Accuracy of Two Digital Implant Impression Systems Based on Confocal Microscopy with Variations in Customized Software and Clinical Parameters Beatriz Giménez, DDS1/Guillermo Pradíes, DDS, PhD2/ Francisco Martínez-Rus, DDS, PhD3/Mutlu Özcan, DDS, Dr Med Dent, PhD4 Purpose: To evaluate the accuracy of two digital impression systems based on the same technology but different postprocessing correction modes of customized software, with consideration of several clinical parameters. Materials and Methods: A maxillary master model with six implants located in the second molar, second premolar, and lateral incisor positions was fitted with six cylindrical scan bodies. Scan bodies were placed at different angulations or depths apical to the gingiva. Two experienced and two inexperienced operators performed scans with either 3D Progress (MHT) or ZFX Intrascan (Zimmer Dental). Five different distances between implants (scan bodies) were measured, yielding five data points per impression and 100 per impression system. Measurements made with a high-accuracy three-dimensional coordinate measuring machine (CMM) of the master model acted as the true values. The values obtained from the digital impressions were subtracted from the CMM values to identify the deviations. The differences between experienced and inexperienced operators and implant angulation and depth were compared statistically. Results: Experience of the operator, implant angulation, and implant depth were not associated with significant differences in deviation from the true values with both 3D Progress and ZFX Intrascan. Accuracy in the first scanned quadrant was significantly better with 3D Progress, but ZFX Intrascan presented better accuracy in the full arch. Conclusion: Neither of the two systems tested would be suitable for digital impression of multipleimplant prostheses. Because of the errors, further development of both systems is required. Int J Oral Maxillofac Implants 2015;30:56–64. doi: 10.11607/jomi.3689 Key words: accuracy, dental implant, digital impression, implant angulation, implant depth, intraoral scanner
T
he fit between a suprastructure and the implant that supports it is considered to be a key factor in the success of implant-supported prostheses. A poor fit may result in tensile, compressive, and bending 1Research
Fellow, Department of Buccofacial Prostheses, Faculty of Odontology, University Complutense of Madrid, Madrid, Spain. 2Professor, Associate Dean, Department of Buccofacial Prostheses, Faculty of Odontology, University Complutense of Madrid, Madrid, Spain. 3Associate Professor, Department of Buccofacial Prostheses, Faculty of Odontology, University Complutense of Madrid, Madrid, Spain. 4Professor, Head of Dental Materials Unit, University of Zürich, Center for Dental and Oral Medicine, Clinic for Fixed and Removable Prosthodontics and Dental Materials Science, Zurich, Switzerland. Correspondence to: Dr Beatriz Giménez, Departamento de estomatología I, Prótesis Bucofacial, Plaza Ramón y Cajal s/n, Ciudad Universitaria, 28040 Madrid, Spain. Email: [email protected] ©2015 by Quintessence Publishing Co Inc.
forces when the prosthesis is connected to the implants. These stresses may remain, even after several years of function, which may result in mechanical complications such as loosening or fracture of the screws, distortion or breakage of the prosthesis, and in some cases even implant fracture.1–3 In addition, because of this misfit, the gap between the suprastructure and the implants may collect microorganisms, resulting in biologic problems in the surrounding tissues.4,5 Passive fit of the suprastructure could prevent these complications. The conventional process for prosthesis fabrication consists of three main steps, namely: intraoral impression, master model fabrication, and laboratory processing. Each step is highly susceptible to distortion; as a consequence, achieving optimal fit is very challenging. The first step in attaining the best possible fit is to reproduce the exact positions of the implants with the digital impression.6 Accuracy of implant impressions is highly affected by the type of impression material, technique, surface modifications of impression copings, fabrication tolerance of implant components,
56 Volume 30, Number 1, 2015 © 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
magnitude of angulations in implant positions with respect to the horizontal crestal plane, and depth of the implants.7 Digital impressions have been introduced using intraoral scanners, the first step toward a fully automated computer-aided design/computer-aided manufacture (CAD/CAM) system that reduces processing steps, adding control to each step.8–11 The level of measurement accuracy of a scanning system largely dictates its potential use and spectrum of indications, especially in implant prosthodontics.12 Several studies of the accuracy of intraoral scanners have been published to test single-unit restorations,13–18 several adjacent teeth,19,20 and full-arch scans.21–25 Limited information is available on the accuracy of intraoral digital impression systems for dental implants24,25 and regarding the factors and clinical parameters that affect their performance, such as the experience of the operator, angulation and position of the implant(s), and quadrant. One other important parameter in accuracy measurements is the software programs associated with the digital impression systems. The software comprises a patient database, a data acquisition environment, and a postprocessing module. The standard tessellation language (STL) data are generated immediately after the scanning procedure using a three-dimensional (3D) Cartesian coordinate system. In principle, STL data should not include a remote postprocessing module to improve the data. However, some manufacturers add a postprocessing module for the obtained STL file that might affect the accuracy of digital impressions. In confocal microscopy combined with moiré effect technology, the focal plane is shifted to a translating movable lens. The light rays, generated by the illumination pattern and reflected by each focal plane, pass through the assembly. The beam guidance is deflected by the beam splitter, in the direction of the detector, where the image of the object is detected in the focal plane. The movable lens is spherical to guarantee the necessary imaging quality for all focal planes. Thus, the focal plane is not actually planar but a curved surface, and the scanned surfaces appear distorted: flat surfaces and straight lines appear curved, and the magnifications and curvatures are different for different positions in the image. These distortions can be compensated because the theoretical distortions are known, having been computed on a reference image. The curvatures can be well approximated by an analytic function, such as, for example, a polynomial.11 The difference between 3D Progress IO Scan (MHT) and ZFX Intrascan (ZFX, Zimmer Dental), which are tested in the present study, relates to the compensations for the distortions. ZFX applies the compensation in the zaxis, and 3D Progress does it in the x- and y-axes.
Fig 1 Resin master model with six implants at locations 27, 25, 22, 12, 15, and 17. The second premolar sites hosted angulated implants, and the lateral incisors hosted implants placed 2 mm subgingivally (left incisor) and 4 mm subgingivally (right incisor). Removable soft tissue allowed proper measurement with CMM.
The objectives of this study were therefore to evaluate the accuracy of two intraoral digital impression scanners based on the same confocal microscopy, combined with the moiré effect technology operating with two software programs with different postprocessing correction modules using a six-implant model, with implants located at different angulations and depths. The following null hypotheses were tested: (1) Accuracy would not be affected by the two intraoral scanner systems; and (2) operator factor, (3) angulation of implants, and (4) depth of implants would not affect the accuracy of the digital impressions.
MATERIALS AND METHODS Master Model
Six implants (Certain 4 × 1/11 mm, Biomet 3i) were placed in the maxilla at the sites of the lateral incisors (12, 22), second premolars (15, 25), and second molars (17, 27) in an edentulous resin model (Frasaco) (Fig 1) as described elsewhere.24,25 This kind of implant was chosen, following metrologic expertise, because of the favorable design of the internal connection; its two flat surfaces allow the measurements to be made in the best possible way using an industrial 3D coordinate measurement machine (CMM) (Mitutoyo Crista Apex). An external connection or a conical connection would result in less accurate “true” values because of the difficulty in measuring such geometries. The implants were placed with the following angulations and depths using a micromilling machine (Cendres & Metaux): (1) the second molar implants at 0 degrees and 0 mm depth (at the gingival margin level); (2) right second premolar implant, with 30 degrees of distal angulation and 0 mm depth; (3) left second The International Journal of Oral & Maxillofacial Implants 57
© 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
a
b
Figs 2a and 2b (a) PEEK scan body (height: 8 mm) used for high-accuracy measurements; (b) view of the complete model with the removable artificial gingiva and the six scan bodies.
was measured to establish the orientation of the implant. In addition, the circumference of the implant was measured to determine the location of its center. Coordinates in the x-, y-, and z-axes expressed the location of each result from these two figures. This procedure was performed three times. A mean of the three measurements performed with the CMM was used as the final “true” location for each implant.
Intraoral Scanners and Scan Protocol
Fig 3 The 3D Progress (MHT) intraoral scanner, which is based on parallel confocal red laser technology.
premolar implant with 30 degrees of mesial angulation and 0 mm depth; (4) right lateral incisor implant at 0 degrees and 4 mm apical to the gingival margin; (5) left lateral incisor implant at 0 degrees and 2 mm apical to the gingival margin. Soft tissue was simulated using silicone (Vestogum, 3M ESPE) to enable accurate measurements of the implants with the CMM. Six high-precision scan bodies, each 8 mm long, were manufactured from polyether ether ketone (PEEK) (Createch Medical) (Figs 2a and 2b).
Coordinate Measurement Machine
An independent laboratory specialized in extremely accurate design and fabrication of CAD/CAM structures (Createch Medical) made the measurements of the master model and assessed the accuracy of the intraoral scanners 3D Progress IO scan (MHT; Fig 3) and ZFX Intrascan (ZFX, Zimmer Dental).24,25 The CMM was used to measure the master model to obtain the actual “true” data regarding the 3D positions of the implants. A high-accuracy touch signal probe with a 1-mm ruby sphere was used to measure the points at the heads of the implants to locate them in the x-, y-, and zaxes of the space. The plane of the implant connection
Scanning was performed with the two systems, which are based on x- and y-axes (3D Progress IO Scan, MHT) and z-axis (ZFX Intrascan, Zimmer Dental) postprocessing correction algorithms. The hardware manufacturer (MHT) provided customized software for both systems (Exoscan-mht-2012-12-19, Exocad). Both scanning technologies are based on confocal microscopy combined with a moiré detection effect that does not necessitate the application of powder to the surfaces to be scanned. This scan technology allows for the capture of 10 to 20 images per second with a capture volume of 10.4 × 9.6 × 18 mm.
Experience and Calibration of Operators
Four operators participated in the study for each system. Two of the operators (operators 1 and 2) had experience with intraoral scanners. Operator 2 had made digital impressions for both systems. This operator had 3 years’ experience working with different digital impression systems (more than 400 scans, including the present ones). The other two operators (operators 3 and 4) had no experience with any intraoral digital scanners. The scanning technique was explained to the inexperienced operators. To calibrate them before the impressions for the study were made, three impressions of the implant study model were made, with supervision by the experienced operators. Each operator made five full-arch impressions, and measurements were made between the right second molar implant (#27) and the others as follows: 27-25, 27-22, 27-12, 27-15, 27-17 (Fig 4). A continuous scan
58 Volume 30, Number 1, 2015 © 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
Figs 4a and 4b (a) The center point of the implant is located in the intersection between the vertical line that crosses the center of the cylinder (scan body) and the perpendicular horizontal plane of the head of the implant. (b) The distances are the lines that join the center point of the left second molar implant with the rest of the implants, resulting in five different distances (27-25, 27-22, 27-12, 27-15, and 27-17) that are the data points for the study.
44.5375 mm
42.0365 mm
46.0436 mm 13.0152 mm
a Figs 5a and 5b (a) The distance deviation was calculated following the formula: STL (distance) – CMM (“true distance”). The result would be the difference in distance from the “real” measurements. (b) The green line represents the distance from the true center point to the center points of the implants captured by the intraoral scanner (STL), and the red line represents the distance measured with the CMM, which constitutes the “real distance” in the present study.
46.0436 mm
b
42.0365 mm 46.0436 mm 13.0152 mm 36.1279 mm 35.7591 mm
36.1279 mm 35.7591 mm
a
around each scan body was practiced. The starting point was always implant no. 27, and the last implant to be scanned was implant no. 17 (right second molar). In contrast to what is required by most intraoral scanners, it was suggested by both manufacturers that the whole scan body not be captured and areas be left without being scanned. This was asked because the scanner created double images while trying to capture the whole scan body, mostly at mesial and distal sites. A minimum registration of 50% of the total scan body was required.
Accuracy Assessment
35.7591 mm
All the data from the CMM and the 3D Progress and ZFX Intrascan systems were imported using industrial reverse-engineering software (Rapidform) that could read the STL files, as has been described previously.24,25 The distances and angles of the center points of the implants were used to evaluate the accuracy of the intraoral scanners. To locate the center point of the implants for both intraoral scan systems, the STL file, the original design of the scan bodies, and the CAD used to manufacture the scan bodies were imported into the reverse-engineering software. The cylinders of the STL data captured by the scanner were isolated and matched one by one with the original CAD designs of the scan bodies. The center line of the cylinder
b
was determined; consequently, the center point of the implant was established (Figs 4a and 4b). The linear distance from the center point of the left second molar implant (#27) was considered as the reference point for measurements, following the “zero method” described elsewhere, in which the center point of the rest of the implants was measured in the CMM data.26 Subsequently, the same procedure was performed for the data obtained from the 3D Progress and ZFX Intrascan systems. Next, the distances between implants obtained with the 3D Progress and ZFX Intrascan systems were compared with the corresponding distances obtained with the CMM. The angulations were measured between the fit plane of implant no. 27 and each fit plane of the other implants measured by the CMM. The same angulation measurement procedure was repeated with the geometries obtained from the intraoral scanners (Figs 5a and 5b). The measurements were not divided into their x-, y-, and z-axis components because the intraoral scanners of the CMM and the CAD cylinders use different coordinate systems. The data sets of digitized measurements are composed of points that are located in a common coordinate system. Each point was defined in terms of x, y, and z coordinates. Concurrently, the points describe a part of the surface of the digitized object. Each single data set produced by independent The International Journal of Oral & Maxillofacial Implants 59
© 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
Table 1 Distance Deviation for Each Between-Implant Distance* Interimplant ZFX Intrascan distance (n = 20) Mean (µm) SD 27-25
–32.7
Table 2 Influence of Operator Experience in Distance Deviations of the Intraoral Scanners
3D Progress Mean (µm)
111.1
27-22
–157
292
27-12
–142.8
487.7
28.8
94
9.3
209.5
164.5
526.3
Inexperienced (n = 50)
–216.7
836.6
484.6
1,057.3
27-17
–150.6
1,080.3
497.4
1,346
.937
Operator type Experienced (n = 50)
27-15 P values
ZFX Intrascan
SD
3D Progress
Mean (µm)
SD
Mean (µm)
SD
–179
601
249
702
–101
705
224
930
P values
.551
.879
.143
*Deviations from the “true” measurements as determined by the CMM.
Table 3 Influence of Implant Depth in the Distance Deviations for the Intraoral Scanners ZFX Intrascan
Table 4 Influence of Implant Angulation in the Distance Deviations for the Intraoral Scanners
3D Progress
ZFX Intrascan
Depth of placement Mean (µm)
SD
SD
Implant angulation
2 mm (n = 40)
–150
397
87
403
Angled (n = 40)
–125
0 mm (n = 60)
–133
782
337
997
Straight (n = 60)
–150
P values
.889
Mean (µm)
.086
measurements received its own coordinate system. If the data were broken down into the x-, y-, and z-axes, errors would be introduced and the data could not be compared reliably. Each impression contained five data points that corresponded to each implant distance measured (27-25, 27-22, 27-12, 27-15, 27-17). Each operator collected 25 data points, yielding 100 data points for each system from the four operators (experienced operators n = 50/inexperienced operators n = 50; angulated implants n = 40/nonangulated implants n = 60; deeply placed implants n = 40/epicrestally placed implants n = 60). The distance deviation was calculated by subtracting the distance between implants obtained with the intraoral scanner from the distance obtained with the CMM (with the latter representing the “true” measurements) (STL – CMM = distance deviation). When the STL distance was longer than the CMM distance, the result was positive, and when the STL distance was shorter than the CMM, then the result was negative.
Statistical Analysis
The data were analyzed using statistical software (Minitab Release 14, Minitab Inc). The measured distances (in microns) between the implants obtained with the 3D Progress IO scan and ZFX Intrascan were compared with the distances between the implants of the “true data” measured with the CMM. The homogeneity of the data for implant distances, operator, and experience were measured (Anderson-Darling and
P values
Mean (µm)
.845
3D Progress
SD
Mean (µm)
SD
596
257
776
693
224
854 .842
Levene tests, α = .05). The differences between experienced and inexperienced operators, implant angulation, and depth effect were compared using the two-sample t test and one-way analysis of variance.
RESULTS The data were normally distributed when the independent variable of “implant distance” was analyzed (P = .989). The distance deviation of the digital impression measurements compared with the reference measurements (CMM) for the implant distances were calculated with one-way ANOVA, and the results are shown in Table 1. Experience of the operator (Table 2), implant depth (Table 3), and implant angulation (Table 4) did not show significant deviation from the true values with either 3D Progress or ZFX Intrascan (two-sample t test). The accuracy obtained in quadrant 1 (for the distances 27-25, 27-22) was significantly better (P = .001) for 3D Progress (Fig 6a). ZFX Intrascan presented better accuracy (P = .002) in quadrant 2 (last scan quadrant, distances 27-12, 27-15, 27-17) (Fig 6b). Significant differences were found between quadrants 1 and 2 for 3D Progress for means (P = .009) and variances (P = .000) (Fig 7a). For ZFX Intrascan, no differences were found for the means between quadrants 1 and 2, but a significant difference was found for the variances (P = .000) (Fig 7b).
60 Volume 30, Number 1, 2015 © 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
Normal –100 100 382.2
140
–100 100
70
120
Quadrant 1 2
100 80
Mean SD 19.05 160.6 382.2 1,028
60 40
% of measurements
% of measurements
Normal
N 40 60
20
Quadrant 1 2
60 50
Mean –94.85 –170.1
40 30
SD 227.0 824.0
N 40 60
20 10 0
0 –2,000 –1,000
0
1,000
–2,400 –1,600
2,000 3,000
Deviation (µm)
a
–800
0
800
1,600
Deviation (µm)
b
Figs 6a and 6b Differences in accuracy and precision between the first quadrant and the second quadrant with the different scanners. The green lines represent ± 100-µm distance deviations. (a) Measurements with 3D Progress. Quadrant 2 showed significantly worse accuracy and repeatability compared to quadrant 1 for means (P = .009) and repeatability (P = .000). (b) Measurements with ZFX Intrascan. Quadrant 2 showed significantly higher variability than the first quadrant (P = .000). There were no differences for the means between the first and second quadrants (P = .505). Device MHT ZFX
.0025 Mean 19.05 –94.85
SD 160.6 227.0
.0005
N 40 40
.0015 .0010
Mean SD 382.2 1,028 –170.1 824.0
.0004 Density
Density
.0020
Device MHT ZFX N 60 60
.0003 .0002 .0001
.0005
.0000
.0000 –800
a
–400
0
400
–2,000 –1,000
800
Deviation (µm)
b
0
1,000 2,000 3,000
Deviation (µm)
Figs 7a and 7b Comparison of ZFX Intrascan and 3D Progress in quadrants 1 and 2. (a) In quadrant 1, 3D Progress showed significantly better accuracy than ZFX Intrascan (for means, P = .0012). No significant difference was found for the variability of both systems in quadrant 1 (P = .156). (b) In quadrant 2, ZFX Intrascan performed significantly better than 3D Progress (P = .002). No differences were found in the repeatability (P = .164).
For quadrant 1, 3D Progress performed significantly better for means (P = .0012), but no differences were found in this quadrant for variability (P = .156). In contrast, for quadrant 2, ZFX performed significantly better for the means (P = .002), and no differences were found between the two systems with respect to variability (P = .164) (Figs 8a and 8b). Figures 9 and 10 display 3D reproductions of the discrepancies in the implant positions measured with the CMM and the intraoral scanners.
DISCUSSION This study evaluated the accuracy of two intraoral digital impression scanners based on parallel confocal laser technology operating with two software programs
with different postprocessing correction modules using a multiple-implant model. Based on the results of this study, since accuracy was significantly affected by the two intraoral scanner systems, but the factors of operator experience and angulation and depth of implants did not affect the accuracy of the digital impressions, the null hypothesis could be partially accepted. In previous studies24,25 with the same multipleimplant model in which the same parameters were analyzed using scanners that were based on active wavefront sampling and parallel confocal technologies (Lava COS, iTero scanner), similar to the findings of this study, the error increased from the first to the last implant scanned. However, the error (distance deviation) range with the two scanners tested in the present study was greater than those seen with the Lava COS and iTero scanners. With the same methodology, The International Journal of Oral & Maxillofacial Implants 61
© 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
-IM PLT .2 7–1 7 Z FX
-IM PLT .2 7–1 5
Z FX
-IM PLT .2 7–1 2
Z FX
-IM PLT .2 7–2 2
Z FX
-IM PLT .2 7–2 5
Z FX
DD
DD
G-I MP LT. 27 –17
G-I MP LT. 27 –15
G-I MP LT. 27 –12
DD
DD
5,000
G-I MP LT. 27 –25 DD G-I MP LT. 27 –22
4,000 UCL = 3,424
Deviation (µm)
3,000 2,000 1,000 0
Mean = –151
–1,000 –2,000 –3,000 LCL = 3,725
–4,000 1
21
41
61
81
101
121
141
161
181
Observation (samples) Fig 8 Amount of error in the implant distance obtained by 3D Progress and ZFX Intrascan depending on the quadrant. Note that the error and variability increased from the first scanned implant to the last one. Each black point on the graph represents the deviation created by the scanner for every implant distance (27-25, 27-22, 27-12, 27-15, and 27-17). Any increased distance between the black point and the zero line indicates that the distance deviation is higher. An increased distance between two consecutive black points indicates increased variability. The deviation of the distance was calculated by subtracting the scanner distance from the STL distance. UCL = upper confidence limit; LCL = lower confidence limit. Figs 9a and 9b Three-dimensional reproductions of the implant positions with the intraoral scanners. (a) Images representing the best ZFX impressions (less distance deviation/error). The silver color corresponds to the CMM data “truth” and the pink color to the ZFX results. (b) The best MHT impression, in which the minimum error was achieved.
a
b
Lava COS showed an average error of –45.02 ± 37.31 µm in the full arch for experienced operators and –4.37 ± 73.47 µm for the inexperienced operators (for the high-accuracy protocol), and the iTero scanner showed a mean error of –32 ± 216.1 µm. The results of the present study showed errors of –150.6 ± 1,080.3 µm with the ZFX Intrascan and 497.4 ± 1,346 µm for the 3D Progress, which are remarkably larger than errors observed in previous studies.24,25 In this study, whereas the inexperienced operators had never used an intraoral scanner before participating in the study, the experienced operators (n = 2) had
experience with both devices and had almost 4 years of experience in digital dentistry. One of the experienced operators had a scan history of more than 400 scans with different systems, including the present one. The other experienced operator only had experience with the 3D Progress or ZFX, with a history of more than 50 scans with each system. In principle, experience is a difficult parameter to measure. However, in previous studies of different intraoral scanners, some influence of experience of the operator in the accuracy of other intraoral scanners was detected. The difference between the systems tested in this study and those of
62 Volume 30, Number 1, 2015 © 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
Figs 10a to 10d Three-dimensional reproductions of the implant positions with the intraoral scanners. (a) Images representing the best ZFX impression (less distance deviation/error). The silver color corresponds to the CMM data (“truth”) and the pink color to the ZFX results. (b) The best MHT impression, in which minimum error was achieved. (c) The worst impression (maximum error) obtained with the ZFX intraoral scanner. (d) Worst impression (maximum error) with the MHT intraoral scanner.
a
b
c
d
the systems examined in previous studies is the inherent error within one system. Because of the inherent high variability within each system, it was not possible to distinguish any influence of clinical parameters, including operator experience. The accuracy of the digital impressions was not significantly affected by the angulation of the implants. The lack of parallelism between implants may result in increased dislodgement of impression material and its subsequent distortion when the impression tray is taken out of the mouth.27 Accurate conventional implant impressions could not be obtained when the implants diverged by 25 degrees.28,29 Splinting of impression copings is often performed to eliminate any rotational movements and to stabilize the impression copings while placing the analogs.30 In multiple-implant situations, especially when the implants have buccal inclinations, splinting may cause deformation of impression material upon removal; thus, it is not recommended.30,31 In this respect, digital impressions that reproduce the 3D positions of the implants by optical means could solve the actual problem that cannot be solved with conventional impressions. Likewise, it has been noted that the deeper an implant is placed subgingivally, the greater the subgingival portion of the impression coping, resulting in less coverage of the impression coping by the impression material and a consequent loss of stability. This might also lead to distortion.6 No differences were found between the two scanning systems in this study for deeply placed implants and those placed at the bone crest. Future examinations of variations in implant depth should focus
on testing different visible proportions of the scan bodies in an attempt to distinguish the accuracy of the digital scan systems at this level. With respect to the quadrant scanned, while 3D Progress performed significantly better in the first quadrant, no differences were found in this quadrant for the variability. In quadrant 2, ZFX performed significantly better, and no differences were found between the two systems in the variability. The reasons for this could be the difference in the method of algorithm correction; in this case, the correction would be more efficient in the z-axis for the ZFX scanner compared to the x- and y-axes for the MHT scanner.11 In this study, a CMM was used to deliver the “true” distances, which were considered as the control group. A study on the accuracy of conventional impressions with a similar methodology studied the four displacements of the implant components that occur from the implant impression to the master cast.32 The total error in the master cast was noted to be 98.5 ± 29.9 µm for the nonsplinted impression coping group and 99.3 ± 28.28 µm for the splinted one. In another study,19 discrepancies of 124 ± 34 µm were reported for the transfer technique, 116 ± 46 µm for the pickup technique, and 80 ± 25 µm for the splinted pickup technique. Although conventional impression materials are considered as the gold standard, when the data are compared to the “true” values, the related problems might be eliminated. However, the error inherent in the present systems was far from the level of conventional standards for multiple-implant impressions, indicating a need for further improvement in software programs. The International Journal of Oral & Maxillofacial Implants 63
© 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
CONCLUSION Based on the results of this study, the type of algorithm correction applied to confocal microscopy combined with the moiré effect had a significant effect in the accuracy of the digital impressions. The 3D Progress instrument performed significantly better in the first quadrant, and ZFX Intrascan provided better accuracy for the second quadrant. The accuracy and variability of both scanners were affected by the quadrant that was being measured. The error increased from the first to the last implant scanned. The cumulative error caused by “stitching” through the full-arch scan was confirmed. No significant effect was found in the deviations related to experience of the operator, implant angulation, or implant depth. Because of the sizes of the errors and the low predictability of the results of the tested scan systems, they could not be considered ideal for multiple-implant impressions, and the software programs require further development.
ACKNOWLEDGMENTS The authors would like to acknowledge Createch Medical Ltd for fabricating the scan bodies, Mikel Gomez Picaza for helping with the measurements and providing advice on high-accuracy methods, and Ricardo García Mata for his assistance with the statistics. The authors have no support or funding to report and declare that no competing interests exist.
REFERENCES 1. Gunne J, Jemt T, Linden B. Implant treatment in partially edentulous patients: A report on prostheses after 3 years. Int J Prosthodont 1994;7:143–148. 2. Jemt T, Linden B, Lekholm U. Failures and complications in 127 consecutively placed fixed partial prostheses supported by Brånemark implants: From prosthetic treatment to first annual checkup. Int J Oral Maxillofac Implants 1992;7:40–44. 3. Wenneberg A, Albrektsson T. Current challenges in successful rehabilitation with oral implants. J Oral Rehabil 2010;38:286–294. 4. Yannikakis S, Prombonas A. Improving the fit of implant prosthetics: An in vitro study. Int J Oral Maxillofac Implants 2013;28:126–134. 5. Aglietta M, Siciliano VI, Zwahlen M, et al. A systematic review of the survival and complication rates of implant-supported fixed dental prostheses with cantilever extensions after an observation period of at least 5 years. Clin Oral Implants Res 2009;20:441–451. 6. Lee H, Ercoli C, Funkenbush PD, Feng C. Effect of subgingival depth of implant placement on the dimensional accuracy of the implant impression: An in vitro study. J Prosthet Dent 2008;99:107–113. 7. Akalin ZF, Ozkan YK, Ekerim A. Effects of implant angulation, impression material, and variation in arch curvature width on implant transfer model accuracy. Int J Oral Maxillofac Implants 2013;28: 149–157. 8. Lee SJ, Gallucci GO. Digital vs. conventional implant impressions: Efficiency outcomes. Clin Oral Implants Res 2012;22:111–115. 9. Howell KJ, McGlumphy EA, Drago C, Knapik G. Comparison of the accuracy of Biomet 3i Encode Robocast Technology and conventional implant impression techniques. Int J Oral Maxillofac Implants 2013;28:228–240.
10. Jo SH, Kim S, Shim JS, Lee KW, Moon HK. Effect of impression coping and implant angulation on the accuracy of implant impressions: An in vitro study. J Adv Prosthodont 2010;2:128–133. 11. Patzelt SB, Emmanouilidi A, Stampf S, Strub JR, Att W. Accuracy of full-arch scans using intraoral scanners. Clin Oral Investig 2014;18: 1687–1694. 12. Mehl A, Ender A, Mörmann W, Attin T. Accuracy testing of a new intraoral 3D camera. Int J Comput Dent 2009;12:11–28. 13. Syrek A, Reich G, Ranftl D, et al. Clinical evaluation of all-ceramic crowns fabricated from intraoral digital impressions based on the principle of active wavefront sampling. J Dent 2010;38:553–559. 14. Henkel GL. A comparison of fixed prostheses generated from conventional vs digitally scanned dental impressions. Compend Contin Educ Dent 2007;28:422–431. 15. Brawek PK, Wolfart S, Endres L, Kirsten A, Reich S. The clinical accuracy of single crowns exclusively fabricated by digital workflow: The comparison of two systems. Clin Oral Invest 2013;17:2119–2125. 16. Seelbach P, Brueckel C, Wöstmann B. Accuracy of digital and conventional impression technique and workflow. Clin Oral Invest 2013;17:1759–1764. 17. Kim SY, Kim MJ, Han JS, Yedi IS, Lim YJ, Kwon HB. Accuracy of dies captured by an intraoral digital impression system using parallel confocal imaging. Int J Prosthodont 2013;26:161–163. 18. Luthardt RG, Loos R, Quaas S. Accuracy of intraoral data acquisition in comparison to the conventional impression. Int J Comput Dent 2005;8:283–294. 19. Güth JF, Keul C, Stimmelmayr M, Beuer F, Edelhoff D. Accuracy of digital models obtained by direct and indirect data capturing. Clin Oral Invest 2013,17:1201–1208. 20. Karl M, Shubinski P, Taylor T. Effect of intraoral scanning on the passivity of fit of implant-supported fixed partial prostheses. Quintessence Int 2012:43:555–563. 21. Ender A, Melh A. Full arch scans: Conventional versus digital impressions—An in-vitro study. Int J Comput Dent 2011;14:11–21. 22. Van der Meer WJ, Andriessen FS, Wismeijer D, Ren Y. Application of intra-oral dental scanners in the digital workflow of implantology. PLoS ONE 2012;7(8):e43312. 23. Ender A, Mehl A. Influence of scanning strategies on the accuracy of digital intraoral scanning systems. Int J Comput Dent 2013;16:11–21. 24. Giménez B, Özcan M, Martínez-Rus F, Pradíes G. Accuracy of a digital impression system based on active wavefront sampling technology for implants considering operator experience, implant angulation and depth. Clin Implant Dent Relat Res 2013 Jul 24. [Epub ahead of print] 25. Giménez B, Özcan M, Martínez-Rus F, Pradíes G. Accuracy of a 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;29:853–862. 26. Jemt T, Hjalmarsson L. In vitro measurements of precision of fit of implant supported frameworks. A comparison between "virtual" and "physical" assessments of fit using two different techniques of measurements. Clin Implant Dental Relat Research 2012;5:175–182. 27. Jang HK, Kim S, Shim JS, Lee KW, Moon HS. Accuracy of impressions for internal-connection implant prosthesis with various divergent angles. Int J Oral Maxillofac Implants 2011;26:1011–1015. 28. Assuncao WG, Filho HG, Zaniquelli O. Evaluation of transfer impressions for osseointegrated implants at various angulations. Implant Dent 2004;13:358–366. 29. Assuncao WG, Tabata LF, Cardoso A, Rocha EP, Gomes EA. Prosthetic transfer impression accuracy evaluation for osseointegrated implants. Implant Dent 2008;17:248–256. 30. Vigolo P, Fonzi F, Majzoub Z, Cordioli G. An evaluation of impression technique for multiple internal connection implant prostheses. J Prosthet Dent 2004;92:470–476. 31. Choi JH, Lim YJ, Yim SH, Kim CW. Evaluation of the accuracy of implant-level impression techniques for internal-connection implant prostheses in parallel and divergent models. Int J Oral Maxillofac Implants 2007;22:761–768. 32. Kim S, Nicholls JI, Han CH, Lee KH, Lee WK. Displacement of implant components from impressions to definitive casts. Int J Oral Maxillofac Implants 2006;21:747–755.
64 Volume 30, Number 1, 2015 © 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
T
he fit between a suprastructure and the implant that supports it is considered to be a key factor in the success of implant-supported prostheses. A poor fit may result in tensile, compressive, and bending 1Research
Fellow, Department of Buccofacial Prostheses, Faculty of Odontology, University Complutense of Madrid, Madrid, Spain. 2Professor, Associate Dean, Department of Buccofacial Prostheses, Faculty of Odontology, University Complutense of Madrid, Madrid, Spain. 3Associate Professor, Department of Buccofacial Prostheses, Faculty of Odontology, University Complutense of Madrid, Madrid, Spain. 4Professor, Head of Dental Materials Unit, University of Zürich, Center for Dental and Oral Medicine, Clinic for Fixed and Removable Prosthodontics and Dental Materials Science, Zurich, Switzerland. Correspondence to: Dr Beatriz Giménez, Departamento de estomatología I, Prótesis Bucofacial, Plaza Ramón y Cajal s/n, Ciudad Universitaria, 28040 Madrid, Spain. Email: [email protected] ©2015 by Quintessence Publishing Co Inc.
forces when the prosthesis is connected to the implants. These stresses may remain, even after several years of function, which may result in mechanical complications such as loosening or fracture of the screws, distortion or breakage of the prosthesis, and in some cases even implant fracture.1–3 In addition, because of this misfit, the gap between the suprastructure and the implants may collect microorganisms, resulting in biologic problems in the surrounding tissues.4,5 Passive fit of the suprastructure could prevent these complications. The conventional process for prosthesis fabrication consists of three main steps, namely: intraoral impression, master model fabrication, and laboratory processing. Each step is highly susceptible to distortion; as a consequence, achieving optimal fit is very challenging. The first step in attaining the best possible fit is to reproduce the exact positions of the implants with the digital impression.6 Accuracy of implant impressions is highly affected by the type of impression material, technique, surface modifications of impression copings, fabrication tolerance of implant components,
56 Volume 30, Number 1, 2015 © 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
magnitude of angulations in implant positions with respect to the horizontal crestal plane, and depth of the implants.7 Digital impressions have been introduced using intraoral scanners, the first step toward a fully automated computer-aided design/computer-aided manufacture (CAD/CAM) system that reduces processing steps, adding control to each step.8–11 The level of measurement accuracy of a scanning system largely dictates its potential use and spectrum of indications, especially in implant prosthodontics.12 Several studies of the accuracy of intraoral scanners have been published to test single-unit restorations,13–18 several adjacent teeth,19,20 and full-arch scans.21–25 Limited information is available on the accuracy of intraoral digital impression systems for dental implants24,25 and regarding the factors and clinical parameters that affect their performance, such as the experience of the operator, angulation and position of the implant(s), and quadrant. One other important parameter in accuracy measurements is the software programs associated with the digital impression systems. The software comprises a patient database, a data acquisition environment, and a postprocessing module. The standard tessellation language (STL) data are generated immediately after the scanning procedure using a three-dimensional (3D) Cartesian coordinate system. In principle, STL data should not include a remote postprocessing module to improve the data. However, some manufacturers add a postprocessing module for the obtained STL file that might affect the accuracy of digital impressions. In confocal microscopy combined with moiré effect technology, the focal plane is shifted to a translating movable lens. The light rays, generated by the illumination pattern and reflected by each focal plane, pass through the assembly. The beam guidance is deflected by the beam splitter, in the direction of the detector, where the image of the object is detected in the focal plane. The movable lens is spherical to guarantee the necessary imaging quality for all focal planes. Thus, the focal plane is not actually planar but a curved surface, and the scanned surfaces appear distorted: flat surfaces and straight lines appear curved, and the magnifications and curvatures are different for different positions in the image. These distortions can be compensated because the theoretical distortions are known, having been computed on a reference image. The curvatures can be well approximated by an analytic function, such as, for example, a polynomial.11 The difference between 3D Progress IO Scan (MHT) and ZFX Intrascan (ZFX, Zimmer Dental), which are tested in the present study, relates to the compensations for the distortions. ZFX applies the compensation in the zaxis, and 3D Progress does it in the x- and y-axes.
Fig 1 Resin master model with six implants at locations 27, 25, 22, 12, 15, and 17. The second premolar sites hosted angulated implants, and the lateral incisors hosted implants placed 2 mm subgingivally (left incisor) and 4 mm subgingivally (right incisor). Removable soft tissue allowed proper measurement with CMM.
The objectives of this study were therefore to evaluate the accuracy of two intraoral digital impression scanners based on the same confocal microscopy, combined with the moiré effect technology operating with two software programs with different postprocessing correction modules using a six-implant model, with implants located at different angulations and depths. The following null hypotheses were tested: (1) Accuracy would not be affected by the two intraoral scanner systems; and (2) operator factor, (3) angulation of implants, and (4) depth of implants would not affect the accuracy of the digital impressions.
MATERIALS AND METHODS Master Model
Six implants (Certain 4 × 1/11 mm, Biomet 3i) were placed in the maxilla at the sites of the lateral incisors (12, 22), second premolars (15, 25), and second molars (17, 27) in an edentulous resin model (Frasaco) (Fig 1) as described elsewhere.24,25 This kind of implant was chosen, following metrologic expertise, because of the favorable design of the internal connection; its two flat surfaces allow the measurements to be made in the best possible way using an industrial 3D coordinate measurement machine (CMM) (Mitutoyo Crista Apex). An external connection or a conical connection would result in less accurate “true” values because of the difficulty in measuring such geometries. The implants were placed with the following angulations and depths using a micromilling machine (Cendres & Metaux): (1) the second molar implants at 0 degrees and 0 mm depth (at the gingival margin level); (2) right second premolar implant, with 30 degrees of distal angulation and 0 mm depth; (3) left second The International Journal of Oral & Maxillofacial Implants 57
© 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
a
b
Figs 2a and 2b (a) PEEK scan body (height: 8 mm) used for high-accuracy measurements; (b) view of the complete model with the removable artificial gingiva and the six scan bodies.
was measured to establish the orientation of the implant. In addition, the circumference of the implant was measured to determine the location of its center. Coordinates in the x-, y-, and z-axes expressed the location of each result from these two figures. This procedure was performed three times. A mean of the three measurements performed with the CMM was used as the final “true” location for each implant.
Intraoral Scanners and Scan Protocol
Fig 3 The 3D Progress (MHT) intraoral scanner, which is based on parallel confocal red laser technology.
premolar implant with 30 degrees of mesial angulation and 0 mm depth; (4) right lateral incisor implant at 0 degrees and 4 mm apical to the gingival margin; (5) left lateral incisor implant at 0 degrees and 2 mm apical to the gingival margin. Soft tissue was simulated using silicone (Vestogum, 3M ESPE) to enable accurate measurements of the implants with the CMM. Six high-precision scan bodies, each 8 mm long, were manufactured from polyether ether ketone (PEEK) (Createch Medical) (Figs 2a and 2b).
Coordinate Measurement Machine
An independent laboratory specialized in extremely accurate design and fabrication of CAD/CAM structures (Createch Medical) made the measurements of the master model and assessed the accuracy of the intraoral scanners 3D Progress IO scan (MHT; Fig 3) and ZFX Intrascan (ZFX, Zimmer Dental).24,25 The CMM was used to measure the master model to obtain the actual “true” data regarding the 3D positions of the implants. A high-accuracy touch signal probe with a 1-mm ruby sphere was used to measure the points at the heads of the implants to locate them in the x-, y-, and zaxes of the space. The plane of the implant connection
Scanning was performed with the two systems, which are based on x- and y-axes (3D Progress IO Scan, MHT) and z-axis (ZFX Intrascan, Zimmer Dental) postprocessing correction algorithms. The hardware manufacturer (MHT) provided customized software for both systems (Exoscan-mht-2012-12-19, Exocad). Both scanning technologies are based on confocal microscopy combined with a moiré detection effect that does not necessitate the application of powder to the surfaces to be scanned. This scan technology allows for the capture of 10 to 20 images per second with a capture volume of 10.4 × 9.6 × 18 mm.
Experience and Calibration of Operators
Four operators participated in the study for each system. Two of the operators (operators 1 and 2) had experience with intraoral scanners. Operator 2 had made digital impressions for both systems. This operator had 3 years’ experience working with different digital impression systems (more than 400 scans, including the present ones). The other two operators (operators 3 and 4) had no experience with any intraoral digital scanners. The scanning technique was explained to the inexperienced operators. To calibrate them before the impressions for the study were made, three impressions of the implant study model were made, with supervision by the experienced operators. Each operator made five full-arch impressions, and measurements were made between the right second molar implant (#27) and the others as follows: 27-25, 27-22, 27-12, 27-15, 27-17 (Fig 4). A continuous scan
58 Volume 30, Number 1, 2015 © 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
Figs 4a and 4b (a) The center point of the implant is located in the intersection between the vertical line that crosses the center of the cylinder (scan body) and the perpendicular horizontal plane of the head of the implant. (b) The distances are the lines that join the center point of the left second molar implant with the rest of the implants, resulting in five different distances (27-25, 27-22, 27-12, 27-15, and 27-17) that are the data points for the study.
44.5375 mm
42.0365 mm
46.0436 mm 13.0152 mm
a Figs 5a and 5b (a) The distance deviation was calculated following the formula: STL (distance) – CMM (“true distance”). The result would be the difference in distance from the “real” measurements. (b) The green line represents the distance from the true center point to the center points of the implants captured by the intraoral scanner (STL), and the red line represents the distance measured with the CMM, which constitutes the “real distance” in the present study.
46.0436 mm
b
42.0365 mm 46.0436 mm 13.0152 mm 36.1279 mm 35.7591 mm
36.1279 mm 35.7591 mm
a
around each scan body was practiced. The starting point was always implant no. 27, and the last implant to be scanned was implant no. 17 (right second molar). In contrast to what is required by most intraoral scanners, it was suggested by both manufacturers that the whole scan body not be captured and areas be left without being scanned. This was asked because the scanner created double images while trying to capture the whole scan body, mostly at mesial and distal sites. A minimum registration of 50% of the total scan body was required.
Accuracy Assessment
35.7591 mm
All the data from the CMM and the 3D Progress and ZFX Intrascan systems were imported using industrial reverse-engineering software (Rapidform) that could read the STL files, as has been described previously.24,25 The distances and angles of the center points of the implants were used to evaluate the accuracy of the intraoral scanners. To locate the center point of the implants for both intraoral scan systems, the STL file, the original design of the scan bodies, and the CAD used to manufacture the scan bodies were imported into the reverse-engineering software. The cylinders of the STL data captured by the scanner were isolated and matched one by one with the original CAD designs of the scan bodies. The center line of the cylinder
b
was determined; consequently, the center point of the implant was established (Figs 4a and 4b). The linear distance from the center point of the left second molar implant (#27) was considered as the reference point for measurements, following the “zero method” described elsewhere, in which the center point of the rest of the implants was measured in the CMM data.26 Subsequently, the same procedure was performed for the data obtained from the 3D Progress and ZFX Intrascan systems. Next, the distances between implants obtained with the 3D Progress and ZFX Intrascan systems were compared with the corresponding distances obtained with the CMM. The angulations were measured between the fit plane of implant no. 27 and each fit plane of the other implants measured by the CMM. The same angulation measurement procedure was repeated with the geometries obtained from the intraoral scanners (Figs 5a and 5b). The measurements were not divided into their x-, y-, and z-axis components because the intraoral scanners of the CMM and the CAD cylinders use different coordinate systems. The data sets of digitized measurements are composed of points that are located in a common coordinate system. Each point was defined in terms of x, y, and z coordinates. Concurrently, the points describe a part of the surface of the digitized object. Each single data set produced by independent The International Journal of Oral & Maxillofacial Implants 59
© 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
Table 1 Distance Deviation for Each Between-Implant Distance* Interimplant ZFX Intrascan distance (n = 20) Mean (µm) SD 27-25
–32.7
Table 2 Influence of Operator Experience in Distance Deviations of the Intraoral Scanners
3D Progress Mean (µm)
111.1
27-22
–157
292
27-12
–142.8
487.7
28.8
94
9.3
209.5
164.5
526.3
Inexperienced (n = 50)
–216.7
836.6
484.6
1,057.3
27-17
–150.6
1,080.3
497.4
1,346
.937
Operator type Experienced (n = 50)
27-15 P values
ZFX Intrascan
SD
3D Progress
Mean (µm)
SD
Mean (µm)
SD
–179
601
249
702
–101
705
224
930
P values
.551
.879
.143
*Deviations from the “true” measurements as determined by the CMM.
Table 3 Influence of Implant Depth in the Distance Deviations for the Intraoral Scanners ZFX Intrascan
Table 4 Influence of Implant Angulation in the Distance Deviations for the Intraoral Scanners
3D Progress
ZFX Intrascan
Depth of placement Mean (µm)
SD
SD
Implant angulation
2 mm (n = 40)
–150
397
87
403
Angled (n = 40)
–125
0 mm (n = 60)
–133
782
337
997
Straight (n = 60)
–150
P values
.889
Mean (µm)
.086
measurements received its own coordinate system. If the data were broken down into the x-, y-, and z-axes, errors would be introduced and the data could not be compared reliably. Each impression contained five data points that corresponded to each implant distance measured (27-25, 27-22, 27-12, 27-15, 27-17). Each operator collected 25 data points, yielding 100 data points for each system from the four operators (experienced operators n = 50/inexperienced operators n = 50; angulated implants n = 40/nonangulated implants n = 60; deeply placed implants n = 40/epicrestally placed implants n = 60). The distance deviation was calculated by subtracting the distance between implants obtained with the intraoral scanner from the distance obtained with the CMM (with the latter representing the “true” measurements) (STL – CMM = distance deviation). When the STL distance was longer than the CMM distance, the result was positive, and when the STL distance was shorter than the CMM, then the result was negative.
Statistical Analysis
The data were analyzed using statistical software (Minitab Release 14, Minitab Inc). The measured distances (in microns) between the implants obtained with the 3D Progress IO scan and ZFX Intrascan were compared with the distances between the implants of the “true data” measured with the CMM. The homogeneity of the data for implant distances, operator, and experience were measured (Anderson-Darling and
P values
Mean (µm)
.845
3D Progress
SD
Mean (µm)
SD
596
257
776
693
224
854 .842
Levene tests, α = .05). The differences between experienced and inexperienced operators, implant angulation, and depth effect were compared using the two-sample t test and one-way analysis of variance.
RESULTS The data were normally distributed when the independent variable of “implant distance” was analyzed (P = .989). The distance deviation of the digital impression measurements compared with the reference measurements (CMM) for the implant distances were calculated with one-way ANOVA, and the results are shown in Table 1. Experience of the operator (Table 2), implant depth (Table 3), and implant angulation (Table 4) did not show significant deviation from the true values with either 3D Progress or ZFX Intrascan (two-sample t test). The accuracy obtained in quadrant 1 (for the distances 27-25, 27-22) was significantly better (P = .001) for 3D Progress (Fig 6a). ZFX Intrascan presented better accuracy (P = .002) in quadrant 2 (last scan quadrant, distances 27-12, 27-15, 27-17) (Fig 6b). Significant differences were found between quadrants 1 and 2 for 3D Progress for means (P = .009) and variances (P = .000) (Fig 7a). For ZFX Intrascan, no differences were found for the means between quadrants 1 and 2, but a significant difference was found for the variances (P = .000) (Fig 7b).
60 Volume 30, Number 1, 2015 © 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
Normal –100 100 382.2
140
–100 100
70
120
Quadrant 1 2
100 80
Mean SD 19.05 160.6 382.2 1,028
60 40
% of measurements
% of measurements
Normal
N 40 60
20
Quadrant 1 2
60 50
Mean –94.85 –170.1
40 30
SD 227.0 824.0
N 40 60
20 10 0
0 –2,000 –1,000
0
1,000
–2,400 –1,600
2,000 3,000
Deviation (µm)
a
–800
0
800
1,600
Deviation (µm)
b
Figs 6a and 6b Differences in accuracy and precision between the first quadrant and the second quadrant with the different scanners. The green lines represent ± 100-µm distance deviations. (a) Measurements with 3D Progress. Quadrant 2 showed significantly worse accuracy and repeatability compared to quadrant 1 for means (P = .009) and repeatability (P = .000). (b) Measurements with ZFX Intrascan. Quadrant 2 showed significantly higher variability than the first quadrant (P = .000). There were no differences for the means between the first and second quadrants (P = .505). Device MHT ZFX
.0025 Mean 19.05 –94.85
SD 160.6 227.0
.0005
N 40 40
.0015 .0010
Mean SD 382.2 1,028 –170.1 824.0
.0004 Density
Density
.0020
Device MHT ZFX N 60 60
.0003 .0002 .0001
.0005
.0000
.0000 –800
a
–400
0
400
–2,000 –1,000
800
Deviation (µm)
b
0
1,000 2,000 3,000
Deviation (µm)
Figs 7a and 7b Comparison of ZFX Intrascan and 3D Progress in quadrants 1 and 2. (a) In quadrant 1, 3D Progress showed significantly better accuracy than ZFX Intrascan (for means, P = .0012). No significant difference was found for the variability of both systems in quadrant 1 (P = .156). (b) In quadrant 2, ZFX Intrascan performed significantly better than 3D Progress (P = .002). No differences were found in the repeatability (P = .164).
For quadrant 1, 3D Progress performed significantly better for means (P = .0012), but no differences were found in this quadrant for variability (P = .156). In contrast, for quadrant 2, ZFX performed significantly better for the means (P = .002), and no differences were found between the two systems with respect to variability (P = .164) (Figs 8a and 8b). Figures 9 and 10 display 3D reproductions of the discrepancies in the implant positions measured with the CMM and the intraoral scanners.
DISCUSSION This study evaluated the accuracy of two intraoral digital impression scanners based on parallel confocal laser technology operating with two software programs
with different postprocessing correction modules using a multiple-implant model. Based on the results of this study, since accuracy was significantly affected by the two intraoral scanner systems, but the factors of operator experience and angulation and depth of implants did not affect the accuracy of the digital impressions, the null hypothesis could be partially accepted. In previous studies24,25 with the same multipleimplant model in which the same parameters were analyzed using scanners that were based on active wavefront sampling and parallel confocal technologies (Lava COS, iTero scanner), similar to the findings of this study, the error increased from the first to the last implant scanned. However, the error (distance deviation) range with the two scanners tested in the present study was greater than those seen with the Lava COS and iTero scanners. With the same methodology, The International Journal of Oral & Maxillofacial Implants 61
© 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
-IM PLT .2 7–1 7 Z FX
-IM PLT .2 7–1 5
Z FX
-IM PLT .2 7–1 2
Z FX
-IM PLT .2 7–2 2
Z FX
-IM PLT .2 7–2 5
Z FX
DD
DD
G-I MP LT. 27 –17
G-I MP LT. 27 –15
G-I MP LT. 27 –12
DD
DD
5,000
G-I MP LT. 27 –25 DD G-I MP LT. 27 –22
4,000 UCL = 3,424
Deviation (µm)
3,000 2,000 1,000 0
Mean = –151
–1,000 –2,000 –3,000 LCL = 3,725
–4,000 1
21
41
61
81
101
121
141
161
181
Observation (samples) Fig 8 Amount of error in the implant distance obtained by 3D Progress and ZFX Intrascan depending on the quadrant. Note that the error and variability increased from the first scanned implant to the last one. Each black point on the graph represents the deviation created by the scanner for every implant distance (27-25, 27-22, 27-12, 27-15, and 27-17). Any increased distance between the black point and the zero line indicates that the distance deviation is higher. An increased distance between two consecutive black points indicates increased variability. The deviation of the distance was calculated by subtracting the scanner distance from the STL distance. UCL = upper confidence limit; LCL = lower confidence limit. Figs 9a and 9b Three-dimensional reproductions of the implant positions with the intraoral scanners. (a) Images representing the best ZFX impressions (less distance deviation/error). The silver color corresponds to the CMM data “truth” and the pink color to the ZFX results. (b) The best MHT impression, in which the minimum error was achieved.
a
b
Lava COS showed an average error of –45.02 ± 37.31 µm in the full arch for experienced operators and –4.37 ± 73.47 µm for the inexperienced operators (for the high-accuracy protocol), and the iTero scanner showed a mean error of –32 ± 216.1 µm. The results of the present study showed errors of –150.6 ± 1,080.3 µm with the ZFX Intrascan and 497.4 ± 1,346 µm for the 3D Progress, which are remarkably larger than errors observed in previous studies.24,25 In this study, whereas the inexperienced operators had never used an intraoral scanner before participating in the study, the experienced operators (n = 2) had
experience with both devices and had almost 4 years of experience in digital dentistry. One of the experienced operators had a scan history of more than 400 scans with different systems, including the present one. The other experienced operator only had experience with the 3D Progress or ZFX, with a history of more than 50 scans with each system. In principle, experience is a difficult parameter to measure. However, in previous studies of different intraoral scanners, some influence of experience of the operator in the accuracy of other intraoral scanners was detected. The difference between the systems tested in this study and those of
62 Volume 30, Number 1, 2015 © 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
Figs 10a to 10d Three-dimensional reproductions of the implant positions with the intraoral scanners. (a) Images representing the best ZFX impression (less distance deviation/error). The silver color corresponds to the CMM data (“truth”) and the pink color to the ZFX results. (b) The best MHT impression, in which minimum error was achieved. (c) The worst impression (maximum error) obtained with the ZFX intraoral scanner. (d) Worst impression (maximum error) with the MHT intraoral scanner.
a
b
c
d
the systems examined in previous studies is the inherent error within one system. Because of the inherent high variability within each system, it was not possible to distinguish any influence of clinical parameters, including operator experience. The accuracy of the digital impressions was not significantly affected by the angulation of the implants. The lack of parallelism between implants may result in increased dislodgement of impression material and its subsequent distortion when the impression tray is taken out of the mouth.27 Accurate conventional implant impressions could not be obtained when the implants diverged by 25 degrees.28,29 Splinting of impression copings is often performed to eliminate any rotational movements and to stabilize the impression copings while placing the analogs.30 In multiple-implant situations, especially when the implants have buccal inclinations, splinting may cause deformation of impression material upon removal; thus, it is not recommended.30,31 In this respect, digital impressions that reproduce the 3D positions of the implants by optical means could solve the actual problem that cannot be solved with conventional impressions. Likewise, it has been noted that the deeper an implant is placed subgingivally, the greater the subgingival portion of the impression coping, resulting in less coverage of the impression coping by the impression material and a consequent loss of stability. This might also lead to distortion.6 No differences were found between the two scanning systems in this study for deeply placed implants and those placed at the bone crest. Future examinations of variations in implant depth should focus
on testing different visible proportions of the scan bodies in an attempt to distinguish the accuracy of the digital scan systems at this level. With respect to the quadrant scanned, while 3D Progress performed significantly better in the first quadrant, no differences were found in this quadrant for the variability. In quadrant 2, ZFX performed significantly better, and no differences were found between the two systems in the variability. The reasons for this could be the difference in the method of algorithm correction; in this case, the correction would be more efficient in the z-axis for the ZFX scanner compared to the x- and y-axes for the MHT scanner.11 In this study, a CMM was used to deliver the “true” distances, which were considered as the control group. A study on the accuracy of conventional impressions with a similar methodology studied the four displacements of the implant components that occur from the implant impression to the master cast.32 The total error in the master cast was noted to be 98.5 ± 29.9 µm for the nonsplinted impression coping group and 99.3 ± 28.28 µm for the splinted one. In another study,19 discrepancies of 124 ± 34 µm were reported for the transfer technique, 116 ± 46 µm for the pickup technique, and 80 ± 25 µm for the splinted pickup technique. Although conventional impression materials are considered as the gold standard, when the data are compared to the “true” values, the related problems might be eliminated. However, the error inherent in the present systems was far from the level of conventional standards for multiple-implant impressions, indicating a need for further improvement in software programs. The International Journal of Oral & Maxillofacial Implants 63
© 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Giménez et al
CONCLUSION Based on the results of this study, the type of algorithm correction applied to confocal microscopy combined with the moiré effect had a significant effect in the accuracy of the digital impressions. The 3D Progress instrument performed significantly better in the first quadrant, and ZFX Intrascan provided better accuracy for the second quadrant. The accuracy and variability of both scanners were affected by the quadrant that was being measured. The error increased from the first to the last implant scanned. The cumulative error caused by “stitching” through the full-arch scan was confirmed. No significant effect was found in the deviations related to experience of the operator, implant angulation, or implant depth. Because of the sizes of the errors and the low predictability of the results of the tested scan systems, they could not be considered ideal for multiple-implant impressions, and the software programs require further development.
ACKNOWLEDGMENTS The authors would like to acknowledge Createch Medical Ltd for fabricating the scan bodies, Mikel Gomez Picaza for helping with the measurements and providing advice on high-accuracy methods, and Ricardo García Mata for his assistance with the statistics. The authors have no support or funding to report and declare that no competing interests exist.
REFERENCES 1. Gunne J, Jemt T, Linden B. Implant treatment in partially edentulous patients: A report on prostheses after 3 years. Int J Prosthodont 1994;7:143–148. 2. Jemt T, Linden B, Lekholm U. Failures and complications in 127 consecutively placed fixed partial prostheses supported by Brånemark implants: From prosthetic treatment to first annual checkup. Int J Oral Maxillofac Implants 1992;7:40–44. 3. Wenneberg A, Albrektsson T. Current challenges in successful rehabilitation with oral implants. J Oral Rehabil 2010;38:286–294. 4. Yannikakis S, Prombonas A. Improving the fit of implant prosthetics: An in vitro study. Int J Oral Maxillofac Implants 2013;28:126–134. 5. Aglietta M, Siciliano VI, Zwahlen M, et al. A systematic review of the survival and complication rates of implant-supported fixed dental prostheses with cantilever extensions after an observation period of at least 5 years. Clin Oral Implants Res 2009;20:441–451. 6. Lee H, Ercoli C, Funkenbush PD, Feng C. Effect of subgingival depth of implant placement on the dimensional accuracy of the implant impression: An in vitro study. J Prosthet Dent 2008;99:107–113. 7. Akalin ZF, Ozkan YK, Ekerim A. Effects of implant angulation, impression material, and variation in arch curvature width on implant transfer model accuracy. Int J Oral Maxillofac Implants 2013;28: 149–157. 8. Lee SJ, Gallucci GO. Digital vs. conventional implant impressions: Efficiency outcomes. Clin Oral Implants Res 2012;22:111–115. 9. Howell KJ, McGlumphy EA, Drago C, Knapik G. Comparison of the accuracy of Biomet 3i Encode Robocast Technology and conventional implant impression techniques. Int J Oral Maxillofac Implants 2013;28:228–240.
10. Jo SH, Kim S, Shim JS, Lee KW, Moon HK. Effect of impression coping and implant angulation on the accuracy of implant impressions: An in vitro study. J Adv Prosthodont 2010;2:128–133. 11. Patzelt SB, Emmanouilidi A, Stampf S, Strub JR, Att W. Accuracy of full-arch scans using intraoral scanners. Clin Oral Investig 2014;18: 1687–1694. 12. Mehl A, Ender A, Mörmann W, Attin T. Accuracy testing of a new intraoral 3D camera. Int J Comput Dent 2009;12:11–28. 13. Syrek A, Reich G, Ranftl D, et al. Clinical evaluation of all-ceramic crowns fabricated from intraoral digital impressions based on the principle of active wavefront sampling. J Dent 2010;38:553–559. 14. Henkel GL. A comparison of fixed prostheses generated from conventional vs digitally scanned dental impressions. Compend Contin Educ Dent 2007;28:422–431. 15. Brawek PK, Wolfart S, Endres L, Kirsten A, Reich S. The clinical accuracy of single crowns exclusively fabricated by digital workflow: The comparison of two systems. Clin Oral Invest 2013;17:2119–2125. 16. Seelbach P, Brueckel C, Wöstmann B. Accuracy of digital and conventional impression technique and workflow. Clin Oral Invest 2013;17:1759–1764. 17. Kim SY, Kim MJ, Han JS, Yedi IS, Lim YJ, Kwon HB. Accuracy of dies captured by an intraoral digital impression system using parallel confocal imaging. Int J Prosthodont 2013;26:161–163. 18. Luthardt RG, Loos R, Quaas S. Accuracy of intraoral data acquisition in comparison to the conventional impression. Int J Comput Dent 2005;8:283–294. 19. Güth JF, Keul C, Stimmelmayr M, Beuer F, Edelhoff D. Accuracy of digital models obtained by direct and indirect data capturing. Clin Oral Invest 2013,17:1201–1208. 20. Karl M, Shubinski P, Taylor T. Effect of intraoral scanning on the passivity of fit of implant-supported fixed partial prostheses. Quintessence Int 2012:43:555–563. 21. Ender A, Melh A. Full arch scans: Conventional versus digital impressions—An in-vitro study. Int J Comput Dent 2011;14:11–21. 22. Van der Meer WJ, Andriessen FS, Wismeijer D, Ren Y. Application of intra-oral dental scanners in the digital workflow of implantology. PLoS ONE 2012;7(8):e43312. 23. Ender A, Mehl A. Influence of scanning strategies on the accuracy of digital intraoral scanning systems. Int J Comput Dent 2013;16:11–21. 24. Giménez B, Özcan M, Martínez-Rus F, Pradíes G. Accuracy of a digital impression system based on active wavefront sampling technology for implants considering operator experience, implant angulation and depth. Clin Implant Dent Relat Res 2013 Jul 24. [Epub ahead of print] 25. Giménez B, Özcan M, Martínez-Rus F, Pradíes G. Accuracy of a 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;29:853–862. 26. Jemt T, Hjalmarsson L. In vitro measurements of precision of fit of implant supported frameworks. A comparison between "virtual" and "physical" assessments of fit using two different techniques of measurements. Clin Implant Dental Relat Research 2012;5:175–182. 27. Jang HK, Kim S, Shim JS, Lee KW, Moon HS. Accuracy of impressions for internal-connection implant prosthesis with various divergent angles. Int J Oral Maxillofac Implants 2011;26:1011–1015. 28. Assuncao WG, Filho HG, Zaniquelli O. Evaluation of transfer impressions for osseointegrated implants at various angulations. Implant Dent 2004;13:358–366. 29. Assuncao WG, Tabata LF, Cardoso A, Rocha EP, Gomes EA. Prosthetic transfer impression accuracy evaluation for osseointegrated implants. Implant Dent 2008;17:248–256. 30. Vigolo P, Fonzi F, Majzoub Z, Cordioli G. An evaluation of impression technique for multiple internal connection implant prostheses. J Prosthet Dent 2004;92:470–476. 31. Choi JH, Lim YJ, Yim SH, Kim CW. Evaluation of the accuracy of implant-level impression techniques for internal-connection implant prostheses in parallel and divergent models. Int J Oral Maxillofac Implants 2007;22:761–768. 32. Kim S, Nicholls JI, Han CH, Lee KH, Lee WK. Displacement of implant components from impressions to definitive casts. Int J Oral Maxillofac Implants 2006;21:747–755.
64 Volume 30, Number 1, 2015 © 2015 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
