European Journal of Orthodontics 33 (2011) 344–349 © The Author 2010.
2010. Published by Oxford University Press on behalf of the European Orthodontic Society.
doi:10.1093/ejo/cjq102 All rights reserved. For permissions, please email: [email protected]
Advance Access Publication 3 December 2010
Two- versus three-dimensional imaging in subjects with
unerupted maxillary canines
Susanna Botticelli, Carlalberta Verna, Paolo M. Cattaneo, Jens Heidmann and
Birte Melsen
Department of Orthodontics, University of Aarhus, Denmark.
Correspondence to: Susanna Botticelli, Department of Orthodontics, School of Dentistry, Faculty of Health
Sciences, University of Aarhus, Vennelyst Boulevard 9, 8000 Aarhus C, Denmark. E-mail: [email protected]
SUMMARY The aim of this study was to evaluate whether there is any difference in the diagnostic
information provided by conventional two-dimensional (2D) images or by three-dimensional (3D) cone
beam computed tomography (CBCT) in subjects with unerupted maxillary canines.
Twenty-seven patients (17 females and 10 males, mean age 11.8 years) undergoing orthodontic treatment
with 39 impacted or retained maxillary canines were included. For each canine, two different digital image
sets were obtained: (1) A 2D image set including a panoramic radiograph, a lateral cephalogram, and the
available periapical radiographs with different projections and (2) A 3D image set obtained with CBCT. Both
sets of images were submitted, in a single-blind randomized order, to eight dentists. A questionnaire was
used to assess the position of the canine, the presence of root resorption, the difficulty of the case, treatment
choice options, and the quality of the images. Data analysis was performed using the McNemar–Bowker test
for paired data, Kappa statistics, and paired t-tests.
Downloaded from by guest on May 30, 2015
The findings demonstrated a difference in the localization of the impacted canines between the
two techniques, which can be explained by factors affecting the conventional 2D radiographs such as
distortion, magnification, and superimposition of anatomical structures situated in different planes of
space. The increased precision in the localization of the canines and the improved estimation of the space
conditions in the arch obtained with CBCT resulted in a difference in diagnosis and treatment planning
towards a more clinically orientated approach.
Introduction
of the unerupted canine with neighbouring teeth and adjacent
In subjects with delayed eruption of maxillary canines, the anatomical structures (Gratt, 1994). DPTs are affected by a
diagnostic procedure can, in the majority of cases, be limited certain degree of distortion in the horizontal plane (Mckee
to clinical inspection and palpation of the alveolar process, et al., 2001, 2002; Yeo et al., 2002). Periapicals provide
considering also occlusal development and somatic maturity. improved visualization of the interdental relationship, but
However, according to Ericson and Kurol (1986 a,b), 8–10 information in the bucco-lingual direction can only be obtained
per cent of patients older than 10 years also require a by combining different projections. DPT and periapicals may
radiographic examination to obtain sufficient information be integrated with other two-dimensional (2D) images: occlusal
regarding the position and angulation of the canine and its radiographs, lateral, and postero-anterior head films. As images
relationship with the neighbouring structures. routinely used to identify impacted teeth all suffer from errors
A study describing the routines of orthodontists and oral implicit in the technique, the application of medical computed
surgeons revealed that 78 per cent used more than two, and tomographic (CT) scanning has been suggested. This method
23 per cent four or more radiographs (Southall and Gravely, allows for a more precise visualization of the relationship with
1989) to describe the position of an ectopic canine and for the neighbouring teeth, but its use has been limited to specific
treatment planning. Clinicians tend to choose the techniques indications since the cost and the irradiation dose are high
that they are most familiar with (Bishara, 1992). Not (Ericson and Kurol, 1987, 1988; Elefteriadis and Athanasiou,
considering the cost–benefit ratio between information 1996; Ericson and Bjerklin, 2001; Ericson et al., 2002).
provided and radiation exposure might lead to an unjustified With the introduction of cone beam computed tomography
radiation dose to the patient. (CBCT; Mozzo et al., 1998), the advantages of the CT scan
Localization of impacted canines can be challenging could be obtained with reduced scanning time, lower irradiation
with conventional radiographic methods, due to distortion, dose, and cost. CBCT utilizes a 2D, or panel detector, which
superimposition of three-dimensional (3D) structures, and allows generation of a 3D data set of the head of the patient
imaging artefacts. Conventional radiographs such as dental using a single rotation of the X-ray source and detector. This
pantomographs (DPT) and periapicals (Clark, 1909) provide technique differs from fan-beam CT scanners, where multiple
information regarding the vertical and mesio-distal relationship ‘slices’ are stacked to obtain a complete 3D image. The value
3D IMAGING OF UNERUPTED CANINES 345
of CBCT in dento-maxillofacial imaging has been recognized
by a number of researchers (Mah and Hatcher, 2003; Ludlow
et al., 2006) and different scanners have been introduced.
The aim of this study was to assess the difference in the
diagnosis and treatment approach to unerupted maxillary
canines based on conventional 2D images and 3D CBCT
image sets.
Subjects and methods
Twenty-seven patients (17 females and 10 males, mean age
11.8 years) undergoing orthodontic treatment at the
Department of Orthodontics, School of Dentistry, University
of Aarhus, Denmark, were included in this study. In total, 39
ectopic canines were examined. The subjects were selected
from patients already enrolled for an additional CBCT
examination because of ectopic position of the tooth or
because of other indications for 3D evaluation. The CBCT
examination was considered supplemental to conventional
radiographic examination during the period of gradual
implementation of the machine at the Department of
Orthodontics. Therefore, this material was considered unique
for further comparative studies. Consent to undergo the
additional radiographic examination and to use the material
Downloaded from by guest on May 30, 2015
for the present investigation was obtained from all patients.
Two different sets of images were available for each canine.
One set comprised a DPT, a lateral head film, and the available
periapical radiographs with different projections, as routinely
used in orthodontic treatment planning. The 2D digital images
(Figure 1) were obtained with the Digora Optime System
(Soredex, Tusuula, Finland). The second set of images was
Figure 1 Example of a two-dimensional data set composed of digital
created from CBCT scans generated with a NewTom 3G radiographs: (A) lateral head film, (B) dental pantomograph, (C) periapical
scanner (Quantitative Radiology s.r.l., Verona, Italy). This radiographs with different projections.
group included a series of static images produced elaborating
the CBCT raw data with dedicated software. Six to eight to estimate the difficulty of the case and the quality of the
different 3D reconstructions were generated with the images visualized on a visual analogue scale, and to choose a
maximum intensity projection and volumetric rendering treatment strategy among given alternatives.
methods, 12–20 axial images, one curved planar reformation The influence of the radiographic method was assessed
(panorex), a series of multiplanar reconstructions (cross- by means of a McNemar–Bowker test for paired data,
sections), two oblique planar reformation recording the bucco- Kappa statistics, and a paired t-test. The differences were
lingual and mesio-distal dimension of the tooth (Figure 2). considered significant at the 5 per cent level. The Statistical
Each image set was assembled in a Power Point® presentation. Package for Social Sciences, version 12.0 (SPSS Inc.,
Each set, either of conventional radiographs or CBCT- Chicago, Illinois, USA), was used.
generated images, was assigned a numerical code and, in a
randomized order, submitted for evaluation to eight different Results
dentists (three at the beginning of their orthodontic Considering a data set as composed of both 2D and 3D
education; two with a moderate degree of experience at the image sets, the total number of data sets for the 39 canines
end of their postgraduate education and three specialists submitted to the eight operators was 312. The results of this
with more than 5 years of experience). investigation divided for parameters are summarized in
The image sets were sent in small batches to ensure that no Table 1. The agreement among the 2D and 3D methods and
dentist had access to both 2D and 3D images of the same canine the description of the systematic difference are described.
at any time. The evaluators were asked to fill in a questionnaire
regarding, in the first part, the position of the canine expressed
Localization of the canine
by different parameters (Stivaros and Mandall, 2000) and to
assess the presence of root resorption on the lateral incisors. In The inclination measured to the midline did not differ
the second part of the questionnaire, the operators were asked significantly when evaluated using the two methods (Figure 3).
346 S. BOTTICELLI ET AL.
Downloaded from by guest on May 30, 2015
Figure 2 Example of the types of images delivered as a three-dimensional (3D) data set: (A) axial view, (B)
curved planar reformation (panorex), (C) transplanar reformation (cross-sections), (D) 3D reconstructions. Several
‘cuts’ from the axial and the cross-section view were provided.
Table 1 Summary of the results: agreement among the two (2D) and three-dimensional methods for the parameters considered.
Significance calculated with the McNemar–Bowker test, the Student’s t-test, and description of the systematic difference.
Parameter Percentage of agreement P Description of systematic difference
Inclination to the midline 74 >0.05 /
Mesio-distal position of the apex 64 0.001 2D indicated less variation in apex position
Vertical level of the clinical crown 66 0.013 2D indicated higher vertical level
Overlap with the lateral incisor 70 0.001 2D indicated less overlap
Labio-palatal position of the crown 68 0.001 2D indicated more palatal position of crown
Labio-palatal position of the apex 65 0.001 2D indicated more palatal position of apex
Root resorption of neighbouring incisor/s 82 0.001 2D indicated less root resorption
Treatment strategy 70 0.008 2D lead to observational strategy
Assessment of difficulty 46 <0.05 2D indicated less difficult treatment
Assessment of image quality 51 <0.05 2D was judged less positively with respect to image quality
3D IMAGING OF UNERUPTED CANINES 347
When assessing the mesio-distal localization of the apex, a reflected the larger overlap when evaluated on the 3D
significant difference was found for the two methods. The images.
tendency of the 2D data set, to score in the region of the first With respect to labio-palatal localization of the crown
premolar was higher, while on the 3D data set, a larger and the apex, paired data analysis revealed significantly
spread was observed. The vertical level of the clinical crown larger labial localization with the 3D method (Table 2).
differed significantly with a tendency towards a higher
Root resorption
position based on the 2D images. The definition of overlap
with the lateral incisor was congruent for the two methods in Significantly more root resorption was found (P = 0.001)
70 per cent of cases. The statistically significant disagreement when using the 3D data set (Table 3).
Downloaded from by guest on May 30, 2015
Figure 3 Bubble graphs showing the level of agreement when evaluating: (a) inclination with respect to the midline, (b) mesio-distal position of the apex, (c)
vertical level of crown height, (d) overlap with the neighbouring incisor defined within categories. The spheres’ size is proportional to the number of data sets
(total = 312). The spheres lying on the diagonal line depict the total agreement between the two methods. (a) The distribution of the disagreement is equal in the
2D and 3D areas of the graph. (b) The apex is seen more in the area of the first premolar following 2D evaluation, while a larger spread is observed on the 3D area
of the graph. (c) The crown is seen more apical following 2D evaluation, as shown by the distribution of larger spheres in the 2D area of the graph. (d) The overlap
is seen as larger following 3D evaluation, as described by the skewed distribution of larger spheres towards the 3D area of the graph.
Table 2 Number of data sets (N = 312) scored as labial or palatal
in the evaluation of the crown and apex labio-palatal position.
Table 3 Number of data sets (N = 312) scored as with or without
root resorption of the adjacent teeth.
Labio-palatal position Labio-palatal position
of the crown of the apex
Resorption
3D 3D
3D Total
Labial Palatal Total Labial Palatal Total
No Yes
2D Labial 143 16 159 56 27 83
Palatal 85 68 153 81 148 229 Resorption 2D No 236 40 276
Total 228 84 312 137 175 312 Yes 14 22 36
Total 250 62 312
The canine crowns and apices were classified as labial more frequently
after three-dimensional (3D) evaluation with respect to two-dimensional A more frequent diagnosis of root resorption was identified after three-
(2D; P = 0.001) as shown in bold. dimensional (3D) evaluation, P = 0.001, shown in bold.
348 S. BOTTICELLI ET AL.
Treatment choice explained by the horizontal distortion, which affects the
image of objects located behind or in front of the focal trough
When asked to make a treatment choice, agreement was
on an DPT image (Yeo et al., 2002). Anatomical structures
found for 70.5 per cent of the cases (Table 4). The lack of
located within the focal trough of a panoramic radiograph
congruence was statistically significant. A more frequent
would appear undistorted, while other objects located in front
choice of an observational–interceptive approach was based
or behind the sharp line are blurred, magnified, or constricted
on the 2D evaluation, while a more active intervention, with
and sometimes not clearly recognizable (Gratt, 1994).
special focus on expansion and space maintenance, was
Clinically, the difference between the two methods
recommended based on the 3D examination.
concerning the vertical level of the clinical crown would
have an influence on the estimated outcome of treatment;
Difficulty of the case the higher the canine position with respect to the occlusal
The assessment of the difficulty of a case differed plane, the longer and more difficult treatment. A more
significantly. The paired sample t-test showed a difference cranial localization was identified following 2D evaluation
at the 95 per cent confidence level, rendering treatment with respect to 3D. This is in accordance with the findings
more difficult based on 3D examination. of Chaushu et al. (1999) who reported that palatally located
canines will be projected higher than labially located
Quality of the images canines on a DPT as the central ray in panoramic radiography
is directed from a slight negative angulation of −7 degrees.
When finally asked whether they found the images The method of examination also influenced the estimation
appropriate for the given diagnostic purpose the respondents of overlap with the adjacent lateral incisor. A larger overlap
preferred the 3D images. The paired sample t-test showed a was scored on the 3D images. This could be due to the
difference at the 95 per cent confidence level; the mean horizontal deformation that affects the DPT, resulting in an
score was higher when evaluating the 3D image sets. increased dispersion of objects in the horizontal plane
Downloaded from by guest on May 30, 2015
(Gratt, 1994). Clinically, in subjects where the overlap is
Discussion larger, such as in upper anterior crowding, the overlap will
appear less severe in two-dimensions.
Comparison of the findings of this study, based on A difference in the perception of the canine position and
conventional and CBCT-generated images, demonstrated a the space conditions in the arch will influence the treatment
difference with respect to localization of the canine apex plan. This is confirmed by the findings, where the evaluators
mesio-distally and of both the apex and crown bucco- suggested active orthodontic treatment more frequently
palatally, vertical localization of the crown, overlap with the based on the 3D image set. Several evaluators completed
lateral incisor, and perception of root resorption. The the questionnaire specifying the need for expansion or other
agreement among the two methods with respect to procedures targeted at maintenance of the leeway space.
inclination of the canine to the midline could be explained This treatment approach has been shown to be more
by the classification distinguishing only between three successful than simple observation (Baccetti et al., 2008).
categories, each comprising a large range of angulations. In the present research, observation was chosen more
In agreement with Chaushu et al. (2004), the present study frequently following 2D evaluation. It can therefore be
demonstrated significant differences with respect to speculated that the use of CBCT allows for a treatment
localization of the apex in a mesio-distal direction, with a choice with a better prognosis.
higher tendency for a score in the intermediate category (first Lack of congruence among the two examinations with
premolar region) following 2D examination. This might be respect to labial or palatal localization of the crown and
apex is of clinical importance. The majority of the canines
in this sample were localized labially or centrally. The
Table 4 Distribution with respect to choice of treatment based
on two (2D) and three-dimensional (3D) evaluation. incorrect diagnosis based on the 2D images can be ascribed
to the fact that several teeth were localized approximately in
the middle of the alveolar crest. The 3D image set allowed
Treatment choice 2D 3D
more precise localization with respect to the lateral incisor
since axial sections were provided. Information on the exact
Extraction of primary canine only 26 12
Observation-no treatment 63 50 position of the crown is relevant when performing surgical
Extraction of the permanent canine 6 15 exposure, while the orthodontist needs to localize the apex
Surgical exposure/orthodontic treatment 211 230 to define the vector of traction.
Surgical transplantation of the canine 6 5
The present study demonstrated that root resorption was
more frequently diagnosed on the 3D image set. However, the
Less observational and more surgical exposure and orthodontic traction
were suggested after 3D evaluation. Lack of congruence between the two resolution of the images obtained with the NewTom does not
methods was statistically significant (P = 0.008). allow for clear depiction of resorption craters either at the
3D IMAGING OF UNERUPTED CANINES 349
cement or dentine level. Therefore, the evaluators were asked of the impacted tooth, which improves diagnosis and
not to measure root resorption but to assess its presence or provides useful information for treatment consultation.
absence. CBCT allows determination of close proximity
between teeth: this might not change evaluation of the prognosis
References
of the resorbed teeth but influences the treatment plan in terms
of determination of the direction of orthodontic traction. Baccetti T, Leonardi M, Armi P 2008 A randomized clinical study of two
interceptive approaches to palatally displaced canines. European Journal
The greatest amount of overlap was also scored on the 3D of Orthodontics 30: 381–385
images and could explain the higher grade of difficulty assigned Bishara S E 1992 Impacted maxillary canines: a review. American Journal
after 3D evaluation. The quality of the images was, as of Orthodontics and Dentofacial Orthopedics 101: 159–171
anticipated, assessed positively for the 3D image set. Further Chaushu S, Chaushu G, Becker A 1999 Reliability of a method for the
improvements in CBCT are occurring both at the hard- and localization of displaced maxillary canines using a single panoramic
radiograph. Clinical Orthodontics and Research 2: 194–199
software level. It is however already possible to ameliorate the
Chaushu S, Chaushu G, Becker A 2004 The role of digital volume
volumetric data exported in DICOM format by elaboration tomography in the imaging of impacted teeth. World Journal of
with other software dedicated to dento-maxillo-facial imaging. Orthodontics 5: 120–132
The quality of the voxels, basic unit in the image, is still inferior Clark C 1909 A method of ascertaining the relative position of unerupted
when compared with a conventional CT scan. teeth by means of film radiographs. Proceedings of the Royal Society of
Medicine, pp. 87–89
The main advantage of CBCT with respect to a CT scan is
Elefteriadis J N, Athanasiou A E 1996 Evaluation of impacted canines by
the reduction in the radiation dose. Ludlow et al. (2006) means of computerized tomography. International Journal of Adult
reported values of effective dose of 36.3 mSv for the NewTom Orthodontics and Orthognathic Surgery 11: 257–264
3G corresponding to four DPTs. In comparison, the radiation Ericson S, Bjerklin K 2001 The dental follicle in normally and ectopically
dose of a lateral and postero-anterior cephalogram can be up erupting maxillary canines: a computed tomography study. Angle
Orthodontist 71: 333–342
to 166 mSv (Mah and Hatcher, 2003).
Ericson S, Kurol J 1986a Longitudinal study and analysis of clinical
In order to understand the associated radiation/cancer supervision of maxillary canine eruption. Community Dentistry and
Downloaded from by guest on May 30, 2015
risk, the dosimetric cost has to be interpreted in terms of Oral Epidemiology 14: 172–176
background equivalent radiation time compared with daily Ericson S, Kurol J 1986b Radiographic assessment of maxillary canine
light irradiation in a sunny place. The NewTom 3G eruption in children with clinical signs of eruption disturbance. European
Journal of Orthodontics 8: 133–140
corresponds to 4–6 days compared with 18 days for a CT
Ericson S, Kurol J 1987 Radiographic examination of ectopically erupting
scan. The results of this study showed that the 3D image set maxillary canines. American Journal of Orthodontics and Dentofacial
obtained with CBCT provides additional information with Orthopedics 91: 483–492
respect to the 2D images. This information, mainly in the Ericson S, Kurol J 1988 CT diagnosis of ectopically erupting maxillary
position of the unerupted canine and its relationship with canines—a case report. European Journal of Orthodontics 10: 115–121
neighbouring structures, has a strong clinical relevance, Ericson S, Bjerklin K, Falahat B 2002 Does the canine dental follicle cause
resorption of permanent incisor roots? A computed tomographic study
which appears to justify the risks of the radiation dose. of erupting maxillary canines. Angle Orthodontist 72: 95–104
Gratt B M 1994 Panoramic radiography. In: Goaz P W, White S C (eds.).
Oral radiology: principles and interpretation, 3rd edn. Mosby, St Louis,
Conclusions pp. 242–244
When comparing the diagnosis of unerupted maxillary Ludlow J B, Davies-Ludlow L E, Brooks S L, Howerton W B 2006 Dosimetry
of 3 CBCT devices for oral and maxillofacial radiology: CB Mercuray,
canines based on 2D and 3D images, the findings showed a NewTom 3G and i-CAT. Dentomaxillofacial Radiology 35: 219–226
difference between the two techniques. Using the 3D image Mah J, Hatcher D 2003 Current status and future needs in craniofacial
set, the crown of the canine was perceived to be more imaging. Orthodontics and Craniofacial Research 6 (suppl 1): 179–182
occlusally positioned than on the 2D images. Furthermore, Mckee I W, Glover K E, Williamson P C, Lam E W, Heo G, Major P W
the 3D images allowed determination of the mesio-distal 2001 The effect of vertical and horizontal head positioning in panoramic
radiography on mesiodistal tooth angulations. Angle Orthodontist 71:
position of the apex, while the bucco-palatal position could 442–451
be assessed with less uncertainty. The overlap with the lateral Mckee I W, Williamson P C, Lam E W, Heo G, Glover K E, Major P W
incisor was perceived as more severe after 3D evaluation. 2002 The accuracy of 4 panoramic units in the projection of mesiodistal
The same was true concerning the presence of root resorption. tooth angulations. American Journal of Orthodontics and Dentofacial
Orthopedics 121: 166–175
In accordance with these findings, a higher degree of difficulty
Mozzo P, Procacci C, Tacconi A, Martini P T, Andreis I A 1998 A new
of the case was judged on the basis of the 3D image set. volumetric CT machine for dental imaging based on the cone-beam
These differences can be explained by factors affecting technique: preliminary results. European Radiology 8: 1558–1564
the conventional 2D radiographs such as distortion, Southall P J, Gravely J F 1989 Vertical parallax radiology to localize an object
magnification, and superimposition of anatomic structures in the anterior part of the maxilla. British Journal of Orthodontics 16: 79–83
situated in different planes. The treatment plan was different Stivaros N, Mandall N A 2000 Radiographic factors affecting the
management of impacted upper permanent canines. Journal of
using 2D and 3D image sets. 3D examination more often Orthodontics 27: 169–173
led to an approach of expansion and orthodontic traction. Yeo D K, Freer T J, Brockhurst P J 2002 Distortions in panoramic
The CBCT gives a perception of the intra-osseous position radiographs. Australian Orthodontic Journal 18: 92–98
