Q U I N T E S S E N C E I N T E R N AT I O N A L
ORTHODONTICS
Accuracy and efficiency of full-arch digitalization and
3D printing: A comparison between desktop model
scanners, an intraoral scanner, a CBCT model scan, and
stereolithographic 3D printing
Christian Wesemann1/Jonas Muallah 1/James Mah, DDS, MS, DMSc 2/Axel Bumann, DDS, PhD3
Objective: The primary objective of this study was to com-
pare the accuracy and time efficiency of an indirect and direct
digitalization workflow with that of a three-dimensional (3D)
printer in order to identify the most suitable method for ortho-
dontic use. Method and Materials: A master model was
measured with a coordinate measuring instrument. The dis-
tances measured were the intercanine width, the intermolar
width, and the dental arch length. Sixty-four scans were taken
with each of the desktop scanners R900 and R700 (3Shape),
the intraoral scanner TRIOS Color Pod (3Shape), and the
Promax 3D Mid cone beam computed tomography (CBCT)
unit (Planmeca). All scans were measured with measuring
software. One scan was selected and printed 37 times on the
D35 stereolithographic 3D printer (Innovation MediTech). The
printed models were measured again using the coordinate
measuring instrument. Results: The most accurate results
Key words: 3D printed models, accuracy, CBCT model scan, desktop model scanner, digitalization, full-arch scans,
intraoral scanner
Since Duret’s first studies of computer-aided dental
medicine in 1971, computer-aided design/comput-
1
Doctoral Student, Charité – Universitätsmedizin Berlin, Campus Benjamin Frank-
lin, Center for Craniofacial Sciences (CC3), Berlin, Germany.
2
Director of Advanced Education Program in Orthodontics, University of Nevada
Las Vegas, School of Dental Medicine, Las Vegas, NV, USA.
3
Director of 3D Imaging, Mesantis 3D Dental-Radiology, Berlin, Germany.
Correspondence: Mr Christian Wesemann, Charité – Universitäts-
medizin Berlin, Campus Benjamin Franklin, Center for Craniofacial Sci-
ences (CC3), Assmanshauser Str. 4–6, 14197 Berlin, Germany. Email:
christian.wesemann@charite.de
VOLUME 48 • NUMBER 1 • JANUARY 2017
Christian
Wesemann
were obtained by the R900. The R700 and the TRIOS intraoral
scanner showed comparable results. CBCT-3D-rendering with
the Promax 3D Mid CBCT unit revealed significantly higher
accuracy with regard to dental casts than dental impressions.
3D printing offered a significantly higher level of deviation
than digitalization with desktop scanners or an intraoral scan-
ner. The chairside time required for digital impressions was
27% longer than for conventional impressions. Conclusion:
Conventional impressions, model casting, and optional digit-
ization with desktop scanners remains the recommended
workflow process. For orthodontic demands, intraoral scanners
are a useful alternative for full-arch scans. For prosthodontic
use, the scanning scope should be less than one quadrant
and three additional teeth. (Quintessence Int 2017;48:41–50;
doi: 10.3290/j.qi.a37130)
er-assisted manufacture (CAD/CAM) has gained in
importance. A wide field of new treatment concepts
has been developed. For all of them an efficient and
accurate digitalization of the initial situation is essen-
tial. Thereby, the devices for digitalization have to meet
various demands for the different dental disciplines.
For orthodontic use, only fast full-arch scans are
required in the majority of cases. The resulting digital
models can be used for simplified cast analysis,1 indi-
41
Q U I N T E S S E N C E I N T E R N AT I O N A L
Wesemann et al
vidually optimized wires (SureSmile), and rapid proto-
typing based therapies with splints (Invisalign). For
CAD/CAM manufactured prosthetic restorations,
mainly small areas with high accuracy are necessary.
However, the accuracy of full-arch scans is of interest
when considering the digital impression taking of mul-
tiple prepared abutments or implants.2
At present, there are three well-established proced-
ures: firstly, direct digitalization using intraoral scan-
ners;3,4 secondly, indirect digitalization of dental
impressions or casts using desktop scanners;5 and
finally, the application of cone beam computed tomog-
raphy (CBCT), which can be used in a similar way to
desktop scanners or to generate the required data from
existing records.6,7 For orthodontic use, an additional
physical model is often needed for treatment and stor-
age. In an indirect workflow, it is already part of the
process. In a direct workflow, it has to be printed with a
three-dimensional (3D) printer afterwards.
The number of studies concentrating on the accu-
racy of full-arch digital models has been increasing in
the last 3 years.3,4,8,9 Studies concentrating on full-arch
scans with intraoral scanners found great deviations at
the distal ends, which resulted from the accumulation
of matching errors in the region of the anterior teeth
caused by the limited complexity of their surface.3,4,10
Therefore, the current recommendation is that the
scanning scope should not exceed a single quadrant
when used in prosthodontics.11,12
Most of these studies use superimposition of a digi-
tal master model and the test model to evaluate the
accuracy.13,14 This method allows a reliable calculation
of the average surface deviation to be made and areas
of high deviation to be located. However, the ability to
calculate the clinical value of the mean surface diver-
gence is limited due to the absence of a reference par-
ameter. Furthermore, most studies are based on only a
small number of cases (n = 5 to 108),14,15 thus limiting
their reliability.
The aim of the present study was to evaluate the
different methods of digitalization using a standardized
master model with reproducible and individually inde-
pendent points of measurement. The accuracy of digital
42
full-arch models should be compared with the accuracy
of 3D printed models to determine when direct or indi-
rect digitalization is preferable in orthodontic use. Fur-
thermore, an analysis of the impact of matching errors
when using an intraoral scanner allows evaluation of
the possible width of the scanning scope for prostho-
dontic demands. The equipment used comprised the
R700 and R900 (3shape) desktop scanners, the TRIOS
Color Pod (3shape) intraoral scanner, the Planmeca
Promax 3D Mid CBCT unit with a Planmeca Romexis 3D
Ortho Studio Advanced software extension (Planmeca),
and the D35 3D-Printer (Innovation MediTech).
METHOD AND MATERIALS
For the design of the master model, a dimension-stable
FotoDent-LED resin model of a eugnathic maxilla was
modified. Using a Metabo Magnum TBE 4512 bench
drill, five holes were made vertically in the underside of
the model in such a way that the intercanine width
(ICW), intermolar width (IMW), and dental arch length
(AL) could be determined. The lengths of these dis-
tances were geared to the average peak of Western
population.16 As the broadest possible dimensions of
an average jaw were to be measured, the maxilla,
which is the larger, was selected for investigation
(Fig 1). The surface was subsequently trimmed parallel
to the underside in order to ensure that all measure-
ment points were located at the same level. The high-
est point of the calculable centerline of each drill was
chosen as a largely independent and therefore repro-
ducible measuring point (Fig 2).15 The master model
was measured in collaboration with the Fraunhofer
Institute IPK Berlin, Center for Micro Production Tech-
nology, using a Zeiss O-Inspect 422 coordinate measur-
ing instrument and Calypso 5.4.20 software (Zeiss). The
uncertainty of measurement (maximum permissible
error for length measurement; MPEE) was 1.9 μm +
L/250 (ISO 10360). This corresponds to a maximum
uncertainty of 2.1 μm at a distance of 50 mm. The drills
were first checked for their roundness in order to verify
that they qualified as an exact base for the measuring
points. The centerlines were then defined and the
VOLUME 48 • NUMBER 1 • JANUARY 2017

Fig 1 Master model with five holes as markers.
respective distances measured. These results were set
as the gold standard. Pilot tests to calculate the number
of cases resulted in n = 37 to prove significant differ-
ences > 10 μm, and n = 64 to prove significant differ-
ences > 5 μm. All devices for digitalization were cali-
brated before use. Sixty-four scans were then per-
formed with each device. The scans were carried out by
a trained operator in nearly standard ambient condi-
tions (22 to 25°C; 1,013.25 hPa). To ensure dimensional
stability, the master model was re-measured using the
same coordinate measuring instrument after the com-
pletion of all scans.
After a training period of 3 days on the TRIOS intra-
oral scanner, the master model was scanned 74 times.
The first 10 scans were discarded to avoid initial error.
The scan path proceeded from occlusal to buccal to
oral in one quadrant, then moved on to the second.17
The scan was repeated if the image transmission was
interrupted by an oversized gap between the scan-
ner head and the teeth, or by rapid movements. This
procedure prevented any faulty image data overlaps
caused by the operator. The scans were imported to
the software OrthoAnalyzer (3shape; Version 2013.1)
and saved as stereolithography (STL) files without fur-
ther processing. In order to simulate a clinical workflow
for scans made by the R700 and R900 desktop scan-
ners, an impression of the master model was taken
and a plaster cast made. The double mix impression
VOLUME 48 • NUMBER 1 • JANUARY 2017
Q U I N T E S S E N C E I N T E R N AT I O N A L
Wesemann et al
Fig 2 Measurements were taken between the highest points of
each centerline.
was created using a vinyl polysiloxane (VPS; Flexi-
time medium flow + putty; Heraeus Kulzer). For the
plaster model, Kanistone KFO special dental stone
type 3 (Kaniedenta) was used. The scans were made
using Scanlt Orthodontics software, version 2013.1,
and imported to OrthoAnalyzer to be saved as STL files.
The Planmeca Promax 3D Mid CBCT unit installed with
Planmeca Romexis 3D Ortho Studio Advanced soft-
ware, version 3.6.0 R, allows impressions and plaster
models to be captured and directly processed into a
STL file. The existing plaster model and an impression
of the master model made of Flexitime medium flow
(the manufacturers’ instructions prescribed the use
of a single phase, homogenous material) were used
for the captures. The CBCT unit was calibrated for
both materials, and 64 captures were taken of each
of them, all according to the recommended settings
70 kV/12.5 mA/22.3 seconds/voxel size 0.1 mm. The
data were stored in STL format.
Analysis of the data was carried out using the mea-
suring software Convince Premium 2012 (3shape). All
STL data were imported and the vertical walls of the
holes were marked. The centerlines of the cylinders
were calculated and the highest point of each center-
line was set as the measurement point. X, y, and z coor-
dinates could thus be defined in a way that could be
reproduced. A 3D point-to-point measuring technique
was used (Fig 3).
43
Q U I N T E S S E N C E I N T E R N AT I O N A L
Wesemann et al

A digital model was subsequently chosen, and

Fig 3 Digital models were measured using Convince Premium
2012 (3Shape) measuring software.
measured values were defined as master values for
testing of the D35 printer. Owing to financial restric-
tions, 37 physical models were printed and measured
using the coordinate measuring instrument. The meas-
ured ranges and the absolute deviations from the mas-
ter model obtained by the different methods were then
compared (Fig 4).
The null hypothesis “the devices digitize and print
distances with equal accuracy” was adopted in case no
significant difference could be shown in their trueness
and precision. Trueness was defined as the proximity of
the measured value to the true value. The dimension of

Master model Digital master model

n=1

n = 2 × 64 n = 64 n = 64 n = 64

D35
(Innovation MediTech)
Promax 3D R900 R700 TRIOS
Mid (Promeca) (3Shape) (3Shape) (3Shape) n = 37

Measured with software Convince Premium Measured with coordinate measuring

instrument (Zeiss O-Inspect 422)

Fig 4 Study design.

44 VOLUME 48 • NUMBER 1 • JANUARY 2017

 Q U I N T E S S E N C E I N T E R N AT I O N A L
Wesemann et al

average trueness was the mean value or the median in
the event of the data not being normally distributed.
Precision was based on the proximity of agreement in a
set of results.18 Precision correlates to the standard
deviation or the interquartile range if the data were not
normally distributed.
Statistical analysis
First of all, ranges and absolute differences were
checked for normal distribution. Thereafter, nonpara-
metric test methods were applied because the data in
part did not follow normal distribution. The Friedman
test, a global test for independent samples, showed a
significant difference between the different devices. A
Wilcoxin post-hoc pairwise test was used to determine
which methods differ significantly. The median as a
measure of central tendency and the quartile range as
a measure of dispersion were used for the evaluation.
Mean value and standard deviation (SD) were not part
of the pairwise comparison. P values smaller than .05
were treated as statistically significant. The calculations
were made using SPSS statistical software (version 22;
IBM) and the support of Medistat.
RESULTS
Using the coordinate measuring instrument, 169 to 200
measuring points were positioned in each hole to
check their roundness. The average diameters were:
1,990.7 μm, 1,991.3 μm, 1,989.3 μm, 1,990.4 μm, and
1,990.8 μm. The deviation between the largest and
smallest holes was thus 2 μm. The maximum deviation
within one hole was 1.6 μm. The calculation of the dis-
tances gave the master values: ICW 32,839 μm, IMW
49,993 μm, and AL 77,679 μm. After the scans were
completed, the master model was again measured and
the following results achieved: ICW 32,840 μm, IMW
49,994 μm, and AL 77,697 μm. The digital master model
was measured with the following results: ICW
32,842 μm, IMW 49,996 μm, and AL 77,655 μm. The
descriptive statistic and the time required for each
method is shown in Table 1.

Table 1 Deviations from reference of each method (μm)

Planmeca Planmeca
R900 R700 Trios Pod plaster impression D35 (20$
(28,000$) ‡ (21,000$) ‡ (27,000$) ‡ (105,000$) ‡ (105,000$) ‡ per jaw) ‡
Chairside 5.5 5.5 7 5.5 5.5 NA
Average time
Laboratory 50 50 NA 50 50 2–5 days†
required*
Digitalization/processing 1.5 3 3 6 6 NA
Median 11 18 25 61 136 23

Intercanine Interquartile range 11 28 31 15 16 23

width Mean 12 21 27 63 135 27
SD 8 18 19 12 30 20
Median 13 27 36 73 188 68

Intermolar Interquartile range 15 25 46 14 25 57

width Mean 14 30 50 75 185 65
SD 10 19 39 22 13 39
Median 17 52 36 231 345 126
Interquartile range 17 55 33 33 43 74
Arch length
Mean 17 50 37 236 336 118
SD 12 31 22 28 39 61
*Average time per jaw in minutes.
†Manufacturer quoted prices (USD) in September 2016.

‡If ordered.

NA, not applicable.

VOLUME 48 • NUMBER 1 • JANUARY 2017 45

Q U I N T E S S E N C E I N T E R N AT I O N A L
Wesemann et al

500 Intercanine width Intermolar width Arch length

Deviations from reference (μm)

400

300

200

100

0
R900

R700

TRIOS

Planmeca model

Planmeca impr

D35

R900

R700

TRIOS

Planmeca model

Planmeca impr

D35

R900

R700

TRIOS

Planmeca model

Planmeca impr

D35
Fig 5 Boxplot of the deviations for intercanine width, intermolar width, and arch length.

The shortest range, ICW, had the highest accuracy meca Promax 3D Mid model scan (P = .108). The maxi-
(Fig 5). The results produced by the R900 scanner mum deviations were: R900, 38 μm; R700, 82 μm;
showed significantly less deviation than its predeces- TRIOS, 165 μm; Planmeca Promax 3D Mid model,
sor, the R700 (P = .002). The R700 scanner was signifi- 118 μm; Planmeca Promax 3D Mid impression, 218 μm;
cantly more accurate than the TRIOS intraoral scanner and the D35, 162 μm.
(P = .048). The highest deviation was found in the The deviations in combined measurement lengths
results of Planmeca Promax 3D Mid (P < .001). Thereby (the AL), are summarized in Fig 5. Again, the R900 pro-
the digitalization of plaster casts was significantly more duced significantly more accurate results than any
accurate than impressions (P < .001). All measurements other method. However, within this directly measured
taken by the Planmeca Promax 3D Mid, for both casts range, the TRIOS had significantly better results than
and impressions, showed reduced values compared to the R700 (P = .11). While the deviations within the first
the master values. The D35 3D printer showed results three distances with TRIOS did not show significant
comparable to the TRIOS (P = .827). The greatest devia- differences, the last distance was significant less accu-
tions were: R900, 34 μm; R700, 98 μm; TRIOS, 73 μm; rate (P < .001). The greatest deviations were found
Planmeca Promax 3D Mid plaster, 97 μm; Planmeca again in the results obtained with the Planmeca,
Promax 3D Mid impression, 316 μm; and D35, 90 μm. whereby the deviations of plaster models were signifi-
With regard to the greater measurement distance cantly more accurate than impressions (P < .001). All
IMW, all devices showed increased deviation (Fig 5). data collected using the Planmeca were smaller than
Congruent with the results shown for the ICW, the R900 the master values. The D35 produced significantly
showed significantly smaller deviations than the R700 fewer accurate results than the R700 and the TRIOS
(P < .001), and the R700 significantly better results than (P < .001), but significantly more accurate results than
the TRIOS (P = .002). TRIOS produced significantly more the Planmeca Promax 3D Mid plaster scan (P < .001).
accurate results than Planmeca Promax 3D Mid plaster The greatest deviations were: R900, 61 μm; R700,
and impressions (P < .001). All the data collected by the 112 μm; TRIOS, 82 μm; Planmeca Promax 3D Mid
Planmeca showed a reduction compared to the master model, 295 μm; Planmeca Promax 3D Mid impression,
values. The D35 did not differ significantly to the Plan- 449 μm; and the D35, 221 μm.

46 VOLUME 48 • NUMBER 1 • JANUARY 2017


DISCUSSION
Due to the increasing use of computer-aided treatment
concept systems, the need for an efficient and, most
importantly, accurate method of digitalizing the initial
dental situation is essential. For this reason, the aim of
the present study was to evaluate the accuracy of full-
arch impressions using different digitalization methods.
The best results were obtained with the desktop
scanner R900. Although there are no reference studies
for the R900, the deviation lay within the range given in
the manufacturer’s instruction manual (accuracy
15 μm). For the previous model, the R700, better results
were obtained for ICW and IMW than for combined
distances (AL). The reason for this discrepancy could be
that most of the range sections showed a negative
deviation, which accumulates and contributes to the
degree of inaccuracy. The previously published values
of a 50 μm average deviation5 can be viewed as compa-
rable if the different measurement methods used are
taken into account. The later model, the R900, thus
appears to be approximately 50% more accurate and
time efficient. Further research is necessary to establish
to what extent these in-vitro measured values, without
saliva or above-average undercut, for example, are
transferrable. However, other in-vitro11,12 and in-vivo
studies9,14 also conclude that the desktop scanner
method remains the most accurate, as long as great
care is put into the selection and handling of dental
impressions and cast materials.
In recent years, an increasing number of intraoral
scanners have been used in the field of prosthodontics
and orthodontics. As many studies of the restorative
aspects of scanning evaluation have restricted them-
selves to limited areas, it is positive that some more
recent studies address the issue of full-arch scans.2,4,10,15
The Lava COS (3M Espe) and the iTero (Cadent) scan-
ners used in these studies show average deviations of
44.9 μm,10 38.0 μm,4 23.5 μm (Lava COS),15 and
61.1 μm,15 49.0 μm,4 32.4 μm (iTero).10 Depending on
the design of the study, the results obtained by the
TRIOS are comparable to those mentioned above
(median 36 μm AL and IMW). The greatest deviations in
VOLUME 48 • NUMBER 1 • JANUARY 2017
Q U I N T E S S E N C E I N T E R N AT I O N A L
Wesemann et al
the results collected by the TRIOS were in the molar
range. There was a similar finding in almost all compa-
rable studies with intraoral scanners.3,4,10,14 Complex
dental surface structures are essential for all intraoral
scanners as their results can only be calculated based
on the difference between closely located object
points. The limited complexity of anterior teeth makes
it difficult to achieve an exact stitching of the scan.3,4,10
Small errors in this area, especially any that influence
the angular relationship, add up to relatively high devi-
ations at the distal ends. Therefore, most studies con-
clude that a scan of a single quadrant is clinically suffi-
cient for prosthodontics,11,12 but a full-arch scan does
not meet the requirements.2 In contrast, the results of
TRIOS demonstrated that matching errors in the an-
terior region did not influence the accuracy of the scan
significantly until reaching the following canine. If a
full-arch scan is taken from the maxillary right first
molar to the left first molar, only the last three teeth
contribute to 46.6% of the deviations. The first nine
teeth indeed show consistently low deviations.
Previous studies, in which study models were
obtained from CBCT data, used patient images to sepa-
rate tooth data. Those study models show a critical
surface solution and an average deviation of at least
1 mm for ICW and IMW.6,7 Due to the fact that a
patient’s exposure to radiation has to be kept “As Low
As Reasonably Achievable” (ALARA), the making of new
images to generate a digital model is contraindicated.
It was, therefore, only possible if previously existing
records were available. In 2014, a new version of the
software Planmeca Romexis 3D Ortho Studio Advanced
was released with an additional feature allowing direct
digitalization of dental impressions and casts and their
formatting into STL files using CBCT. The patient is not
exposed to any radiation during this procedure. When
compared to the reference value, the collected data
showed a small reduction independent of the object
placement, which is why a system-related error can be
assumed. Similar reductions were also found in other
studies that conducted CBCT measurements.19,20 One
theory that could explain this deviation, and which is
common in the fields of radiology and anthropology,
47
Q U I N T E S S E N C E I N T E R N AT I O N A L
Wesemann et al
involves the partial volume averaging (PVA) effect.21,22
Voxels that are located completely inside an object,
replicate object density well. However, voxels that are
located at intersections and cover both air and object,
show an average density. This average value can be
close to the value of air, but also resembles the value of
the object, depending on the location. Depending on
the threshold value settings, it thus can lead to the
object appearing smaller due to an error during conver-
sion into a surface polygon net (STL format). It remains
to be seen if an improved threshold value leads to
more precise results.
As far as can be established, no general definition of
the accuracy required for a digital dental model as
being clinically sufficient has yet been made. For dental
AL, a deviation of up to 100 μm for restorative works
has been rated critical.10 Hayashi et al5 classified punc-
tual deviations of 100 μm as trivial in the orthodontic
field. The American Board of Orthodontics, however,
claims that in an “objective grading system for dental
casts,” deviations of up to 500 μm are clinically accept-
able with regard to the categories of “alignment” and
“marginal ridges.”23 From a physiologic point of view,
the responses of the periodontal space to deflections
differing in height can provide valuable data on clini-
cally required accuracy. The mean tactile sensitivity
threshold of vital teeth is approximately 15 μm.24 If a
tooth is moved in a horizontal direction by less than
this amount, the subliminal irritation is not registered
and only long-term adaption of the periodontal liga-
ment occurs. Assuming that later work affects a mini-
mum of two teeth, and that the deflection of both
teeth is on average the same, the result is an accept-
able total deflection of maximum 30 μm for the entire
jaw. In addition to this, digital inaccuracies of less than
30 μm should be pursued as the gold standard with
regard to the greatest measured lengths AL and IMW.
Larger deflections are registered by mechanoreceptors,
but as long as displacement remains below the intrinsic
mobility of the tooth, no impulses for bone remodeling
are conducted by the desmodontal fibers.25 Minimal
intrinsic tooth mobility is exhibited by molars and aver-
ages 70 μm,25 allowing a maximum value of 140 μm to
48
be deduced, which applies to the entire jaw. Greater
movement will result in bone resorption. If the deflec-
tion condenses the periodontal space by less than half,
the blood flow persists with the result of direct bone
resorption on the pressure side. Deflections compress-
ing the periodontal space by more than half cause cir-
culatory disorders and lead to indirect resorption,
which may be associated with root resorption.25 The
periodontal space is approximately 250 μm in areas of
the highest deflection,26 and a horizontal deflection of
around 125 μm per tooth or 250 μm for the entire jaw
can thus be viewed as the cut-off value between indi-
rect and direct absorption. Based on this, full-arch scans
with deviations of less than 30 μm were classed as
“excellent,” less than 140 μm as “very good,” less than
250 μm as “acceptable,” and above 250 μm as “insuffi-
cient.” These values should be regarded as a simplified
working hypothesis. Consequentially, the following
conclusions can be drawn: Despite inaccuracies occur-
ring during impression and casting, the indirect digita-
lization by the R900 showed “excellent” results. The
results of the TRIOS and R700 were clinically compara-
ble and rated “very good.” If only the first nine teeth of
a full-arch scan and the ICW are taken into account, the
results of TRIOS are “excellent.” The indirect digitaliza-
tion of plaster models using the CBCT device Promax
3D Mid was “acceptable,” but the digitalization of
impressions appeared “insufficient.” Printed models
made by the D35 produced “very good” results com-
pared to the digital master model, while the entire
direct workflow with an intraoral scan and a 3D print-
out showed an “acceptable” accuracy.
Despite the potential errors that may occur during
the taking of impressions and the casting of models,
indirect digitalization with the R900 showed the high-
est level of accuracy. It must be taken into account that
VPS was used instead of alginate in this study. Ender
and Mehl10 found deviations of 13.0 ± 2.9 μm for VPS
and 37.7 ± 34.9 μm for alginate. Considering these
additional deviations, the differences between the
R900 and TRIOS can be neglected. The dimensional
stability of printed models does not conform to ISO
6873:2013 for type 3 and type 4 dental gypsum, but it is
VOLUME 48 • NUMBER 1 • JANUARY 2017

comparable to type 2 plaster. Generating digital mod-
els directly is much faster, but it takes a longer chairside
time. In vitro, 27% more chairside time was needed
compared to the impression taking with VPS. Grunheid
et al3 reported 7.5 minutes for two alginate impressions
and 20.5 minutes to scan both jaws intraorally. At the
same time, 73.3% of patients preferred alginate instead
of intraoral scanning.3 As seen in previous studies,4,10,14
the results obtained prove direct digitalization and 3D
printing to be a possible alternative. However, new
technologies should not only be comparable, but also
offer improvements over conventional methods to gain
wide acceptance in daily practice. This sort of major
improvement has not yet been demonstrated for the
orthodontic use of intraoral scanners. One reason could
be that common intraoral scanners are primarily
designed for prosthodontics. To scan a small area with
excellent resolution, the use of small scanners, which
offer good handling, is preferred. However, orthodontic
demands are different. Here it would be useful to
record the entire dentition with a larger scanner, ideally
in only one capture. Matching errors could be avoided
and time used more efficiently. Without doubt, intra-
oral scanners have great potential. However, the
devices have to be highly customized to meet the
requirements of single disciplines, in order to fully uti-
lize their capability.
Deviations, possibly caused by the methodology,
that require further consideration include the measure-
ment inaccuracy (MPEE) of the Zeiss O-Inspect 422
coordinate measuring instrument, which is 1.9 μm +
L/250, a radius inaccuracy of 0.8 μm, and a dimensional
variance of the master model over time. This variance
measured 1 μm for IMW and ICW and was in accor-
dance with the MPEE. AL deviations of up to 18 μm
were measured and must be taken into account.
CONCLUSION
The results obtained in the present study suggest that
the process of making a conventional impression, then
model casting, and finally optional digitalization with a
modern desktop scanner is the best choice for ortho-
VOLUME 48 • NUMBER 1 • JANUARY 2017
Q U I N T E S S E N C E I N T E R N AT I O N A L
Wesemann et al
dontic use. If a CBCT device is available, it can be used
for the indirect digitalization of plaster models for digi-
tal storage, diagnosis, and planning. This method also
provides an acceptable level of accuracy for the manu-
facturing of orthodontic appliances. Direct digitaliza-
tion with intraoral scanners is demonstrated to be a
suitable alternative, although it requires more chairside
time and does not result in a higher level of accuracy
than an indirect workflow with desktop scanners. If it is
possible to manufacture the appliance in a completely
digital workflow, intraoral scanners are highly suitable.
If a physical working model is required, it should be
manufactured conventionally and afterwards digi-
talized. If a full-arch digital model is needed for pros-
thodontic use, an indirect digitalization with the desk-
top scanner R900 showed the most accurate results.
The intraoral scanner TRIOS Color Pod showed clinically
acceptable results if the scanning scope is less than a
quadrant and three additional teeth. Indirect digitaliza-
tion with a CBCT unit does not appear to be accurate
enough for prosthodontic demands.
REFERENCES
1. Rheude B, Sadowsky PL, Ferriera A, Jacobson A. An evaluation of the use of
digital study models in orthodontic diagnosis and treatment planning. Angle
Orthod 2005;75:300–304.
2. Vandeweghe S, Vervack V, Dierens M, De Bruyn H. Accuracy of digital impres-
sions of multiple dental implants: an in vitro study (Epub ahead of print, 6 May
2016). Clin Oral Implants Res doi: 10.1111/clr.12853.
3. Grunheid T, McCarthy SD, Larson BE. Clinical use of a direct chairside oral
scanner: an assessment of accuracy, time, and patient acceptance. Am J
Orthod Dentofacial Orthop 2014;146:673–682.
4. Patzelt SB, Emmanouilidi A, Stampf S, Strub JR, Att W. Accuracy of full-arch
scans using intraoral scanners. Clin Oral Investig 2014;18:1687–1694.
5. Hayashi K, Sachdeva AU, Saitoh S, Lee SP, Kubota T, Mizoguchi I. Assessment
of the accuracy and reliability of new 3-dimensional scanning devices. Am J
Orthod Dentofacial Orthop 2013;144:619–625.
6. Lightheart KG, English JD, Kau CH, et al. Surface analysis of study models
generated from OrthoCAD and cone-beam computed tomography imaging.
Am J Orthod Dentofacial Orthop 2012;141:686–693.
7. Tarazona B, Llamas JM, Cibrian R, Gandia JL, Paredes V. A comparison
between dental measurements taken from CBCT models and those taken
from a digital method. Eur J Orthod 2013;35:1–6.
8. Ender A, Mehl A. Full arch scans: conventional versus digital impressions: an
in-vitro study. Int J Comput Dent 2011;14:11–21.
9. Akyalcin S, Dyer DJ, English JD, Sar C. Comparison of 3-dimensional dental
models from different sources: diagnostic accuracy and surface registration
analysis. Am J Orthod Dentofacial Orthop 2013;144:831–837.
10. Ender A, Mehl A. In-vitro evaluation of the accuracy of conventional and dig-
ital methods of obtaining full-arch dental impressions. Quintessence Int
2015;46:9–17.
49
Q U I N T E S S E N C E I N T E R N AT I O N A L
Wesemann et al
11. Su TS, Sun J. Comparison of repeatability between intraoral digital scanner
and extraoral digital scanner: an in-vitro study. J Prosthodont Res 2015;59:
236–242.
12. Güth JF, Runkel C, Beuer F, Stimmelmayr M, Edelhoff D, Keul C. Accuracy of
five intraoral scanners compared to indirect digitalization (Epub ahead of
print, 12 Jul 2016). Clin Oral Investig.
13. Ender A, Mehl A. Accuracy of complete-arch dental impressions: a new meth-
od of measuring trueness and precision. J Prosthet Dent 2013;109:121–128.
14. Flugge TV, Schlager S, Nelson K, Nahles S, Metzger MC. Precision of intraoral
digital dental impressions with iTero and extraoral digitization with the iTero
and a model scanner. Am J Orthod Dentofacial Orthop 2013;144:471–478.
15. van der Meer WJ, Andriessen FS, Wismeijer D, Ren Y. Application of intra-oral
dental scanners in the digital workflow of implantology. PLoS One
2012;7:e43312.
16. Bishara SE, Jakobsen JR, Treder J, Nowak A. Arch width changes from 6 weeks
to 45 years of age. Am J Orthod Dentofacial Orthop 1997;111:401–409.
17. Müller P, Ender A, Joda T, Katsoulis J. Impact of digital intraoral scan strategies
on the impression accuracy using the TRIOS Pod scanner. Quintessence Int
2016;47:343–349.
18. Joint Committee for Guides in Metrology. International vocabulary of metrol-
ogy: Basic and general concepts and associated terms (VIM), 3rd edition.
JGCM, 2012:37–39.
50
19. Baumgaertel S, Palomo JM, Palomo L, Hans MG. Reliability and accuracy of
cone-beam computed tomography dental measurements. Am J Orthod
Dentofacial Orthop 2009;136:19–25.
20. Lascala CA, Panella J, Marques MM. Analysis of the accuracy of linear measure-
ments obtained by cone beam computed tomography (CBCT-NewTom).
Dentomaxillofac Radiol 2004;33:291–294.
21. Spoor CF, Zonneveld FW, Macho GA. Linear measurements of cortical bone
and dental enamel by computed tomography: applications and problems.
Am J Phys Anthropol 1993;91:469–484.
22. Tarazona-Alvarez P, Romero-Millan J, Penarrocha-Oltra D, Fuster-Torres MA,
Tarazona B, Penarrocha-Diago M. Comparative study of mandibular linear
measurements obtained by cone beam computed tomography and digital
calipers. J Clin Exp Dent 2014;6:e271–e274.
23. Casko JS, Vaden JL, Kokich VG, et al. Objective grading system for dental casts
and panoramic radiographs. American Board of Orthodontics. Am J Orthod
Dentofacial Orthop 1998;114:589–599.
24. Jang KS, Kim YS. Comparison of oral sensory function in complete denture
and implant-supported prosthesis wearers. J Oral Rehabil 2001;28:220–225.
25. Radlanski R-J. Curriculum Orale Struktur- und Entwicklungsbiologie. Berlin:
Quintessenz, 2011.
26. Wichelhaus A. Kieferorthopädie- Therapie Band 1 (Grundlegende Behand-
lungskonzepte). Stuttgart: Thieme, 2012.
VOLUME 48 • NUMBER 1 • JANUARY 2017