See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/325900041

Accuracy of printed casts generated from digital implant impressions versus

stone casts from conventional implant impressions: A comparative in vitro
study

Article  in  Clinical Oral Implants Research · June 2018

DOI: 10.1111/clr.13297

CITATIONS READS

7 211

6 authors, including:

Khaled El Rafie Yukio Kudara

Tufts University Tufts University
8 PUBLICATIONS   125 CITATIONS    12 PUBLICATIONS   173 CITATIONS   

SEE PROFILE SEE PROFILE

Panos Papaspyridakos
Columbia University
74 PUBLICATIONS   1,758 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

A Meta-Analysis of Complications During Sinus Augmentation Procedure View project

Digital Workflow in Complete Arch Implant Prostheses View project

All content following this page was uploaded by Panos Papaspyridakos on 20 March 2020.

The user has requested enhancement of the downloaded file.

 | |
Received: 18 December 2017    Revised: 24 April 2018    Accepted: 2 May 2018

DOI: 10.1111/clr.13297

ORIGINAL RESEARCH

Accuracy of printed casts generated from digital implant

impressions versus stone casts from conventional implant
impressions: A comparative in vitro study

Bahaa Alshawaf1 | Hans-Peter Weber1 | Matthew Finkelman2 | Khaled El

Rafie1 | Yukio Kudara1 | Panos Papaspyridakos1

1
Division of Postgraduate
Prosthodontics, Tufts University School of
Dental Medicine, Boston, Massachusetts
2
Department of Public Health and
Community Service, Tufts University School
of Dental Medicine, Boston, Massachusetts
Correspondence
Panos Papaspyridakos, Division of
Postgraduate Prosthodontics, Tufts
University School of Dental Medicine, 1
Kneeland Street, Boston 02111, MA.
Email: panpapaspyridakos@gmail.com
Funding information
Tufts University School of Dental Medicine
Abstract
Objectives: The aim of this in vitro study was to compare the accuracy of printed
implant casts from digital impressions with two intra‐oral scanners (IOS) to stone
casts from conventional impressions. The hypothesis was that printed casts would be
more accurate than stone casts from conventional impressions.
Materials and methods: A mandibular stone cast with Kennedy class II edentulism
was fabricated using two internal connection tissue‐level implants at 30 degrees to
each other (Replace Select RP, Nobel Biocare) to serve as master. Digital impressions
(n = 10) were made with the white light (WL) and Active Wavefront Sampling tech‐
nology (AWST) IOS. The resultant standard tessellation language (STL) datasets were
used to print implant casts through stereolithography (SLA) prototyping. The con‐
ventional casts (n = 10) were produced with splinted open tray impression technique
and polyether material in type IV stone. The master cast and all groups were digitized
with lab reference scanner. The test groups STL datasets were superimposed to mas‐
ter cast STL in inspection software (Geomagic control 2015) to calculate root‐mean‐
square error.
Results: The conventional, WL IOS and AWST IOS groups had mean values of
53.49 μm (SD 9.47), 108.09 μm (SD 9.59) and 120.39 μm (SD 5.91), respectively. The
Shapiro–Wilk test showed no evidence of nonnormality (p = 0.131) and Levene’s test
showed no evidence of heterogeneity of variance (p = 0.518). The one‐way ANOVA
demonstrated a statistically significant difference (p < 0.001). Tukey’s honest signifi‐
cant difference (HSD) showed statistically significant differences between all groups:
for the comparison of AWST IOS and WL IOS, the p‐value was 0.009, and the p‐val‐
ues of the other post hoc tests were <0.001.
Conclusion: Printed casts generated from digital impressions for partially edentulous
posterior mandibular arches had inferior accuracy to conventional stone casts fabri‐
cated from splinted open tray impressions. The printed casts from WL IOS had better
accuracy compared to AWST IOS.

KEYWORDS
dental implants, digital dentistry, digital implant impressions, implant impressions, partial
edentulous, printed casts, stone casts

Clin Oral Impl Res. 2018;29:835–842. wileyonlinelibrary.com/journal/clr   © 2018 John Wiley & Sons A/S. |  835
Published by John Wiley & Sons Ltd
 |
836       ALSHAWAF et al.
1 |  I NTRO D U C TI O N
Achieving passive fit of implant prosthesis is identified as an im‐
portant factor for the long‐term success of implant‐supported fixed
dental prostheses (IFDP) (Jemt, 1996; Papaspyridakos et al., 2012).
Although passive fit is yet to be clearly defined by clinical criteria,
clinical studies have reported that the acceptable misfit between
implant platform and prosthesis varies from 59 to 150 μm (Jemt &
Book, 1996; Papaspyridakos et al., 2012). Various factors can con‐
tribute to misfit of the IFDP, such as the accuracy of implant im‐
pression, the master cast fabrication, and the prosthesis fabrication
procedures (Papaspyridakos et al., 2014).
An accurate master cast is directly dependent on the accuracy of
the implant impression, among other factors as well (Papaspyridakos
et al., 2014). Understanding the impact of these factors allows the
fabrication of an accurate master cast, which reduces the possibility
of framework misfit. The introduction of digital impressions using
intra‐oral optical scanner (IOS) into the fields of fixed and implant
prosthodontics aims to aid in achieving this goal while it carries ad‐
vantages, namely the elimination of tray selection, reduced risks
of distortion during impression making, pouring, disinfecting, and
shipping to the laboratory, potentially increased patient comfort
and acceptance and finally electronic storage as digital information,
leading to better efficiency and reduced costs (Papaspyridakos et
al., 2014,2018). In a digital workflow for the fabrication of implant
prostheses, deviations in accuracy (trueness and precision) have
been reported when making a digital impression in similar fashion
to the conventional approach. A systematic review found that digital
impressions resulted in better marginal and internal fit than conven‐
tional for tooth‐supported single crowns and FDPs (Chochlidakis et
al., 2016). Regarding implant‐supported prostheses, there are less
available data currently, but there is a constant increase of published
data on the accuracy of digital impressions with IOS systems, and
the results seem promising for both partially and fully edentulous
arches (Amin et al., 2017; Basaki, Alkumru, Souza, & Finer, 2017;
Chew et al., 2017; Chia et al., 2017; Lee, Betensky, Gianneschi, &
Gallucci, 2015; Marghalani et al., 2018; Papaspyridakos et al., 2016;
Papaspyridakos, Rajput, Kudara, & Weber, 2017; Vandeweghe,
Vervack, Dierens, & Bruyn, 2017).
The digital impression leads to a Standard Tesselation Language
(STL) file. This STL represents a virtual cast that can be used for a
complete digital workflow or can be used to generate a physical cast
through printing or milling (Cho, Schaefer, Thompson, & Guentsch,
2015; Ender & Mehl, 2015; Joda & Brägger, 2015; Joda, Ferrari, &
Brägger, 2017; Patzelt, Bishti, Stampf, & Att, 2014). Researchers
investigated in vitro the amount of three dimensional (3‐D) devi‐
ations produced in digital workflow from various steps including
scanning, milling implant casts from the STL files, and fabricating
single implant crowns (Koch, Gallucci, & Lee, 2016). Their results
revealed that a cumulative error could be observed throughout
the procedure, including the milling of the cast and concluded
that there were less 3‐D deviations to fabricate the single crown
through complete digital workflow (Koch et al., 2016). An in vitro
study by Lin et al. compared conventional casts to milled casts for
partially edentulous posterior mandibular arches with two implants
in four different angulations. Their results showed that conven‐
tional stone casts were always more accurate than milled casts
(Lin, Harris, Elathamna, Abdel‐Azim, & Morton, 2015). Marghalani
et al. compared the accuracy of digital vs. conventional implant im‐
pressions for partially edentulous posterior mandibular arches with
F I G U R E 1   Workflow of the present
study
ALSHAWAF et al.
two implants with 30 degrees angulation and reported similar ac‐
curacy but did not print or mill casts out of the digital impressions
(Marghalani et al., 2018).
Printing of dental casts from STL files that were generated with
IOS digital impressions is gaining popularity. To the authors’ knowl‐
edge, there are no studies comparing printed casts versus stone
casts from digital and conventional implant impressions, respec‐
tively. The aim of this in vitro study was to compare the accuracy
of printed casts generated from digital implant impressions with a
white light (WL) IOS (CEREC Omnicam) and an Active Wavefront
Sampling technology (AWST) IOS (True Definition) to stone casts
from conventional impressions for a partially edentulous mandibular
arch scenario. The hypothesis was that printed casts would be more
accurate than stone casts from conventional impressions.
2 | M ATE R I A L S A N D M E TH O DS
The flowchart describes the different stages of the study and its de‐
sign (Figure 1). A partially edentulous mandibular cast was fabricated
with two tissue‐level internal connection implants (Replace Select,
Nobel Biocare) to simulate a clinical situation of a partially edentu‐
lous posterior mandible (Kennedy class II) to be restored with a 3‐
unit IFDP. The two implants were at 30‐degree angulation to each
other that could only be fabricated in a specialized facility with clear
acrylic resin (Model Plus Inc., Grayslake, IL, USA). Due to the use of
clear acrylic resin, it was not possible to digitize this cast by digital
scanning; hence, a stone cast was fabricated using custom open tray,
and polyether impression (Impregum, 3M ESPE) material (Marghalani
et al., 2018). The open tray impression copings were splinted with
urethane dimethacrylate‐based visible light‐cured resin (Triad gel)
before the impression was taken. Implant replicas were attached
to the impression copings, and the impression was poured with low
expansion stone (0.09%) type four (Resin Rock, Whipmix Corp). The
stone cast was used as the master cast (control) in this study. Four
locating notches were made on the cast to standardize the process
by which conventional impressions were taken (Amin et al., 2017;
Papaspyridakos et al., 2016).
2.1 | Group 1: Conventional open tray splinted
implant level impression
Ten open tray splinted implant level impressions were taken with
polyether impression material (Impregum, 3M ESPE) by one clinical
operator. To standardize the procedure, implant impression copings
were attached with urethane dimethacrylate‐based visible light‐
cured resin bars (Triad gel) at standardized dimensions (Marghalani
et al., 2018). This was achieved by filling drinking straws with triad
gel, light curing them, and allowing the gel to set for 24 hr before
the procedure. The bars were sectioned with thin disks, and then
attached to the impression copings with a minimum amount of Triad
gel. The open tray splinted impression was taken with polyether im‐
pression material in the fabricated custom tray (Figure 2).
|
      837
F I G U R E 2   Conventional splinted impression coping technique
2.2 | Group 2: Digital impressions with white light
IOS (CEREC Omnicam)
Scan bodies (NT Trading) were tightened to the Replace Select
(Nobel Biocare) implants on the master cast with 10 Ncm
(Marghalani et al., 2018). Digital impression was taken with WL
IOS and exported as STL file. The process of connecting scan bod‐
ies and taking digital impression was repeated 10 times to pro‐
duce 10 STL files from WL IOS. This process was performed by
one operator.
2.3 | Group 3: Digital impressions with Active
Wavefront Sampling technology IOS (True Definition)
Scan bodies (NT Trading) were tightened to the Replace Select
(Nobel Biocare) implants on the master cast with 10 Ncm (Figure 3).
Digital impression was taken with AWST IOS and exported as STL
file. The process of connecting scan bodies and taking digital impres‐
sion was repeated ten times to produce ten STL files from AWST
IOS. The STL files were saved (Figure 4). This process was performed
by one operator.
2.4 | Cast fabrication procedures
The conventional splinted impressions from group 1 were poured
in type IV stone, which has low expansion (0.09%) (Resin Rock,
Whipmix Corp, Louisville, KY, USA). The stone casts were saved.
The STL files obtained from digital implant impressions from
groups 2 and 3 were processed and printed by one dental techni‐
cian. All casts were printed in polyurethane using the same printing
machine (Varseo 3D printing system, BEGO, Germany), which uti‐
lizes SLA technology and is reported to have 25 μm of resolution ac‐
cording to the manufacturer. The implant replicas (NT Trading Digital
Model Analog) were retrofitted into the printed casts so that they
could be used as working casts from the digital pathway (Figure 5),
and the printed casts were saved.
 |
838       ALSHAWAF et al.
F I G U R E 3   Digital implant impression technique
F I G U R E 4   STL file from digital impression with True Definition
scanner
2.5 | Digitization of the stone and printed casts
from the three groups
Before any measurements were recorded, all casts were stored at
room temperature for one week. A high‐resolution extra‐oral ref‐
erence scanner with 10 μm accuracy (Activity 880 scanner; Smart
Optics, Bochum, Germany) was calibrated according to manu‐
facturer’s instructions before digital scanning (Amin et al., 2017;
Papaspyridakos et al., 2016). This reference scanner was used to
digitize the three test groups. The 3‐D transformation functionality
of this scanner allows it to capture the 3‐D position of implants. It
utilizes a white light camera, which is capable of capturing multiple
pictures and transforming them into a 3‐D image (Amin et al., 2017;
Marghalani et al., 2018).
Scan bodies (NT Trading) were tightened with 10 Ncm on the
first test cast and then scanned with the digital scanner. The scan
bodies were transferred from the first cast to the second cast for
scanning. This process was repeated for all 10 casts of the three
groups.
To minimize the possible effect of scan bodies, the same scan
bodies were transferred from one implant in the first cast to the
F I G U R E 5   Printed cast from digital impression with True
Definition scanner
same implant in the other ten casts of each group in order (Amin et
al., 2017; Marghalani et al., 2018). These scan bodies had titanium
interface to resist deformation over time. An independent examiner
labeled the casts with random numbers before digital scanning. The
numbering reference sheet was saved, and all scanning procedures
were performed by the same operator, who was blinded to the type
of cast. The STL files that were produced were compared to the mas‐
ter cast.
2.6 | STL superimposition procedures
The STL dataset of each cast from the test groups were super‐
imposed to the STL dataset of the control (master cast) using in‐
spection software (Figures 6 and 7) (Geomagic Control 2015, 3D
Systems). Geomagic software was utilized to calculate the differ‐
ence between each cast through the use of the root‐mean‐square
(RMS) error.
2.7 | Power calculation
The software package nQuery Advisor (version 7.0) was used to per‐
form a power calculation. Based on the findings of Marghalani et al.
(2018), a sample size of n = 10 per group was sufficient to achieve
power >99%, using a significance level of α = 0.05.
2.8 | Statistical analysis
Descriptive data of the RMS error (means and SDs) were calculated
for each group (conventional, WL, AWST). The Shapiro–Wilk test
was used to assess normality of the data. Levene’s test was used to
assess homogeneity of variance. The comparison between the three
groups was carried out via one‐way ANOVA. Tukey’s honest signifi‐
cant difference (HSD) for post hoc comparisons was used to com‐
pare each group to the others. SPSS Version 22 (IBM Corp., 2013)
was used in the analysis.
ALSHAWAF et al.
F I G U R E 6   Superimposition of STL files of master cast and cast
generated from conventional splinted impression
F I G U R E 7   Superimposition of STL files of master cast and
printed cast generated from digital impression with True Definition
scanner
3 | R E S U LT S
3‐D deviations as measured by RMS (mean ± SD) for the three im‐
plant groups found that the conventional cast group had the low‐
est mean value (53.49 ± 9.47 μm), followed by the WL cast group
(108.09 ± 9.59 μm). The AWST cast group had the highest mean
value (120.39 ± 5.91 μm).
The Shapiro–Wilk test on the data showed no evidence of non‐
normality (p = 0.131) and Levene’s test showed no evidence of
heterogeneity of variance (p = 0.518). The one‐way ANOVA demon‐
strated a statistically significant difference (p < 0.001). Tukey’s HSD
reported significant post hoc group differences, with the conven‐
tional group having lower 3‐D deviations compared to both WL
(p < 0.001) and AWST (p < 0.001), and WL was also significantly
lower than AWST (p = 0.009). These group differences are depicted
in Figure 8. The conventional dental cast group had a minimum of
39.41 μm and a maximum of 72.82 μm. The AWST dental cast group
had a minimum of 110.37 μm and a maximum of 131.25 μm. The WL
dental cast group had a minimum of 94.43 μm and a maximum of
130.12 μm.
|
      839
4 | D I S CU S S I O N
The objective of the present study was to compare the accuracy of
printed implant casts from digital impressions with two IOS systems
to stone casts from conventional impressions for a Kennedy class II
scenario. To the authors’ knowledge, this is the first study comparing
printed casts vs. stone casts from digital and conventional implant
impressions, respectively. The hypothesis was that printed casts
from digital impressions would be more accurate than stone casts
from conventional impressions.
The findings of this in vitro study show that the stone casts from
the conventional splinted implant impression technique had less 3‐D
deviations than the printed casts from the digital impression tech‐
nique. The difference was statistically significant. The mean value of
deviations for the conventional group was about 53.49 μm whereas
for both printed groups it was more than double. The recorded mean
value for WL IOS was 108.09 μm and the mean value for AWST IOS
was 120.39 μm. The difference in mean value of 3‐D deviations be‐
tween the two printed cast groups and the master (control) was also
statistically significant.
To the authors’ knowledge, there are no other studies that com‐
pare the accuracy of printed casts to conventional casts (Kennedy
Class II scenario with two implants), but there are several studies
that compared milled casts to conventional casts. A study by Lin et
al. compared conventional casts to milled casts for partial edentulous
with two implants in four different angulations. Their results showed
that conventional casts were always more accurate than milled casts
(Lin et al., 2015). In another in vivo study, the digital implant impres‐
sion technique inferior 3‐D accuracy when compared with the con‐
ventional splinted for Kennedy Class II scenario (Alsharbaty, Alikhasi,
Zarrati, & Shamshiri, 2018). The 3‐D deviations associated with the
TRIOS digital implant impressions were not clinically acceptable to
fabricate well‐fitting FDPs for partially edentulous patients. An in
vitro study (Lee et al., 2015) compared the accuracy of conventional
cast for single implant scenario to milled casts from i‐Tero IOS with
the use of scan body. Their results showed a statistically significant
difference only in the vertical displacement with −88 μm for con‐
ventional casts and 93 μm for milled casts. These results cannot be
directly compared to the present study because of differences in
prototyping technology, while it seems that a specific pattern of dis‐
placement might be the source of deviation in digitally produced im‐
plant casts (Lee et al., 2015). An in vitro study by Karl et al. compared
the strain produced from cement‐retained prosthesis fabricated in
stone casts from cementable abutment‐level impressions vs. printed
casts from IOS impressions. The authors used prefabricated abut‐
ments instead of scan bodies and an old type of IOS (Lava C.O.S) to
obtain STL files for the printed cast group. The conventional group
was produced by open tray technique and polyether impression ma‐
terial. Zirconia IFDPs were fabricated on both groups and the strain
produced by each IFDP was measured on patient cast with strain
gauge attachments. The results showed no statistically significant
difference between the two groups in terms of generated strain
(Karl, Graef, Schubinski, & Taylor, 2012). This result might indicate
 |
840       ALSHAWAF et al.
that the 3‐D deviations between conventional stone and printed
implant casts might not have clinical significance in terms of strain
in cement‐retained prosthesis. Nonetheless, major differences exist
between the two studies such as the angulation of implants, not
using scan bodies, cementable abutment‐level impression, and the
use of a currently discontinued older type of IOS (Lava C.O.S.), that
make comparisons to the present study impossible.
In the present study, the conventional cast group was produced
by a protocol similar to other studies that reported minimal deviations
when using open tray splinted impression technique, polyether im‐
pression materials and type IV low expansion stone (Papaspyridakos
et al., 2014). Nonetheless, the 3‐D implant deviations of conven‐
tional casts were slightly higher in this study with mean deviation
of 53.49 μm compared to another study which reported mean devi‐
ation of 29.8 μm (Sabouhi, Bajoghli, & Abolhasani, 2015). This could
be due to multiple differences between the two studies, such as op‐
erator technique, master cast undercuts, difference in implants an‐
gulation and implant position within the dental arch. If nonsplinted
impression technique was used then more discrepancy would be ex‐
pected in the conventional group (Papaspyridakos et al., 2014).
In the present study, two printed groups were fabricated, one
from WL digital impressions and the second from AWST ones. The
results revealed a statistically significant difference between the
two groups with less 3‐D deviations in the WL group. A previous
study investigated the difference in accuracy of IOS STL file be‐
tween WL and AWST for the same clinical scenario. Although their
results showed less 3‐D deviations for the WL group compared to
the AWST group, there was no statistically significant difference
(Marghalani et al., 2018). However, this study did not print casts out
of the STL files and the investigators only compared the accuracy of
STL files. The propagation and accumulation of error in digital work‐
flow from scanning to milling of a partially edentulous arch with sin‐
gle implant scenario has been shown (Koch et al., 2016). The STL
file accuracy is comparable to or even superior to the conventional
approach, and this finding has also been reported in several studies
with Kennedy class II scenarios (Chew et al., 2017; Chia et al., 2017;
F I G U R E 8   Box‐plot of the root‐mean‐
square values from all three test groups
Marghalani et al., 2018). The present study takes one step further
and reports the outcomes when the actual STL files are used to print
physical casts. The results might support the accumulation of error
that were noticed from the IOS stage to the printing stage in the
current study. More research is needed to identify the influencing
factors.
In regards to the accuracy assessment, the comparison between
the accuracy of the three groups and control cast was carried out by
superimposition in Geomagic software. This method was described
by multiple previous articles to calculate the 3‐D deviations between
two items in terms of RMS error in micrometers (Lin et al., 2015;
Patzelt et al., 2014; Sabouhi et al., 2015). The Geomagic software
allowed the superimposition of the test STL files to control cast STL
file through best‐fit match algorithm. Using this method of super‐
imposition can introduce inaccuracy due to the cancellation of the
positive and negative deviations to each other, which results in a
reduced estimation of the actual deviation from the reference cast.
To eliminate this inaccuracy, root‐mean‐square error was used to
calculate the 3‐D deviations. The limitation of using RMS error for
measuring the discrepancy is the lack of details regarding the pattern
of displacement.
This study exhibited a number of limitations. The conditions for
obtaining conventional and digital impression are easily controlled
in in vitro studies, which might not be similar in clinical situations.
The presence of saliva, blood, and gingival crevicular fluid will in‐
fluence the results. In the present in vitro study, the investigation
was limited to only one scenario of number of implants, position,
and angulation (Kennedy class II, 30 degrees angulation). All of these
variables are identified as influencing factors in the accuracy of
both conventional and printed casts. Studies for dentate scenarios
showed more discrepancy toward the distal of the arch compared to
the mesial (Ender & Mehl, 2015). Another limitation is the effect of
scan bodies on printed casts. However, the length and geometry of
the scan bodies were standardized, and they had titanium interface
to resist deformation over time. The scan bodies were not torqued to
avoid possible deformation to the screw. Further research is needed
ALSHAWAF et al.
to identify the effect of these factors on printed implant casts. In
terms of printing, the angulation of the prototype within the printing
machine could affect the implant position in the cast. A comparison
of vertical, horizontal, and oblique printing has not yet been per‐
formed. In this study, all printing was standardized in the horizontal
direction. Another limitation is the comparison method, which re‐
lied only on RMS error. During the digitization procedure with the
reference scanner, the use of the same scan bodies for all the casts
will standardize the scan body geometry and illuminate any possi‐
ble manufacturing error effect. The scan bodies were manufactured
with titanium connection, and the screws were not torqued to pre‐
vent deformation.
For single implant crowns and short‐span IFDPs, the use of
IOS systems and CAD/CAM technology offers the possibility of a
digital workflow (Joda & Brägger, 2016a,2016b; Joda et al., 2017).
It has been well documented for single crowns (Joda & Brägger,
2016a,2016b; Joda et al., 2017; Joda, Katsoulis, & Brägger, 2016;
Schepke, Meijer, Kerdijk, & Cune, 2015). The milled prosthesis gen‐
erated from the STL file is cemented to a titanium insert (Ti‐base)
without a master cast (complete digital workflow) or on a milled/
printed cast (semi‐digital workflow) and the prosthesis is screw‐re‐
tained to the implant level. It can also be cement retained on a CAD/
CAM abutment generated from the STL file. During a complete dig‐
ital workflow, the cast is not necessary as the prosthesis and the
titanium insert or the abutment can be cemented without a cast or
even intraorally (Chun et al., 2015). For the full arch fixed implant
rehabilitation, a complete digital workflow is currently not feasible
and instead a combination of digital and conventional workflow is
the current status (Papaspyridakos et al., 2017, 2018).
The investigated clinical scenario in this study is very common
and impacts a large number of patients (Papaspyridakos, 2015). To
ensure adequate longevity of the implant prosthesis and limit the
risk of biologic and technical complications, the prosthesis must
be fabricated from an accurate master cast to achieve passive fit
(Papaspyridakos et al., 2014). In the present study, the accuracy of
printed casts was investigated and 3‐D printing method had larger
amount of deviations compared to conventional casts. No studies are
available at the moment to indicate the amount of framework misfit
as a result of deviations in the master cast. Hence, the deviations de‐
tected in the printed casts might still be within the acceptable range
for clinical use. This can be investigated in future research by fabri‐
cating an implant framework, assessing the actual fit of the frame‐
work and correlating potential misfit to the 3‐D deviations. Future
research is also needed to investigate the effect of digital workflow
on the accuracy of printed implant casts. In addition, no comparison
between printed and milled implant casts has yet been conducted
and published in the literature, which warrants further investigation.
5 | CO N C LU S I O N
Within the limitations of this in vitro study, the following conclusions
may be drawn:
|
      841
1. Stone casts generated from conventional splinted implant im‐
pressions had significantly less 3‐D deviations compared to
printed casts from digital impressions for the scenario of Kennedy
class II mandibular edentulism with 2 implants.
2. The printed casts from different IOS systems showed a statisti‐
cally significant difference in terms of implant 3‐D deviations. WL
demonstrated superior accuracy compared to AWST scanner.
AC K N OW L E D G E M E N T S
The present study was supported by the Division of Postgraduate
Prosthodontics at Tufts University School of Dental Medicine. It was
performed by Dr Bahaa Alshawaf in partial fulfillment of the require‐
ments for his Master of Science degree at Tufts University School of
Dental Medicine.
C O N FL I C T O F I N T E R E S T
The authors do not have any financial interest in the companies
whose materials are included in this article.
ORCID
Panos Papaspyridakos  http://orcid.org/0000-0002-8790-2896
REFERENCES
Alsharbaty, M. H. M., Alikhasi, M., Zarrati, S., & Shamshiri, A. R. (2018).
A clinical comparative study of 3‐dimensional accuracy between
digital and conventional implant impression techniques. Journal of
Prosthodontics, https://doi.org/10.1111/jopr.12764. [Epub ahead of
print]
Amin, S., Weber, H. P., Finkelman, M., El Rafie, K., Kudara, Y., &
Papaspyridakos, P. (2017). Digital vs. conventional full‐arch implant
impressions: A comparative study. Clinical Oral Implants Research,
28, 1360–1367. https://doi.org/10.1111/clr.12994.
Basaki, K., Alkumru, H., De Souza, G., & Finer, Y. (2017). Accuracy of
digital vs conventional implant impression approach: A three‐di‐
mensional comparative in vitro analysis. International Journal of Oral
and Maxillofacial Implants, 32, 792–799. https://doi.org/10.11607/
jomi.5431.
Chew, A. A., Esguerra, R. J., Teoh, K. H., Wong, K. M., Ng, S. D., & Tan,
K. B. (2017). Three‐dimensional accuracy of digital implant impres‐
sions: Effects of different scanners and implant level. International
Journal of Oral and Maxillofacial Implants, 32, 70–80. https://doi.
org/10.11607/jomi.4942.
Chia, V. A., Esguerra, R. J., Teoh, K. H., Teo, J. W., Wong, K. M., & Tan,
K. B. (2017). In vitro three‐dimensional accuracy of digital im‐
plant impressions: The effect of implant angulation. International
Journal of Oral and Maxillofacial Implants, 32, 313–321. https://doi.
org/10.11607/jomi.5087.
Cho, S. H., Schaefer, O., Thompson, G. A., & Guentsch, A. (2015).
Comparison of accuracy and reproducibility of casts made by dig‐
ital and conventional methods. Journal of Prosthetic Dentistry, 113,
310–315. https://doi.org/10.1016/j.prosdent.2014.09.027.
Chochlidakis, K. M., Papaspyridakos, P., Geminiani, A., Chen, C. J., Feng,
I. J., & Ercoli, C. (2016). Digital versus conventional impressions
842       |
for fixed prosthodontics: A systematic review and meta‐anal‐
ysis. Journal of Prosthetic Dentistry, 116, 184–190. https://doi.
org/10.1016/j.prosdent.2015.12.017.
Chun, H. J., Yeo, I. S., Lee, J. H., Kim, S. K., Heo, S. J., Koak, J. Y., & Lee,
S. J. (2015). Fracture strength study of internally connected zir‐
conia abutments reinforced with titanium inserts. International
Journal of Oral and Maxillofacial Implants, 30, 346–350. https://doi.
org/10.11607/jomi.3768.
Ender, A., & Mehl, A. (2015). In‐vitro evaluation of the accuracy of
conventional and digital methods of obtaining full‐arch den‐
tal impressions. Quintessence International, 46, 9–17. https://doi.
org/10.3290/j.qi.a32244.
IBM Corp. (2013). IBM SPSS statistics for windows, version 22.0. Armonk,
NY: IBM Corp.
Jemt, T. (1996). In vivo measurements of precision of fit involving im‐
plant‐supported prostheses in the edentulous jaw. International
Journal of Oral and Maxillofacial Implants, 11, 151–158.
Jemt, T., & Book, K. (1996). Prosthesis misfit and marginal bone loss
in edentulous implant patients. International Journal of Oral and
Maxillofacial Implants, 11, 620–625.
Joda, T., & Brägger, U. (2015). Time‐efficiency analysis comparing digital
and conventional workflows for implant crowns: A prospective clinical
crossover trial. International Journal of Oral and Maxillofacial Implants,
30, 1047–1053. https://doi.org/10.11607/jomi.3963.
Joda, T., & Brägger, U. (2016a). Time‐efficiency analysis of the treatment
with monolithic implant crowns in a digital workflow: A random‐
ized controlled trial. Clinical Oral Implants Research, 27, 1401–1406.
https://doi.org/10.1111/clr.12753.
Joda, T., & Brägger, U. (2016b). Patient‐centered outcomes comparing
digital and conventional implant impression procedures: A random‐
ized crossover trial. Clinical Oral Implants Research, 27, e185–e189.
https://doi.org/10.1111/clr.12600.
Joda, T., Ferrari, M., & Brägger, U. (2017). Monolithic implant‐sup‐
ported lithium disilicate (LS2) crowns in a complete digital work‐
flow: A prospective clinical trial with a 2‐year follow‐up. Clinical
Implant Dentistry and Related Research, 19, 505–511. https://doi.
org/10.1111/cid.12472.
Joda, T., Katsoulis, J., & Brägger, U. (2016). Clinical fitting and adjustment
time for implant‐supported crowns comparing digital and conven‐
tional workflows. Clinical Implant Dentistry and Related Research, 18,
946–954. https://doi.org/10.1111/cid.12377.
Karl, M., Graef, F., Schubinski, P., & Taylor, T. (2012). Effect of intraoral
scanning on the passivity of fit of implant‐supported fixed dental
prostheses. Quintessence International, 43, 555–562.
Koch, G. K., Gallucci, G. O., & Lee, S. J. (2016). Accuracy in the digital
workflow: From data acquisition to the digitally milled cast. Journal
of Prosthetic Dentistry, 115, 749–754. https://doi.org/10.1016/j.
prosdent.2015.12.004.
Lee, S. J., Betensky, R. A., Gianneschi, G. E., & Gallucci, G. O. (2015).
Accuracy of digital versus conventional implant impressions. Clinical
Oral Implants Research, 26, 715–719. https://doi.org/10.1111/
clr.12375.
Lin, W. S., Harris, B. T., Elathamna, E. N., Abdel‐Azim, T., & Morton, D.
(2015). Effect of implant divergence on the accuracy of definitive
casts created from traditional and digital implant‐level impres‐
sions: An in vitro comparative study. International Journal of Oral
and Maxillofacial Implants, 30, 102–109. https://doi.org/10.11607/
jomi.3592.
Marghalani, A., Weber, H. P., Finkelman, M., El Rafie, K., Kudara, Y., &
Papaspyridakos, P. (2018). Digital versus conventional impres‐
sion techniques for partially edentulous arches: An accuracy
View publication stats
ALSHAWAF et al.
evaluation. Journal of Prosthetic Dentistry, 119, 574–579. https://doi.
org/10.1016/j.prosdent.2017.07.002.
Papaspyridakos, P. (2015). Implant success rates for single crowns and
fixed partial dentures in general dental practices may be lower than
those achieved in well‐controlled university or specialty settings.
Journal of Evidence‐Based Dental Practice, 15, 30–32. https://doi.
org/10.1016/j.jebdp.2014.12.003.
Papaspyridakos, P., Benic, G. I., Hogsett, V. L., White, G. S., Lal, K., & Gallucci,
G. O. (2012). Accuracy of implant casts generated with splinted and
non‐splinted impression techniques for edentulous patients: An op‐
tical scanning study. Clinical Oral Implants Research, 23, 676–681.
https://doi.org/10.1111/j.1600-0501.2011.02219.x.
Papaspyridakos, P., Chen, C. J., Gallucci, G. O., Doukoudakis, A., Weber,
H. P., & Chronopoulos, V. (2014). Accuracy of implant impressions
for partially and completely edentulous patients: A systematic re‐
view. International Journal of Oral and Maxillofacial Implants, 29,
836–845. https://doi.org/10.11607/jomi.3625.
Papaspyridakos, P., Gallucci, G. O., Chen, C. J., Hanssen, S., Naert, I., &
Vandenberghe, B. (2016). Digital versus conventional implant im‐
pressions for edentulous patients: Accuracy outcomes. Clinical Oral
Implants Research, 27, 465–472. https://doi.org/10.1111/clr.12567.
Papaspyridakos, P., Kang, K., DeFuria, C., Amin, S., Kudara, Y., & Weber, H.
P. (2018). Digital workflow in full‐arch implant rehabilitation with seg‐
mented minimally veneered monolithic zirconia fixed dental prosthe‐
ses: 2‐year clinical follow‐up. Journal of Esthetic Restorative Dentistry,
30, 5–13. https://doi.org/10.1111/jerd.12323.
Papaspyridakos, P., Rajput, N., Kudara, Y., & Weber, H. P. (2017). Digital
workflow for fixed implant rehabilitation of an extremely atrophic
edentulous mandible in three appointments. Journal of Esthetic
Restorative Dentistry, 29, 178–188. https://doi.org/10.1111/
jerd.12290.
Patzelt, S. B., Bishti, S., Stampf, S., & Att, W. (2014). Accuracy of com‐
puter‐aided design/computer‐aided manufacturing‐generated
dental casts based on intraoral scanner data. Journal of American
Dental Association, 145, 1133–1140. https://doi.org/10.14219/
jada.2014.87.
Sabouhi, M., Bajoghli, F., & Abolhasani, M. (2015). Evaluation of the
three‐dimensional accuracy of implant impression techniques in
two simulated clinical conditions by optical scanning. International
Journal of Oral and Maxillofacial Implants, 30, 26–34. https://doi.
org/10.11607/jomi.3487.
Schepke, U., Meijer, H. J., Kerdijk, W., & Cune, M. S. (2015). Digital ver‐
sus analog complete‐arch impressions for single‐unit premolar
implant crowns: Operating time and patient preference. Journal
of Prosthetic Dentistry, 114, 403–406. https://doi.org/10.1016/
j.prosdent.2015.04.003.
Vandeweghe, S., Vervack, V., Dierens, M., & De Bruyn, H. (2017).
Accuracy of digital impressions of multiple dental implants: An in
vitro study. Clinical Oral Implants Research, 28, 648–653. https://
doi.org/10.1111/clr.12853.
How to cite this article: Alshawaf B, Weber H-P, Finkelman
M, El Rafie K, Kudara Y, Papaspyridakos P. Accuracy of
printed casts generated from digital implant impressions
versus stone casts from conventional implant impressions: A
comparative in vitro study. Clin Oral Impl Res. 2018;29:835–
842. https://doi.org/10.1111/clr.13297