Comparison of Three-Dimensional Accuracy of Digital and Conventional Implant Impressions Effect of Interimplant Distance in An Edentulous Arch PDF

This peer-reviewed, accepted manuscript will undergo final editing and production prior to publication in JOMI.
Comparison of Three-Dimensional Accuracy of Digital and Conventional

Implant Impressions: Effect of Interimplant Distance in an Edentulous Arch
Ming Yi Tan, BDS, MDSa; Sophia Hui Xin Yee, BDS, MDSb; Keng Mun Wong, BDS, MSDc;
Ying Han Tan, BDS, MScd; Keson Beng Choon Tan, BDS, MSDe
a
Instructor, Faculty of Dentistry, National University of Singapore
11, Lower Kent Ridge Road,
Singapore 119083,
Republic of Singapore
b
Registrar, Khoo Teck Puat Hospital, Singapore
90 Yishun Central,
Singapore 768828,
c
Adjunct Senior Lecturer, Faculty of Dentistry, National University of Singapore
Singapore 119083,
d
Adjunct Staff, Faculty of Dentistry, National University of Singapore
Singapore 119083,
e
Associate Professor, Faculty of Dentistry, National University of Singapore
Singapore 119083,
Correspondence
Keson B. Tan,
Faculty of Dentistry, National University of Singapore,
11 Lower Kent Ridge Road,
Singapore 119083,
Republic of Singapore.
Fax: +65 67785742
Email: dentanbc@nus.edu.sg
Submitted December 18, 2017; Accepted July 3, 2018.
1

© 2018 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.
ABSTRACT
Purpose: This study compared the three-dimensional (3D) accuracy of conventional
impressions with digital impression systems (intraoral scanners and dental laboratory scanners)
for two different interimplant distances in maxillary edentulous arches. Materials and
Methods: Six impression systems comprising one conventional impression material
(Impregum), two intraoral scanners (TRIOS and True Definition), and three dental laboratory
scanners (Ceramill Map400, inEos X5, and D900) were evaluated on two completely
edentulous maxillary arch master models (A and B) with six and eight implants, respectively.
Centroid positions at the implant platform level were derived using either physical or virtual
probe hits with a coordinate measuring machine. Comparison of centroid positions between
master and test models (n = 5) defined linear distortions (dx, dy, dz), global linear distortions
(dR), and 3D reference distance distortions between implants (ΔR). The two-dimensional (2D)
angles between the central axis of each implant to the x- or y-axes were compared to derive
absolute angular distortions (Absdθx, Absdθy). Results: Model A mean dR ranged from 8.7 ±
8.3 µm to 731.7 ± 62.3 µm. Model B mean dR ranged from 16.3 ± 9 µm to 620.2 ± 63.2 µm.
Model A mean Absdθx ranged from 0.021 ± 0.205 degrees to –2.349 ± 0.166 degrees, and
mean Absdθy ranged from –0.002 ± 0.160 degrees to –0.932 ± 0.290 degrees. Model B mean
Absdθx ranged from –0.007 ± 0.076 degrees to –0.688 ± 0.574 degrees, and mean Absdθy
ranged from –0.018 ± 0.048 degrees to –1.052 ± 0.297 degrees. One-way analysis of variance
(ANOVA) by Impression system revealed significant differences among test groups for dR and
ΔR in both models, with True Definition exhibiting the poorest accuracy. Independent samples
t tests for dR, between homologous implant location pairs in Model A versus B, revealed the
presence of two to four significant pairings (out of seven possible) for the intraoral scanner
systems, in which instances dR was larger in Model A by 110 to 150 µm. Conclusion:
Reducing interimplant distance may decrease global linear distortions (dR) for intraoral scanner
systems, but had no effect on Impregum and the dental laboratory scanner systems. Impregum
consistently exhibited the best or second-best accuracy at all implant locations, while True
Definition exhibited the poorest accuracy for all linear distortions in both Models A and B.
Impression systems could not be consistently ranked for absolute angular distortions. Int J Oral
Maxillofac Implants 2018. doi: 10.11607/jomi.6855
Keywords: 3D accuracy, Implant, CMM, Digital impression, intra-oral scanner, dental

laboratory scanner
2

INTRODUCTION
Digital dentistry appears to have reached a level adequate for clinical utility. For the
impression making step, digitized mapping of intra-oral architecture promises to eliminate the
problems of conventional techniques. Intra-oral scanner systems are generally advocated for
dentate arches or short edentulous spans, but there is mounting interest in extending their
application to longer spans. Image acquisition is accomplished by manoeuvring a handheld
intra-oral scanner around relevant structures in the oral cavity, and these captured images are
integrated to produce a digital three-dimensional (3D) replica of scanned structures.
Dental laboratory scanner systems similarly allow operators to adopt a digital workflow. The
stone model obtained from a conventional implant impression is scanned with a dental
laboratory scanner to create a virtual model. However, problems inherent to conventional
impressions remain.
The subsequent design and milling processes are similar for intra-oral scanner and dental
laboratory scanner systems. Resultant prostheses may be further characterized manually if
needed.
Digital technology can provide much-anticipated technical, clinical, and procedural benefits,
but a recent consensus conference highlighted the scarcity of appropriate clinical
documentation to support widespread use of CAD/CAM technology in implant
reconstructions1. There is an urgent need for good-quality long-term studies to validate the
efficacy and effectiveness of digital options. However, new digital workflow parameters have
been reported on but may not be directly relevant to clinical questions, and attempts to
standardize terminology and communication are still in their infancy. The American College of
Prosthodontists (ACP) Digital Dentistry Glossary Development Task Force acknowledged the
dynamism of this field, and predicted the need for annual review of their newly minted
glossary2.
Conventional impressions remain essential in implant prosthodontics, with several studies

validating the superiority of polyether or addition polymerized silicones3-5.
With edentulous arches, Gherlone et al6 found that All-on-Four® prostheses fabricated using
either conventional models or intra-oral scans were comparable in terms of implant survival
and crestal bone loss at 12 months. Papaspyridakos et al7 found that for a five-implant
edentulous mandibular model, both the conventional impression technique with splinted
impression copings and digital impression with the 3Shape TRIOS intra-oral scanner were
more accurate than conventional impressions with non-splinted impression copings.
3

Vandeweghe et al8 evaluated four intra-oral scanner systems on a six-implant edentulous

mandibular model and found that 3M True Definition and 3Shape TRIOS fared better than
Cerec Omnicam and Lava COS. In the latter two studies, the investigation of implant
impressions in edentulous arches was limited to mandibular arches. Implants were placed
within the interforaminal area, simulating the common clinical scenario of full-arch implant-
supported fixed prostheses in patients with early loss of posterior teeth. The inter-implant
distances in both studies were not clearly stated, but average mandibular arch lengths
interforaminally would result in implants placed relatively close together. This proximity may
have skewed the results towards better accuracy if adjacent implants could be captured by the
limited image field of the respective intra-oral scanners.
The verdict on digital implant impression accuracy remains inconclusive. Many studies did not
perform direct comparisons between conventional implant impressions and the digital
alternatives (intra-oral scanner or dental laboratory scanner)8. Comparisons between studies
could not be made due to the heterogeneous study conditions and methodology. A pertinent
problem with the studies on edentulous arches was the lack of information on inter-implant
distance. Furthermore, many studies that utilized the superimposition method for distortion
analysis7,8 did not compare test samples directly with the physical master models, but with
digitized master models, the derivation of which could introduce errors.
The purpose of this study, therefore, was to compare the 3D accuracy of implant positions
obtained via a conventional impression system with that obtained from digital impression
systems (intra-oral scanner or dental laboratory scanner) for two different inter-implant
distances in maxillary edentulous arches.
MATERIALS AND METHODS
Six impression systems, comprising one conventional impression material, two intra-oral
scanners and three dental laboratory scanners, were evaluated. The features of each system are
summarized in Table 1 and the experimental outline is illustrated in Figure 1.
Fabrication of Master Models
Two completely edentulous maxillary master models (Figure 2) that replicated the clinical
situation for full-arch implant-supported fixed prostheses were fabricated with heat-
polymerized polymethyl-methacrylate (PMMA) (Lucitone 199 Denture Base Resin, Dentsply).
Master Model A comprised six parallel implants (I, J, K, L, M and N) spaced along the arch
4

with approximately 20mm inter-implant distance. Master Model B comprised eight parallel
implants (P, Q, R, S, T, U, V and W) spaced approximately 13mm apart.
Implants used were Straumann Bone Level implants RC (021.6412, 4.8 mmD and 12.0 mmL,
Straumann) with the apical 9.0mm secured with PMMA. Polyvinyl siloxane (PVS) impression
material (Aquasil Ultra LV, Dentsply Caulk) simulated the peri-implant soft tissue around the
coronal 3.0mm of each implant.
Three Grade 5 silicon nitride ball bearings (8057222, Tsubaki Nakashima Co) of 8.0mm
diameter and manufacturer-specified sphericity of 0.2µm were secured on the hard palate. 3D
distances between these ball bearings were monitored to ensure that the master models
remained dimensionally stable throughout the study. Each master model was secured with
PMMA to a custom aluminum block. The peripheral flat planes of this block served as vertical
stops for the custom impression trays (Tray Resin II, Shofu Inc). The aluminum block also
facilitated securing of the master models to the worktable of the Coordinate Measuring
Machine (CMM) (Global Silver Performance 7.10.7, Brown and Sharpe). Master models were
stored at 20°C for 30 days after completion to ensure that the models were dimensionally stable
before proceeding.
Fabrication of Conventional Impression test models
Five polyether (Impregum) impressions, dispensed from a Pentamix Automatic Mixing Unit
(3M ESPE), were made of each master model via the open-tray impression technique and with
torque application of 15Ncm15-17 to the impression copings (RC impression post, 025.4202,
Straumann) which were splinted with pattern resin (GC Pattern Resin LS, GC Corp)
incrementally placed on a floss scaffold. To minimise the effect of polymerization shrinkage,
the pattern resin segments on individual impression copings were kept separate initially and
final connection of the residual gap was done with a small increment. On removal of each
impression, implant analogs (Bone Level Implant Analog, 025.4101, Straumann) were
attached to the impression copings. Care was taken to prevent rotation of impression copings
within the impression material by stabilising the implant analog. Type IV dental stone models
(Silky Rock, Whip Mix Co.) were poured using the manufacturer’s recommended
powder:water ratio, removed after one hour, and stored at 20ºC for at least seven days to
mitigate the effects of dental stone expansion.
Scanning procedure
Three types of scan bodies were used. Implant positions in test groups TRIOS, True Definition
and D900 were registered using new precision-milled Core3D scan bodies (Core Scanbody,
5

2077, Core3D Centres). Test group Ceramill Map400 utilized Range three Kit b Scan bodies
(792322, Amann Girrbach) and the corresponding Range three Kit a/b screws (792341, Amann
Girrbach). Test group inEos X5 utilized a two-piece assembly comprising a 2-CONnect
abutment for bridges and bars (2-CONnect L 810 M, Sirona) that inserts into the implant
analog, and an inPost scan body (inPost for 2-CONnect KS61, 6551639, Sirona), which was
hand-tightened on the abutment.
Dental Laboratory Scanner
The respective scan bodies for the dental laboratory scanner systems were secured with 10Ncm
torque to the implant analogs in the test stone models derived from Impregum conventional
impressions, and one laboratory scan of each stone model was done using all three dental
laboratory scanner systems. Additional scans were prescribed at specific locations on the stone
model if the initial scan was deemed to be deficient.
Intra-oral Scanner
Five intra-oral scans of each master model were performed using the intra-oral scanner systems
(TRIOS and True Definition), following insertion of the assigned scan bodies with 10Ncm
torque. Because True Definition required the use of a powder spray (3M High Resolution
Scanning Spray, 3M ESPE), TRIOS scans were conducted first to avoid powder contamination
of the master models. Both intra-oral scanner systems have recommended scan strategies for
dentate arches with natural teeth surfaces providing unique features for the stitching of images.
However, these scan strategies do not work well for the edentulous arches with identical scan
bodies and the manufacturers do not have published recommendations for edentulous arches.
Help was sought from the respective manufacturers and modified scan strategies were adopted
with approval. The distal-most scan body in quadrant one was scanned first. Each scan body
was captured fully in the buccal and palatal directions, before the scan progressed further
anteriorly. Increasing amounts of the palate were captured during the palatal scan of each scan
body.
CMM measurements of Master Models and Test Models
Implant positions were determined by a CMM with manufacturer-specified measurement

accuracy of 2µm. For physical models comprising the two master models and the ten stone
models derived from Impregum conventional impressions, physical probe hits by the CMM
spring-loaded ruby ball stylus of 1.0mm diameter and 20.0mm shank length (Renishaw Plc)
allowed derivation of the Cartesian coordinates of datum points. The CMM metrology
software (PC-DMISTM CAD++ Version 2013 MR1, Wilcox Associates Inc.) then constructed
6

geometric forms from these measured datum points. Virtual models from the intra-oral scanner
and dental laboratory scanner test groups were exported as STL files from their respective
software suites and imported into the same CMM metrology software, where virtual probe hits
were made in the same manner as on physical models.
For master model implants and stone model implant analogs, the coronal implant platform was
measured via six probe hits to define a plane, while the internal conical portion was measured
via ten probe hits at two levels to define a cone. The centroid was found by constructing a
pierce point between the central axis of the internal cone and the coronal flat plane (Figure 3).
For virtual model implants, the top platform of the virtual scan body was measured via four
virtual probe hits to define a plane. The axial portion of each scan body was measured via eight
virtual probe hits at two levels to define a cylinder (TRIOS, True Definition, inEos X5 and
D900) or cone (Ceramill Map400). CMM measurements confirmed that for the three scan
body types from Core3D, Amann Girrbach and Sirona, the nominal scan body heights were
10.00mm, 10.40mm and 12.18mm respectively. The virtual implant centroid was defined by
constructing a pierce point between the central axis of the virtual scan body cylinder or cone
and a constructed plane offset by the magnitude of the respective scan body heights, apical to
the top platform plane (Figure 3).
Each geometric feature was measured three times to minimize the magnitude of errors. A
verification check using the CMM software tolerance criterion functions was conducted after
each measurement. The tolerance value for all physical geometric features was set at 6µm,
while that for virtual geometric features was set at 20µm. The higher tolerance limit in virtual
models served to compensate for minor surface irregularities inherent in STL files. 3D
distances between the three centroids of the same constructed feature were calculated, and a
limit of acceptable deviation was set at 10µm for both physical and virtual models.
The alignment set up for Model A consisted of XY-plane formed by implants I, K, N; X-axis
formed by implants I and N; and origin of local coordinate system determined by the centroid
position of implant I. The alignment set up for Model B consisted of XY-plane formed by
implants P, S, W; X-axis formed by implants P and W; and origin of local coordinate system
determined by the centroid position of implant P (Figure 1).
Distortion parameters
Distortion values were derived by calculating the difference in 3D coordinate values between
master and test models. Three linear distortions (dx, dy, dz) were computed for each implant,
indicating distortion along the x-, y-, and z-axes respectively. Global linear distortion (dR) is
7

defined by the formula dR = √(dx2 + dy2 + dz2).
Two absolute angular distortions (Absdθx and Absdθy), occurring about the x- and y-axes
respectively, were computed. Absdθx reflects anteroposterior angular distortions, while Absdθy
reflects mediolateral angular distortions.
3D reference distances (R) between the centroids of implants in test models were compared
with those in the master models to derive 3D reference distance distortions (ΔR).
To allow comparison between Models A and B, homologous implant location pairs were
assigned (Table 2).
The dependent variables were dx, dy, dz, dR, Absdθx, Absdθy and ΔR, as detailed above for each
of the six- or eight-implant models. One-way ANOVA and post-hoc Tukey HSD test were
used for the comparison of the six impression systems in each of the inter-implant distance
Models (A or B), by implant location. Independent samples t-tests between homologous
implant location pairs were conducted to compare Model A with Model B. Significance was
reported at a level of alpha = 0.05. The standard deviations of the measured variables reflect
the precision of each impression system. All statistical calculations were performed using a
statistical software (SPSS Statistics v21.0, IBM).
RESULTS
The 3D linear distortions (dx, dy, dz) as well as the global linear distortion (dR) for Models A
and B are shown in Figures 4 and 5 and Tables 3 and 4. Model A mean dR ranged from
8.7±8.3µm to 731.7±62.3µm. Model B mean dR ranged from 16.3±9µm to 620.2±63.2µm
The absolute angular distortions (Absdθx and Absdθy) for Models A and B are shown in
Figures 6 and 7 and Tables 3 and 4. Model A mean Absdθx ranged from 0.021±0.205º to -
2.349±0.166º, while mean Absdθy ranged from -0.002±0.160º to -0.932±0.290º. Model B
mean Absdθx ranged from -0.007±0.076º to -0.688±0.574º, while mean Absdθy ranged from -
0.018±0.048º to -1.052±0.297º.
Mean 3D reference distance distortions (ΔR) for Models A and B are shown in Figures 8 and
9.
The data was initially submitted to the Shapiro-Wilk test to check for the assumption of
normality, and it was not rejected. One-way ANOVA by Impression system revealed
significant differences among the test groups for dR and ΔR in both Models A and B for all
8

implant locations except the origin. For Absdθx, significant differences among the test groups
were found for implants K and M in Model A, as well as implants P, T, U and W in Model B.
For Absdθy, significant differences among the test groups were found for implants I, K, M and
N in Model A, as well as R, S, T, V and W in Model B. Tables 5 and 6 summarize the results
of the one-way ANOVA and Tukey HSD procedures performed.
Independent samples t-tests for dR, between homologous implant location pairs in Model A
versus Model B revealed presence of two to four significant pairings out of seven possible
pairings for test groups Impregum, TRIOS, True Definition, and inEos X5 (Table 7). dR was
larger in Model A than in B for the intra-oral scanner systems by 110-150µm. dR was larger in
Model B than in A for Impregum and inEos X5, but by a much smaller extent of less than
32µm. Inter-implant distance did not affect Absdθx or Absdθy in any consistent manner (Table
7).
DISCUSSION
Implant impression systems may be broadly categorized into three groups: (1) conventional
impression; (2) digital impression with intra-oral scanner; (3) digital impression with dental
laboratory scanner. In general, the systems utilising intra-oral scanners were found to fare
poorer, although this reached significance only for True Definition in most instances. These
trends were applicable to both the six-implant (Model A) and eight-implant (Model B)
configurations.
Linear distortions
The results showed significantly poorer performance of True Definition for all linear
distortions. Another interesting finding related to True Definition was the consistent negative
linear distortions in the x- and y-directions for both Models A and B. Negative dx and dy
indicated ‘shrinkage’ in virtual model arch width and height respectively, which can lead to
fabrication of a smaller framework.
Absolute angular distortions
The impression systems could not be consistently ranked for absolute angular distortions.
Negative Absdθx values indicate posterior tilting of implants towards the palate, and negative
Absdθy values indicate the tilting of implants towards the patient’s right (Figures 2A and 2B).
9

Except for Impregum, it was observed that the greatest absolute angular distortions typically
occurred at Implant K for Model A and Implants S or W for Model B. Implants K and S are
diagonally placed with respect to their respective origins (I or P), while Implants W and N are
located across the arch from origins (I or P). The consistency in sites affected by the greatest
absolute angular distortions hint at possible challenges when using digital systems to register
diagonally placed implants or when crossing the dental arch. However, this did not reach
statistical significance.
3D reference distance distortions
The mathematical derivation of the global linear distortion (dR) in itself will always derive a
positive value. Therefore, 3D reference distance distortions, ΔR, between all implant locations
to the respective origins (I in Model A and P in Model B) were used to reveal increases or
decreases in R.
The results showed significantly poorer performance of True Definition for all 3D reference
distance distortions. The consistently negative 3D reference distance distortion values
correspond to the negative linear distortions (dx and dy) that were described earlier. This
implied shrinkage of virtual models when True Definition was utilized, and this distortion
ranged from -268 to -709µm in Model A and from -151 to -602µm in Model B.
D900 showed a similar tendency for negative 3D distance distortions for some reference
distances, but the magnitude of distortion was much lower with ΔR ranging from -60 to +3µm
in Model A and from -4 to -35µm in Model B. The remaining four impression systems
displayed an opposite trend, with distortions tending to be positive, implying an expansion of
the resultant model.
Mean ΔR results for TRIOS in Model A appeared counter-intuitive, with mean ΔR decreasing
as the magnitude of the reference distances increased. The results were verified to be correct
and taking into account the large standard deviations, ΔR for TRIOS for the various 3D
reference distances were actually not significantly different.
Comparison between Models A and B
For Impregum, mean dR was found to be statistically greater in Model B than in Model A for
three out of seven possible homologous implant location pairs. The magnitude of this
difference was, however, less than 16µm and unlikely to be clinically significant.
Differences in global linear distortion magnitude between Models A and B were detected more
readily for the intra-oral scanner systems. Of the homologous implant location pairs that
10

showed significance, the distortions noted for Model A were consistently greater than Model B
by 110-150µm. The limited image fields of the respective intra-oral scanner systems may have
been a limitation in image acquisition and stitching of the longer edentulous spans in Model A.
For the dental laboratory scanner systems, significant homologous implant location pairs were
noted only for inEos X5, for which mean dR was found to be significantly greater in Model B
than in Model A for three out of seven possible pairings. The magnitude of this difference was
greater than for Impregum, which may be because a conventional impression was required in
the workflow for dental laboratory scanner systems. Nonetheless, the difference between
models was less than 32µm and unlikely to be clinically significant.
Comparison with previous studies
3D reference distance distortions for Impregum were found to be comparable to the results
shown by Wee et al4 who compared four linear distances between test and master models
replicating the clinical scenario of an edentulous arch with five implants. For polyether, mean
deviation was reported as 16.2±8.8µm. In the current study, ΔR for Impregum ranged from
1.9±10.3µm to 23.4±9.5µm in Model A and ranged from 6.0±2.3µm to 35.5±9.2µm in Model
B.
Papaspyridakos et al7 reported that median 3D deviation was 5–13µm for conventional
impressions with splinted impression copings, 13–132µm for conventional impressions with
non-splinted impression copings and 8–29µm for digital impression with a 3Shape TRIOS
intra-oral scanner. Vandeweghe et al8 reported mean deviation for Lava COS, 3M True
Definition, Cerec Omnicam and 3Shape TRIOS to be 112±25µm, 35±12µm, 61±23µm and
28±7µm respectively. However, these values should not be compared directly with the present
study, due to fundamental differences in methodologies for assessment of accuracy. The
methodologies previously reported have utilized linear distance measurements18; scanning of
calibrated objects of known dimensions19; measuring fit of final prostheses20; superimposition
of digital images21; and use of a CMM15,16. The above two studies on multiple implants for the
edentulous arch7,8 compared test and control groups by superimposing test scans with reference
scans, hence all surfaces of each scan body contributed to the measured distortion. On the other
hand, this study utilized the CMM for distortion analysis at the implant platform level. For
Model A, dR for TRIOS ranged from 36.3±28.6µm to 172.5±74.2µm, while that for True
Definition ranged from 291.6±79.8µm to 731.7±62.3µm. For Model B, dR for TRIOS ranged
from 40.2±19.4µm to 148.0±35.4µm, while that for True Definition ranged from
160.9±20.1µm to 620.2±63.2µm.
11

Considerations for selecting implant impression system for the edentulous arch
The dental laboratory scanner systems’ fabrication workflow incorporates a conventional

impression. By calculating the mean differences in 3D reference distance distortions between
Impregum and the dental laboratory scanner systems, we may quantify the magnitude of
distortion resulting from the digital laboratory scanning step only. Reference distances I-L and
P-T in Models A and B respectively were analyzed to determine accuracy between diagonally
placed implants in an edentulous arch, while reference distances I-N and P-W were analyzed to
determine cross-arch accuracy (Table 8). The magnitude of distortion resulting from the dental
laboratory scanner scanning step alone for Ceramill Map400, inEos X5 and D900 was found to
range from 13.8 to -28.4µm for I-L; 28.7 to -19.2µm for P-T; 14.9 to -79.9µm for I-N; and 6.7
to -57.7µm for P-W.
This study demonstrated that impression system contributed to varying levels of distortion.
However, the ensuing fabrication steps also contribute to overall prosthesis accuracy. Tan et
al22, using an edentulous mandibular master cast with five parallel implants, quantified
distortions in full-arch implant prosthesis castings. Overall linear displacements for the five
cylinders (implants) were found to be 20.2±14.5µm. Overall angular displacements for all five
cylinders were -0.098±0.109º for dθx and 0.020±0.097º for dθy. Unfortunately, studies
evaluating CAD/CAM fabrication of full-arch implant prostheses using similar distortion
measurement methodology are lacking. Örtorp et al23 compared conventional castings with
milled frameworks, and concluded that the milled frameworks exhibited better fit and
precision. However, distortions were calculated by the superimposition technique and the
results cannot be directly extrapolated to the current study. Further investigation is warranted
to identify the direction and magnitude of distortion at each stage of CAD/CAM prosthesis
fabrication. The clinician or technician may then balance positive distortions with negative
distortions to minimize overall inaccuracy.
Software compatibility with a variety of scan bodies and implant systems influences utility.
Two of the investigated dental laboratory scanner systems (Ceramill Map400 and inEos X5)
required use of manufacturer-specified scan bodies, while the investigated intra-oral scanner
systems (TRIOS and True Definition) and one dental laboratory scanner (D900) allowed use of
third-party scan bodies. Impression systems that allow use of third-party scan bodies may
simplify a clinic’s inventory. However, it was noted that the standard deviations for those
impression systems that utilize third-party scan bodies were larger, though these trends did not
reach statistical significance. One speculation for this lower precision is a lack of optimisation
by the various scanners possibly leading to poorer scan body recognition.
12

Fit of scan body into the implant / implant analog is another potential source of error; the
coaxiality of scan body and implant will influence angular distortions in particular. All three
scan bodies used in this study were equally challenged in this aspect, as none were produced
by the implant manufacturer. Stimmelmayr et al24 evaluated the reproducibility of implant scan
body fit, but the protocol included digitization of models with their attached scan bodies.
Errors inherent to this digitization process may obscure the true distortions that had resulted
from poor fit of scan bodies to the implant or implant analog. More specific investigations are
required to determine the effect of scan body fit on scan accuracy.
Due to the lack of precise instructions from manufacturers on the recommended torque
magnitude for scan body insertion, pilot studies were conducted to investigate the effect of
torque magnitude on scan body distortion and thus scan body height. Five scan bodies from
each test group were tested on a single implant analog. Each scan body was secured with
increasing torque magnitudes of 5, 10 and 15Ncm. Six probe hits were used for the registration
of the top flat surface of the scan body, while eight probe hits were used to register the external
cylindrical (TRIOS, True Definition, inEos X5 and D900) or conical (Ceramill Map400)
surfaces. The pierce point between the central axis and the top plane of each scan body defined
the ‘scan body centroid’. Scan bodies were subsequently removed to allow for CMM
registration of the implant analog coronal flat plane and internal conical surface; these
registrations then allowed for the pierce point to be calculated, deriving the ‘analog centroid’.
The 3D distance between each scan body centroid and the analog centroid was calculated. At
the same time, coaxiality values were monitored.
Mean coaxiality values for Core3D scan bodies were found to be much higher than those of the
Amann Girrbach and Sirona scan bodies (Figure 10). These values were also much higher than
those found in an earlier study15,16 that utilized an earlier version of the Core3D scan bodies.
The older Core3D scan body was entirely constructed with PEEK, but the newer version used
in this study consisted of a lower metal segment which seats on the coronal flat plane of each
implant. Possible improper mating of PEEK to metal during the scan body fabrication process
may have resulted in a non-cylindrical scan body and consequently poorer coaxiality with
respect to the corresponding implant. This may also explain the counterintuitive trends in scan
body height for the Core3D scan body as torque magnitude increased (Figure 11). Amann
Girrbach and Sirona scan body heights decreased as torque magnitude increased, indicating
compression of scan bodies. Interestingly, this compression was not observed for the Core3D
scan bodies; mean scan body height decreased when torque application increased from 5 to
10Ncm, but increased when torque application was increased from 10 to 15Ncm. These
13

findings hint at innate scan body distortions that occur and highlight obvious differences
between scan body brands. The contribution of the scan body towards overall accuracy of the
acquired virtual models therefore should be taken into account.
In addition, two studies demonstrated that the machining tolerance of implant components is
measurable and should be considered in all accuracy studies. Ma et al25 evaluated two
generations of Nobel Biocare implant components, and found that tolerances ranged from 22–
100µm. Braian et al26 utilized external- and internal-hexagon Biomet 3i implants. Horizontal
machining tolerances on prefabricated gold abutments for external-hexagon implants was
44±9µm, while that for internal-hexagon implants was 58±28µm. When castable plastic
abutments were used, horizontal tolerance for the external-hexagon implants was 12±89µm,
while that for internal-hexagon implants was 86±47µm.
It appears that the dental laboratory scanner systems studied are appropriate for the derivation
of implant models in the fully-edentulous patient. The conventional impression provides
excellent 3D accuracy and precision, and errors attributable to the subsequent digital laboratory
scanning step are minor. Dental laboratory scanner systems provide the option for a digital
workflow, with the most important advantages being fabrication consistency and the
possibility of using zirconia and other CAD/CAM materials.
True Definition exhibited the poorest accuracy and does not appear to be appropriate for the
investigated clinical scenario. The overall scan procedure for True Definition was found to be
more difficult to achieve compared to TRIOS, which may reflect limitations in system
capability. Use of a powder spray for True Definition may also contribute to errors, as it is
impossible to evenly coat all areas of interest. TRIOS was not found to be significantly
different from other test groups. However, for dR, the results indicate slightly greater distortion
of TRIOS as compared to Impregum and the dental laboratory scanner systems. Intra-oral
scanner systems may still be utilized if there are clear clinical contraindications to the use of
the conventional impression system, such as patient intolerance towards conventional
impression materials. In this situation, factors such as the presence of unique soft tissue
morphology may increase the accuracy of intra-oral scanning, and attempts should be made to
compensate for non-characteristic morphology. Lee27 has recently suggested that unique
shapes drawn on the palate with pressure-indicating paste may enhance the stitching of images.
Limitations
The in-vitro nature of this study eliminated a number of patient- and operator-related factors
that could have affected results. All impressions (conventional or intra-oral scans) were made
14

on master models with no surrounding saliva, soft tissue and oral musculature and thus there
was uninhibited access and visibility. While the relative ease of the impression step was
equally applicable to all systems in this study, the increase in difficulty in-vivo is unlikely to
be of equal extent.
Factors such as patient comfort also could not be evaluated. Wismeijer et al28 suggests that
intra-oral scanners were preferred due to the unpleasant taste of the conventional impression
material and the preparation required for the conventional impression. However, patients of the
same study perceived the duration of the intra-oral scan more negatively. Moreover, some
intra-oral scanner systems such as the 3M True Definition Scanner require use of a surface
scanning spray, application of which increases time needed and discomfort experienced.
Parallel implants were chosen over angulated ones, which have been reported to affect scan
accuracy16. This allowed for a reduction of experimental variables in this study, thus focusing
on the role of inter-implant distance on impression system accuracy. However, clinically it is
likely that implants in the maxilla are angulated and this should be a variable to be further
investigated.
The time required to master each implant impression system may vary within the same
operator and among different operators. Therefore, no limitations were imposed on the number
of tries allowed for each impression system in this study. The single operator in-charge of
performing all conventional impressions, intra-oral scans and dental laboratory scanner scans
underwent training provided by the respective suppliers, and was guided by other operators
more experienced and familiar with each impression system. Conventional impressions were
only accepted if they were of reasonable quality, in that there was proper fusion of impression
material, complete capture of the area of interest, and absence of gross irregularities, bubbles
and separation of impression material from the tray. The digital impressions were evaluated
using different criteria. Virtual models were checked using the respective instantaneous
previews to ensure that scan images were properly stitched. Further, the reference scan bodies
had to be captured in their entirety with good detail reproduction; scans with incomplete or
distorted scan bodies were rejected. The results of this study reflect the accuracy of implant
impression systems based on impressions or scans that would have been accepted by any
operator, regardless of familiarity or experience, using widely-accepted and reasonable criteria.
Operator familiarity thus did not pose a true limitation to this study.
Centroid position determination depended on the accuracy of measured geometric features and
therefore strict tolerance criteria were implemented. It may be argued that this level of
15

scanning accuracy is unnecessary, as the various software are still able to determine implant
position via a ‘best-fit’ algorithm. Unfortunately, the accuracy of the algorithms present in
each software cannot be evaluated, as the proprietary data manipulation steps are not reported
by any manufacturer.
The current study simulated six-implant and eight-implant situations in the edentulous maxilla.
These two scenarios were chosen for their relatively common indication in completely
edentulous patients requesting reconstruction via fixed prostheses. Future studies evaluating an
alternative number of implants (e.g. two or four implants) in the edentulous maxilla will
further aid clinicians who wish to attempt digital impressions for the fabrication of a
removable prosthesis assisted by two or more implants, or for the fabrication of an ‘All-On-
Four®’ fixed prosthesis.
CONCLUSIONS
This in-vitro study examined the effect of implant impression system and inter-implant
distance on the 3D accuracy of implant positions in the resultant stone or virtual models.
Within the limitations of this study, the following conclusions were made:
1. Impregum consistently exhibited the lowest or second lowest dR at all implant locations.
2. True Definition exhibited the poorest accuracy for all linear distortions (dx, dy, dz and
dR) and 3D reference distance distortion (ΔR) parameters in both Models A and B, but
not for absolute angular distortions (Absdθx and Absdθy).
3. Excluding True Definition, there was no significant difference among the remaining five
impression systems for linear distortion parameters (dx, dy, dz and dR) in both Models A
and B.
4. The 6 impression systems could not be consistently ranked for angular distortions.
5. Reducing inter-implant distance may decrease global linear distortions (dR) for intra-oral
scanner systems, but had no effect on Impregum and the dental laboratory scanner
systems.
ACKNOWLEDGEMENTS
The authors thank Mr. Chan Swee Heng and Mr. Tok Wee Wah for technical assistance. We
also thank Straumann, Core3D Centres, 3Shape, 3M ESPE, Amann Girrbach and Sirona for
providing system-specific training and support.
16

REFERENCES
[1] Hämmerle CH, Cordaro L, van Assche N, Benic GI, Bornstein M, Gamper F,
Gotfredsen K, Harris D, Hürzeler M, Jacobs R, Kapos T, Kohal RJ, Patzelt SB, Sailer I,
Tahmaseb A, Vercruyssen M, Wismeijer D. Digital technologies to support planning,
treatment, and fabrication processes and outcome assessments in implant dentistry.
Summary and consensus statements. The 4th EAO consensus conference 2015. Clin
Oral Implants Res. 2015;26:97-101.
[2] Grant GT, Campbell SD, Masri RM, Andersen MR, American College of
Prosthodontists Digital Dentistry Glossary Development Task Force. Glossary of
digital dental terms: American College of Prosthodontists. J Prosthodont. 2016;25:2-9.
[3] Lee H, So JS, Hochstedler JL, Ercoli C. The accuracy of implant impressions: a
systematic review. J Prosthet Dent. 2008;100:285-91.
[4] Wee AG. Comparison of impression materials for direct multi-implant impressions. J
Prosthet Dent. 2000;83:323-31.
[5] Hoods-Moonsammy VJ, Owen P, Howes DG. A comparison of the accuracy of

polyether, polyvinyl siloxane, and plaster impressions for long-span implant-supported
prostheses. Int J Prosthodont. 2014;27:433-8.
[6] Gherlone E, Capparé P, Vinci R, Ferrini F, Gastaldi G, Crespi R. Conventional versus

digital impressions for “all-on-four” restorations. Int J Oral Maxillofac Implants.
2016;31:324-30.
[7] Papaspyridakos P, Gallucci GO, Chen CJ, Hanssen S, Naert I, Vandenberghe B. Digital
versus conventional implant impressions for edentulous patients: accuracy outcomes.
Clin Oral Implants Res. 2016;27:465-72.
[8] Vandeweghe S, Vervack V, Dierens M, De Bruyn H. Accuracy of digital impressions

of multiple dental implants: an in vitro study. Clin Oral Implants Res. 2017;28:648-53.
[9] 3Shape. 2013. 3Shape TRIOS Color Digital Impression Solution. [ONLINE] Available
at: https://www.3shape.com/en/knowledge-center/news-and-press/press-
releases/2013/3shape-launches-trios-colour. (Accessed 18 March 2018)
[10] 3M. 2015. 3M True Definition Scanner Technical Specifications. [ONLINE] Available
at: http://multimedia.3m.com/mws/media/919861O/3m-true-definition-scanner-
technical-specifications.pdf. (Accessed 10 January 2017).
17

[11] Amann Girrbach. 2017. Amann Girrbach - Products - Ceramill Map400. [ONLINE]
Available at: https://www.amanngirrbach.com/en/products/scan/ceramill-map400/.
(Accessed 10 January 2017).
[12] Dental Compare. 2016. Ceramill Map 400 Scanner from AmannGirrbach GmbH.
[ONLINE] Available at: http://www.dentalcompare.com/4723-Dental-Laboratory-3D-
Scanning-Systems/4459992-Ceramill-Map-400-Scanner/. (Accessed 10 January 2017).
[13] Sirona. 2013. The inEos X5: A multi-talent for the dental lab. [ONLINE] Available at:
http://www.sirona.com/en/news-events/news-press/digital-dentistry-news-
detail/27118/. (Accessed 10 January 2017).
[14] Dental Equipment Store. 2017. 3Shape D900 Series. [ONLINE] Available at:
http://www.dentalequipmentcenter.com/3shape-dental.html. (Accessed 10 January
2017).
[15] Chew AA, Esguerra RJ, Teoh KH, Wong KM, Ng SD, Tan KB. Three-dimensional
accuracy of digital implant impressions: effects of different scanners and implant level.
Int J Oral Maxillofac Implants. 2017;32:313-21.
[16] Chia VA, Esguerra RJ, Teoh KH, Teo JW, Wong KM, Tan KB. In Vitro Three
Dimensional Accuracy of Digital Implant Impressions: The Effect of Implant
Angulation. Int J Oral Maxillofac Implants 2017;32:313-21.
[17] Teo JW, Tan KB, Nicholls JI, Wong KM, Uy J. Three-dimensional accuracy of plastic
transfer impression copings for three implant systems. Int J Oral Maxillofac Implants
2014;29:577-84.
[18] Chandran DT, Jagger DC, Jagger RG, Barbour ME. Two- and three-dimensional
accuracy of dental impression materials: effects of storage time and moisture
contamination. Biomed Mater Eng. 2010;20:243-9.
[19] Del Corso M, Aba G, Vazquez L, Dargaud J, Dohan Ehrenfest DM. Optical three-
dimensional scanning acquisition of the position of osseointegrated implants: an in
vitro study to determine method accuracy and operational feasibility. Clin Implant Dent
Relat Res. 2009;11:214-21.
[20] Baig MR, Tan KB, Nicholls JI. Evaluation of the marginal fit of a zirconia ceramic
computer-aided machined (CAM) crown system. J Prosthet Dent. 2010;104:216-27.
[21] Ender A, Mehl A. Accuracy of complete-arch dental impressions: a new method of

measuring trueness and precision. J Prosthet Dent. 2013;109:121-8.
18

[22] Tan KB, Rubenstein JE, Nicholls JI, Yuodelis RA. Three-dimensional analysis of the
casting accuracy of one-piece, osseointegrated implant-retained prostheses. Int J
Prosthodont. 1993;6:346-63.
[23] Örtorp A, Jemt T, Bäck T, Jälevik T. Comparisons of precision of fit between cast and
CNC-milled titanium implant frameworks for the edentulous mandible. Int J
Prosthodont. 2003;16:194-200.
[24] Stimmelmayr M, Güth JF, Erdelt K, Edelhoff D, Beuer F. Digital evaluation of the
reproducibility of implant scanbody fit – an in vitro study. Clin Oral Investig.
2012;16:851-6.
[25] Ma T, Nicholls JI, Rubenstein JE. Tolerance measurements of various implant

components. Int J Oral Maxillofac Implants. 1997;12:371-5.
[26] Braian M, De Bruyn H, Fransson H, Christersson C, Wennerberg A. Tolerance

measurements on internal- and external-hexagon implants. Int J Oral Maxillofac
Implants. 2014;29:846-52.
[27] Lee JH. Improved digital impressions of edentulous areas. J Prosthet Dent.
2017;117:448-9.
[28] Wismeijer D, Mans R, van Genuchten M, Reijers HA. Patients’ preferences when
comparing analogue implant impressions using a polyether impression material versus
digital impressions (Intraoral Scan) of dental implants. Clin Oral Implants Res.
2014;25:1113-8.
19

TABLES
Table 1 Features of test groups
System Impregum, TRIOS9 True Ceramill inEos X513 D90014

PentaTM Soft Definition10 Map40011,12
3M ESPE 3Shape 3M ESPE Amann Sirona 3Shape
Girrbach
Category Conventional Digital Digital

(intra-oral scanner) (dental laboratory scanner)
Material / Polyether Optical Active Strip light Strip light Triangulation

technology sectioning wavefront projection projection using 4 x 5
sampling with blue megapixel
light cameras and
blue LED
Impression RC Core Core Range three inPost Scan Core
coping / Scan impression Scanbody, Scanbody, Kit b Scan Body Scanbody,
body post (2077) (2077) Body (6551639) (2077)
(025.4202) (792322) on 2-
with screw CONnect
(792341) abutment (L
810 M)
Straumann Core3D Core3D Amann Sirona Core3D
Centres Centres Girrbach Centres
Need powder? No No Yes No No No
Registers No Yes No No No Yes

colour?
Manufacturer- NA Not Not 20 µm 12 µm 15 µm

specified specified specified
accuracy
20

Table 2 Homologous implant location pairs
Homologous implant location pairs

(Model A ó Model B) Rationale
JóQ Implant in right premolar region

JóR Implant in right premolar region
KóS Most anterior, right implant location
LóT Most anterior, left implant location
MóU Implant in left premolar region
MóV Implant in left premolar region
NóW Direct cross-arch implant location
Table 3 Model A global linear, absolute angular and 3D reference distance distortions
Impregum TRIOS True Ceramill inEos X5 D900

Definition Map400
Mean I 0.0 ( 0) 0.0 ( 0) 0.0 ( 0) 0.0 ( 0) 0.0 ( 0) 0.0 ( 0)
dR, µm J 10.9 ( 5.0) 172.5 (74.2) 291.6 ( 79.8) 52.8 (25.3) 25.4 ( 8.6) 62.3 (14.4)
(SD) K 15.6 ( 6.6) 153.6 (40.7) 482.6 ( 93.1) 49.9 (13.1) 54.2 (45.5) 53.9 ( 9.1)
L 24.6 (10.2) 143.4 (70.7) 716.6 (116.0) 49.8 (20.3) 53.3 (14.5) 59.5 (11.5)
M 24.4 ( 9.6) 100.0 (57.9) 731.7 ( 62.3) 42.5 (12.7) 43.9 (13.9) 87.2 (31.1)
N 20.1 ( 9.3) 36.3 (28.6) 677.1 ( 50.3) 8.7 ( 8.3) 35.0 (19.2) 59.8 (40.0)
Mean I -0.273 (0.069) -0.676 (0.338) -0.875 (0.866) -0.827 (0.185) -0.618 (0.071) -0.537 (0.269)
Absdθx, J -0.035 (0.163) 0.021 (0.205) -0.037 (0.520) -0.359 (0.078) -0.129 (0.154) -0.150 (0.175)
º (SD) K -0.096 (0.133) -2.189 (0.453) -2.249 (1.100) -2.349 (0.166) -1.620 (0.539) -1.944 (0.671)
L 0.153 (0.133) -0.096 (0.241) -0.243 (0.684) -0.073 (0.074) 0.054 (0.039) -0.159 (0.130)
M 0.276 (0.250) 0.306 (0.106) -0.063 (0.414) -0.036 (0.305) 0.224 (0.081) -0.437 (0.606)
N 0.151 (0.399) -0.313 (0.374) -0.437 (1.302) -0.757 (0.240) -0.350 (0.319) -0.269 (0.358)
Mean I -0.089 (0.069) 0.215 (0.292) 0.469 (0.320) 0.611 (0.351) 0.533 (0.133) 0.295 (0.208)
Absdθy, J -0.075 (0.127) -0.141 (0.304) -0.219 (0.272) -0.229 (0.302) -0.233 (0.071) -0.222 (0.051)
º (SD) K -0.039 (0.082) -0.932 (0.290) -0.715 (0.360) -0.774 (0.180) -0.537 (0.158) -0.347 (0.187)
L 0.138 (0.135) 0.136 (0.241) -0.143 (0.245) 0.125 (0.098) 0.064 (0.059) 0.220 (0.349)
M 0.322 (0.106) 0.186 (0.072) -0.304 (0.346) 0.031 (0.266) 0.106 (0.127) 0.472 (0.362)
N 0.061 (0.151) 0.110 (0.201) -0.155 (0.257) 0.307 (0.116) -0.002 (0.160) 0.446 (0.462)
Mean I-J 1.9 (10.3) 166.8 ( 78.0) -267.5 ( 85.4) 28.9 (36.1) 11.1 ( 9.3) -12.3 (37.0)
ΔR, µm I-K 11.1 (12.6) 119.0 ( 61.4) -480.9 ( 92.9) 29.5 (32.9) 25.4 (10.8) 2.7 (32.6)
(SD) I-L 22.0 (12.2) 71.2 (107.8) -703.9 (121.1) 35.8 (31.9) 45.4 (20.2) -6.4 (39.1)
I-M 23.4 ( 9.5) 26.4 ( 39.6) -709.2 ( 66.8) 30.5 (24.6) 40.1 (15.1) -49.8 (23.1)
I-N 20.1 ( 9.3) 13.3 ( 47.4) -677.1 ( 50.3) -4.8 (11.6) 35.0 (19.2) -59.8 (40.0)
SDs in parantheses.
21

Table 4 Model B global linear, absolute angular and 3D reference distance distortions
Impregum TRIOS True Ceramill inEos X5 D900

Definition Map400
Mean P 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0)
dR, µm Q 16.3 ( 9.0) 40.2 (19.4) 160.9 ( 20.1) 48.9 ( 7.3) 23.4 ( 9.9) 68.7 (27.6)
(SD) R 19.1 ( 9.8) 40.8 (36.3) 369.1 ( 64.5) 30.2 ( 7.4) 31.0 ( 5.7) 62.7 (35.3)
S 27.3 (10.6) 73.5 (35.9) 464.8 (100.0) 36.2 (13.5) 48.0 ( 9.9) 46.1 (25.3)
T 35.3 ( 8.4) 148.0 (35.4) 611.1 ( 85.2) 56.1 (15.8) 60.3 (15.0) 65.6 (23.9)
U 39.3 ( 5.8) 75.3 (31.7) 620.2 ( 63.2) 46.3 (14.7) 65.4 (10.7) 47.1 (27.9)
V 34.8 ( 2.5) 110.4 (86.6) 582.1 ( 61.5) 37.4 (21.5) 64.7 ( 4.7) 53.0 (42.8)
W 35.5 ( 9.2) 59.9 (47.9) 567.8 ( 30.5) 32.3 (25.7) 66.4 ( 5.9) 73.4 (24.0)
Mean P -0.151 (0.142) 0.532 (0.326) 0.511 (0.379) 0.596 (0.141) 0.136 (0.170) 0.140 (0.371)
Absdθx, Q 0.022 (0.239) -0.112 (0.214) 0.021 (0.157) -0.153 (0.187) -0.034 (0.031) 0.226 (0.314)
º (SD) R 0.205 (0.147) 0.432 (0.305) 0.130 (0.303) 0.226 (0.261) 0.134 (0.276) 0.306 (0.261)
S -0.243 (0.417) -0.476 (0.252) -0.688 (0.574) -0.459 (0.222) -0.349 (0.105) -0.312 (0.321)
T -0.038 (0.269) 0.330 (0.340) 0.335 (0.242) 0.329 (0.112) 0.160 (0.054) 0.041 (0.123)
U 0.233 (0.112) -0.361 (0.313) -0.135 (0.304) -0.109 (0.146) -0.032 (0.067) -0.146 (0.130)
V -0.110 (0.130) -0.108 (0.293) -0.260 (0.183) 0.069 (0.137) -0.182 (0.046) -0.227 (0.075)
W -0.007 (0.076) 0.232 (0.164) 0.226 (0.162) 0.448 (0.141) 0.224 (0.101) 0.161 (0.071)
Mean P -0.091 (0.116) 0.150 (0.188) 0.185 (0.451) 0.284 (0.244) -0.018 (0.048) -0.153 (0.342)
Absdθy, Q 0.162 (0.091) 0.255 (0.055) 0.144 (0.178) 0.484 (0.123) 0.328 (0.022) 0.196 (0.447)
º (SD) R -0.066 (0.102) 0.377 (0.081) 0.706 (0.547) 0.706 (0.141) 0.136 (0.252) 0.387 (0.104)
S -0.084 (0.129) -0.812 (0.330) -1.052 (0.297) -0.795 (0.112) -0.526 (0.139) -0.376 (0.175)
T 0.067 (0.152) 0.117 (0.141) 0.145 (0.068) 0.435 (0.104) 0.154 (0.180) 0.155 (0.200)
U 0.267 (0.605) 0.161 (0.107) 0.083 (0.150) 0.493 (0.176) 0.178 (0.068) 0.043 (0.109)
V -0.053 (0.100) 0.333 (0.314) 0.134 (0.288) 0.632 (0.187) 0.180 (0.084) 0.207 (0.149)
W 0.080 (0.146) -0.714 (0.315) -0.784 (0.340) -0.839 (0.296) -0.877 (0.381) -0.707 (0.240)
Mean P-Q 6.0 ( 2.3) 12.6 ( 17.2) -151.1 (32.8) 22.6 (10.3) 14.6 ( 7.5) -34.7 (28.8)
ΔR, µm P-R 8.6 ( 8.5) -9.1 ( 28.9) -365.3 (68.7) 9.8 (15.7) 27.6 ( 9.1) -16.8 (40.6)
(SD) P-S 16.9 ( 9.5) 28.7 ( 69.0) -463.4 (99.9) 29.1 (10.9) 42.2 (14.0) -4.2 (26.3)
P-T 28.6 (11.6) 38.8 (126.6) -596.4 (90.3) 50.2 (20.9) 57.3 (16.1) 9.4 (28.9)
P-U 31.2 (11.2) 33.3 ( 41.6) -602.5 (70.0) 37.8 (20.9) 62.6 (12.2) 9.5 (45.9)
P-V 30.5 ( 5.9) 69.8 (109.2) -574.2 (63.0) 23.4 (31.0) 62.3 ( 5.1) 23.9 (61.6)
P-W 35.5 ( 9.2) 50.1 ( 60.4) -567.8 (30.5) 28.8 (30.4) 66.4 ( 5.9) -22.2 (81.8)
SDs in parantheses.
22

Table 5 Summary of One-way ANOVA and Tukey HSD for global linear and absolute angular
distortions, by Implant Location.
Variable Model Implant p-value Statistical significant subsets

Location Impregum TRIOS True Ceramill inEos D900
Definition Map400 X5
dR A I .000 - - - - - -
J .000 a b c a a a
K .000 a b c a a a
L .000 a b c ab ab ab
M .000 a a b a a a
N .000 a a b a a a
B P .000 - - - - - -
Q .000 a ab c ab a b
R .000 a a b a a a
S .000 a a b a a a
T .000 a b c a a a
U .000 a a b a a a
V .000 a a b a a a
W .000 a a b a a a
Absdθx A I .243 a a a a a a
J .265 a a a a a a
K .000 b a a a a a
L .392 a a a a a a
M .018 b b ab ab ab a
N .372 a a a a a a
B P .001 a b b b ab ab
Q .109 a a a a a a
R .471 a a a a a a
S .426 a a a a a a
T .034 a a a a a a
U .005 b a ab ab ab ab
V .054 ab ab a b ab ab
W .000 a ab ab b ab a
Absdθy A I .002 a ab b b b ab
J .812 a a a a a a
K .000 c a ab ab ab bc
L .168 a a a a a a
M .001 b b a ab ab b
N .013 ab ab a ab ab b
B P .105 a a a a a a
Q .131 a a a a a a
R .000 a ab b b a ab
S .000 d ab a abc bc cd
T .010 a a a b ab ab
U .161 a a a a a a
V .001 a ab a b a a
W .000 b a a a a a
23

Table 6 Summary of One-way ANOVA and Tukey HSD for 3D reference distance (ΔR)
distortions
Variable Model Implant p-value Statistical significant subsets

Location Impregum TRIOS True Ceramill inEos D900
Definition Map400 X5
ΔR A I-J .000 b c a b b b
I-K .000 b c a b b b
I-L .000 b b a b b b
I-M .000 c c a c c b
I-N .000 c c a bc c b
B P-Q .000 c c a c c b
P-R .000 b b a b b b
P-S .000 b b a b b b
P-T .000 b b a b b b
P-U .000 b b a b b b
P-V .000 b b a b b b
P-W .000 b b a b b b
24

Table 7 Summary of independent samples t-test for homologous implant location pairs
between Models A and B. Significantly different test groups listed.
Homologous dR Absdθx Absdθy

implant
location A>B B>A A>B B>A A>B B>A
pairs
(A ó B)
JóQ TRIOS TRIOS Impregum
True Definition True Definition
JóR TRIOS TRIOS inEos X5
KóS
LóT True Definition
MóU True Definition Impregum True Definition

inEos X5 D900
MóV True Definition Impregum True Definition True Definition inEos X5

inEos X5 D900
Nó W True Definition Impregum True Definition Impregum

inEos X5 Ceramill Map400 True Definition
inEos X5 inEos X5
Table 8 3D distance distortion attributable to laboratory scanning
Reference distance Mean 3D Mean 3D distance Mean 3D distance

distance distortion in dental distortion
distortion in laboratory scanner systems attributable to
Impregum (µm) laboratory scanning
(µm) (µm)
Diagonal
I-L 22.0 Ceramill Map400 35.8 13.8
(Model A) inEos X5 45.4 23.4
D900 - 6.4 -28.4
P-T 28.6 Ceramill Map400 50.2 21.6
(Model B) inEos X5 57.3 28.7
D900 9.4 -19.2
Cross-arch
I-N 20.1 Ceramill Map400 - 4.8 -24.9
(Model A) inEos X5 35.0 14.9
D900 -59.8 -79.9
P-W 35.5 Ceramill Map400 28.8 - 6.7
(Model B) inEos X5 66.4 30.9
D900 -22.2 -57.7
25

FIGURES
Figure 1 Experimental Outline, with 6 tested systems
26

y
K L
J M
III
x
I N
(0,0,0) I II
B
z
y
S T
R U
Q V
III
x
P W
(0,0,0) I II
Figure 2 A, Schematic of Master Model A with six implants (I, J, K, L, M, N). B, Schematic of
C Master Model B with eight implants (P, Q, R, S, T, U, V, W). I-II-III, silicon nitride
ball bearings. Local coordinate system axes orientation with origin at centroid of
Implant I in Model A and Implant P in Model B respectively. X-axis is defined as I to
N centroids and P to W centroids respectively.
27

Scanbody central axis
A
Scanbody top platform plane
Distance between planes = scanbody height
Implant centroid

Implant platform plane
Implant internal cone
Implant central axis
B C D
Figure 3 A, implant centroid was defined by constructing a pierce point between the central
axis of the virtual scanbody / implant internal cone and the implant platform plane.
B, scanbody cylinder (TRIOS, True Definition, inEos X5, D900) or cone (Ceramill
Map400) was defined by eight virtual probe hits at two levels. C, scanbody top plane
was defined by four virtual probe hits. D, virtual scanbody with measured features.
28

800
600
Model A linear distortions, µm
400
200
-200
-400
-600
-800
I J K L M N I J K L M N I J K L M N I J K L M N
dx dy dz dR
Impregum TRIOS True DeBinition
Ceramill Map400 inEos X5 D900

Figure 4 Model A linear distortions, µm, by implant location.
29

800
600
Model B linear distortions, µm
400
200
-200
-400
-600
-800
P Q R S T U V W P Q R S T U V W P Q R S T U V W P Q R S T U V W
dx dy dz dR
Impregum TRIOS True DeBinition
Ceramill Map400 inEos X5 D900

Figure 5 Model B linear distortions, µm, by implant location.
30

1
Model A absolute angular

0.5
0
distortions, º
-0.5
-1
-1.5
-2
-2.5
I J K L M N I J K L M N
AbsdθX AbsdθY
Impregum TRIOS True DeBinition Ceramill Map400 inEos X5 D900

Figure 6 Model A absolute angular distortions, °, by implant location.
31

1
Model B absolute angular

0.5
distortions, º 0
-0.5
-1
-1.5
-2
-2.5
P Q R S T U V W P Q R S T U V W
AbsdθX AbsdθY
Figure 7 Model B absolute angular distortions, °, by implant location.
32

300
Model A 3D reference distance

200
100
distortion (ΔR), µm 0
-100
-200
-300
-400
-500
-600
-700
-800
I-J I-K I-L I-M I-N
ΔR
Figure 8 Model A 3D reference distance distortions, µm
33

300
Model B 3D reference distance

200
100
distortion (ΔR), µm 0
-100
-200
-300
-400
-500
-600
-700
-800
P-Q P-R P-S P-T P-U P-V P-W
ΔR
Figure 9 Model B 3D reference distance distortion, µm
34

140
Mean Coaxiality, μm 120
100
80
Core3D
60 Amann Girrbach
Sirona
40
20
0
5 10 15
1 2 3
Torque Magnitude, Ncm
Figure 10 Mean Coaxiality of scan bodies (µm), by System, variation with torque magnitude.
Error bars = SD.
60
Deviation from nominal scan body
50
40
30
height, μm
20
10 Core3D
0 Amann Girrbach
-10 Sirona
-20
-30
-40
5 10 15
1 2 3
Torque Magnitude, Ncm
Figure 11 Deviation from nominal scan body height (µm), by System, variation with torque
magnitude. Error bars = SD.
35

Comparison of Three-Dimensional Accuracy of Digital and Conventional Implant Impressions Effect of Interimplant Distance in An Edentulous Arch PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Comparison of Three-Dimensional Accuracy of Digital and Conventional Implant Impressions Effect of Interimplant Distance in An Edentulous Arch PDF

Uploaded by

Copyright:

Available Formats

This peer-reviewed, accepted manuscript will undergo final editing and production prior to publication in JOMI.

Comparison of Three-Dimensional Accuracy of Digital and Conventional

Submitted December 18, 2017; Accepted July 3, 2018.

Keywords: 3D accuracy, Implant, CMM, Digital impression, intra-oral scanner, dental

Conventional impressions remain essential in implant prosthodontics, with several studies

Vandeweghe et al8 evaluated four intra-oral scanner systems on a six-implant edentulous

MATERIALS AND METHODS

Fabrication of Master Models

Fabrication of Conventional Impression test models

Dental Laboratory Scanner

CMM measurements of Master Models and Test Models

Implant positions were determined by a CMM with manufacturer-specified measurement

defined by the formula dR = √(dx2 + dy2 + dz2).

Absolute angular distortions

3D reference distance distortions

Comparison between Models A and B

Comparison with previous studies

The dental laboratory scanner systems’ fabrication workflow incorporates a conventional

[5] Hoods-Moonsammy VJ, Owen P, Howes DG. A comparison of the accuracy of

[6] Gherlone E, Capparé P, Vinci R, Ferrini F, Gastaldi G, Crespi R. Conventional versus

[8] Vandeweghe S, Vervack V, Dierens M, De Bruyn H. Accuracy of digital impressions

[21] Ender A, Mehl A. Accuracy of complete-arch dental impressions: a new method of

[25] Ma T, Nicholls JI, Rubenstein JE. Tolerance measurements of various implant

[26] Braian M, De Bruyn H, Fransson H, Christersson C, Wennerberg A. Tolerance

System Impregum, TRIOS9 True Ceramill inEos X513 D90014

Category Conventional Digital Digital

Material / Polyether Optical Active Strip light Strip light Triangulation

Need powder? No No Yes No No No

Registers No Yes No No No Yes

Manufacturer- NA Not Not 20 µm 12 µm 15 µm

Table 2 Homologous implant location pairs

Homologous implant location pairs

JóQ Implant in right premolar region

Impregum TRIOS True Ceramill inEos X5 D900

Impregum TRIOS True Ceramill inEos X5 D900

Variable Model Implant p-value Statistical significant subsets

Variable Model Implant p-value Statistical significant subsets

Homologous dR Absdθx Absdθy

JóR TRIOS TRIOS inEos X5

LóT True Definition

MóU True Definition Impregum True Definition

MóV True Definition Impregum True Definition True Definition inEos X5

Nó W True Definition Impregum True Definition Impregum

Table 8 3D distance distortion attributable to laboratory scanning

Reference distance Mean 3D Mean 3D distance Mean 3D distance

Figure 1 Experimental Outline, with 6 tested systems

Distance between planes = scanbody height

Implant internal cone

Implant central axis

Impregum TRIOS True DeBinition

Ceramill Map400 inEos X5 D900

Impregum TRIOS True DeBinition

Ceramill Map400 inEos X5 D900

Figure 5 Model B linear distortions, µm, by implant location.

Model A absolute angular

Impregum TRIOS True DeBinition Ceramill Map400 inEos X5 D900

Model B absolute angular

Impregum TRIOS True DeBinition Ceramill Map400 inEos X5 D900

Figure 7 Model B absolute angular distortions, °, by implant location.

Model A 3D reference distance

Impregum TRIOS True DeBinition Ceramill Map400 inEos X5 D900

Figure 8 Model A 3D reference distance distortions, µm

Model B 3D reference distance