purpose of this study was to compare the accuracy of 3D-printed casts obtained from digital files with conventional gypsum casts with implant analogs. The null hypothesis was that no significant difference would be found in the analog position of the 3D-printed casts and the conventional gypsum casts. MATERIAL AND METHODS A master patient model was created from dentate typodont. The maxillary left central incisor was removed, and a Straumann Bone Level RC implant was positioned for a screw-retained prosthesis (Fig. 1). A laboratory scanner
Buda et al
-
2018
3
Figure 1. Master patient model. A, Fabrication. B, Occlusal view.
(D2000; 3Shape) with an accuracy of 5 mm (ISO 12836) was used for the scanning. A Straumann Mono Scanbody was connected to the master model implant and scanned 20 times (Fig. 2). The average of the 20 STL files was used to create a master patient file and served as the control master patient for all comparisons (Fig. 3). 3Shape Dental Design and Implant Studio software were used to design an implant analog anterior model. Two 3D printing systems were evaluated: the Statasys Objet500 (an industrial Polyjet production system with a vertical layer thickness of 16 mm) and the Formlab 2 (a budget SLA vat system with a vertical layer thickness of 50 mm). These systems were used to fabricate 2 different printed casts (groups OBJ and FORM). A conventional gypsum case with an implant analog was made from an open-tray impression with an open-tray fixture level impression coping made with medium viscosity and heavybody impression material (Virtual; Ivoclar Vivadent AG). Initially, 10 scans were obtained from the master patient model to evaluate the precision of the system. To establish baseline data (control group), the implant position was scanned with a high-accuracy laboratory scanner (± 5 mm) (D2000; 3Shape) to obtain 20 reference STL files for comparison. The ninth STL file was used as the master STL file because it was the closest to the average of the 20 STL files. Ten polyvinylsiloxane impressions were made with fixture-level copings and the open-tray technique (Fig. 3).12 An impression coping (regular connection open-tray impression coping; Straumann AG) was then tightened to 10 Ncm in the patient master implant model. A perforated stock tray was filled with heavy-body impression material, and medium-body material (Virtual; Ivoclar Vivadent AG) was injected around the impression coping. The impressions were made and allowed to polymerize in a chamber with 95% relative humidity and a constant temperature of 37 C. Implant analogs (regular connection; Straumann AG) were connected to the impression
Buda et al
Figure 2. Scanbody in place for laboratory scanning.
copings, and type 4 die stone (Fuji rock EP; GC America) was poured into the impressions (group GYP) (Fig. 4A). In the experimental group OBJ, the master STL file was used to fabricate stereolithographic models (n=10) with a 3D printer (Object 500; Stratasys). This system is a polyjet 3D printer with a layer thickness of 16 mm, which does not require light polymerizing at the end of the process. After making the printed cast, a repositionable implant analog (regular connection; Straumann AG) was inserted into the cast (Fig. 4B). The printed casts were then scanned with the same D2000 high-accuracy laboratory scanner, and a second set of data was obtained. In the experimental group FORM, the master STL file was used to fabricate stereolithographic casts (n=10) with an economical 3D printer (Formlabs II; Formlabs) (Fig. 4C). This system is a vat SLA 3D printer with a variable layer thickness from 25 to 50 to 100 mm. After the cast was printed, a repositionable implant analog (regular connection; Straumann AG) was inserted into the cast. The printed casts were then scanned with the same D2000 laboratory scanner, and a third set of data was obtained.
THE JOURNAL OF PROSTHETIC DENTISTRY
4
Volume
-
Issue
-
Master Model PVS Impression
Laboratory Scanner
Conventional gypsum cast with implant analog
Printed cast (stratasys)
Printed cast (formlab)
STL
STL file conventional
STL file printed 1
STL file printed 2
Control
GYP
OBJ
FORM
Figure 3. Study design. FORM, Formlabs Formlab 2; GYP, conventional gypsum cast; OBJ, Statasys Objet500; PVS, polyvinyl siloxane; STL, standard tessellation language.
All sets of STL files were superimposed with highaccuracy superimposition software (Convince software; 3Shape) working with a best-fit alignment algorithm to compare and measure differences between STL files. The differences among the control and test groups were indicated in a color- coded gradient (Fig. 5) to illustrate the discrepancy from the master patient model STL file. The center of the implant analog was measured in vertical and horizontal axis directions x, y, and z (Fig. 6). Differences were calculated among groups GYP, OBJ, FORM and control. In addition to the implant position in the x, y, and z coordinates, rotational position change around the vertical axis13 and tilting of the implant were measured. Each cast was scanned 5 times with the D2000 scanner, and 5 STL files were obtained for a total of 50 STL files each for groups GYP, OBJ, and FORM. The results obtained were the distance between the center of the scan body and the proprietary point for the 3Shape scanning software. Differences among the 3 reference groups were evaluated for statistical significance with 1-way ANOVA, and a post hoc t test with Bonferroni correction was used to detect any statistical difference among all 3 groups (a=.05). RESULTS Accuracy was measured as the discrepancy from the reference model. For vertical displacement of the implant body, group OBJ had the lowest value at e30 ±24 mm, followed by group FORM at e40 ±27 mm, and group GYP at +70 ±25 mm. Groups OBJ and FORM were statistically different from GYP (P<.001) (Table 1). For horizontal displacement of the implant shoulder, group OBJ had the lowest value of 85 ±12 mm, followed by group GYP at 103 ±25 mm and group FORM at 235 ±42 mm, and they were statistically different (P<.005). For horizontal
THE JOURNAL OF PROSTHETIC DENTISTRY
displacement of the implant apex, group OBJ was 123 ±25 mm and not different from group GYP at 136 ±40 mm but significantly different (P<.001) from group FORM at 326 ±54 mm (Table 1). For implant body tilting in the vertical axis from the true vertical implant axis position, significant differences were found between GYP versus OBJ and OBJ versus FORM (P<.001). Analysis of implant orientation or rotational position change around the vertical axis demonstrated that for GYP and FORM, there was no difference from the master patient control model. However, OBJ was significantly different (P<.001) from GYP and FORM with an approximately 2.5-degree rotation (Table 2). DISCUSSION The null hypothesis that no significant difference would be found between the implant analog definitive cast fabrication systems was rejected. The results showed that the printed cast obtained with the industrial poly-jet technology production system performed better or equally to the gypsum cast in terms of 3D implant positioning except for rotational position change around the vertical axis. Such results demonstrate that 3D-printed casts obtained with industrial polyjet technology are a useful and reliable alternative to traditional gypsum casts. The level of accuracy obtained with the economical 3D printer was inferior to that obtained with the gypsum cast, especially at the implant analog shoulder. Unfortunately, the budget 3D printer tested in this study was not sufficiently accurate to be used for the production of single-implant analog definitive casts. Discrepancies in the vertical axis for the 3D printing systems were e30 or e40 mm, while the gypsum cast discrepancy was +70 mm. In the laboratory fabrication
Buda et al
-
2018
5
Figure 5. Color-coded superimposition with 3Shape Convince software.
Figure 6. Horizontal and vertical axes measured at implant shoulder and implant apex.
Table 1. Displacement of implant analog for cast fabrication groups, mean ±standard deviation (mm) Vertical Displacement
Implant Shoulder Displacement
Implant Apex Displacement
GYP
70 ±25
56 ±25
73 ±40
OBJ
30 ±24
29 ±12
60 ±25
FORM
e40 ±27
188 ±42
263 ±54
<.001
<.001
<.001
Group
P
Figure 4. A, Implant analog gypsum cast. B, Statasys Objet500 3Dprinted implant analog cast. C, Formlabs Formlab 2 3D-printed implant analog cast.
process, this would translate to an implant crown fabricated on the gypsum cast being 70 mm short of the actual clinical situation, and with the 2 printing systems producing a crown that is 30 to 40 mm too long. The 30 to 40 mm error of the printed casts is probably clinically imperceptible. Whether a laboratory-produced error of a crown being 70 mm short is clinically significant is unknown. The horizontal displacement values were greater, ranging from approximately 140 to 330 mm, and would
Buda et al
GYP, conventional gypsum cast; OBJ, Statasys Objet500; FORM, Formlabs Formlab 2.
most likely be clinically perceptible at the time of a clinical evaluation of the crown. The horizontal discrepancy of group OBJ was not different from that of group GYP, and therefore the OBJ system should be considered equal to or better than the GYP system. The actual clinical relevance of these printing system discrepancies is yet to be determined because the level of clinical acceptable discrepancy in the x, y, and z vectors is still undefined. Future research will evaluate more 3D printing systems and milled cast systems.
THE JOURNAL OF PROSTHETIC DENTISTRY
6
Volume
Table 2. Mean ±standard deviation of tilting and rotational position change around the vertical axis error of implant analog for cast fabrication group (degrees) Tilting on y-axis
Rotational Position Change Around y-axis
GYP
0.158 ±0.083
0.03 ±0.22
OBJ
0.024 ±0.085
2.39 ±0.52
FORM
0.269 ±0.233
0.44 ±0.36
<.001
<.001
Group
P
GYP, conventional gypsum cast; OBJ, Statasys Objet500; FORM, Formlabs Formlab 2.
CONCLUSIONS Based on the findings of this in vitro study, the following conclusions were drawn: 1. The results for the variables of vertical and horizontal displacement of the implant platform and implant apex, tilting on the y-axis, and implant orientation demonstrated significant differences in accuracy among the implant analog cast fabrication systems. 2. The Polyjet industrial printing system was more accurate than the conventional gypsum implant analog cast. REFERENCES 1. Andersson M, Oden A. A new all-ceramic crown. A dense-sintered, highpurity alumina coping with porcelain. Acta Odontol Scand 1993;51:59-64. 2. Andersson M, Carlsson L, Persson M, Bergman B. Accuracy of machine milling and spark erosion with a CAD/CAM system. J Prosthet Dent 1996;76: 187-93. 3. Prithviraj DR, Pujari ML, Garg, Shruthi DP. Accuracy of the implant impression obtained from different impression materials and techniques: review. J Clin Exp Dent 2011;3:106-11.
THE JOURNAL OF PROSTHETIC DENTISTRY
-
Issue
-
4. Ramsay C, Ritter R. Utilization of digital technology for fabrication of definitive Implant-supported restorations. J Esthet Restor Dent 2012;24:299-309. 5. Lee SJ, Betensky RA, Gianneschi GE, Gallucci GO. Digital vs conventional implant impressions: efficacy outcomes. Clinical Oral Implants Res 2013;24:111-5. 6. Guth JF, Keul C, Stimmelmayr M, Beuer F, Edelhoff D. Accuracy of digital models obtained by direct and indirect technique data capturing. Clinical Oral Investig 2013;17:1201-8. 7. Van der Meer WJ, Andriessen FS, Wismeijer D, Ren Y. Application of intraoral dental scanners in the digital workflow of implantology. PLoS One 2012;7:e43312. 8. Rudolph H, Luthardt RG, Walter MH. Computer-aided analysis of the influence of digitizing and surfacing on the accuracy in dental CAD/CAM technology. Comput Biol Med 2007;37:579-87. 9. Gimenez B, Mutlu O, Martinez F, Pradies G. Accuracy of digital impression system based on parallel confocal laser technology for implants with consideration of operator experience and implant angulation and depth. Int J Oral Maxillofac Implants 2014;29:853-62. 10. Gimenez B, Pradies G, Martinez F, Mutlu O. Accuracy of two digital implant impression systems based on confocal microscopy with variations in customised software and clinical parameters. Int J Oral Maxillofac Implants 2015;30:56-64. 11. Lee SJ, Betensky RA, Gianneschi GE, Gallucci GO. Accuracy of digital versus conventional implant impressions. Clin Oral Implant Res 2015;26: 715-9. 12. Lee H, So JS, Hochstedler JL, Ercoli C. The accuracy of implant impression: a systematic review. J Prosthet Dent 2008;100:285-91. 13. Vigolo P, Majzoub Z, Cordioli G. In vitro comparison of master cast accuracy for single tooth implant replacement. J Prosthet Dent 2000;83: 562-6. Corresponding author: Dr John A. Sorensen 1959 NE Pacific St Box 357456 Seattle, WA 98195-7456 Email: [email protected] Acknowledgements The authors thank B&B Dental Ceramic Arts and 3DBioCAD for printing the study casts, use of their scanning systems, and dedicated help in performing this study. The authors also thank the Tylman Research Grant Committee, American Academy of Fixed Prosthodontics, for sponsoring the research project. Copyright © 2018 by the Editorial Council for The Journal of Prosthetic Dentistry.
Buda et al
Accuracy of 3-dimensional computer-aided manufactured single-tooth implant definitive casts Michele Buda, DDS, MSD,a Manuel Bratos, DDS, MSD, CDT,b and John A. Sorensen, DMD, PhD, FACPc
ABSTRACT Statement of problem. The integration of the digital workflow into routine prosthodontic practice for single-tooth implant surgery and fixed prosthesis fabrication has occurred at a remarkable pace in the last 5 years. With the greater demands of esthetics and precision, the definitive implant analog cast must ensure accurate implant positioning as well as an accurate relationship to adjacent teeth. Purpose. The purpose of this in vitro study was to evaluate the accuracy of the 3-dimensional (3D) implant position of definitive casts produced by 3D printing and analog technology. Material and methods. A master patient model was created from a dentate typodont. The maxillary left central incisor was removed, and a Straumann RC implant was positioned for a screw-retained prosthesis. A laboratory scanner with an accuracy of 5 mm was used for all scanning. A scanbody was connected to the master model implant and scanned to create a master patient file, which served as the control master patient for all comparisons. The two 3D printing systems used for this study were the Statasys Objet500 (group OBJ), an industrial Polyjet production system, and the Formlabs Formlab 2 (group FORM), a budget SLA Vat system. In addition, a conventional gypsum cast (group GYP) with an implant analog was made with elastomeric impression material. With a sample size of 10 per group, each gypsum cast and 2 printed group casts were scanned with the D2000 laboratory scanner 5 times per cast. Convince software (3Shape) was used for 3D analysis to calculate accuracy. The following variables were measured: implant analog vertical displacement, horizontal displacement of implant platform and apex, degree of tilting in the vertical axis, and rotational position change around the vertical axis. Means and standard deviations were calculated for trueness. One-way ANOVA and the post hoc t test with Bonferroni correction were used to investigate any significant differences among the experimental groups (a=.05). Results. For vertical displacement of the implant body, group OBJ had the lowest value of e30 ±24 mm. The values obtained for OBJ and FORM were significantly different from that obtained for GYP (P<.05). For horizontal displacement of the implant shoulder, Group OBJ had the lowest value, 85 ±12 mm, and the difference among these groups was significantly different (P<.05). The value for horizontal displacement of the implant apex was 123 ±25 mm for group OBJ and not significantly different from that obtained for group GYP (136 ±40 mm) but significantly different from that obtained for group FORM (326 ±54 mm). Also, the analysis of implant body tilting in the vertical axis showed significant differences between the values obtained for groups GYP and OBJ and between the values obtained for groups OBJ and FORM. With regard to implant rotational position change around the vertical axis, the values obtained for the gypsum cast and group FORM were not statistically different from those obtained for the master patient control model (P>.05). However, the implant orientation of group OBJ was significantly different from the orientation of groups GYP and FORM (P<.05). The actual clinical relevance of these printing system discrepancies is yet to be determined because the level of clinical acceptable discrepancy in the x, y, and z vectors is still undefined. Conclusions. This study showed statistically significant differences in accuracy among the implant analog cast fabrication systems; however; the level of clinical acceptable discrepancy is still undefined. Although further research is needed, this study supports the conclusion that the Polyjet industrial printing system was more accurate than the conventional implant analog gypsum cast. (J Prosthet Dent 2018;-:---)
Supported by Tylman Research Grant 2016-2017, American Academy of Fixed Prosthodontics. a Former postgraduate student and graduate, Department of Restorative Dentistry, School of Dentistry, University of Washington, Seattle, Wash; and Private practice, Treviso, Italy. b Affiliate Lecturer, Department of Restorative Dentistry, School of Dentistry, University of Washington, Seattle, Wash. c Professor, Department of Restorative Dentistry, Associate Dean for Graduate Studies, and Director, Biomimetics Biomaterials Biophotonics Biomechanics & Technology Laboratory, School of Dentistry, University of Washington, Seattle, Wash.
THE JOURNAL OF PROSTHETIC DENTISTRY
1
2
Volume
Clinical Implications The industrial 3-dimensional printing system proved more accurate than conventional elastomeric impressions and gypsum casts. Currently the cost of these high-quality printing systems is excessive except for large commercial laboratories. As with all technology, capital costs will decrease, making them affordable and sufficiently precise.
Applications of computer-aided design and computeraided manufacturing (CAD-CAM) to implant prosthodontics have been increasingly used to fabricate implant frameworks and monolithic restorations.1,2 Without an intraoral scanner, the gold-standard method used to record implant location and orientation requires an intraoral impression, from which a definitive cast is obtained.3 The definitive casts are usually scanned with an optical laboratory scanner, and a framework or restoration is then digitally designed and milled by a multi-axis milling machine with different restorative materials.4 Recently, with the rapid development of optical scanning technology, digital impressions have become more popular for implant-supported dental prostheses as well as conventional tooth-supported prostheses.5 A digital impression can be created with 2 different methods6: scanning a conventional definitive cast (indirect technique) or with an intraoral scanner (direct technique). The workflow for fabricating an implantsupported prosthesis could be entirely digital or combined with traditional laboratory procedures.7 When obtained with an intraoral scanner, the position of the implants is transferred to CAD-CAM machinery with a digital standard tessellation language (STL) file to fabricate the prosthesis. In this situation, a physical cast is lacking. Nevertheless, in most situations, physical casts are still needed so that they can be mounted on articulators to evaluate occlusion and adjust interproximal contacts. To allow technicians and dentists to perform these essential steps, 3-dimensional (3D) printed or milled casts can be obtained from the initial STL file. Accuracy describes closeness to the real dimensions of the object. Few studies have investigated the accuracy of 3D-printed or milled casts. Nonetheless, accuracy might depend on both the printing resolution and the 3D positioning of the implant analog in the cast relative to the teeth and soft tissues.8 A consensus is lacking regarding the accuracy of printing alone or printing combined with intraoral scanning for STL data acquisition. The accuracy of digital impressions has been compared with conventional impressions involving dental implants.9,10 Lee et al11
THE JOURNAL OF PROSTHETIC DENTISTRY
-
Issue
-
evaluated the accuracy of milled casts by using a laboratory scanner instead of a coordinate-measuring machine (CMM) to determine the true measurements as the control. Milled casts and gypsum casts were scanned with a laboratory scanner, and STL files were compared with reference file with Geomagic superimposition software. The 3D coordinate axes were defined, and the STL data sets from the gypsum and the digitally milled casts were individually aligned to the reference data set by a repeated best-fit algorithm based on the selected surfaces of the maxillary left first premolar and molar. The authors found a statistically significant difference between the gypsum and milled cast in the vertical axis of the implant position. In previous studies, both linear measurements between casts and CMMs have been used. The limitation of linear measurements is the uncertain replication of references points, whereas CMMs are unable to measure anatomic shapes. Therefore, superimposition of STL files obtained from a high-accuracy scanner offers a more accurate and consistent way to measure differences. The term ‘3D printing’ refers to the process of joining materials together to form objects from 3D model data. It is used to describe a wide variety of technologies known as additive manufacturing. According to the American Society for Testing and Materials, there are 7 categories of additive manufacturing: vat photopolymerization, material jetting, binder jetting, material extrusion, power bed fusion, sheet lamination, and direct energy deposition. Vat photopolymerization and material jetting are used to manufacture polymeric objects used in dentistry. In the vat photopolymerization system, an ultraviolet laser polymerizes a liquid photopolymer. The build platform is lowered into the vat as each layer polymerizes. In material jetting technology, inkjet print heads are used to jet liquid photopolymers onto a build platform. The material is immediately polymerized by ultraviolet lamps and solidified, allowing layers to be built up. In this study, 3D-printed casts were included because they are most often used by dental laboratories and are less expensive to produce than their milled counterparts. The purpose of this study was to compare the accuracy of 3D-printed casts obtained from digital files with conventional gypsum casts with implant analogs. The null hypothesis was that no significant difference would be found in the analog position of the 3D-printed casts and the conventional gypsum casts. MATERIAL AND METHODS A master patient model was created from dentate typodont. The maxillary left central incisor was removed, and a Straumann Bone Level RC implant was positioned for a screw-retained prosthesis (Fig. 1). A laboratory scanner
Buda et al
-
2018
3
Figure 1. Master patient model. A, Fabrication. B, Occlusal view.
(D2000; 3Shape) with an accuracy of 5 mm (ISO 12836) was used for the scanning. A Straumann Mono Scanbody was connected to the master model implant and scanned 20 times (Fig. 2). The average of the 20 STL files was used to create a master patient file and served as the control master patient for all comparisons (Fig. 3). 3Shape Dental Design and Implant Studio software were used to design an implant analog anterior model. Two 3D printing systems were evaluated: the Statasys Objet500 (an industrial Polyjet production system with a vertical layer thickness of 16 mm) and the Formlab 2 (a budget SLA vat system with a vertical layer thickness of 50 mm). These systems were used to fabricate 2 different printed casts (groups OBJ and FORM). A conventional gypsum case with an implant analog was made from an open-tray impression with an open-tray fixture level impression coping made with medium viscosity and heavybody impression material (Virtual; Ivoclar Vivadent AG). Initially, 10 scans were obtained from the master patient model to evaluate the precision of the system. To establish baseline data (control group), the implant position was scanned with a high-accuracy laboratory scanner (± 5 mm) (D2000; 3Shape) to obtain 20 reference STL files for comparison. The ninth STL file was used as the master STL file because it was the closest to the average of the 20 STL files. Ten polyvinylsiloxane impressions were made with fixture-level copings and the open-tray technique (Fig. 3).12 An impression coping (regular connection open-tray impression coping; Straumann AG) was then tightened to 10 Ncm in the patient master implant model. A perforated stock tray was filled with heavy-body impression material, and medium-body material (Virtual; Ivoclar Vivadent AG) was injected around the impression coping. The impressions were made and allowed to polymerize in a chamber with 95% relative humidity and a constant temperature of 37 C. Implant analogs (regular connection; Straumann AG) were connected to the impression
Buda et al
Figure 2. Scanbody in place for laboratory scanning.
copings, and type 4 die stone (Fuji rock EP; GC America) was poured into the impressions (group GYP) (Fig. 4A). In the experimental group OBJ, the master STL file was used to fabricate stereolithographic models (n=10) with a 3D printer (Object 500; Stratasys). This system is a polyjet 3D printer with a layer thickness of 16 mm, which does not require light polymerizing at the end of the process. After making the printed cast, a repositionable implant analog (regular connection; Straumann AG) was inserted into the cast (Fig. 4B). The printed casts were then scanned with the same D2000 high-accuracy laboratory scanner, and a second set of data was obtained. In the experimental group FORM, the master STL file was used to fabricate stereolithographic casts (n=10) with an economical 3D printer (Formlabs II; Formlabs) (Fig. 4C). This system is a vat SLA 3D printer with a variable layer thickness from 25 to 50 to 100 mm. After the cast was printed, a repositionable implant analog (regular connection; Straumann AG) was inserted into the cast. The printed casts were then scanned with the same D2000 laboratory scanner, and a third set of data was obtained.
THE JOURNAL OF PROSTHETIC DENTISTRY
4
Volume
-
Issue
-
Master Model PVS Impression
Laboratory Scanner
Conventional gypsum cast with implant analog
Printed cast (stratasys)
Printed cast (formlab)
STL
STL file conventional
STL file printed 1
STL file printed 2
Control
GYP
OBJ
FORM
Figure 3. Study design. FORM, Formlabs Formlab 2; GYP, conventional gypsum cast; OBJ, Statasys Objet500; PVS, polyvinyl siloxane; STL, standard tessellation language.
All sets of STL files were superimposed with highaccuracy superimposition software (Convince software; 3Shape) working with a best-fit alignment algorithm to compare and measure differences between STL files. The differences among the control and test groups were indicated in a color- coded gradient (Fig. 5) to illustrate the discrepancy from the master patient model STL file. The center of the implant analog was measured in vertical and horizontal axis directions x, y, and z (Fig. 6). Differences were calculated among groups GYP, OBJ, FORM and control. In addition to the implant position in the x, y, and z coordinates, rotational position change around the vertical axis13 and tilting of the implant were measured. Each cast was scanned 5 times with the D2000 scanner, and 5 STL files were obtained for a total of 50 STL files each for groups GYP, OBJ, and FORM. The results obtained were the distance between the center of the scan body and the proprietary point for the 3Shape scanning software. Differences among the 3 reference groups were evaluated for statistical significance with 1-way ANOVA, and a post hoc t test with Bonferroni correction was used to detect any statistical difference among all 3 groups (a=.05). RESULTS Accuracy was measured as the discrepancy from the reference model. For vertical displacement of the implant body, group OBJ had the lowest value at e30 ±24 mm, followed by group FORM at e40 ±27 mm, and group GYP at +70 ±25 mm. Groups OBJ and FORM were statistically different from GYP (P<.001) (Table 1). For horizontal displacement of the implant shoulder, group OBJ had the lowest value of 85 ±12 mm, followed by group GYP at 103 ±25 mm and group FORM at 235 ±42 mm, and they were statistically different (P<.005). For horizontal
THE JOURNAL OF PROSTHETIC DENTISTRY
displacement of the implant apex, group OBJ was 123 ±25 mm and not different from group GYP at 136 ±40 mm but significantly different (P<.001) from group FORM at 326 ±54 mm (Table 1). For implant body tilting in the vertical axis from the true vertical implant axis position, significant differences were found between GYP versus OBJ and OBJ versus FORM (P<.001). Analysis of implant orientation or rotational position change around the vertical axis demonstrated that for GYP and FORM, there was no difference from the master patient control model. However, OBJ was significantly different (P<.001) from GYP and FORM with an approximately 2.5-degree rotation (Table 2). DISCUSSION The null hypothesis that no significant difference would be found between the implant analog definitive cast fabrication systems was rejected. The results showed that the printed cast obtained with the industrial poly-jet technology production system performed better or equally to the gypsum cast in terms of 3D implant positioning except for rotational position change around the vertical axis. Such results demonstrate that 3D-printed casts obtained with industrial polyjet technology are a useful and reliable alternative to traditional gypsum casts. The level of accuracy obtained with the economical 3D printer was inferior to that obtained with the gypsum cast, especially at the implant analog shoulder. Unfortunately, the budget 3D printer tested in this study was not sufficiently accurate to be used for the production of single-implant analog definitive casts. Discrepancies in the vertical axis for the 3D printing systems were e30 or e40 mm, while the gypsum cast discrepancy was +70 mm. In the laboratory fabrication
Buda et al
-
2018
5
Figure 5. Color-coded superimposition with 3Shape Convince software.
Figure 6. Horizontal and vertical axes measured at implant shoulder and implant apex.
Table 1. Displacement of implant analog for cast fabrication groups, mean ±standard deviation (mm) Vertical Displacement
Implant Shoulder Displacement
Implant Apex Displacement
GYP
70 ±25
56 ±25
73 ±40
OBJ
30 ±24
29 ±12
60 ±25
FORM
e40 ±27
188 ±42
263 ±54
<.001
<.001
<.001
Group
P
Figure 4. A, Implant analog gypsum cast. B, Statasys Objet500 3Dprinted implant analog cast. C, Formlabs Formlab 2 3D-printed implant analog cast.
process, this would translate to an implant crown fabricated on the gypsum cast being 70 mm short of the actual clinical situation, and with the 2 printing systems producing a crown that is 30 to 40 mm too long. The 30 to 40 mm error of the printed casts is probably clinically imperceptible. Whether a laboratory-produced error of a crown being 70 mm short is clinically significant is unknown. The horizontal displacement values were greater, ranging from approximately 140 to 330 mm, and would
Buda et al
GYP, conventional gypsum cast; OBJ, Statasys Objet500; FORM, Formlabs Formlab 2.
most likely be clinically perceptible at the time of a clinical evaluation of the crown. The horizontal discrepancy of group OBJ was not different from that of group GYP, and therefore the OBJ system should be considered equal to or better than the GYP system. The actual clinical relevance of these printing system discrepancies is yet to be determined because the level of clinical acceptable discrepancy in the x, y, and z vectors is still undefined. Future research will evaluate more 3D printing systems and milled cast systems.
THE JOURNAL OF PROSTHETIC DENTISTRY
6
Volume
Table 2. Mean ±standard deviation of tilting and rotational position change around the vertical axis error of implant analog for cast fabrication group (degrees) Tilting on y-axis
Rotational Position Change Around y-axis
GYP
0.158 ±0.083
0.03 ±0.22
OBJ
0.024 ±0.085
2.39 ±0.52
FORM
0.269 ±0.233
0.44 ±0.36
<.001
<.001
Group
P
GYP, conventional gypsum cast; OBJ, Statasys Objet500; FORM, Formlabs Formlab 2.
CONCLUSIONS Based on the findings of this in vitro study, the following conclusions were drawn: 1. The results for the variables of vertical and horizontal displacement of the implant platform and implant apex, tilting on the y-axis, and implant orientation demonstrated significant differences in accuracy among the implant analog cast fabrication systems. 2. The Polyjet industrial printing system was more accurate than the conventional gypsum implant analog cast. REFERENCES 1. Andersson M, Oden A. A new all-ceramic crown. A dense-sintered, highpurity alumina coping with porcelain. Acta Odontol Scand 1993;51:59-64. 2. Andersson M, Carlsson L, Persson M, Bergman B. Accuracy of machine milling and spark erosion with a CAD/CAM system. J Prosthet Dent 1996;76: 187-93. 3. Prithviraj DR, Pujari ML, Garg, Shruthi DP. Accuracy of the implant impression obtained from different impression materials and techniques: review. J Clin Exp Dent 2011;3:106-11.
THE JOURNAL OF PROSTHETIC DENTISTRY
-
Issue
-
4. Ramsay C, Ritter R. Utilization of digital technology for fabrication of definitive Implant-supported restorations. J Esthet Restor Dent 2012;24:299-309. 5. Lee SJ, Betensky RA, Gianneschi GE, Gallucci GO. Digital vs conventional implant impressions: efficacy outcomes. Clinical Oral Implants Res 2013;24:111-5. 6. Guth JF, Keul C, Stimmelmayr M, Beuer F, Edelhoff D. Accuracy of digital models obtained by direct and indirect technique data capturing. Clinical Oral Investig 2013;17:1201-8. 7. Van der Meer WJ, Andriessen FS, Wismeijer D, Ren Y. Application of intraoral dental scanners in the digital workflow of implantology. PLoS One 2012;7:e43312. 8. Rudolph H, Luthardt RG, Walter MH. Computer-aided analysis of the influence of digitizing and surfacing on the accuracy in dental CAD/CAM technology. Comput Biol Med 2007;37:579-87. 9. Gimenez B, Mutlu O, Martinez F, Pradies G. Accuracy of digital impression system based on parallel confocal laser technology for implants with consideration of operator experience and implant angulation and depth. Int J Oral Maxillofac Implants 2014;29:853-62. 10. Gimenez B, Pradies G, Martinez F, Mutlu O. Accuracy of two digital implant impression systems based on confocal microscopy with variations in customised software and clinical parameters. Int J Oral Maxillofac Implants 2015;30:56-64. 11. Lee SJ, Betensky RA, Gianneschi GE, Gallucci GO. Accuracy of digital versus conventional implant impressions. Clin Oral Implant Res 2015;26: 715-9. 12. Lee H, So JS, Hochstedler JL, Ercoli C. The accuracy of implant impression: a systematic review. J Prosthet Dent 2008;100:285-91. 13. Vigolo P, Majzoub Z, Cordioli G. In vitro comparison of master cast accuracy for single tooth implant replacement. J Prosthet Dent 2000;83: 562-6. Corresponding author: Dr John A. Sorensen 1959 NE Pacific St Box 357456 Seattle, WA 98195-7456 Email: [email protected] Acknowledgements The authors thank B&B Dental Ceramic Arts and 3DBioCAD for printing the study casts, use of their scanning systems, and dedicated help in performing this study. The authors also thank the Tylman Research Grant Committee, American Academy of Fixed Prosthodontics, for sponsoring the research project. Copyright © 2018 by the Editorial Council for The Journal of Prosthetic Dentistry.
Buda et al