G Model

JPOR 432 No. of Pages 9

journal of prosthodontic research xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Journal of Prosthodontic Research

journal homepage: www.elsevier.com/locate/jpor

Original article

Three-dimensional evaluation of marginal and internal fit of

3D-printed interim restorations fabricated on different finish line
designs
Nawal Alharbia,b,* , Saud Alharbic , Vincent M.J.I. Cuijpersd, Reham B. Osmane,f ,
Daniel Wismeijerg
a
Department of Oral Implantology and Prosthetic Dentistry, Academic Centre for Dentistry Amsterdam (ACTA), University of Amsterdam and Vrije Universiteit
Amsterdam, Gustav Mahlerlaan 3004, 1081LA Amsterdam, Netherlands
b
Department of Prosthetic Dental Science, King Saud University, College of Dentistry, Riyadh, Saudi Arabia
c
Prince Sultan Military Medical City, Riyadh, Saudi Arabia
d
Department of Biomaterials, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands
e
Removable Prosthodontics Department, Faculty of Dentistry, Cairo University, Egypt
f
Department of Oral Implantology and Prosthetic Dentistry, Academic Centre for Dentistry Amsterdam (ACTA), Universiteit van Amsterdam and Vrije
Universiteit Amsterdam, The Netherlands
g
Department of Oral Implantology and Prosthetic Dentistry, Academic Centre for Dentistry Amsterdam (ACTA), Universiteit van Amsterdam and Vrije
Universiteit Amsterdam, The Netherlands

A R T I C L E I N F O A B S T R A C T

Article history:
Received 15 February 2017 PurposeTo evaluate the influence of fabrication method and finish line design on marginal and internal fit
Received in revised form 10 July 2017 of full-coverage interim restorations.
Accepted 23 September 2017 MethodsFour typodont models of maxillary central-incisor were prepared for full-coverage
Available online xxx restorations. Four groups were defined; knife-edge (KE), chamfer (C), rounded-shoulder (RS),
rounded-shoulder with bevel (RSB). All preparations were digitally scanned. A total of 80 restorations
Keywords: were fabricated; 20 per group (SLA/3D-printed n = 10, milled n = 10). All restorations were positioned on
3D-SLA printing the master die and scanned using micro-computed tomography. The mean gaps were measured digitally
Finish line design
(ImageJ). The results were compared using MANOVA (a = .05).
Marginal and internal fit
ResultsInternal and marginal gaps were significantly influenced by fabrication method (P = .000) and
Micro-CT
Additive manufacturing finish-line design (P = .000). 3D-Printed restorations showed statistically significant lower mean gap
compared to milled restorations at all points (P = .000). The mean internal gap for 3D-printed restorations
were 66, 149, 130, 95 mm and for milled restorations were 89, 177, 185, 154 mm for KE, C, RS, RSB
respectively. The mean absolute marginal discrepancy in 3D-printed restorations were (30, 41, 30,
28 mm) and in milled restorations were (56, 54, 52, 38 mm) for KE, C, RS, RSB respectively.
ConclusionsThe fabrication methods showed more of an influence on the fit compared to the effect of
the finish-line design in both milled and printed restorations. SLA-printed interim restorations exhibit
lower marginal and internal gap than milled restorations. Nonetheless, for both techniques, all values
were within the reported values for CAD/CAM restorations.
Significance3D-printing can offer an alternative fabrication method comparable to those of milled
restorations.
© 2017 Japan Prosthodontic Society. Published by Elsevier Ltd. All rights reserved.

1. Introduction

In contemporary prosthodontics, interim restorations are used

* Corresponding author at: Department of Oral Implantology and Prosthetic
Dentistry, Academic Centre for Dentistry, Amsterdam (ACTA), University of
Amsterdam and Vrije Universiteit Amsterdam, Gustav Mahlerlaan 3004, 1081 LA
Amsterdam, The Netherlands.
E-mail address: nawalmurshed@gmail.com (N. Alharbi).
as a transitional phase to evaluate the functional and esthetical
outcomes [1–4]. The restorations should offer adequate fit to
ensure mechanical stability and durability of the restoration and
thus the health of the surrounding tissues [5,6]. Lack of adequate fit
can result in plaque accumulation, cement microleakage, marginal

https://doi.org/10.1016/j.jpor.2017.09.002
1883-1958/ © 2017 Japan Prosthodontic Society. Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: N. Alharbi, et al., Three-dimensional evaluation of marginal and internal fit of 3D-printed interim restorations
fabricated on different finish line designs, J Prosthodont Res (2017), https://doi.org/10.1016/j.jpor.2017.09.002
G Model
JPOR 432 No. of Pages 9
2 N. Alharbi et al. / journal of prosthodontic research xxx (2017) xxx–xxx
discoloration, poor esthetics, teeth sensitivity, caries and peri-
odontal diseases [5]. Wide variation in the definition of adequate
fit and a clinically acceptable marginal gap exists in the literature
and can be attributed to different study designs including different
restorative materials, examination method, finish line designs and
fabrication methods [7]. A recent systematic review on the fit of
CAD/CAM restorations fabricated from different materials reported
an absolute marginal gap (AMG) of crowns to be in the range from
10–110 mm with the majority of values less than 80 mm. In the
same context, the internal gap at axial part was 23–154 mm and at
occlusal part was 45–219 mm [8].
With the ongoing developments in the field of data acquisition
and manufacturing processes, digital technologies are gaining
more acceptances. Additive manufacturing (AM) is being increas-
ingly applied in the field of prosthodontics. Dental restorations can
be fabricated using resin based AM techniques both directly and
indirectly; stereolithography and or digital light processing (SLA/
DLP) [2,4,9–13]. Both techniques use light/laser to cure a
photosensitive liquid polymer layer by layer, following a specific
path of the designed model.
Limited studies have evaluated the marginal and/or internal fit
of AM copings fabricated using SLA/DLP techniques. Kim et al. [10]
found that the mean marginal gap of Co–Cr crowns was
significantly larger when fabricated indirectly using SLA technique
(96 mm) compared to traditional, lost-wax fabrication method
(67 mm). The authors also reported an axial and an occlusal gap of
84 mm and 114 mm respectively for the indirectly manufactured
SLA crowns [10]. In contrast, Park et al. [2] found PMMA DLP-
fabricated implant restorations to have better marginal fit
compared to milled and conventionally fabricated counterparts
fabricated from PEEK and PMMA respectively. Nevertheless, the
authors found the fit of the three manufacturing methods to be
within the clinically acceptable range.
Munoz et al. [12] used DLP technique to fabricate wax patterns
for indirect manufacturing of cast gold crowns and revealed that
the marginal gap was significantly larger for DLP fabricated
patterns compared to the milled or manually fabricated wax
patterns.
In the same context, different finish line designs have been
shown to influence the marginal gap of the dental restorations.
Comlekoglu et al. [14] compared the marginal gap of zirconia
crowns with knife-edge, mini-chamfer, chamfer and rounded
Fig. 1. Flow of study design.
Please cite this article in press as: N. Alharbi, et al., Three-dimensional evaluation of marginal and internal fit of 3D-printed interim restorations
fabricated on different finish line designs, J Prosthodont Res (2017), https://doi.org/10.1016/j.jpor.2017.09.002
shoulder finish line designs and found the lowest absolute
marginal discrepancy values with knife-edge design (87 mm)
compared to mini-chamfer (114 mm), chamfer (144 mm) and
rounded shoulder (114 mm) finish line designs. Euán et al. [15]
found lower mean marginal gap values for Lava All-Ceramic
System with round shoulder finish-line design (52 mm) compared
to chamfer counterpart (64 mm). In contrast, Tsitrou et al. [16] and
Akbar et al. [17] revealed no significant difference in margnial gap
of dental restorations with shoulder and chamfer finish line
designs. The margin gap of CEREC 3 system composite resin crowns
was 94 mm in chamfer finish line and 91 mm with shoulder finish
line [16].
To the authors’ best knowledge, no previous study has
evaluated the internal and marginal adaptation of 3D-printed
directly fabricated restorations using the SLA technique with
different finish line designs. Therefore, the aim of this study was to
evaluate the influence of fabrication method and finish line design
on the marginal and internal fit of full coverage interim dental
restorations in two digital workflows (3D-printing versus milling).
Our null hypothesis is that there is no difference in marginal and
internal fit between 3D-printing and milled restorations in
different finish line designs.
2. Materials and methods
2.1. Study design
Four typodont models of maxillary central incisors were
prepared to receive full coverage dental restorations; each with
a different finish line design. The four finish line designs were;
knife-edge (KE), chamfer (C), rounded shoulder (RS), rounded
shoulder with bevel (RSB). In both digital workflows, the prepared
models were digitally scanned and then the data was exported as a
standard tessellation language (STL) format to a CAD software for
the design and the fabrication of full coverage restorations (n = 80).
For each finish line design, the restorations were fabricated using
3D-printing (n = 10) and milling (n = 10) techniques. All restora-
tions were fitted on the master die without cementation and
scanned using micro-computed tomography (Skyscan 1072,
Bruker microCT, Kontich) [18]. The marginal and internal gaps
were measured using digital imaging processing software (ImageJ
1.51; NIH). The flow of the study design is shown in Fig. 1.
G Model
JPOR 432 No. of Pages 9

N. Alharbi et al. / journal of prosthodontic research xxx (2017) xxx–xxx 3

2.2. Crown preparation

The models were prepared following the basic principles of

crown preparation [19]. All models were prepared with 2 mm
incisal reduction. Four finish lines designs were used; KE, 1 mm (C),
1 mm (RS) and 1 mm (RSB) [19]. All teeth were prepared with the
use of silicon index [20], together with the use of magnifying loupe
(30) for better visualization. The preparations were done using
diamond burs kit (LOT:22483; Horico) under continuous water-
cooling. The prepared models were scanned using desktop optical
scanner (Medit Identica BlueTM; Renishaw). The scanner was
calibrated following the manufacturer’s instructions. A thin layer
of anti-reflective powder coating was sprayed at a fixed distance of
20 cm for all the specimens to enhance the scanning procedure
(Helling 3D Scan Spray; Helling GmbH) [21,22]. The digital files
were exported in STL format, which were used for the design of
restorations.

2.3. Restorations design and CAD/CAM fabrication

The restorations were designed using 3-Shape Dental SystemTM

CAD solution version 2015. The design parameters were standard-
ized for all groups. A cement gap of 30 mm was defined [23].
Additional vertical space of 80 mm and a horizontal space of 30 mm
was added as recommended by an experienced dental technician.
One trained technician printed the restorations using stereo-
lithography-based 3D-printer (DW028D, DWS) in a hybrid
composite resin material (Temporis1, shade A2, LOT: 040725;
DWS). The thickness of build layer was .05 mm and the maximum
laser speed was 5000 mm/s [24]. The printed specimens were
cleaned with 95% ethanol for one minute and post-cured using
ultraviolet curing unit (S2; DWS) for 30 min as per the
manufacturer’s instructions [24]. The milled restorations were
fabricated using 5-axis milling machine (Wissner Ltd.; Germany)
in PMMA-based acrylate resin (Polycon1 ae; Straumann; shade A2)
with a bur diameter of 1 and 3 mm following manufacturer’s
recommendation [25]. All specimens were then stored in a dry,
lightproof box and tested within ten days of the manufacturing
process.
2.4. Micro-CT scanning and analysis
Before scanning, all restorations were visually examined for any
manufacturing defects, using optical magnification loupe 3.5.
Each restoration was manually placed on the master die without
cementation and without any manual adjustment [18]. The
position of each crown was visually checked before scanning;
where the crown was stable without any rotation or displacement
[18].
The specimens were scanned using a commercially available
desktop X-ray micro-tomography system scanner (Skyscan 1072,
Fig. 3. Slicing protocol for all scanned restorations.
Bruker micro-CT, Kontich). A custom holder was fabricated to
standardize the position of the specimens during scanning and to
ensure that the long axis of the tooth was perpendicular to the
scanning beam. The specimen holder involves a transverse
cylindrical hole, which allows the X-ray beam to pass through.
The specimen position was verified on the screen by a visible well
defined, sharp and clear circle formed when the X-ray beam pass
and run parallel to the transverse circular hole within the holder
(Fig. 2). A scan resolution of 11 mm was defined for all specimens
where the whole crown and coronal 1/3 of the root were included
in one scan. Scans were recorded at X-ray source set to 100 kV and
98 mA, a 180 sample rotation and 10-frames averaging, a rotation
step of .90 , an exposure time of 3.8 s and a 1 mm thick aluminum
filter to optimize image contrast.
After scanning, a cone beam reconstruction was performed to
obtain 11 mm pixels, 2D-images with NRecon v.1.6.9.18 software.
The ring artifact correction was set to 20 and beam hardening
correction to 4%. For the analysis, six slices were selected from each
specimen; three in mesio-distal (MD) and three in labio-palatal
(LP) direction. The distance between each of the selected
measuring points was 60 slices in LP and 110 slices in MD
direction (Fig. 3). The slicing of the images was performed with
CTanalyzer v.1.14 software. A total of 480 images were transferred
to Image processing software for analysis.
One blinded and trained examiner performed the measure-
ments of all images (n = 480) by using Image processing software
(ImageJ 1.51; NIH). A second examiner was consulted for any

Fig. 2. Custom-made specimen holder with visible circle indicating that X-rays run parallel to the transverse hollow cylinder.

Please cite this article in press as: N. Alharbi, et al., Three-dimensional evaluation of marginal and internal fit of 3D-printed interim restorations
fabricated on different finish line designs, J Prosthodont Res (2017), https://doi.org/10.1016/j.jpor.2017.09.002
G Model
JPOR 432 No. of Pages 9
4 N. Alharbi et al. / journal of prosthodontic research xxx (2017) xxx–xxx
Fig. 4. Internal and marginal fit measurement points on; (A) labio-palatal (LP) sections, (B) mesio-distal (MD) sections, (C) marginal fit measurement points on both (LP) and
(MD) sections.
difficulty encountered. Prior to each reading, the measurement
tool was calibrated by measuring the pixel size; one pixel was equal
to 11 mm. All measurements were taken under the magnification of
400. The internal and marginal fit was measured at several defined
points on MD and LP sections as shown in Fig. 4. The internal fit
involves both the incisal gap and mid-axial gap measured at the
middle of mid-axial wall. Further for better interpretation of the
results, the data of the incisal gap and mid-axial gap were reported
separately. The marginal fit was reported by using the vertical gap
(VG), horizontal gap (HG) and absolute marginal discrepancy (AMD).
Each point was measured three times and the average was used. For
each tooth, a total of 51 measurements (15 for internal fit and 36 for
marginal fit) were collected as follows; on each BL slice: (2 mid-
axial) + (1 incisal) + (2 VG) + (2 HG) + (2 AMD) = 9. Measurement on
each MD slice: (2 mid-axial) + (2 VG) + (2 HG) + (2 AMD) = 8. Thus, the
total measurements per tooth: (9  3 slices) + (8  3 slices) = 51.
2.5. Examiner calibration
Examiner calibration was done prior to the study. The blinded
examiner (A.S) was calibrated against an experienced examiner (A.
N). A total of 20 2D-images were used for the calibration procedure.
The internal and marginal gaps were measured on all 2D-images.
After 10 days, the gaps were re-measured. Intra-examiner and
inter-examiner reliability were analyzed using interclass correla-
tion (ICC).
2.6. Statistical analysis
The mean and standard deviations of internal and marginal
gaps were analyzed using descriptive statistics. The data were
checked for normality distribution and equivalence of variances.
The results were compared using multivariate analysis of variance
(MANOVA). A follow-up simple main effect test with Bonferroni
corrections was used to evaluate the difference between the
fabrication methods in each finish line design (a = .05). One-way
ANOVA was used to analyze the influence of finish line design on
the marginal and internal gap of the restorations in each
fabrication method separately.
3. Results
The intra-examiner reliability was .98 and the absolute
agreement between the two examiners was .97 and thus
eliminating any possible measurement or examiner bias.
Please cite this article in press as: N. Alharbi, et al., Three-dimensional evaluation of marginal and internal fit of 3D-printed interim restorations
fabricated on different finish line designs, J Prosthodont Res (2017), https://doi.org/10.1016/j.jpor.2017.09.002
The mean and standard deviation of the marginal and
internal gap of all specimens are shown in Tables 1 and 2.
The result of MANOVA was reported using Pillai’s trace
multivariate test as the assumption of homogeneity of vari-
ance-covariance matrices was violated (Box’s M test, P = .000)
[26]. The mean internal and marginal gap were significantly
influenced by fabrication method (P = .000) and finish line design
(P = .000). Yet the influence of fabrication method was more as
shown by the Partial Eta squared value of (.96) compared to (.66)
for finish line design. MANOVA showed statistically significant
interaction effect between fabrication methods and finish line
design, F (18,207) = 8.15, P = .000; Pillai’s trace = 1.245. Step-down
univariate analysis showed that the interaction between fabrica-
tion methods and finish line designs was statistically significant at
incisal, internal, mid-axial, VG and HG. On the contrary, at AMD,
the fabrication method showed similar effect in different finish
line designs (P = .133). The 3D-Printed restorations showed lower
estimated mean marginal and internal gap values compared to
milled restorations. The difference was statistically significant at
all points (P = .000). Amongst all points, the incisal gap was the
largest in both 3D-printed and milled restorations (Tables 1 and 2)
(Fig. 5).
MANOVA simple main effect showed that the mean internal
gap, mid-axial and AMD of 3D-printed restorations were
significantly lower than that of milled restorations for all finish
lines designs (P <.05). The difference at incisal gap was significant
in all finish lines except in KE design (P = .662), whereas for the HG
the difference was significant in all finish line designs except in C
design (P = .219). The difference at VG was significant in all finish
lines except RSB design (P = .142).
The results of one-way ANOVA revealed that the mean internal
gap and AMD were significantly different between different finish
line designs (P = .000) in 3D-printed restorations, whereas in
milled restorations only mean internal gap was significantly
different in different finish lines (Table 3). Tukey post hoc showed
that in 3D-printed restorations the lowest mean internal gap was
in KE design, the difference between different finish lines was
statistically significant (P <.05) (Fig. 6A). Similarly, in milled
restorations the lowest mean internal gap was in KE design. The
difference was statistically significant (P = .000) between all
designs except S and C (P = .249). The lowest AMD for 3D-printed
restorations was with RSB design (Fig. 6B). The difference between
the groups was not significant except for C design (P <.05).
Likewise, the lowest mean AMD in milled restorations was with
RSB finish line (P >.05) (Fig. 7A and B).
G Model
JPOR 432 No. of Pages 9

N. Alharbi et al. / journal of prosthodontic research xxx (2017) xxx–xxx 5

Table 1
Mean and standard deviation (SD) of internal gap (sum of incisal gap + mid-axial gap).

Fabrication methods Finish line Specimens number (N) Mean (mm) Standard deviation (SD)
Incisal gap 3D-printed KE 10 93 14
C 10 213 31
RS 10 210 6
RSB 10 161 15
Total 40 169 52
Milling KE 10 97 17
C 10 271 18
RS 10 259 19
RSB 10 208 11
Total 40 209 71
Total KE 20 95 15
C 20 242 38
RS 20 235 29
RSB 20 184 27
Total 80 189 65

Mid-axial gap 3D-printed KE 10 39 5

C 10 45 3
RS 10 51 4
RSB 10 29 2
Total 40 41 9
Milling KE 10 80 6
C 10 84 7
RS 10 111 9
RSB 10 99 7
Total 40 94 14
Total KE 20 59 22
C 20 64 20
RS 20 81 32
RSB 20 64 36
Total 80 67 29

Internal gap 3D-printed KE 10 66 8

C 10 149 15
RS 10 130 4
RSB 10 95 7
Total 40 110 33
Milling KE 10 89 8
C 10 177 10
RS 10 185 11
RSB 10 154 7
Total 40 151 39
Total KE 20 77 14
C 20 163 19
RS 20 158 29
RSB 20 124 31
Total 80 131 42

Knife edge (KE), chamfer (C), rounded shoulder (RS), rounded shoulder with bevel (RSB).

4. Discussion In the literature, several methods are reported to evaluate the

This in vitro study evaluated the effect of fabrication methods in
two digital workflows; 3D-printing versus milling and four finish
line designs (KE, C, RS, RSB) on the marginal and internal fit of full
coverage interim dental restorations. Based on the results, the null
hypothesis that no difference exists in marginal and internal fit
between 3D-printing and milled restorations in different finish line
designs was rejected.
In this study, only a maxillary central incisor model was
evaluated to eliminate the influence of different tooth geometries
[27]. In addition, all restorations were analyzed without
cementation, which is not a true representation of the actual
clinical situation. Nonetheless, it has been shown that the use of
cement may lead to improper seating of the restoration and thus
increases the marginal and internal gap [7]. Furthermore, the
radiopaque nature of the cement may result in radiation artifacts
during the procedure of micro-CT scanning, an artifact that may
impede the measurements of the internal and marginal gaps
[18,28].
marginal and internal fit of the restorations [7]. In this study,
micro-CT was used. This approach offers several advantages of
being non-destructive, and allows for quantitative measurements
in three dimensions [18,28]. However, shortcoming of this method
is the radiation artifacts resulting from the difference in coefficient
of radiation absorption among different materials [28]. A custom-
made holder was fabricated to standardize the position of the
specimens during the scanning procedures. The maximum
magnification to visualize the complete crown resulted in pixel
size of 11 mm.
During the reslicing, the locations of the 2D-slices were selected
so that the measurement points were widely distributed on the
tooth without any distortion of the image. (Fig. 3) The measure-
ments of the selected points on the 2D-slices allowed for 3D-
analysis of the fit of the restorations. Marginal fit was evaluated by
measuring AMD, VG and HG to have an in-depth view of the
marginal misfit of the restorations [29]. The number of measure-
ment points for marginal gap was 36, which has been shown to
maintain the precision level of the measurements [7,28–30].

Please cite this article in press as: N. Alharbi, et al., Three-dimensional evaluation of marginal and internal fit of 3D-printed interim restorations
fabricated on different finish line designs, J Prosthodont Res (2017), https://doi.org/10.1016/j.jpor.2017.09.002
G Model
JPOR 432 No. of Pages 9

6 N. Alharbi et al. / journal of prosthodontic research xxx (2017) xxx–xxx

Table 2
Mean and standard deviation (SD) of marginal gap.

Fabrication methods Finish line Specimens number (N) Mean (mm) Standard deviation (SD)
Vertical gap (VG) 3D-printed KE 10 16 10
C 10 25 6
RS 10 21 4
RSB 10 26 10
Total 40 22 8
Milling KE 10 39 21
C 10 40 5
RS 10 34 13
RSB 10 19 4
Total 40 33 15
Total KE 20 27 20
C 20 33 9
RS 20 27 12
RSB 20 23 8
Total 80 28 13

Horizontal gap (HG) 3D-printed KE 10 25 9

C 10 22 4
RS 10 20 3
RSB 10 25 4
Total 40 23 6
Milling KE 10 31 8
C 10 25 3
RS 10 41 8
RSB 10 33 4
Total 40 32 8
Total KE 20 28 9
C 20 23 4
RS 20 30 12
RSB 20 30 6
Total 80 28 8

Absolute marginal discrepancy (AMD) 3D-printed KE 10 30 10

C 10 41 5
RS 10 30 5
RSB 10 28 2
Total 40 32 8
Milling KE 10 56 25
C 10 54 5
RS 10 52 16
RSB 10 38 3
Total 40 50 16
Total KE 20 43 23
C 20 48 8
RS 20 41 5
RSB 20 33 6
Total 80 41 15

Knife edge (KE), chamfer (C), rounded shoulder (RS), rounded shoulder with bevel (RSB).

Internal fit was evaluated by measuring the incisal and mid-axial
gaps, yet data of the mid-axial and incisal gaps were reported
separately to provide details of the internal misfit of the
restorations. The results were analyzed using MANOVA rather
than multiple two-way ANOVA’s to minimize the inflation of type I
error [26,31].
The results showed that 3D-printed restorations exhibit
significantly lower marginal and internal gap values than their
milled counterparts. The inferior marginal and internal fit of milled
restorations may be attributed to errors resulting from the
tolerance of milling burs [32,33]. Nevertheless, the mean AMG
marginal gaps of both 3D-printed and milled restorations were
within the reported values for the CAD/CAM restorations (10–
110 mm) [8].
Among all measurement points, the largest gap in both
fabrication methods was found at the incisal part (169 mm for
3D-printing and 209 mm for milling). This finding was in
accordance with previous studies [2,27,34–37]. In the present
study, the measured incisal gap was 169 mm and 209 mm which
were greater than the programmed incisal cement space by
1.5 and 1.8 for both the printed and milled restorations
respectively. This finding was different than the results of other
reported studies. Kokubo et al. [34] reported an incisal gap of
170 mm, which is 3 to 4 larger than the programmed cement
space. Other reports revealed an incisal gap and an occlusal gap
that ranged from 110 to 204 mm, that was 5 to 6 more than the
programmed cement space [2,35–37]. The reason for the increased
incisal gap and whether it is related to incorporated spaces during
the design process still merits further investigation.
Several design parameters including programmed cement gaps
are proposed based on the required cement thickness and are
mostly recommended based on operators’ and technicians’
experience and the available software [9,33,37]. The programmed
cement thickness reported in literature varied from no gap [11], to
10 mm [3], till 15 mm with an additional vertical space of 65 mm
and a horizontal space of 50 mm [38], 30 mm [2,10,23], 50 mm [34],
60 mm [1], till 85 mm [9]. Hoang et al. [9] evaluated the
reproducibility of 25 mm, 45 mm, 65 mm, 85 mm and 105 mm
cement gap of DLP 3D-printed resin copings and found that the
85 mm cement gap was the most reproducible [9]. In the same

Please cite this article in press as: N. Alharbi, et al., Three-dimensional evaluation of marginal and internal fit of 3D-printed interim restorations
fabricated on different finish line designs, J Prosthodont Res (2017), https://doi.org/10.1016/j.jpor.2017.09.002
G Model
JPOR 432 No. of Pages 9

N. Alharbi et al. / journal of prosthodontic research xxx (2017) xxx–xxx 7

Fig. 5. Estimated mean of internal and marginal gap in 3D-printed and milled restorations.

context, the results of the current study revealed that the mid-axial
gap in 3D-printed restorations was 41 mm, which is lower than the
programmed cement space (60 mm). Thus, it may be recom-
mended that the programmed axial gap be increased to enhance
the marginal fit though this still requires to be confirmed with
further studies (Fig. 6A and B). This lack of consistency between
measured axial gaps and programed incorporated cement space
may be due to an error occurring during the slicing procedure of
the STL file. Such an error can minimize the accuracy of the fitting
surface of printed restorations compared to the designed model.
On the contrary, the larger mid-axial gap of the milled restorations
was attributed to errors from the tolerance of milling burs [32].
Similarly, the incisal gap in KE design was lower than the
programmed incisal space in both fabrication methods. Vertical
displacement of the restoration is unlikely to be the reason
considering the values of the measured vertical gap (Table 1).
The results of MANOVA showed that the finish line design is less
influential on the marginal and internal fit of interim restorations
compared to the fabrication method. Several studies evaluated the
influence of finish line design on marginal and internal fit of milled
dental restorations with controversial findings [14–17]. This can be
viewed in the light of different experimental designs, different
material, fabrication and analysis methods.
The findings of this study revealed that in both fabrication
methods KE design has the least internal gap and RSB design
showed the least AMD. The difference in AMD between different
finish line designs was not significant in milled restorations. In 3D-
printed restorations, the C design showed significantly larger AMD
than the other finish line designs. This may be attributed to the
incremental layer pattern of build process in the 3D printing
technology. The curved surface of the axio-gingival line angle in C
design may cause an increased stair-stepping error during the
slicing of the STL file. Therefore, the resultant increase in the values
of the AMD recorded with the C design in the printed group.
Although the results revealed that KE design exhibit the least
internal gap in both fabrication methods, it cannot be recom-
mended for current clinical practice. KE design represents a very
conservative approach, which may limit the choice of restorative
material. In addition, the visibility of KE finish line on digital or
analogue model might not be optimal. Gingival bulk has been

Table 3
The influence of finish line designs on internal and AMD (mm) in both fabrication methods using One-way ANOVA.

Fabrication methods Sum of squares df Mean square F Significance

3D-printing Internal gap Between groups .040 3 .013 145 .000
Within groups .003 36 .000
Total .044 39
Absolute marginal discrepancy Between groups .001 3 .000 9.43 .000
Within groups .001 36 .000
Total .003 39

Milling Internal gap Between groups .057 3 .019 209 .000

Within groups .003 36 .000
Total .061 39
Absolute marginal discrepancy Between groups .002 3 .001 2.83 .052
Within groups .008 36 .000
Total .010 39

Please cite this article in press as: N. Alharbi, et al., Three-dimensional evaluation of marginal and internal fit of 3D-printed interim restorations
fabricated on different finish line designs, J Prosthodont Res (2017), https://doi.org/10.1016/j.jpor.2017.09.002
G Model
JPOR 432 No. of Pages 9

8 N. Alharbi et al. / journal of prosthodontic research xxx (2017) xxx–xxx

Fig. 6. Three dimensional reconstructed micro-CT image of the 3D-printed restoration in; (A) KE finish-line, (B) RSB finish-line.

Fig. 7. Influence of finish line design on internal fit and AMD in; (A) 3D-printing, (B) in milling. Different letters indicate statistical significance between different finish line
design.

reported in some cases with KE design due to insufficient room for
the restorative material [14]. On the contrary, RSB design showed
the lowest AMD values in both fabrication methods, as well as an
internal gap following the KE design (Fig. 6A and B). Improved
marginal adaptation associated with reduced AMD values com-
bined with acceptable internal fit values observed may be an
indication for recommending RSB design for SLA-3D printed
restorations.
However, generalizability of the results should be regarded
with caution. In this study, only one clinically available material
and one system for each fabrication method was applied. Whether
the interpretation given for the findings of this study can be
applied to different materials and different systems still needs to
be evaluated especially considering that such an interim restora-
tion would still be followed with a definitive restoration. In
addition, more studies are still needed to evaluate the influence of
cementation procedure as well as the single versus multi-unit fixed
partial dentures. The optimal cement gap, extra vertical and
horizontal space values that should be incorporated during
designing of the restoration are also parameters that still need
to be further assessed.
5. Conclusion
Within the limitation of this in vitro study, fabrication methods
showed more of an influence on the marginal and internal fit of
dental restorations compared to the effect of the design of the
finish line. Restorations fabricated using 3D-printing techniques
exhibited lower marginal and internal gap than restorations
fabricated using milling techniques. Nevertheless, in both fabrica-
tion methods, all values were within the reported values for CAD/
CAM restorations.
Acknowledgments
The authors would like to thank DWS, Italy for supplying the
specimens used in this study. Additionally, the authors would like

Please cite this article in press as: N. Alharbi, et al., Three-dimensional evaluation of marginal and internal fit of 3D-printed interim restorations
fabricated on different finish line designs, J Prosthodont Res (2017), https://doi.org/10.1016/j.jpor.2017.09.002
G Model
JPOR 432 No. of Pages 9
N. Alharbi et al. / journal of prosthodontic research xxx (2017) xxx–xxx
to thank Irene H.A. Aartman (Assistant Professor in statistics at
ACTA) for her statistical advice. This study was supported by a
scholarship grant number 2/302626 from King Saud University,
Riyadh, Kingdom of Saudi Arabia.
References
[1] Khng KYK, Ettinger RL, Armstrong SR, Lindquist T, Gratton DG, Qian F. In vitro
evaluation of the marginal integrity of CAD/CAM interim crowns. J Prosthet
Dent 2016;115:617–23.
[2] Park J-Y, Lee J-J, Bae S-Y, Kim J-H, Kim W-C. In vitro assessment of the marginal
and internal fits of interim implant restorations fabricated with different
methods. J Prosthet Dent 2016;116:536–42.
[3] Abdullah AO, Tsitrou EA, Pollington S. Comparative in vitro evaluation of CAD/
CAM vs conventional provisional crowns. J Appl Oral Sci 2016;24:258–63.
[4] Joo H-S, Park S-W, Yun K-D, Lim H-P. Complete-mouth rehabilitation using a
3D printing technique and the CAD/CAM double scanning method: a clinical
report. J Prosthet Dent 2016;116:3–7.
[5] Burns DR, Beck DA, Nelson SK. A review of selected dental literature on
contemporary provisional fixed prosthodontic treatment: report of the
committee on research in fixed prosthodontics of the academy of fixed
prosthodontics. J Prosthet Dent 2003;90:474–97.
[6] Tuntiprawon M, Wilson PR. The effect of cement thickness on the fracture
strength of all-ceramic crowns. Aust Dent J 1995;40:17–21.
[7] Nawafleh NA, Mack F, Evans J, Mackay J, Hatamleh MM. Accuracy and
reliability of methods to measure marginal adaptation of crowns and FDPs: a
literature review. J Prosthodont 2013;22:419–28.
[8] Boitelle P, Mawussi B, Tapie L, Fromentin O. A systematic review of CAD/CAM
fit restoration evaluations. J Oral Rehabil 2014;41:853–74.
[9] Hoang LN, Thompson GA, Cho S-H, Berzins DW, Ahn KW. Die spacer thickness
reproduction for central incisor crown fabrication with combined computer-
aided design and 3D printing technology: an in vitro study. J Prosthet Dent
2015;113:398–404.
[10] Kim K-B, Kim J-H, Kim W-C, Kim J-H. In vitro evaluation of marginal and
internal adaptation of three-unit fixed dental prostheses produced by
stereolithography. Dent Mater J 2014;33:504–9.
[11] Kim D-Y, Jeon J-H, Kim J-H, Kim H-Y, Kim W-C. Reproducibility of different
arrangement of resin copings by dental microstereolithography: evaluating
the marginal discrepancy of resin copings. J Prosthet Dent 2017;117:260–5,
doi:http://dx.doi.org/10.1016/j.prosdent.2016.07.007.
[12] Munoz S, Ramos V, Dickinson DP. Comparison of margin discrepancy of
complete gold crowns fabricated using printed, milled, and conventional
hand-waxed patterns. J Prosthet Dent 2017;118:89–94, doi:http://dx.doi.org/
10.1016/j.prosdent.2016.09.018.
[13] Sancho-Puchades M, Fehmer V, Hämmerle C, Sailer I. Advanced smile
diagnostics using CAD/CAM mock-ups. Int J Esthet Dent 2014;10:374–91.
[14] Comlekoglu M, Dundar M, Özcan M, Gungor M, Gokce B, Artunc C. Influence of
cervical finish line type on the marginal adaptation of zirconia ceramic crowns.
Oper Dent 2009;34:586–92.
[15] Euán R, Figueras-Álvarez O, Cabratosa-Termes J, Oliver-Parra R. Marginal
adaptation of zirconium dioxide copings: influence of the CAD/CAM system
and the finish line design. J Prosthet Dent 2014;112:155–62.
[16] Tsitrou EA, Northeast SE, van Noort R. Evaluation of the marginal fit of three
margin designs of resin composite crowns using CAD/CAM. J Dent
2007;35:68–73.
[17] Akbar JH, Petrie CS, Walker MP, Williams K, Eick JD. Marginal adaptation of
Cerec 3 CAD/CAM composite crowns using two different finish line
preparation designs. J Prosthodont 2006;15:155–63.
Please cite this article in press as: N. Alharbi, et al., Three-dimensional evaluation of marginal and internal fit of 3D-printed interim restorations
fabricated on different finish line designs, J Prosthodont Res (2017), https://doi.org/10.1016/j.jpor.2017.09.002
9
[18] Borba M, Cesar PF, Griggs JA, Della Bona Á. Adaptation of all-ceramic fixed
partial dentures. Dent Mater 2011;27:119–26.
[19] Goodacre CJ, Campagni WV, Aquilino SA. Tooth preparations for complete
crowns: an art form based on scientific principles. J Prosthet Dent
2001;85:363–76.
[20] Martínez-Rus F, Suárez MJ, Rivera B, Pradíes G. Evaluation of the absolute
marginal discrepancy of zirconia-based ceramic copings. J Prosthet Dent
2011;105:108–14.
[21] Alharbi N, Osman R, Wismeijer D. Factors influencing the dimensional
accuracy of 3D-printed full coverage dental restorations using stereolithog-
raphy technology. Int J Prosthodont 2016;29:503–10.
[22] Osman R, Alharbi N, Wismeijer D. Build angle, does it have an influence on the
accuracy of 3D-printed dental restorations using digital light-processing
technology. Int J Prosthodont 2017;30:182–8.
[23] Wu JC, Wilson PR. Optimal cement space for resin luting cements. Int J
Prosthodont 1994;7:209–15.
[24] DWS. Dental lab additive manufacturing systems. Italy: DWS; 2013.
[25] AG IS. Straumann CARES tooth-born prosthetic procedures. Switzerland:
Straumann; 2012. p. 10–30.
[26] Tabachnick BG, Fidell LS. Using multivariate statistics. 5th ed. United States:
Allyn and Bacon; 2008. p. 243–68.
[27] Boening KW, Wolf BH, Schmidt AE, Kästner K, Walter MH. Clinical fit of Procera
All Ceram crowns. J Prosthet Dent 2000;84:419–24.
[28] Mously HA, Finkelman M, Zandparsa R, Hirayama H. Marginal and internal
adaptation of ceramic crown restorations fabricated with CAD/CAM technol-
ogy and the heat-press technique. J Prosthet Dent 2014;112:249–56.
[29] Holmes JR, Bayne SC, Holland GA, Sulik WD. Considerations in measurement of
marginal fit. J Prosthet Dent 1989;62:405–8.
[30] Groten M, Axmann D, Pröbster L, Weber H. Determination of the minimum
number of marginal gap measurements required for practical in vitro testing. J
Prosthet Dent 2000;83:40–9.
[31] Haase RF, Ellis MV. Multivariate analysis of variance. J Couns Psychol
1987;34:404–13.
[32] Örtorp A, Jönsson D, Mouhsen A, Vult von Steyern P. The fit of cobalt–
chromium three-unit fixed dental prostheses fabricated with four different
techniques: a comparative in vitro study. Dent Mater 2011;27:356–63.
[33] Beuer F, Schweiger J, Edelhoff D. Digital dentistry: an overview of recent
developments for CAD/CAM generated restorations. Br Dent J 2008;204:505–
11.
[34] Kokubo Y, Nagayama Y, Tsumita M, Ohkubo C, Fukushima S, Steyern P. Clinical
marginal and internal gaps of In-Ceram crowns fabricated using the GN-I
system. J Oral Rehabil 2005;32:753–8.
[35] Martins LM, Lorenzoni FC, Melo AOd, Silva LMd, Oliveira JLGd, Oliveira PCGd,
et al. Internal fit of two all-ceramic systems and metal-ceramic crowns. J Appl
Oral Sci 2012;20:235–40.
[36] Bayramog lu E, Özkan YK, Yildiz C. Comparison of marginal and internal fit of
press-on-metal and conventional ceramic systems for three-and four-unit
implant-supported partial fixed dental prostheses: an in vitro study. J Prosthet
Dent 2015;114:52–8.
[37] Shamseddine L, Mortada R, Rifai K, Chidiac JJ. Marginal and internal fit of
pressed ceramic crowns made from conventional and computer-aided design/
computer-aided manufacturing wax patterns: an in vitro comparison. J
Prosthet Dent 2016;116:242–8.
[38] Anunmana C, Charoenchitt M, Asvanund C. Gap comparison between single
crown and three-unit bridge zirconia substructures. J Adv Prosthodont
2014;6:253–8.