Journal of Orthodontics, Vol.

39, 2012, 151–159

SCIENTIFIC Digital versus plaster study models:

SECTION
How accurate and reproducible are
they?
Neilufar Abizadeh1,2, David R. Moles1, Julian O’Neill2 and Joseph H. Noar1
1
Eastman Dental Institute and Kettering General Hospital, UK; 2Kettering General Hospital, UK

Objective: To compare measurements of occlusal relationships and arch dimensions taken from digital study models with those
taken from plaster models.
Design: Laboratory study
Setting: The Orthodontic Department, Kettering General Hospital, Kettering, UK
Methods and materials: One hundred and twelve sets of study models with a range of malocclusions and various degrees of
crowding were selected. Occlusal features were measured manually with digital callipers on the plaster models. The same
measurements were performed on digital images of the study models. Each method was carried out twice in order to check for
intra-operator variability. The repeatability and reproducibility of the methods was assessed.
Results: Statistically significant differences between the two methods were found. In 8 of the 16 occlusal features measured, the
plaster measurements were more repeatable. However, those differences were not of sufficient magnitude to have clinical
relevance. In addition there were statistically significant systematic differences for 12 of the 16 occlusal features, with the
plaster measurements being greater for 11 of these, indicating the digital model scans were not a true 1 : 1 representation of the
plaster models.
Conclusions: The repeatability of digital models compared with plaster models is satisfactory for clinical applications, although
this study demonstrated some systematic differences. Digital study models can therefore be considered for use as an adjunct to
clinical assessment of the occlusion, but as yet may not supersede current methods for scientific purposes.
Key words: Digital study models, orthodontic study models, reliability, accuracy, reproducibility

Received 26 February 2010; accepted 13 March 2012

Introduction holography, stereophotogrammetry, photography, digi-

Plaster study models form an essential part of the
orthodontic patient record. They act as a diagnostic tool
for the clinician as well as providing a progressive record
of treatment. Their use is widespread but is associated
with several problems, mainly storage, breakage and
loss. For medico-legal purposes, the British Dental
Association suggest that records should be kept for a
minimum of 11 years after completion of treatment.1
Hence a systematic storage system for all patient models
needs to be put into place, demanding storage facilities.
McGuinness and Stephens2 in 1992 identified a major
problem with the storage of study models in hospital
units throughout the United Kingdom.
These problems have encouraged investigation of
alternative storage methods. These include, photocopying,
tized study models and more recently, cone beam CT
scanning. The majority of these techniques have proven to
be less successful than plaster study models and have their
own individual limitations. Hence, plaster models remain
the method of choice for many scientific and clinical
applications.
Digitized study models were introduced in the 1990s3,4
following advances in digital technology. Continuing
development of hardware and software has reduced
the costs of the process,4 both in terms of the scanning
equipment and software required. Advances in both
scanning and CAD technology has allowed virtual
models to be produced by numerous companies across
the world.3 Digital study models are produced by a laser
scan or a combination of laser scanning and stereo-
photogrammetry, to produce a three-dimensional digital

Address for correspondence: N. Abizadeh, Eastman Dental

Institute and Kettering General Hospital, London, UK.
Email: neilufar.abizadeh@gmail.com
# 2012 British Orthodontic Society DOI 10.1179/1465312512Z.00000000023
 152 Abizadeh et al. Scientific Section JO September 2012

scan of a study model, which can be rotated in any plane

through 360 degrees. Current applications allow linear
measurements, treatment planning,5,6 PAR scoring,7,8
Bolton analysis9 and space analyses to be carried out.
Digital models have been shown to be as valid as plaster
casts10–13 and appear to have adequate reliability to be
used in a clinical setting for most purposes.13 However,
it is not clear whether they have sufficient accuracy to be
applied to scientific research.
ESM Digital Solutions (3-ShapeTM, Copenhagen,
Denmark),14 are a company based in Ireland who
produce digital study models based on laser scanning of
plaster models. Their system is yet to be validated. The
aim of the study was to compare measurements of
occlusal relationships and arch dimensions taken from
ESM digital study models with those taken from plaster
models. The specific null hypotheses tested were that (1)
there were no differences in the random error and
Figure 1 The inside of the laser scanner with a plaster model in
spread of measurements taken from the digital models position
compared with those taken from the plaster models and
(2) there were no systematic differences in the measure-
ments taken from the digital models compared with following categories based on the level of crowding in
those taken from the plaster models. at least one arch: mild or none (0–3 mm); moderate (3–
6 mm); severe (.6 mm). Equal numbers of models in
the three incisor and crowding classifications were
Materials and methods selected. The models were placed in a random order
Ethical approval to undertake the study was gained and anonymized using a number code.
from Nottingham Research Ethics Committee (Reference
number: 08/H0403/94, 17/09/2008). Study models were Digital study models
selected from the archives of the Orthodontic Depart-
Models were scanned using a R250 Scanner (3-ShapeTM,
ment, Kettering General Hospital, Kettering, UK from a Copenhagen, Denmark). The R250 scanner consists of
collection scheduled for destruction, as they had been two high resolution charged coupled device cameras,
taken from patients who had completed treatment over one laser projector and one articulating table (Figure 1).
11 years ago. Selection was based upon the following During this process, the model is placed in the scanner
inclusion critera: (1) complete adult dentition from first and articulated through various orientations as the two
molar to first molar in both dental arches; (2) crowding high resolution charged coupled device cameras capture
present in at least one dental arch. Models presenting multiple images of the points projected onto the model
with the following features were excluded: (1) patients surface by the laser. Images were saved as JPEG files
with hypodontia or erupted supernumerary teeth; (2) and imported into the specialized software provided by
carious, heavily restored or hypoplastic teeth, with ESM Digital Solutions, on a desktop computer. These
altered anatomy; (2) cleft palate; (4) models with voids, digital models were a three-dimensional image of the
fractured teeth or any other damage. plaster models that could be rotated in any direction to
To ensure that a full range of occlusal anomalies was get a full view of the model from any angle. Measure-
measured, the models were categorized according to ments were undertaken to a precision of 0.01 mm
incisor relationship and amount of crowding. The (Figure 2). The software included a magnifying func-
incisor relationship was classified according to the tion, which was used to zoom in on parts of the digital
British Standards Institute classification of malocclusion.15 image for ease of landmark identification.
The amount of crowding was determined by measuring
the contact point displacements from the mesial aspect Plaster study models
of the first molar on one side of the arch to the mesial
aspect of the contra-lateral first molar, using a stainless Manual measurements were carried out on the plaster
steel ruler. The models were then placed in the models using digital dial callipers (Mitutoyo Absolute
 JO September 2012 Scientific Section A comparison of digital vs plaster models 153

standard deviations or variances) rather than a differ-

Figure 2 Measurement of intercanine and intermolar widths
using the ESM digital solutions software
Digimatic, Mitutoyo UK, Andover, UK) to a precision
of 0.01 mm (Figure 3).
The measurements taken are shown in Table 1. The
measurements were taken by a single operator (NA) to
eliminate inter-operator variability. Ten models were
measured at a time to minimize operator fatigue. To
determine intra-operator error all the measurements
were repeated on the existing digital models and plaster
models after one month.
Statistical analysis
In the absence of a clear indication of what would
constitute a clinically relevant difference in ‘error’
measurements, i.e. what would be a relevant difference
in the relative spread of observations (measured by
ence in means, it was not appropriate to attempt a
formal a priori sample size calculation. The sample
was therefore a convenience sample based on similar
published research undertaken within the department.16
Data were entered onto an ExcelH spreadsheet (v12,
2007; Microsoft Corp, USA) and statistical analysis was
undertaken using Stata package (Version 10, StataCorp
LP, USA) for statistical analysis.
Intra-technique repeatability (i.e. measurements of
consistency of duplicate measurements ‘within’ techni-
ques) was assessed by calculating the coefficients of
repeatability for each technique separately and by using
variance ratio tests. For reproducibility (i.e. inter-
technique comparisons ‘between’ digital and plaster
measurements) the mean differences were calculated and
paired t-tests were carried out to look for the presence
of systematic errors (bias) between the techniques.
To determine random error and agreement between
the techniques, standard deviations of the differences
between the techniques and Bland and Altman plots
were examined. The plaster models were taken to be the
gold standard, so if a difference was found it was
considered to be due to the digital technique. Statistical
significance was set at P,0.05.
Results
The study sample comprised of 112 sets of plaster
models and 111 sets of digital models. One of the digital

Table 1 Measurements taken from plaster and digital models.

Upper arch length and lower arch length:
Upper and lower intermolar width:
Upper and lower intercanine width:
Overjet:
Overbite:
Anterior open bite:
Centre line discrepancy:
Crown height of upper left central incisor:
Crown height of lower right central incisor:
Crown height of upper left first premolar:
Crown height of lower right first premolar:
Crown height of upper left first molar:
Crown height of lower right first molar:
Total of the greatest mesio-distal width recorded for each tooth from first molar to
first molar, viewed directly above the occlusal surface.
Measured from the tip of the mesiobuccal cusp of the right first molar to the tip of the
mesiobuccal cusp of the left first molar as viewed directly above the occlusal surface.
Measured from the tip of the right cusp to the tip of the left cusp also measured occlusally.
Measured from the labial surface of the mandibular incisor to the labial surface of the
maxillary incisor. Where the labial inclinations of the maxillary incisors differed, the
maximum overjet was recorded.
The maximum amount of vertical overlap of either mandibular central incisor by its
corresponding maxillary central incisor.
The maximum amount of vertical distance between the maxillary incisal tip and mandibular
incisal tip.
The horizontal distance between upper and lower centre-lines.
Measured from the midpoint of the cervical margin of the tooth to the incisal edge.
Measured from the midpoint of the cervical margin of the tooth to the incisal edge.
Measuring the greatest crown height from the midpoint of the cervical margin to the
bucco-occlusal tip.
Measuring the greatest crown height from the midpoint of the cervical margin to the
bucco-occlusal tip.
Measuring the greatest crown height from the cervical margin to the bucco-occlusal tip.
Measuring the greatest crown height from the cervical margin to the bucco-occlusal tip.
 154 Abizadeh et al. Scientific Section JO September 2012

model files (a Class 2 case) was found to be corrupted
and could not be opened for assessment. The distribu-
tion of incisor relationships was 38 Class 1, 38 Class 2
(both Division I and II) and 36 Class 3. Within the Class
1 and Class 2 groups there were 13 mildly crowded
cases, 13 moderately crowded cases and 12 severely
crowded cases. In the Class 3 group there was an equal
number of mild, moderate and severely crowded cases.
Intra-technique repeatability
of the differences (between the first and second readings)
obtained from the plaster study models were signifi-
cantly lower for eight measurements. Whereas the
coefficients of repeatability and the standard deviation
of the differences obtained from the digital study
models were significantly lower for four measurements.
Although statistically significant differences between the
coefficients of repeatability for the two methods were
found, these are unlikely to be clinically significant as
the magnitude of the differences was small.

Table 2 shows the repeatability for the two methods. The
coefficients of repeatability for the first and second readings
from the plaster models ranged from 0.45 mm for upper
left central incisor vertical height to 2.96 mm for upper
arch length measurement. The coefficients of repeatability
for the first and second readings from the digital models
ranged from 0.66 mm for upper inter-canine width to
2.96 mm for upper arch length measurements.
The results of the variance ratio tests to examine
differences in the coefficients of repeatability between
the two techniques are also shown in Table 2. There
was no statistically significant difference in the repeat-
ability of the two methods for four measurements. The
coefficients of repeatability and the standard deviation
Inter-technique reproducibility – systematic error
Table 3 shows the mean differences between the measure-
ments obtained of the same models using the two
techniques and a paired t-test examining the differences.
The plaster method produced significantly higher readings
for 11 measurements. The digital technique produced
significantly higher readings for only one measurement
and there were no significant differences between the two
techniques for four measurements. The significant sys-
tematic differences ranged from 0.14 mm for upper inter-
canine width to 1.58 mm for the lower right 6 vertical
height measurement. The differences were generally small
and probably not of clinical significance.

Table 2 Intra-technique repeatability – the standard deviation of the differences and coefficient of repeatability between the first and second
readings of each technique and comparisons between techniques using the variance ratio test.

PLASTER random error DIGITAL random error

Coefficient of Coefficient of Variance ratio

Parameter SD of differences repeatability SD of differences repeatability test P-value

No differences
U Arch length 1.45 2.84 1.51 2.96 0.673
Anterior openbite 0.92 1.80 0.59 1.16 0.084
U Intercanine width 0.38 0.37 0.34 0.66 0.177
L Intercanine width 0.39 0.76 0.42 0.82 0.465
Plaster better than digital
L Arch length 1.12 2.20 1.49 2.92 0.003
Overjet 0.40 0.78 0.51 1.00 0.010
U Intermolar width 0.46 0.89 0.61 1.19 0.003
UL1 Vertical height 0.23 0.45 0.47 0.93 ,0.001
UL4 Vertical height 0.29 0.58 0.45 0.89 ,0.001
LR4 Vertical height 0.34 0.66 0.76 1.49 ,0.001
UL6 Vertical height 0.27 0.53 0.35 0.69 0.005
LR6 Vertical height 0.43 0.84 0.86 1.69 ,0.001
Digital better than plaster
Overbite 0.58 1.14 0.42 0.82 0.002
L Intermolar width 0.91 1.78 0.60 1.18 ,0.001
Centreline discrepancy 0.86 1.69 0.36 0.70 ,0.001
LR1 Vertical height 0.37 0.72 0.30 0.58 0.023
 JO September 2012 Scientific Section A comparison of digital vs plaster models 155

Table 3 Inter-technique reproducibility – systematic error (mean differences between the readings of the two techniques and paired t-test) and
random error (SD of differences) and limits of agreement (Bland and Altman17).

Systematic error Random error Limits of agreement (systematic and random error)

Mean differences P value from SD of the

Parameter (plaster minus digital) paired t-test differences Lower limit (mm) Upper limit (mm)

No difference in
systematic error
between techniques
Overjet 20.01 0.872 0.55 21.09 1.07
Anterior openbite 0.29 0.322 0.93 21.53 2.11
L Intermolar width 0.07 0.495 1.04 21.97 2.10
Centreline discrepancy 0.11 0.196 0.90 21.65 1.87
Plaster greater than
digital measurement
U Archlength 1.15 ,0.001 1.70 22.18 4.49
L Archlength 0.50 0.004 1.76 22.95 3.94
Overbite 0.67 ,0.001 0.71 20.73 2.07
U Intermolar width 0.15 0.014 0.61 21.05 1.34
U Intercanine width 0.14 ,0.001 0.37 20.59 0.86
UL1 Vertical height 0.53 ,0.001 0.37 20.20 1.26
LR1 Vertical height 0.47 ,0.001 0.54 20.60 1.54
UL4 Vertical height 0.67 ,0.001 0.50 20.31 1.65
LR4 Vertical height 1.09 ,0.001 0.72 20.32 2.50
UL6 Vertical height 0.75 ,0.001 0.42 20.07 1.56
LR6 Vertical height 1.58 ,0.001 0.82 20.02 3.18
Digital greater than
plaster
measurement
L Intercanine width 20.17 ,0.001 0.50 21.15 0.82

Inter-technique reproducibility – random error and
agreement
Table 3 shows the standard deviation of the differences
between the readings obtained by the two techniques, as
well as the limits of agreement. These are also shown
graphically in Figures 4–7. The Bland and Altman plots
for the upper (Figure 4) and lower arch (Figure 5)
length measurements show a wide range for the 95%
limits of agreement (22.18 to 4.49 mm for upper arch
length measurement and 22.95 to 3.94 mm for lower
arch length measurement).
There was evidence of a systematic error, with the
measurements from the plaster models greater that those
from the digital models. The Bland and Altman plots for
upper inter-molar width (Figure 6) and upper inter-
canine width (Figure 7), again showed a positive shift,
indicating the measurements from the plaster models
were consistently greater than those from the digital
models (21.05 to 1.34 mm for upper inter-molar width
measurement and 20.59 to 0.86 mm for upper inter-
canine width measurements). The Bland and Altman
plot for the lower inter-molar width measurement
(Figure 8) shows no systematic error between techni-
ques, whereas the plot for the lower inter-canine width
measurement (Figure 9) demonstrated a negative shift
of the graph, indicating the digital measurements were
greater than the plaster measurements.
Discussion
This study was carried out to assess the reliability of
digital study models produced by ESM Digital Solutions
from plaster study models. There are other studies
assessing alternative systems available worldwide, but as
this company is the first in the British Isles, and with
increasing interest in digital study models by orthodon-
tic practices and hospital departments; it was felt
necessary to perform the task of validating this system
independently.
The use of plaster models and digital callipers for
study model analysis is considered the gold standard for
taking measurements.8,10,11,18 Therefore, it was expected
that the plaster method would have better repeatability
 156 Abizadeh et al.
Figure 3 Measurement of intercanine width using digital callipers
for many of the parameters. The difference in the
coefficients of repeatability between the two methods
did not exceed 1 mm for any of the parameters and both
methods may be considered to be repeatable with
Figure 4 Bland and Altman plot for upper arch length
measurement (plaster and digital methods)
Scientific Section JO September 2012
Figure 5 Bland and Altman plot for lower arch length
measurement (plaster and digital methods)
acceptable differences between the two techniques. The
results of studies by Zilberman et al.,10 Quimby et al.,12
Dalstra and Melsen13 and Keating et al.19 also found
good reliability between the two methods. The coeffi-
cients of repeatability were relatively small and these
differences would not be considered clinically relevant.
The upper and lower arch length measurements both
showed significant bias/systematic error. The limits of
agreement were relatively wide, which is similar to
results from Redlich et al.20 who found up to 3 mm of
discrepancy between calliper measurements and linear
measurements taken using the digital method.
It is possible that plaster models are not a true
representation of the actual tooth measurements due to
dimensional changes in the impression materials and
stone during preparation; however, impression materials
have been validated over many years. The fact that
Figure 6 Bland and Altman plot for upper intermolar width
measurement (plaster and digital methods)
 JO September 2012
Figure 7 Bland and Altman plot for upper intercanine width
measurement (plaster and digital methods)
plaster models have been used to construct appliances,
which fit accurately in the mouth, is the ultimate
validation of accuracy and it is reasonable to consider
them as a gold standard. Indeed, Dalstra and Melsen13
compared measurements taken from digitized stone
models derived from alginate impressions poured
immediately or after 3–5 days and found no statistically
significant differences in measurements.
The systematic error indicated that the measurements
taken from the digital models were generally smaller
than those taken from the plaster models. As the same
models were used for scanning and plaster model
measurements, this indicates that the digital model is
not a true one-to-one representation and shrinkage of
digital models has occurred during the scanning process.
Errors occur during computer processing of the
gathered data, which aims to retain detail, but remove
Figure 8 Bland and Altman plot for lower intermolar width
measurement (plaster and digital methods)
Scientific Section A comparison of digital vs plaster models 157
Figure 9 Bland and Altman plot for lower intercanine width
measurement (plaster and digital methods)
artefacts. They also occur during the merger of all scans
to produce a single composite image.19,21 This study had
similar results and findings to other studies assessing
digital study models. Santoro et al.11 also found that
digital measurements were consistently smaller com-
pared to plaster measurements using the OrthoCADTM
system and it was suggested that this was due to alginate
shrinkage during postage of the impressions to
OrthoCADTM. They concluded that since the shrinkage
was uniform the diagnostic capacity of the software is
not affected, especially when comparing proportional
measurements. Mullen et al.9 also found differences in
the amount of tooth structure measured in the maxillary
and mandibular arches. As in the current study, they
found measurements to be smaller when measured on
the digital system compared with the plaster method.
The measurements carried out in this study have not
taken into consideration any errors due to the capturing
process of the digital image, as the images were only
scanned once and then re-measured for the purpose of
this study. This indicates that the digital errors could
potentially be greater. In hindsight, digitizing a selection
of the sample twice to determine the accuracy of the
capture technique could have been carried out.
The strengths of this study include the large sample
size, considerably greater than other published studies,18
with a range of different malocclusions, including dif-
ferent amounts of crowding. Hence it represented a
broad population, which is encountered in orthodontic
practice. The main source of error in the study was
almost certainly due to landmark identification. It was
difficult to select the same points for measurement on
the model each time. This was noticeable in both upper
and lower arch measurements. This has been found in
other studies9,10,20 and will also affect segmental arch
 158 Abizadeh et al. Scientific Section
length measurement.12 Houston22 stated that accuracy
of landmark selection from cephalometric radiographs
contributes greatly to random error. Repeated landmark
location using either method can differ and the
operator’s opinion of the precise location of a point
may vary at random. This directly affects the repeat-
ability and has been reported by other researchers.9,18,20
Inaccuracies with the digital technique could also be
partly attributed to difficulty in locating the greatest
mesio-distal width of the tooth for arch length assessment.18
In addition, the contact points using digitized systems
are less well defined and this is exacerbated by crowding.9,18
Despite the rotation functions and magnification in the
software, accurate point location remains difficult.9,10,20
Parameters such as inter-canine and inter-molar widths,
where landmark identification was easier, showed good
repeatability and reproducibility when considering both
techniques. This was also found in studies carried out by
Asquith et al.18 and Zilberman et al.10 In contrast, overbite
measurement using the digital system was more repeatable
than the plaster model measurement, probably due to the
use of the magnification tool on the software and the small
cursor point being easier to use in comparison with the
bulkier calliper ends. The angle of the calliper ends placed
on the model and also the plane in which the plaster model
is viewed can contribute to inaccuracies in overbite
measurement.11 This problem does not occur when
using the digital measuring tools. The issue of land-
mark registration could have been overcome if the
models had been marked with two distinct points for an
easier linear measurement. This was carried out by
Asquith et al.18 By providing such landmarks for
measurements errors due to landmark location might
have been reduced, however in this study we were
interested in the clinical use of this system and therefore
the potential random errors have been taken into
consideration.
One of the main uses of study models is for treatment
planning, and whilst this was not investigated formally
in this study, there was a perception that the digital
method of measurement was easier to use than plaster,
which is in agreement with Stevens et al.8 The digital
measurements were rapidly performed9,23 and also had
the advantage of automatic calculation of the Bolton
ratio, tooth width analysis, and space analysis (using
the Tanaka and Johnston and Moyers techniques) if
required. Resultant calculations such as arch length and
overjet were stored automatically on the system for
future reference. There are also the obvious benefits of
ease of storage and retrieval associated with digital
models. A busy hospital orthodontic department or
practice would gain from having easy storage and
recollection of digital models, which would never be
JO September 2012
physically damaged or lost. The issue of storage with
digital study models is of a different nature. Regular
backing up of digital information is required to ensure
information is not lost. Records of a digitized manner
are easier to organize, file, locate and search allowing
flexibility. These factors allow archives to be kept
indefinitely and would become ideal for audit, peer
review and research.
The main problem using digital systems is adjusting
to viewing a three dimensional object in two dimen-
sions. This does require some orientation. In addition,
cross bites are difficult to visualize and rotation of the
model is required fully to comprehend the magnitude
of the cross bite. This problem was also reported
by Stevens et al.8 Until the user is familiar with the
system, location of points and planes can be difficult.18
Therefore a period of practice is advised. Another issue
is concerned with data protection of electronic files. As
with all other patient records, access to digital study
models should be password protected and for added
security transfer of scanned models should take place
with a numbered reference only so the model remains
anonymous.
In summary, measurements taken on a digitized study
model using the ESM Digital Solutions system did show
some significant differences from the manually mea-
sured plaster models. However, many of the differences
were quite small and it seems that either method would
be appropriate for clinical use for the majority of
parameters. Although there were significant systematic
and random errors between the two methods, these are
clinically irrelevant and the advantages digital models
have in a clinical scenario, makes them a suitable
replacement for plaster models. As long as they are used
alongside clinical information and other diagnostic
tools, digital models can prove to be clinically very
useful.
Conclusions
N There were significant differences between the two
methods, with the plaster measurements appearing
more repeatable for 8 of the 16 parameters. How-
ever, these differences would not all be of clinical
relevance.
N There were significant systematic errors for 12 of the
16 parameters, with the digital measurements being
significantly less for 11 of these.
N Overall: repeatability of digital models compared with
plaster models appears reasonable and therefore,
digital models could be considered for use as an
adjunct to clinical findings.
 JO September 2012 Scientific Section
N Digital study models have a role for clinical applica-
tions although cannot yet replace other current
methods for scientific research.
Contributors
Neilufar Abizadeh was responsible for study design;
sample and data collection; analysis and drafting of the
paper and final approval of the article. Professor David
Moles was responsible for logistic, administrative, and
technical support and data interpretation; critical revision,
and final approval of the article. Mr Julian O’Neill and Mr
Joseph Noar were responsible for study design, analysis
and drafting of paper, critical revision, and final approval
of the article. Neilufar Abizadeh is the guarantor.
References
1. British Dental Association. Advice sheet B1: Ethics in
dentistry. British Dental Association, 2000.
2. McGuinness NJ, Stephens CD. Storage of orthodontic
study models in hospital units in the U.K. Br J Orthod 1992:
19; 227–32.
3. Joffe L. OrthoCAD: digital models for a digital era. J
Orthod 2004; 31: 344–47.
4. Mah J. The evolution of digital study models. J Clin Orthod
2007; 41: 557–61.
5. 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:
300–04.
6. Whetten JL, Williamson PC, Heo G, Varnhagen C, Major
PW. Variations in orthodontic treatment planning decisions
of Class II patients between virtual 3-dimensional models
and traditional plaster study models. Am J Orthod
Dentofacial Orthop 2006; 130: 485–91.
7. Mayers M, Firestone AR, Rashid R, Vig KW. Comparison
of peer assessment rating (PAR) index scores of plaster and
computer-based digital models. Am J Orthod Dentofacial
Orthop 2005; 128: 431–34.
8. Stevens DR, Flores-Mir C, Nebbe B, Raboud DW, Heo G,
Major PW. Validity, reliability and reproducibility of plaster
vs. digital study models: comparison of peer assessment
rating and Bolton analysis and their constituent measure-
ments. Am J Orthod Dentofacial Orthop 2006; 129: 794–803.
A comparison of digital vs plaster models 159
9. Mullen SR, Martin CA, Ngan P, Gladwin M. Accuracy of
space analysis with emodels and plaster models. Am J
Orthod Dentofacial Orthop 2007; 132: 346–52.
10. Zilberman O, Huggare JA, Parikakis KA. Evaluation of the
validity of tooth size and arch width measurements using
conventional and three-dimensional virtual orthodontic
models. Angle Orthod 2003; 73: 301–06.
11. Santoro M, Galkin S, Teredesai M, Nicolay OF, Cangialosi
TJ. Comparison of measurements made on digital and
plaster models. Am J Orthod Dentofacial Orthop 2003; 124:
101–05.
12. Quimby ML, Vig KWL, Rashid RG, Firestone AR. The
accuracy and reliability of measurements made on compu-
ter-based digital models. Angle Orthod 2004; 74: 298–303.
13. Dalstra M, Melsen B. From alginate impressions to digital
virtual models: accuracy and reproducibility. J Orthod 2009;
39: 36–39.
14. ESM Digital Solutions. Available at: www.esmdigitalsolu-
tions.com. Accessed on 01 January 2008.
15. British Standards Institution. British Standard Incisor
Classification. British Standards Institute 1983. Glossary
of Dental terms BS4492.
16. Tan SSW, Ahmed S, Moles DR, Cunningham SJ. Picture
archiving and communications systems: a study of the
reliability of orthodontic cephalometric analysis. Eur J
Orthod 2011; 33: 537–43.
17. Bland JM, Altman DG. Statistical methods for assessing
agreement between two methods of clinical measurement.
Lancet 1986; 1: 307–10.
18. Asquith J, Gillgrass T, Mossey P. Three-dimensional
imaging of orthodontic models: a pilot study. Eur J
Orthod 2007; 29: 517–22.
19. Keating AP, Knox J, Bibb R, Zhorov AI. A comparison of
plaster, digital and reconstructed study model accuracy.
J Orthod 2008; 35: 191–201.
20. Redlich M, Weinstock T, Abed Y, Schneor R, Holdstein Y,
Fischer A. A new system for scanning, measuring and
analyzing dental casts based on a 3D holographic sensor.
Ortho Craniofac Res 2008; 11: 90–95.
21. Mah J, Hatcher D. Current status and future needs in
craniofacial imaging. Orthod Craniofac Res 2003; 6: 10–16.
22. Houston WJB. The analysis of errors in orthodontic
measurements. Am J Orthod 1983; 83: 382–90.
23. Gracco A, Buranello M, Cozzani M, Siciliani G. Digital
and plaster models: a comparison of measurements and
times. Prog Orthod 2007; 8: 252–59.