Seminars in Orthodontics

EDITOR -IN-CHIEF
Elliott M. Moskowitz, DDS, MSd

EDITORIAL BOARD
EDITOR-IN-CHIEF EMERITUS
P. Lionel Sadowsky, DMD, BDS, DipOrth, MDent

Mani Alikhani, New York, NY (2017) Peter Ngan, Morgantown, WV (2015)

Rolf G. Behrents, St. Louis, MO (2017) Perry M. Opin, Milford, CT (2017)
S. Jay Bowman, Portage, MI (2017) Jae Hyun Park, Mesa, AZ (2017)
James Caveney, Wheeling, WV (2015) Sheldon Peck, Newton, MA (2014)
John Grubb, Chula Vista, CA (2015) William R. Proffit, Chapel Hill, NC (2015)
Greg Huang, Seattle, WA (2014) Eugene Roberts, Indianapolis, IN (2015)
Robert J. Isaacson, Edina, MN (2015) Emile Rossouw, Rochester, NY (2017)
Laurance Jerrold, Brooklyn, NY (2017) David L. Turpin, Federal Way, WA (2017)
Lysle E. Johnston, Jr., Eastport, MI (2015) James L. Vaden, Cookeville, TN (2015)
Donald R. Joondeph, Bellevue, WA (2015) Robert L. Vanarsdall, Jr., Philadelphia, PA (2015)
Robert G. Keim, Los Angeles, CA (2017) Katherine Vig, Columbus, OH (2017)
Richard Kleefield, Norwalk, CT (2015) Christos Vlachos, Homewood, AL (2014)
Steven J. Lindauer, Richmond, VA (2015) Timothy T. Wheeler, Gainesville, FL (2015)
James A. McNamara, Jr., Ann Arbor, MI (2017) Leslie A. Will, Boston, MA (2017)

INTERNATIONAL
Adrian Becker, Jerusalem, Israel (2017) Rakesh Koul, Lucknow, India (2017)
Jose´ Alexandre Bottrel, Rio de Janeiro, Brazil (2015) Birte Melsen, Aarhus, Denmark (2017)
Theodore Eliades, Nea Ionia, Greece (2014) Antony McCollum, Bryanston, South Africa (2015)
W.G. Evans, Johannesburg, South Africa (2017) Eliakim Mizarahi, Ilford, England (2015)
Jorge Faber, Brasilia, Brazil (2017) Bjørn Øgaard, Oslo, Norway (2017)
Joseph Ghafari, Beirut, Lebanon (2017) Nikolaos Pandis, Corfu, Greece (2017)
Vicente Hernandez, Alicante, Spain (2017) Pratik K. Sharma, London, UK (2017)
Nigel Hunt, London, England (2015) George Skinazi, Paris, France (2015)
Haluk Iseri, Ankara, Turkey (2017) John C. Voudouris, Toronto, Canada (2017)
Roberto Justus, Mexico City, Mexico (2015) William A. Wiltshire, Winnipeg, Canada (2015)
Sanjivan Kandasamy, Midland, WA, Australia (2017) Björn U. Zachrisson, Oslo, Norway (2015)
 Seminars in Orthodontics
VOL 21, NO 4 DECEMBER 2015

Advances in CBCT Diagnostics with Orthodontic Treatment:

Interpretation and Manipulation
Onur Kadioglu, DDS, MS
Guest Editor

■ Introduction 231
Onur Kadioglu

■ Implementation of ultra-low-dose CBCT for routine 2D orthodontic diagnostic

radiographs: Cephalometric landmark identification and image quality assessment 233
Budi Kusnoto, Pardeep Kaur, Abdelrahman Salem, Zheng Zhang,
Maria Therese Galang-Boquiren, Grace Viana, Carla A. Evans, Robert Manasse,
Richard Monahan, Ellen BeGole, Ayas Abood, Xiao Han, Emil Sidky, and
Xiaochuan Pan

■ Assessment of image quality in maxillofacial cone-beam computed tomography

imaging 248
Casey Gamache, Jeryl D. English, Anna M. Salas-Lopez, John Rong, and
Sercan Akyalcin

■ 3D analysis and clinical applications of CBCT images 254

Mohamed Bayome, Jae Hyun Park, YoonJi Kim, and Yoon-Ah Kook

■ 3-Dimensional cone-beam computed tomography superimposition: A review 263

Jae Hyun Park, Kiyoshi Tai, and Payam Owtad

■ Diagnosis and evaluation of skeletal Class III patients with facial asymmetry for
orthognathic surgery using three-dimensional computed tomography 274
Sung-Jin Kim, Hyoung-Seon Baik, Chung-Ju Hwang, and Hyung-Seog Yu

■ Comparison of the anterior alveolar bony changes of moderately crowded cases

treated either with extraction or non-extraction orthodontic treatment 283
Tanner Cook, Fräns Currier, Onur Kadioglu, and Thomas Griffin
 Seminars in Orthodontics
VOL 21, NO 4 DECEMBER 2015

Introduction

A paradigm shift necessitates a major or

radical change in a scientific thought pat-
tern. In order for any newness to be qualified as a
times and the ability to limit the exposed area to
smaller fields of view, resulted in significantly
lower amounts of radiation. Additionally, clin-
paradigm shift in our specialty, it needs to dra- icians are reminded to use the ALARA (As Low
matically alter the practice of orthodontics. Since As Reasonable Achievable) principle and con-
the new millennium, orthodontics has witnessed sider limiting the types and/or number of
multiple advancements. These took place in the radiographs needed with the availability of CBCT
biology of tooth movement, with accelerated images.
cellular response; biomechanics, with temporary When invited as the guest editor for this issue,
anchorage devices; and in imaging, in the form I was not only honored but also excited to revisit
of three-dimensional applications, which the current concepts in 3D imaging as this pro-
includes scanning, printing, and radiography. vided an opportunity to update on the advances
Within the hands of innovative-minded ortho- both in application and methodology. When I
dontic practitioners, these advancements evolved reached out to those I consider experts in this
and flourished, and this still continues today. The field, I was delighted by their positive response.
question then is, which one of those advance- There are six articles presented in this issue.
ments would qualify to be labeled as a paradigm The first two articles describe and discuss
shift? methods, which can aid clinicians and manu-
Considering the changes that have taken place facturers in reducing the radiation dose. In the
in our thought processes via multiple viewing first one, authors from the University of Illinois at
possibilities the Cone-Beam CT is offering, 3D Chicago, have implemented an algorithm that
imaging may be the one with the most impact in helps reduce the radiation and evaluated the
our daily lives. The ability to evaluate facial and quality of 2D images created through this
lingual alveolar bone, airway, skeletal asymme- process. In the second article, authors from the
tries, dental anatomy/morphology/pathology University of Texas in Houston, evaluated
along with the temporomandibular joint are a accuracy utilizing various settings that lower the
few among these possibilities. They are assisted by dose.
the evolving capabilities of the hardware and The next three articles presented fine reviews
imaging software along with techniques that have with status updates for various analyses that are
been recently developed for proper use of these possible through CBCT. Dr. Jae Park and his co-
applications. authors from A.T. Still University in Arizona,
Aside from its initial cost, this fine technology discuss current concepts on 3D superimposition
came with a significant obstacle that is higher as well as advantages and limitations. The other
levels of radiation dose. Although initially con- two articles are written by our Korean colleagues
sidered relatively high, dose levels have reduced from Yonsei University and the Catholic Uni-
with most new machines and it seems that pattern versity of Korea. These authors described 3D
will continue, as this has been a main focus for cephalometric analyses for soft and hard tissue
the manufacturers. Advances in the sensor sen- through new landmarks and planes and presented
sitivity with decreased noise has allowed for utilization of CBCT for skeletal asymmetries.
higher quality images than can be obtained with The final article is from the University of
lower levels of mA. This, coupled with faster scan Oklahoma and is part of their ongoing studies on
alveolar bone changes before and after various
Published by Elsevier Inc. types of orthodontic treatments. The manuscript
http://dx.doi.org/10.1053/j.sodo.2015.07.008 utilized for this issue compares the effects of

Seminars in Orthodontics, Vol 21, No 4, 2015: pp 231–232 231

232

extraction versus non-extraction treatment, stay with multiple future improvements planned.
and is an important contribution that highlights Whether one can call it a paradigm shift or not,
the limits of the bony housing; a topic that has however, is a question only the fourth dimension,
been a subject of much debate within the last time, will be able to tell.
decade. I appreciate the opportunity and thank the
I hope the readers will be able to find valuable editor-in-chief and Elsevier for allowing me to
information in this issue. Attempts are being serve as the guest editor for this issue.
made to strengthen the place of 3D imaging in
Orthodontics and Dentofacial Orthopedics Onur Kadioglu, DDS, MS
worldwide. There is no doubt CBCT is here to Guest Editor
Implementation of ultra-low-dose CBCT for
routine 2D orthodontic diagnostic radiographs:
Cephalometric landmark identification and
image quality assessment
Budi Kusnoto, Pardeep Kaur, Abdelrahman Salem, Zheng Zhang,
Maria Therese Galang-Boquiren, Grace Viana, Carla A. Evans,
Robert Manasse, Richard Monahan, Ellen BeGole, Ayas Abood,
Xiao Han, Emil Sidky, and Xiaochuan Pan

Since its adoption more than two decades ago in the field of dentistry, three-
dimensional (3D) cone-beam computed tomography (CBCT) has rapidly
gained popularity as an imaging tool for three-dimensional visualization,
diagnosis, and treatment simulation. The objective of this two-part study
was to assess landmark identification, as well as the diagnostic value of
images obtained using an ultra-low-dose reduced projection (sparse) views
algorithm Adaptive-Steepest-Descent-Projection-Onto-Convex-Sets (ASD-
POCS) applied to existing dental CBCT data. ASD-POCS was developed for
CBCT studies, producing multiple CBCT data files with 12.5% (39), 25% (76),
and 50% (150) of the original number of projection views (which is in direct
proportion with the lowering of radiation dose). Assessment of 2D landmark
identification derived from CBCT data at different projection views was
conducted. Assessment of diagnostic quality was studied by evaluating
radiographs of various projection views on a visual analog scale by different
dental specialists. In conclusion, this study found no statistically significant
differences in the quality of images at 25% (76) projection views as compared
to 100% (300) projection views. The implementation on ASD-POCS algorithm
by the CBCT manufacturers in the future may be beneficial for clinicians in
reducing the amount of patient radiation exposure when obtaining routine
orthodontic diagnostic radiographs. (Semin Orthod 2015; 21:233–247.) &
2015 Elsevier Inc. All rights reserved.

Department of Orthodontics, University of Illinois at Chicago, Introduction

801 S Paulina St. Rm 131, Chicago, IL 60612; Department of
one-beam computed tomography (CBCT)
Radiology, The University of Chicago, Chicago, IL; Department of
Oral Medicine and Diagnostic Sciences, University of Illinois at
Chicago, Chicago, IL; Department of Radiology, University of Illinois
C is a three-dimensional imaging technique
used in dentistry and medicine. CBCT has
at Chicago, Chicago, IL; Department of Radiation and Cellular become an increasingly important tool in treat-
Oncology, The University of Chicago, Chicago, IL.
Address correspondence to Budi Kusnoto, DDS, MS, Department ment planning and diagnosis in implant den-
of Orthodontics, University of Illinois at Chicago, 801 S Paulina St tistry, endodontics, orthodontics, and other
Rm 131, Chicago, IL 60612. E-mail: [email protected] dental fields.
This work was supported in part by NIH R01, Grant no. The benefits of CBCT in orthodontics include
EB000225. increased accuracy of image geometry, and
The work of image reconstruction was reported from: Department measurements by eliminating the image magni-
of Radiology, The University of Chicago, 5841 South Maryland Ave,
fication, overlapping, and distortion commonly
MC2026, Room I-SB-05J, Chicago, IL 60637. Tel.: þ1 773 702
1293.
encountered in 2D radiographs. Potential uses of
& 2015 Elsevier Inc. All rights reserved.
CBCT are precise localization of ectopic teeth
1073-8746/15/1801-$30.00/0 and measurement of unerupted tooth sizes,
http://dx.doi.org/10.1053/j.sodo.2015.07.001 assessment of root resorption, identification and

Seminars in Orthodontics, Vol 21, No 4, 2015: pp 233–247 233

234 Kusnoto et al

quantification of asymmetry, visualization of Materials and methods

airway abnormalities, and even providing the
Assessment of cephalometric landmarks
imaging data to support treatment simulation
identification
and technology-aided treatment.1
The routine use of CBCT in orthodontics is an The experimental sample in this study consisted
area of debate. One of the concerns with the of four CBCT scans of patients (IRB Protocol
routine use of CBCT is the effect of ionizing 2013-1120) from the archives of a university
radiation on growing children. This has resulted in orthodontic clinic. The Frankfort Horizontal
the development of general guidelines to manage plane was parallel to the floor during image
justification, optimization, and referral criteria for acquisition. The CBCT scans were obtained
users of dental CBCT. Such guidelines were from i-CAT Next Generation scanner (Imaging
adopted by the American Association of Ortho- Sciences International, Hatfield, PA) using a
dontists in 2010: “Although the AAO recognizes 0.3 mm voxel size scan with diameter of 16 cm
that while there may be clinical situations where a and height of 13 cm at 5 mA and 120 kVp.
cone-beam computed tomography (CBCT) The original Digital Imaging and Communica-
radiograph may be of value, the use of such tions in Medicine (DICOM) files of the four
technology is not routinely required for ortho- patients were de-identified. The ASD-POCS algo-
dontic radiography.”2 In addition, the American rithm was used to reduce the number of projection
Academy of Oral and Maxillofacial Radiology views on the original CBCT via iterative recon-
(AAOMR) recently issued the clinical recommen- struction producing multiple CBCT images. The
dations regarding the use of CBCT in orthodontics. original i-CAT image is composed of 300 projection
The article summarizes the potential benefits and views. From the full data set of 300 projection views
risks of maxillofacial CBCT for use in orthodontic in the i-CAT scan, new CBCT images were
diagnosis, treatment and outcome assessment reconstructed with same 300 and reduced 150, 76,
following the ALARA (As Low As Reasonably and 39 numbers of projection views producing four
Achievable) principle.3 CBCT images as shown in Fig. 1. In this study, those
The CBCT industry has improved on this different projection views were named UC300,
technology over two decades of its use in dentistry. UC150, UC76, and UC39.
With technologic advances such as the use of flat Raw data were stored in DICOM format. The
panel detectors, the availability of different fields of CBCT data were imported into Dolphin 3D
view, and the advances in computer science steps software (Version 11.7, Dolphin Imaging,
have been made towards the reduction of radiation Chartsworth, CA) and used throughout this
exposure by CBCT technology.4 study. Four CBCT’s were reconstructed for each
In the field of medical radiology, recon- subject. Dolphin 3D was used to construct 2D
structed algorithms (RA) have been applied to lateral cephalograms and posterior–anterior
CBCT scans.5 Equipment using the ultra-low- cephalograms. The standard lateral and poste-
dose CBCT reconstruction algorithm integrating rior–anterior (PA) cephalometric landmarks
Barzilai–Borwein step-size calculation cuts the were then analyzed for accuracy of landmark
radiation dose while maintaining the image identification by the investigators with a mean
quality.6 These algorithms have not yet been error of less than 1.5 mm. Landmark identi-
applied to reducing radiation exposure from fication among different projection views for
CBCTs in orthodontic radiographs. each single landmark on a given model was
investigated. The five different data sets,
including one full projection views data set and
Purpose of the study
four reduced number of projection views data
The purpose of this study is to find a minimum sets, for each one of the four different patients
number of projection views in 3D CBCT derived by were traced in both lateral and PA cephalometric
the ASD-POCS algorithm needed to provide radiographs. Landmarks were compared on both
clinically acceptable results for routine 2D ortho- the x and y axes to the control (i-CAT 300)
dontic diagnostic radiographs (identification landmarks. The method is described in Fig. 2.
of 2D cephalometric landmarks identification and Using a grid, the image center (0,0) on the x
assessment of diagnostic quality of the 2D image). and y coordinates was located at the landmark
 Implementation of ultra-low-dose CBCT for routine 2D orthodontic diagnostic radiographs 235

Figure 1. Obtaining the reduced projections views and 2D radiographs using ASD-POCS algorithm and Dolphin
3D software.

Figure 2. Comparisons between sparse (reduced) projections views data sets and the full data set.
236 Kusnoto et al

Figure 3. Lateral and PA cephalometric landmarks.

Sella for the lateral cephalometric radiographs spreadsheet was made of all values of x and y
and at the landmark Crista Galli for the PA coordinates of the thirty landmarks. After all the
cephalometric radiographs. Nine skeletal and six measurements were obtained, landmark loca-
dental landmarks were chosen for each lateral tions in the cephalograms derived from the ASD-
and PA cephalometric radiograph derived from POCS data sets were compared to those of the
each CBCT scan (a total of 30 landmarks). The full data set i-CAT 300 as a control. The absolute
landmarks used are summarized in Fig. 3 and distances between the landmarks of the algo-
Table 1. rithm reconstructions and those of the i-CAT 300
The tracings of the various 2D reconstructed control were then calculated. The distances
images were superimposed on Sella–Nasion (S–N) between landmarks were added together for a
line to determine the accuracy of the identification single data set divided by the total number of the
of the landmarks and their location by single landmarks to calculate the average distance from
operator. Only a 1.5 mm deviation was accepted. the control.
Using the transfer structure function in Dolphin
3D imaging, the S–N landmarks (lateral radio-
graph) and Crista Galli perpendicular to Zygomatic Assessment of image quality
left and right (PA radiograph) were transferred Two CBCT scans (image 1 and image 2) acquired
throughout the series of 2D projections to ensure using the i-CAT Next Generation scanner
consistency in the reference plane for super- (Imaging Sciences International, Hatfield, PA)
imposition both orientation and registration. from two patients were selected and de-identified
Due to the small sample size, (n ¼ 4), in (IRB Protocol 2013–0815). A single 3601 rotation
this feasibility study, inter- and intra-observer of the x-ray tube for 20 s on “Full” FOV 16 cm
reliability and accuracy testing were not con- (diameter)  13 cm (height) was chosen with
ducted. After the landmarks were obtained, a patients positioned in the machine and Frankfort

Table 1. List of dento-skeletal lateral and PA cephalometric landmarks

Lateral Cephalometric Landmarks Posteroanterior Cephalometric Landmarks

Skeletal landmarks Dental landmarks Skeletal landmarks Dental landmarks

(1) Sella (S) (1) U1 tip (U1i) (1) Orbitale R (Or R) (1) U1s contact (U1s)
(2) Nasion (N) (2) U1 root (U1r) (2) Orbitale L (Or L) (2) L1s contact (L1s)
(3) Porion (Po) (3) L1 tip (L1i) (3) Jugal R (J R) (3) UR6 MB cusp (UR6)
(4) Orbitale (Or) (4) L1 root (L1r) (4) Jugal L (J L) (4) UL6 MB cusp (UL6)
(5) Pt point (Pt) (5) U6 MB cusp (U6) (5) Condylion R (Cd R) (5) LR6 MB cusp (LR6)
(6) A point (A) (6) L6 MB cusp (L6) (6) Condylion L (Cd L) (6) LL6 MB cusp (LL6)
(7) B point (B) (7) Gonion R (Go R)
(8) Gonion (Go) (8) Gonion L (Go L)
(9) Menton (Me) (9) Menton (Me)
 Implementation of ultra-low-dose CBCT for routine 2D orthodontic diagnostic radiographs 237

horizontal parallel to the floor. Primary recon- E5-2603 and resolution 1440  900 at 60 Hz
struction of the data was done by the CBCT operated at 32 bit. Five monitors were controlled
machine after the data acquisition using 300 by single main computer through a graphics card
projection views. Raw data were stored in DICOM (AMD Radeon HD 7870 2 GB GDDR5). All
format. The scan exposure parameters were monitors had similar display settings and were
120 kVp, 5 mA, and 0.3 mm voxel size. The calibrated similarly. The computers were set up
reconstruction algorithm ASD-POCS was used to in a room where lighting could be dimmed
construct four CBCT scans from each original during the radiographic examination.
scan captured by the i-CAT machine. The Reconstructed two-dimensional radiographic
reconstructed CBCT scans had the same number images were projected on the monitors using
of projection views as the i-CAT (300) and Microsoft PowerPoint 2013 (Microsoft, Redmond,
reduced number of projection views to 150, 76, WA). The reconstructed 2D radiographs from
and 39 as shown in Fig. 4. CBCT image reconstructions with full number of
The data from each CBCT image: one original projection views (300), and reduced number of
and four reconstructed CBCT were exported projection views (150, 76, and 39) were randomized
from the Xoran Cat (i-CAT manufacturer’s and projected on the six different monitors (Fig. 8).
software) in DICOM and imported into the One i-CAT scan reconstructed image from
Dolphin 3D software. each set was duplicated and projected on the first
Three reconstructions from each CBCT were monitor as control. CBCT Data from two patients
performed producing these two-dimensional (image 1 and image 2) were shown in six sets of six
radiographs: images to the participants. Overall, 36 images were
1. Reconstruction of a lateral cephalometric shown to all the participants in the same order.
(LC) projection from the CBCT dataset using The primary investigator was blinded as to the
Dolphin 3D software named as LC300, LC150, order of display of the 2D radiographs. A research
LC76, and LC39 (Fig. 5). assistant randomized the 2D radiographs.
2. Reconstruction of a panoramic (Pano) About 73 subjects were recruited from the
projection from the CBCT dataset using Dolphin faculty and residents of the departments of
3D software named as Pano300, Pano160, orthodontics, oral and maxillofacial surgery, and
Pano76, and Pano39 (Fig. 6). oral and maxillofacial radiology to participate in
3. Reconstruction of a posteroanterior (PA) the survey. The raters consisted of 29 ortho-
cephalogram from CBCT dataset using Dolphin dontists, eight oral surgeons, 11 surgery resi-
3D software named as PA300, PA150, PA76, and dents, and 25 orthodontic residents. In addition,
PA39 (Fig. 7). two experienced oral and maxillofacial radiol-
Two-dimensional images were displayed on six ogists participated in the survey; their responses
20.1 in (1909W Dell computer model Precision were recorded but not analyzed as part of this
T5600) screens with processor 1 Intel Xeon CPU study due to small sample size. All participants

Figure 4. The reconstructed CBCT scans with different level of projection views i-CAT (300) and reduced number
of projections views 150, 76, and 39.
238 Kusnoto et al

Figure 5. Reconstruction of a planar lateral cephalometric projection from the CBCT dataset using Dolphin 3D
software.

Figure 6. Reconstruction of a curved planar panoramic projection from the CBCT dataset using Dolphin 3D
software.
 Implementation of ultra-low-dose CBCT for routine 2D orthodontic diagnostic radiographs 239

Figure 7. Reconstruction of a planar posteroanterior cephalogram from CBCT dataset using Dolphin 3D software.

utilize CBCT studies for specific conditions and amount for all the participants; however, sub-
not routinely for all patients. sequently they were asked for their preferred
Each participant was given a paper ques- light setting and none of the participants asked to
tionnaire to fill out with six sets of questions. change the original setting.
Each set had two questions: the first question Each set of six images were displayed on the
asked if the displayed image was diagnostically monitors in the same order for all participants.
acceptable and the second question asked the All images were evaluated for the following:
participant to rank each image on a visual analog
scale of 1 to 10 (1 ¼ poor quality and 10 ¼ best 1. Panoramic (Pano) images—for diagnostic qual-
quality). Room lighting was dimmed the same ity in a general screening (i.e., Is the image of

Figure 8. Illustration of five monitors controlled by one main computer using graphics Card displaying 36 images
in random sequence generated by random table.
240 Kusnoto et al

diagnostic quality to view bone levels, trabecu- of the three imaging techniques. McNemar tests
lation, and any pathology/dental anomalies?).7 were used to investigate the preference differ-
2. Lateral cephalogram (LC) images—for diagnostic ences between pairs of image reconstructions.
quality to establish an orthodontic diagnosis Friedman tests were used to assess if there were
(i.e., Is the image of diagnostic quality to differences among the ranks of the image quality
locate landmarks such as sella, nasion, gonion, of the five different image reconstructions for all
menton, and the relative positions of the three imaging techniques. Wilcoxon tests were
maxilla to the cranial base, maxilla to the used to test rank differences between pairs of
mandible, and the teeth in relation to their image reconstructions (Fig. 9).
respective jaw).7 Data analysis was conducted using SPSS v.20
3. Posteroanterior (PA) cephalogram—for their diag- (Chicago, IL).
nostic quality in locating jugal point and bony
and teeth midline asymmetries.
Results
Assessment of cephalometric landmark
The data collection included two different sets
identification
of data: nominal and ordinal. For the nominal
data set, a “yes” preference was recorded as 1 and The major indicator used in this study to compare
a “no” preference as 0. For the ordinal data set, and determine the identification of the landmarks
data were collected for the ranking (1–10) on a was the error distance between the 2D cephalo-
visual analog scale (VAS) for different types of 2D metric landmarks locations (x,y) identified of the
x-ray images composed of different numbers of UC reconstructions at various projection views
projection views. An examination of the raw data and the corresponding landmarks derived from
showed that nonparametric tests were appro- the i-CAT control. Examination of the variables
priate to analyze the data set. Cochran tests were being investigated in this study revealed average
used to assess if there were differences among the distances from the control ranging from 0.71 to
preferences of the image quality for five different 1.03 for the lateral cephalometric and from 0.72 to
projection views image reconstructions for each 1.25 for the PA cephalometric radiographs,

Figure 9. Experiment flow diagram.

 Implementation of ultra-low-dose CBCT for routine 2D orthodontic diagnostic radiographs 241

respectively. The distance of paired differences Table 3. Summary of mean error (in millimeter unit)
between compression rates falls into the range for horizontal and vertical direction of 2D lateral and
from 0.09 to 0.12 for the lateral cephalometric and PA cephalometric landmark location derived from 3D
CBCTS produced by ASD-POCS algorithm
from 0.13 to 0.24 for the PA cephalometric
radiograps, respectively (Table 2). ASD-POCS
iCAT 300 300 Comparisons
Data collected from the lateral cephalometric
radiograph of UC300 data set derived from the Landmark X Y X Y Dx Dy Dist
first model show absolute values of x coordinates
1 A Point 47.6 35.9 47.8 35.6 0.2 0.3 0.3
and y coordinates of both i-CAT and UC300 data 2 B Point 49.0 72.6 49.7 73.5 0.7 0.9 1.1
set, the differences in x coordinates and y coor- 3 Gonion 3.7 60.0 2.7 61.6 0.9 1.6 1.9
dinates, and the absolute distances (Table 3). 4 L1 Root 47.0 71.6 47.8 72.0 0.8 0.5 1.0
5 L1 Tip 54.9 54.6 55.3 54.9 0.5 0.3 0.5
The same procedure was performed for the PA 6 L6 34.7 53.4 35.3 54.7 0.6 1.3 1.4
cephalometric radiograph (Table 4, Fig. 10). The Occlusal
average distance between the landmarks 7 Menton 42.2 87.8 41.3 87.9 0.8 0.1 0.8
8 Nasion 50.1 10.2 50.0 10.2 0.1 0.0 0.1
increases gradually when moving through the 9 Orbitale 42.7 12.6 43.0 12.8 0.2 0.2 0.3
reconstructions starting from UC300 to UC39. 10 PT Point 11.0 10.7 10.9 12.4 0.1 1.7 1.7
The only exception is between UC76 and UC39 11 Porion 18.7 10.8 18.7 11.1 0.1 0.3 0.3
12 Sella 0.0 0.0 0.0 0.0 0.0 0.0 0.0
on the lateral cephalometric tracing where the 13 U1 Root 46.1 38.5 46.2 37.6 0.1 0.8 0.8
UC39 shows shorter average distance than the 14 U1 Tip 55.0 53.4 55.5 53.7 0.4 0.3 0.5
UC76. After calculating all the distances of 15 U6 32.3 53.0 32.8 53.8 0.4 0.8 1.0
Occlusal
the landmarks (see last column in Table 3),
the average values of the error distances were
calculated for every single data set on the 2D when the average is calculated for all the land-
lateral cephalometric radiographs derived from marks of a single reconstruction, the results show
the 3D images of patients. that none of the reconstructions exceeded the
Many landmarks showed distances that 1.5 mm limit. This is illustrated in Fig. 10. This
exceeded the 1.5 mm error margin when com- indicates that all the reconstructions are valid for
pared to the iCAT control and hence considered tracing and 2D cephalometric analysis.
inaccurate according to our standards. However, When analyzing the maximum distance
between the landmarks of the projection views
data set and those of the full data i-CAT controls,
Table 2. Descriptive statistics of 2D lateral and PA
cephalometric landmark location derived from 3D it is expected that the maximum distance would
CBCTS produced by ASD-POCS algorithm increase when moving from higher quality
images (UC300) to lower quality images (UC39)
Paired
Variables N Average Differences where the poor quality makes it much harder to
accurately trace the landmarks resulting in lon-
Pair 1 Lateral ceph 300i 4 0.7100 ger distances between the corresponding land-
Lateral ceph 150i 4 0.8350 0.12500
marks. The results in this study confirm this in
Pair 2 Lateral ceph 150i 4 0.8350 the lateral cephalometric values except for a
Lateral ceph 79i 4 1.1275 0.29250
small difference between the UC76 and UC39
Pair 3 Lateral ceph 79i 4 1.1275 measurements. When looking at the PA cepha-
Lateral ceph 33i 4 1.0325 0.09500
lometric values, a big discrepancy is observed
Pair 4 PA ceph 300i 4 0.7200
between the UC300 and UC150.

PA ceph 150i 4 0.8475 0.12750

Assessment of image quality
Pair 5 PA ceph 150i 4 0.8475 In this study using 2D reconstruction from 3D
data set of images 1 and 2, the panoramic
PA ceph 79i 4 1.0050 0.15750 radiograph constructed at 300 projection views
was consistently preferred and had higher mean
Pair 6 PA ceph 79i 4 1.0050 rank than the one reconstructed by the i-CAT
PA ceph 33i 4 1.2475 0.24250 software. In both rank and preference, panoramic
242 Kusnoto et al

Table 4. The average deviations of lateral and posteroanterior cephalometric landmarks at different levels of
projections views ASD-POCS data sets (in millimeters)
Lateral Cephalometric Landmarks Posteroanterior Cephalometric Landmarks

Average Model 300 150 79 33 Model 300 150 79 33

Model 1 1.10 1.11 1.06 1.26 Model 1 0.96 1.18 1.40 1.56
Model 2 1.08 0.97 1.45 1.33 Model 2 0.57 0.86 0.66 1.21
Model 3 0.87 1.05 1.61 1.28 Model 3 0.63 0.66 0.95 0.82
Model 4 1.09 1.17 1.66 1.50 Model 4 0.72 0.69 1.01 1.40
Mean 1.04 1.07 1.44 1.34 Mean 0.72 0.85 1.00 1.25

Max Model 1 3.40 3.20 5.80 4.90 Model 1 3.40 2.20 3.90 4.20
Model 2 3.10 2.70 2.80 3.10 Model 2 2.60 2.80 2.50 3.00
Model 3 2.20 3.70 5.00 5.10 Model 3 1.60 1.70 4.70 2.50
Model 4 3.40 3.20 5.10 5.30 Model 4 3.10 2.30 3.50 4.80
Mean 3.03 3.20 4.68 4.60 Mean 2.68 2.25 3.65 3.63

Min Model 1 0.00 0.00 0.00 0.00 Model 1 0.00 0.00 0.00 0.00
Model 2 0.00 0.00 0.00 0.00 Model 2 0.00 0.00 0.00 0.00
Model 3 0.00 0.00 0.00 0.00 Model 3 0.00 0.00 0.00 0.00
Model 4 0.00 0.00 0.00 0.00 Model 4 0.00 0.00 0.00 0.00
Mean 0.00 0.00 0.00 0.00 Mean 0.00 0.00 0.00 0.00

radiographs at 150 projection views and i-CAT preference, both images showed that at projec-
were not different statistically in image 1. When tion views of 76, the lateral cephalogram is not
assessing clinician preferences, in image 2, at 76 statistically significant different from the one
projection views the panoramic radiographs were generated by the i-CAT software. The study
not statistically different from the i-CAT software. found that the image quality of the lateral
So, this shows that the projection views could be cephalogram with reduced number of projection
reduced to 150 (image 1) and 76 (image 2) while views (76) was similar to i-CAT. Both PA Ceph
maintaining the image quality. images show that 300 projection views were
In lateral cephalogram images, image 1 at 300 consistently ranked higher than the PA ceph of i-
had a higher mean rank of preference than CAT software. Among the pairs, PA Ceph 150 was
the images produced by i-CAT software. In shown to be not statistically significant from PA

Figure 10. The average deviations of lateral and PA cephalometric landmarks at different reduced projections
views of ASD-POCS data sets (in millimeter).
 Implementation of ultra-low-dose CBCT for routine 2D orthodontic diagnostic radiographs 243

ceph i-CAT showing that image quality of PA parameters for different FOVs. For the 6″ FOV
Ceph 150 projection views is similar to the PA used for bitewings, the mA could be reduced to
Ceph derived from i-CAT software (Table 5). 5 mA and for 9″ and 12″ FOVs, it could be lessened
to 2 mA while maintaining the image quality.
The field of view determines the overall
Discussion amount of ionizing radiation. Since the effective
dose of the ionizing radiation is computed from a
Cone-beam computed tomography has its place
weighted summation of effective dose to exposed
in the diagnosis and treatment planning of
organs, adding or removing some organs from
complex cases. The use of CBCT is increasing as
the path of the x-ray beam will affect the effective
practitioners are adding a third dimension to the
dose. For example, limiting the area of exposure
diagnosis and treatment planning of normal
to a smaller FOV of either jaw reduces the dose
patients. The measurements are changing from
compared to the exposure of the full craniofacial
lines to volumes. The real question is whether
area. In orthodontics, if practitioners are not
this third dimension should be used for all
using CBCT for routine purposes, it is possible to
normal patients despite the radiation dose.
reduce the dose without reducing the radiation
CBCT radiation dosage is affected by scan time;
exposure by exposing only the region of interest
the scanner settings used for the radiation expo-
without exposing the entire craniofacial area.
sure (120 kVp, 5 mA and 0.3 mm voxel); pulsed
Ludlow and Walker9 showed that reducing
radiation versus continuous beam; amount, type,
exposure parameters such as scan time, mA,
and shape of the x-ray radiation filter; full 3601
and number of projection views could reduce
rotation versus 1801 rotation; and the size of the
the radiation dose dramatically. The study
FOV (limited or full).8 CBCT systems vary in the
recommended quickscanþ protocols for low
availability of operator-controlled settings.
exposure scans; however, reported that the
There have been technological advancements
quality of the images reconstructed using a low
both in software and hardware since the first
exposure scan were significantly deteriorated when
commercially available CBCT machine. Advance-
compared to standard protocol. In the medical
ments in hardware (radiation sensor technology
literature, problems of insufficient data linked to
and x-ray generation, ability to change the expo-
sparse (reduced) projection views or gaps in the
sure parameters, availability of different FOVs,
projection data of CT and CBCT have been studied.
decreasing the scan time), software (computer
Insufficient data produce artifacts when the
technology and reconstruction algorithms), and
standard filtered back projection (FBP) algorithm
operator knowledge have made great contributions
is used. These artifacts have been resolved using
towards decreasing the radiation exposure. The
other mathematical algorithms. Algorithms resolve
most common problem faced in decreasing the
this issue either by interpolating or extrapolating
radiation dose is a decrease in the image quality.
the missing data from the existing data, or by using
Kwong et al.7 have shown that radiation exposure
iterative algorithms from existing measurements.10
can be lessened by changing the exposure
This same principle can be used to intentionally
create insufficient data to reduce the ionizing
Table 5. Mean ranks of quality assessment of recon-
structed orthodontic radiographs derived from CBCT radiation exposure and retrieve the data using an
on a visual analog scale (Friedman test) algorithm. The image quality of the images
produced using the algorithms could be
Posterior–
Panoramic Lateral Anterior evaluated for their clinical use.
Image Cephalogram Cephalogram The current study was designed to evaluate
Number of Image Image Image Image Image Image one of these methods. As mentioned above, one
Projections Views 1 2 1 2 1 2 of the major methods that can be utilized to
reduce the amount of ionizing radiation to
33 1.42 1.06 1.16 1.71 1.29 1.16 orthodontic patients is to reduce the number of
79 1.89 2.57 3.21 2.62 2.55 2.75
150 3.58 3.76 2.86 3.93 3.53 3.14 projections used to construct the 3D volume. The
300 4.21a 4.64a 3.95a 3.97a 4.04a 4.23a ASD-POCS algorithm was utilized for this pur-
i-CAT 3.90 2.97 3.82 2.76 3.59 3.71 pose. Using the algorithm, four reduced pro-
a
Number is highest mean rank. jection views data sets were created and used
244 Kusnoto et al

to derive 2D lateral and PA cephalometric from different projection views of ASD-POCS

radiographs. The ability to accurately locate reconstructions shows a value that exceeds the
cephalometric landmarks on these 2D images 1.5 mm error margin. But when looking at the
was studied as whether these various projection original data used to calculate the average, some
views data sets are clinically diagnostic. The dis- data actually exceeded the 1.5 mm margin
tance between the landmarks located on the (Fig. 11B).
cephalometric radiographs derived from the Furthermore, one column of the data illus-
reduced projection views data sets were compared trated in Fig. 11B can be further broken down
to those of the full data set. The reduced projec- into more detailed data. Every column in Fig. 10
tion views data set was judged to be clinically represents the average of all the landmarks used
accurate if the average distance of all the land- to calculate the data of a specific orientation
marks on that data set was within 1.5 mm from the (lateral or PA) of a specific model (1, 2, 3, and 4)
corresponding landmarks on the full data set. using a specific projection views data set (UC300,
A clinically acceptable reference mean of 1.5 UC150, UC76, and UC39). Fig. 12 shows the
millimeters (mm) of accuracy was used. Therefore, detailed data of landmarks contained in the
if the absolute distance between landmark tracings lateral cephalometric radiograph of model 3
was equal to 1.5 mm or was less than 1.5 mm, the using UC300 data set. Evident in this column of
landmark location was considered clinically accu- data, even though the mean of the landmarks in
rate. This 1.5 mm margin of error is based on several this data set is far below the 1.5 mm limit, some
studies that confirm this number as the reference individual landmark distances used to calculate
mean used for accuracy. In the study conducted by the mean exceed the limit.
Baumrind and Frantz, the results showed that errors In conclusion, looking at the mean values in
in landmark identification are too great to be Fig. 12, one can easily assume that none of the
ignored even when tracing the same lateral ceph- landmarks exceeded our accuracy limit of 1.5 mm.
alometric images. But again, inexperienced first-year However, many of these landmarks actually did.
orthodontic residents were asked to trace the images Using the average alone to describe the data can
in their study. This might contribute to the mag- lead to false interpretations.
nitude of errors we see in that study. They also found As discussed before and shown in Fig. 2 and
that the magnitude of error varies greatly from Table 1, only 15 landmarks were selected and
landmark to landmark.11 In another study, used in the lateral cephalometric analysis. In a
Baumrind et al.12 found an average positioning separate data sheet, the original 65 lateral
error of approximately 1.5 mm. This was also cephalometric landmarks were used as well.
confirmed by Sagun et al. who compared the The same conclusion derived from this data set
relative accuracies of three computerized growth but with larger standard deviation. This confirms
prediction methods based on lateral cephalograms. the preliminary results.
That study showed a clinically acceptable range The differences found in the image quality of
between 1.6 mm and 1.4 mm.13 Additionally, a study radiographs with different number of projection
by Toepel-Sievers and Fischer-Brandies,14 which views in our study could be attributed to sub-
tested the validity of Ricketts’s growth prediction, jective analysis of the images. Quality assessment
considered length measurements to be clinically of two-dimensional images reconstructed from
useful if the absolute error was less than 1.8 mm. CBCT by reconstruction algorithm ASD-POCS
Calculating the average can help simplify the has shown that this algorithm has the ability to
process of analysis especially for large amounts of reduce the number of projection views for clin-
data. It is often tempting to overlook the many ical diagnosis. The use of existing algorithm for
details in the data and rely on the mean value. reconstruction of images has shown to degrade
However, by doing this, a large amount of image quality of images with reduced number of
information can be very easily obscured and projection views.8 ASD-POCS improved the
unconsciously deviates the argument into a image quality of radiographs reconstructed from
totally different direction than what the data CBCT with same number of projection views, i.e.,
really illustrate. UC 300 had higher mean rank and was preferred
As shown in Figs. 10 and 11A, none of the than i-CAT control. The minimum number of
lateral or PA cephalometric radiographs derived projection views reconstructed radiographs
 Implementation of ultra-low-dose CBCT for routine 2D orthodontic diagnostic radiographs 245

Figure 11. (A) The average deviations of lateral and PA cephalometric landmarks of the four subjects (n ¼ 4) and
(B) average deviation of lateral and PA cephalometric landmarks for individual subjects.

(Pano, LC, and PA Ceph) UC 39 consistently had Clinical application of reduced radiation
a lower rank and was not accepted for their ASD-POCS in repeated imaging
diagnostic quality. The number of projection
views could be reduced to 150 most of the time Orthodontists, surgeons (e.g., craniofacial, oral
for diagnostic ability of 2D radiographs but also and maxillofacial, plastic and reconstructive) as
to 76 for lateral cephalogram landmark identi- well as pediatric dentists are among the dental
fication. It is worth noting that the lateral specialties that manage craniofacial deformities
cephalogram required a lower number of pro- and malocclusions in newborn, children and
jection views to be regarded as diagnostic. One young adults. Sometimes abnormal development
possibility is that majority of the participants have of the dental and craniofacial skeletal region
undergone orthodontic training, which focuses (such as openbites, under-developed and con-
on rigorous cephalometric tracing and analyses, stricted maxillas, mandibular deformities, and
thus landmark identification is easier than that impacted teeth) requires longitudinal follow up
with the less commonly used PA Ceph. usually via a longitudinal series of radiographs
246 Kusnoto et al

Figure 12. Detailed landmark values for UC300 data set derived from one of the subjects.

(2D cephalometric radiographs). The success of individual utilize standard and well-calibrated
growth guidance/modification of dento-skeletal methods of assessment (currently through uti-
and craniofacial structures in a growing lization of 2D cephalometric radiographs).
 Implementation of ultra-low-dose CBCT for routine 2D orthodontic diagnostic radiographs 247

Cephalometric analysis is used to study the dental quality to the i-CAT control. This study suggests
and skeletal relationships of the head when that the use of reconstruction algorithm ASD-
planning and guiding orthodontic treatment for POCS can produce good image quality, con-
problems ranging from straightforward ortho- sequently reducing patient radiation exposure.
dontic cases to more complex orthognathic/
craniofacial cases involving surgical procedures.
Since first available in the 1930s, most cephalo- References
metric analyses have been two-dimensional. 1. Larson BE. Cone-beam computed tomography is the
Other inherent errors such as variability in sub- imaging technique of choice for comprehensive ortho-
ject’s head positioning during cephalometric dontic assessment. Am J Orthod Dentofacial Orthop. 2012;
radiographic acquisition and magnification error 141(4). http://dx.doi.org/10.1016/j.ajodo.2012.02.009.
2. Halazonetis DJ. Cone-beam computed tomography is not
compounded the overall error of the 2D ceph-
the imaging technique of choice for comprehensive
alometric projection data. The development and orthodontic assessment. Am J Orthod Dentofacial Orthop.
utilization of CBCT will indeed reduce those 2012;141(4). http://dx.doi.org/10.1016/j.ajodo.2012.02.010.
compounded errors thus increasing accuracy. 3. American Academy of Oral and Maxillofacial Radiology.
In this study, assessments of ultra-low-dose Clinical recommendations regarding use of cone beam
computed tomography in orthodontics. Position state-
radiation derived from implementation of ASD-
ment by the American Academy of Oral and Maxillofacial
POCS algorithm opens a possibility of using Radiology. Oral Surg Oral Med Oral Pathol Oral Radiol.
CBCT as standard imaging protocol in the near 2013;116:238–257. http://dx.doi.org/10.1016/j.OOOO.
future with significantly reduced radiation 2013.06.002.
exposure when only 2D views are needed (as in 4. Scarfe WC, Farman AG. What is cone-beam CT and how
routine orthodontic diagnostic radiographs). does it work?Dent Clin North Am 2008;52:707–730.
5. Qiu W, Pengpan T, Smith ND, et al. Evaluating iterative
algebraic algorithms in terms of convergence and image
Conclusions quality for cone beam CT. Comput Methods Programs
Biomed. 2013;109:313–322.
This study showed that there is no statistically 6. Park JC, Song B, Kim JS, et al. Fast compressed sensing-
significant mean difference in the majority of based CBCT reconstruction using Barzilai–Borwein for-
mulation for application to on-line IGRT. Med Phys.
landmarks in cephalometric radiographs derived
2012;39:1207–1217.
from ASD-POCS reconstructed CBCTs with 300, 7. Kwong JC, Palomo JM, Landers MA, et al. Image quality
150, 76, and 39 projection views. The mean 2D produced by different cone-beam computed tomography
cephalometric landmark identification measure- settings. Am J Orthod Dentofacial Orthop. 2008;139:317–327.
ments in the radiographs did not exceed the 8. Brooks SL. CBCT dosimetry: orthodontic considerations.
Semin Orthod. 2009;15:14–18.
1.5 mm clinically acceptable reference measure-
9. Ludlow JB, Walker C. Assessment of phantom dosimetry
ment set to evaluate the accuracy of the tracings. and image quality of i-CAT FLX cone-beam computed
These findings may influence companies and tomography. Am J Orthod Dentofacial Orthop. 2013;144:
clinicians to consider decreasing radiation dose 802–817.
and time of exposure to better serve the patients. 10. LaRoque SJ, Sidky EY, Pan X. Accurate image recon-
The qualitative assessment of images con- struction from few-view and limited-angle data in dif-
fraction tomography. J Opt Soc Am A Opt Image Sci Vis.
structed using ASD-POCS reconstruction algo- 2008;25:1772–1782.
rithm also produced useful results. For all 11. Baumrind S, Frantz R. The reproducibility of head film
reconstructed 2D radiographs: panoramic, lateral measurements: 1. landmark identification. Am J Orthod.
cephalometric, and posterior–anterior cephalo- 1971;60:111–126.
metric, ASD-POCS image reconstructions with 12. Baumrind S, Miller D, Molthen R. The reliability of head
film measurements. 3. Tracing superimposition. Am J
300 projection views produced similar or better
Orthod. 1976;70:617–644.
results when compared to the i-CAT control in 13. Sagun M, Kusnoto B, Evans C, et al. Assessment of
terms of clinician’s individual preference. In the Computerized Cephalometric Growth Prediction: A Compar-
panoramic radiograph and posterior–anterior ison of Three Methods. University of Illinois at Chicago,
cephalogram, the images reconstructed from the ProQuest, UMI Dissertations Publishing, 2012.
p. 1513864.
ASD-POCS 300 and 150 had similar image quality
14. Toepel-Sievers C, Fischer-Brandies H. Validity of the
to the i-CAT control. In the lateral cephalogram computer-assisted cephalometric growth prognosis VTO
reconstructions, the images reconstructed from (visual treatment objective) according to Ricketts. J Orofac
the ASD-POCS 300, 150, and 76 had similar image Orthop. 1999;60:185–194.
Assessment of image quality in maxillofacial
cone-beam computed tomography imaging
Casey Gamache, Jeryl D. English, Anna M. Salas-Lopez, John Rong, and
Sercan Akyalcin

The aim was to evaluate image quality produced by a cone-beam CT (CBCT)

scanner under various imaging modes. A modulation transfer function (07-
538, Nuclear Associates, Carle Place, NY) was placed in an anthropomorphic
head phantom (RS-250, Radiology Support Devices, Long Beach, CA). Overall,
67 CBCT scans were obtained by modifying the entire range of kilovolt (kV)
and milliampere (mA) settings, using the CS 9300 CBCT scanner (Carestream
Health, Inc., Rochester, NY). A total of 24 individual raters rated the images to
evaluate line pairs visible. Upon completion of the evaluation process, 4 kV
and mA combinations (85 kV  12 mA, 90 kV  4.0 mA, 60 kV  6.3 mA, and
65 kV  2.5 mA) were selected for further investigation. In all 10 dry skulls
were scanned to measure the defect size of simulated root resorption in 4
different modes. Intraclass correlation (ICC) analysis was used to compare
the agreement between the defect sizes measured from the scans compared
to the actual physical volume. Low kV settings (60 kV) used in combination
with moderate-to-high mA settings (4.0–15 mA) produced the best-perceived
image quality, resulting in up to 56% reduction in the total radiation amount
(mGy cm2) generated. High kV settings (85–90 kV) used in combination with
high mA settings (8.0–15 mA) produced the lowest image quality. All 4
selected imaging modalities presented a high level of accuracy in detection of
root resorption volume compared to actual physical volume (ICCs between
0.93 and 0.98). Radiation exposure from CBCT scans can be reduced while
maintaining adequate image quality and diagnostic accuracy by performing
scans using low kV and moderate-to-high mA settings. (Semin Orthod 2015;
21:248–253.) Published by Elsevier Inc.

Introduction construct study casts, and stereolithographic

D
models from a single scan, and guide the way
ental cone-beam computed tomography
we treat and visualize our patients.2 CBCT offers
(CBCT) is rapidly evolving and is becoming
enhanced diagnostic precision compared to con-
increasingly popular among dental specialists,
ventional multislice CT3 and also supplements or
now with over 3000 units installed in the United
replaces traditional orthodontic records for
States.1 With today’s technology, it is possible to
patients with significant skeletal asymmetries,
produce several types of radiographic images,
impacted canines, and root resorption.4 One
of the biggest controversies with the use of CBCT
Private Practice, Yakima, WA; Department of Orthodontics, in dental imaging is the amount of generated
School of Dentistry, The University of Texas Health Science Center at
Houston, Houston, TX; Department of Imaging Physics, MD
radiation. Although there is a lack of definitive
Anderson Cancer Center, Houstan, TX. data, largely due to the fact that there are
Address correspondence to Sercan Akyalcin, DDS, PhD, Depart- almost 60 CBCT scanners available on the
ment of Orthodontics, School of Dentistry, The University of Texas market,5 recent reviews signify the importance
Health Science Center at Houston, Houston, 7500 Cambridge St,
of generated radiation by CBCT scanning as a
Suite 5130, Houston, TX 77054. E-mail: [email protected].
edu
cause for concern.6,7
Published by Elsevier Inc.
CBCT scanning is often used for investigation
1073-8746/15/1801-$30.00/0 of external root resorption because of its ability
http://dx.doi.org/10.1053/j.sodo.2015.07.002 to produce accurate, anatomically correct images

Seminars in Orthodontics, Vol 21, No 4, 2015: pp 248–253 248

 Assessment of image quality in maxillofacial CBCT imaging 249

of the root structure.8 The detection of root

resorption on lateral incisors caused by the
adjacent impacted canines was found to be
highly accurate with no significant differences
in detection of severity of root resorption
between various CBCT scanners.9 Furthermore,
when comparing CBCT and panoramic imaging
for detection of lateral incisor root resorption
associated with impacted maxillary canines, only
30% of the panoramic images but 50% of CBCT
images detected root resorption.10
When compared to medical CT with a matched
radiation exposure, CBCT scanning has a com-
parable high-contrast resolution but inferior low-
contrast resolution.11 The poor low-contrast res-
olution results in difficulty differentiating between
the various soft tissues in the image. The differ-
ences in image quality between the medical CT
and CBCT can be decreased considerably by use
of improved beam hardening and scatter cor-
rection.12 Additionally, reducing the milliampere
(mA) setting during scans limits the amount
of generated radiation but also dramatically
degrades image quality due to excessive noise;
however, noise in the reconstructed CBCT images
acquired with a low mA can be greatly suppressed
after sinogram domain image processing without Figure 1. Anthropometric phantom used in this study.
noticeable sacrifice of spatial resolution.13 It is of
crucial importance to determine the optimal called Modulation Transfer Function (MTF), also
scanning settings of kilovolt (kV) and mA that known as spatial frequency response. The higher
will allow adequate image quality and yet help the image sharpness and detail, the higher the
minimize the amount of radiation exposure in line pair score should be. An MTF device (07-538,
maxillofacial CBCT imaging. Nuclear Associates, Carle Place, NY) was used to
The purpose of this study was to evaluate the allow for visualization of line pairs on the CBCT
perceived image quality of the scans produced by scans. The anthropometric phantom was seated
the CS 9300 CBCT scanner (Carestream Health, on a 41.5-in tall structure that was firmly placed
Inc., Rochester, NY) under varying exposure on the floor and immobilized. The phantom’s
parameters and to compare whether there is a chin rested on the chin support of the Kodak
difference in the diagnostic accuracy of 4 select 9300’s 3D scanning head support, while the
kV and mA combinations. stabilizing side arms were also engaged to keep
the phantom immobilized while the scans were
performed.
Materials and methods
The MTF was placed between the 2 most infe-
An anthropometric phantom (RS-250, Radiology rior slices, corresponding roughly to the Frankfort
Support Devices, Long Beach, CA) composed of a horizontal. A level was used to help align the
real, unidentifiable human skull surrounded by a phantom parallel to the floor in both anterior–
tissue-like composite was utilized. All CBCT scans posterior and medial–lateral directions. Test
were obtained using the CS 9300 CBCT scanner scans were performed using the manufacturer-
(Carestream Health, Inc., Rochester, NY). The recommended settings for a medium-sized adult
phantom was comprised of 4 slices that stacked on (90 kV, 4 mA, 898 mGy cm2, 6.4 s, 250 m voxel size)
top of each other (Fig. 1). Resolution and per- to evaluate if the line pairs of the MTF could be
ceived sharpness was evaluated by a parameter visualized in a single axial slice in the current
250 Gamache et al

Table 1. Scan settings (mA  kV) evaluated in this computer. The raters marked the box that
study with the corresponding total radiation exposure corresponded to the number of line pairs per
in mGy cm2 millimeter that they were able to see for each of
kV the numbered scans.
mA 90 85 80 75 70 65 60 Intraclass correlation analysis (ICC) was per-
formed between the 2 groups of raters and a high
15   2436 2016 1628 1270 944 level of agreement was found between the groups
12  2309 1949 1613 1302 1016 755 in the evaluation of the line pairs (ICCs ranged
10 2246 1924 1624 1344 1085 847 629
8 1797 1540 1299 1075 868 677 503 between 0.78 and 0.98). Therefore, the mean
6.3 1415 1212 1023 847 684 533 396 values of the visible line pair evaluations were
5 1123 962 812 672 543 423 315 calculated according to the total sample.
4 898 770 650 538 434 339 252
3.2 719 616 520 430 347 271 201 The lowest (85 kV  12 mA) and highest
2.5 561 481 406 336 271 212 157 (60 kV  6.3 mA) rated scans were selected for
2 449 385 325 269 217 169 126 the follow-up tests. In addition to these 2 settings,
the manufacturer-suggested, medium adult scan
position of the phantom and MTF. The position of setting (90 kV  4.0 mA) as well as the lowest
the phantom and MTF were adjusted until the line radiation exposure setting (65 kV  2.5 mA) from
pairs of the MTF could be visualized properly in a the group of 10 highest-rated scans were included
single axial slice. in the follow-up set of scans. For these scans 10 dry
After establishing the proper positioning of human skulls were used. Each of these skulls had
the phantom and MTF, 67 scans (Table 1) were at least 1 tooth (4 premolars, 3 incisors, 2 molars,
performed by altering the kV and mA without and 1 canine) that was readily removable with the
having the need to move the phantom and/or root surface intact. A small, 1 mm diameter bur
MTF to allow for maximum consistency through- with a flat end was used to create small (1 mm
out the imaging process. The kV was adjusted in deep and 1 mm wide) defect in the apical third of
increments of 5 kV from the 90 kV maximum to the root of each tooth. Each of the 10 skulls was
the 60 kV minimum, while the mA was adjusted scanned with the 4 aforementioned settings.
through every possible increment from the After all 40 scans were completed, the physical
15 mA maximum to the 2.0 mA minimum. The defects were measured and recorded using a
scanner had pre-set limitations on the kV and mA digital caliper in increments of 0.01 mm. The
combinations to prevent an excessive amount of physical volume was calculated manually by using
generated radiation; therefore, 3 possible scans the 3 linear measurements obtained with the
were eliminated: (1) 90 kV, 15 mA; (2) 90 kV, digital caliper. The digital volume of each
12 mA; and (3) 85 kV, 15 mA. The scan size defect was measured and recorded using the
(17 cm  11 cm), voxel size (250 m), and scan Carestream 3D CBCT software in increments
time (6.40 s) were kept constant to minimize the of 0.1 mm. The computed CBCT volume was
variables involved and to reproduce conditions
similar to a scan needed for orthodontic imaging.
The full radiation exposure (mGy cm2) from the
resultant 67 image scans were also recorded
using the readings on the scanner (Table 1).
Each scan was opened and the same axial slice
(Fig. 2) displaying the MTF was located, as
identified by the axial millimeter measurement
in the Carestream 3D CBCT software
(Carestream Health, Inc., Rochester, NY). Each
of the 67 axial slices containing the MTF was
saved into a personal computer. Overall, 12
orthodontic residents and 12 orthodontic
faculty members that were blinded to the
settings evaluated each of the scans without a Figure 2. Axial slices with the MTF device used for line
time limitation in a dark room on a personal pair evaluation.
 Assessment of image quality in maxillofacial CBCT imaging 251

calculated manually by using the 3 linear meas- Table 3. Agreement between the physical size of the
urements obtained with the CBCT software. root resorption defect and the select scans
Results of the 2 different measurement methods Scanning Setting ICC
were then compared with the use of ICCs.
85 kV  12 mA 0.95
90 kV  4.0 mA 0.93
Results 60 kV  6.3 mA 0.98
65 kV  2.5 mA 0.96
Table 2 presents the line pair evaluations of the
67 scans. The best-perceived images had low kV
settings in combination with moderate-to-high such as CBCT imaging, continue to play an ever-
mA settings. The highest-rated scan had scanning increasing role in the diagnosis and treatment
parameters of 60 kV and either 6.3 or 15 mA and planning in orthodontic cases, especially in those
equaled to a mean of 1.170 line pairs/mm. with significant skeletal asymmetries, impacted
According to the manufacturer, this produces a canines, and root resorption.4 However, the ability
total radiation exposure of 396 mGy cm2 and to select the optimal scanning parameters to
944 mGy cm2, respectively. The lowest-rated obtain the best image quality while minimizing
image quality was found when high kV settings the amount of generated radiation is a constant
were used in combination with high mA settings. challenge. The benefits of CBCT imaging range
The lowest-rated scan had scanning parameters from improved diagnosis and surgical accuracy in
of 85 kV and 12 mA and equaled to a mean of skeletal asymmetry cases14 to detection of even
0.991 line pairs/mm. According to the manu- mild levels of root resorption15 that help the
facturer, this produces a total radiation exposure orthodontist to predict future damage to teeth
of 2309 mGy cm2. On the other hand, the with continued treatment. Several studies have
manufacturer-recommended kV and mA settings previously evaluated external root resorption, but
of 90 kV  5.0 mA, 90 kV  4.0 mA, 85 kV  none has attempted to determine the scanning
4.0 mA, and 80 kV  4.0 mA produced total parameters that allow adequate image quality for
radiation exposures of 1123, 898, 770, and detection and measurement of the root defect,
650 mGy cm2, respectively. while minimizing the amount of generated
Comparison of the actual root defect size with a radiation. We hypothesized that selecting the
caliper, as the gold standard, to the digital estima- highest settings of the CBCT scanner does not
tion of the root defect volume obtained from the 4 necessarily enhance image quality. Accordingly,
CBCT scans with select scan parameters revealed a this study evaluated the full range of scanning
near perfect agreement, with the ICCs ranging from parameters on the CS 9300 CBCT scanner to
0.935 to 0.980 (Table 3). determine which settings would allow the best
combination of image quality and generated
radiation.
Discussion The first part of the study gave insight as to the
As technology rapidly moves forward, both inside relationship between the image quality produced
and outside of dentistry, 3D imaging techniques, by the CS 9300 CBCT and the various combi-
nations of kV and mA available. This led to the
Table 2. Line pair evaluation of the scans finding that better image quality, defined as the
kV ability to differentiate contrasting line pairs, can
mA 90 85 80 75 70 65 60 be found while using lower radiation exposure
via low kV and moderate-to-high mA settings as
15   1.004 1.037 1.062 1.100 1.170 compared to the manufacturer-recommended
12  0.991 1.037 1.045 1.070 1.120 1.166 settings that actually produced higher levels of
10 0.995 1.037 1.037 1.033 1.066 1.112 1.120
8 1.000 1.037 1.033 1.054 1.070 1.087 1.145 radiation exposure. These results are consistent
6.3 1.037 1.012 1.033 1.083 1.075 1.125 1.170 with the findings of Xu et al.16 who evaluated the
5 1.016 1.012 1.054 1.037 1.104 1.116 1.129 same scanner used in our study and found that
4 1.016 1.037 1.045 1.054 1.087 1.100 1.133
3.2 1.016 1.008 1.054 1.070 1.091 1.116 1.095 the various scanning programs, including sinus
2.5 1.020 1.037 1.058 1.079 1.112 1.116 1.062 and TMJ scans, could be manually modified with
2 1.045 1.062 1.045 1.083 1.095 1.083 1.058 regard to kV and mA to maintain adequate image
252 Gamache et al

quality but reduce the amount of generated caries detection, then periapical radiographs are
radiation by about 30%. still recommended as the diagnostic accuracy for
The second part of the study evaluated the caries detection with CBCT is less than with
ability of the CS 9300 CBCT scanner to detect and conventional periapical radiographs.22 There-
measure simulated root resorption. The results fore, it is still suggested that the CBCT scanning
showed that nearly perfect correlation exists technology is only used in select cases as opposed
between measurement of simulated root resorp- to being the standard for all patients due to the
tion with a digital caliper, the gold standard, and increased radiation exposure with limited benefit
measurement of the same simulated root resorp- in cases without facial asymmetry, impactions,
tion with the Carestream 3D CBCT imaging soft- root resorption, or interdisciplinary needs.
ware. These findings are consistent with those of While the results of this study showed nearly
Alqerban et al.9 who found that all 6 of the CBCT perfect agreement between the CBCT measure-
scanners tested, all using their pre-set parameters ments and digital caliper measurements, it is
that were held constant throughout scans but important to note that these measurements were
differed between the different scanners, had high taken on a phantom and dry skulls that were
accuracy with no significant differences between motionless. In clinical practice, all CBCT scans will
them in the detection of the severity of root have inherent distortion from slight patient move-
resorption. In a research attempt by Liedke et al.,17 ment that would reduce the quantification accuracy
it was reported that root resorption defects can be of the CBCT software.23,24 Additionally, a live
detected better when using high (200 m) resolution patient may require slightly different parameter
and medium (300 m) resolution CBCT images as settings because of radiation attenuation in the soft
compared to low (400 m) resolution CBCT images. tissues.25 Future research should evaluate the ability
All CBCT images evaluated in our study had of CBCT software to obtain highly accurate image
a resolution of 250 m, which may explain the quality in living subjects while utilizing low kV and
high level of agreement of the measurements moderate-to-high mA CBCT settings.
performed on the scans when compared to
manual measurements. Since root resorption is
most difficult to detect in the apical third of the Conclusions
root,18–20 our results showed that the CS 9300 The total radiation exposure from CBCT scans
scanner is fully capable of diagnosing the defect can be reduced by while maintaining adequate
size, using both the manufacturer-recommended image quality using low kV and moderate-to-high
and user-defined settings of kV and mA. mA settings rather than the manufacturer-
According to our findings, the amount of total recommended settings as long as the individual
radiation exposure for a 17 cm  11 cm CBCT characteristics of patients, i.e., attenuation allows
scan could be reduced by 56% by changing the for such modifications.
scanning parameters for a medium-sized indi- CS 9300 CBCT scanner was able to perform
vidual from 90 kV and 4 mA (898 mGy cm2) to highly reproducible results in estimating the
60 kV and 6.3 mA (396 mGy cm2). These results simulated root resorption in the apical third of
show that it is possible to obtain descent-quality the root regardless of the kV and mA settings.
CBCT images while limiting the amount of
generated radiation through thoughtful selec-
tion of scanning parameters. References
If the patient is in need of a full set of periapical 1. Hatcher D. Operational principles for cone-beam com-
radiographs, which, depending on the type of puted tomography. J Am Dent Assoc. 2010;141:3S–6S.
2. Seet K, Barghi A, Yartsev S, Van Dyk J. The effects of field-
collimation used, can alone expose the patient to of-view and patient size on CT numbers from bone-beam
just as much radiation as a CBCT,21 then the computed tomography. Phys Med Biol. 2009;54:6251–6262.
CBCT may be a better choice to replace the 3. Suomalainen A, Vehmas T, Kortesniemi M, Robinson S,
combination of panoramic, cephalometric, and Peltola J. Accuracy of linear measurements using dental
cone beam and conventional multislice computed
full set of periapical radiographs. This holds true
tomography. Dentomaxillofac Radiol. 2008;37:10–17.
if the periapical radiographs are for closer 4. Kapila S, Conley R, Harrell W Jr. The current status of
examination of bone level. However, if the cone beam computed tomography imaging in orthodon-
main purpose of the periapical radiographs is tics. Dentomaxillofac Radiol. 2011;40:24–34.
 Assessment of image quality in maxillofacial CBCT imaging 253

5. Available at: http://3dorthodontist.com/CBCT_Machines. 16. Xu J, Reh D, Carey J, Mahesh M, Siewerdsen J. Technical

html. Accessed 19.02.13. assessment of a cone-beam CT scanner for otolaryngology
6. Baumrind S. The road to three-dimensional imaging in imaging: image quality, dose, and technique protocols.
orthodontics. Semin Orthod. 2011;17:2–12. Med Phys. 2012;39:4932–4942.
7. Scholz R. The radiology decision. Semin Orthod. 2011;17:15–19. 17. Liedke G, da Silveira H, da Silveira H, Dutra V,
8. Lund H, Gröndahl K, Gröndahl H. Cone beam computed de Figueiredo J. Influence of voxel size in the diagnostic
tomography for assessment of root length and marginal ability of cone beam tomography to evaluate simulated
bone level during orthodontic treatment. Angle Orthod. external root resorption. J Endod. 2009;35:233–235.
2010;80:466–473. 18. da Silveira H, da Silveira H, Liedke G, Lermen C,
9. Alqerban A, Jacobs R, Fieuws S, Nackaerts O, Willems G. Dos Santos R, de Figueiredo J. Diagnostic ability of
Comparison of 6 cone-beam computed tomography computed tomography to evaluate external root resorp-
systems for image quality and detection of simulated tion in vitro. Dentomaxillofac Radiol. 2007;36:393–396.
canine impaction-induced external root resorption in 19. Stephanopoulos G, Mikrogeorgis G, Lyroudia K. Assess-
maxillary lateral incisors. Am J Orthod Dentofacial Orthop. ment of simulated internal resorption cavities using
2011;140:129–139. digital and digital subtraction radiography: a comparative
10. Alqerban A, Jacobs R, Souza P, Willems G. In-vitro study. Dent Traumatol. 2011;27:344–349.
comparison of 2 cone-beam computed tomography 20. Lermen C, Liedke G, da Silveira H, da Silveira H, Mazzola
systems and panoramic imaging for detecting simulated A, de Figueiredo J. Comparison between two tomo-
canine impaction-induced external root resorption in graphic sections in the diagnosis of external root
maxillary lateral incisors. Am J Orthod Dentofacial Orthop. resorption. J Appl Oral Sci. 2010;18:303–307.
2009;136(764):e1–11. 21. Silva M, Wolf U, Heinicke F, Bumann A, Visser H, Hirsch E.
11. Yu L, Vrieze T, Brueswitz M, et al. Dose and image quality Cone-beam computed tomography for routine orthodontic
evaluation of a dedicated cone-beam CT system for high- treatment planning: a radiation dose evaluation. Am J Orthod
contrast neurologic applications. Am J Roentgenol. Dentofacial Orthop. 2008;133(640):e1–e5.
2010;194:W193–W201. 22. Valizadeh S, Tavakkoli M, Vasigh H, Azizi Z, Zarrabian T.
12. Elstrom U, Muren L, Petersen J, Grau C. Evaluation of Evaluation of cone beam computed tomography (CBCT)
image quality for different kV cone-beam CT acquisition system: comparison with intraoral periapical radiography
and reconstruction methods in the head and neck region. in proximal caries detection. J Dent Res Dent Clin Dent
Acta Oncol. 2011;50:908–917. Prospects. 2012;6:1–5.
13. Wang J, Li T, Liang Z, Xing L. Dose reduction for 23. Hanzelka T, Foltán R, Horká E, Sedý J. Reduction of the
kilovoltage cone-beam computed tomography in radia- negative influence of patient motion on quality of CBCT
tion therapy. Phys Med Biol. 2008;53:2897–2909. scan. Med Hypotheses. 2010;75:610–612.
14. Hsu S, Gateno J, Bell R, et al. Accuracy of a computer- 24. Miracle A, Mukherji S. Conebeam CT of the head and
aided surgical simulation protocol for orthognathic neck, part 1: physical principles. Am J Neuroradiol.
surgery: a prospective multicenter study. J Oral Maxillofac 2009;30:1088–1095.
Surg. 2013;71:128–142. 25. Schafer S, Stayman J, Zbijewski W, Schmidgunst C,
15. Levander E, Malmgren O. Evaluation of the risk of root Kleinszig G, Siewerdsen J. Antiscatter grids in mobile
resorption during orthodontic treatment: a study of C-arm cone-beam CT: effect on image quality and dose.
upper incisors. Eur J Orthod. 1988;10:30–38. Med Phys. 2012;39:153–159.
3D analysis and clinical applications of CBCT
images
Mohamed Bayome, Jae Hyun Park, YoonJi Kim, and Yoon-Ah Kook

Cone-beam computed tomography (CBCT) images is an essential element in

diagnosing and treatment planning patients in need of orthodontic and/or
orthognathic surgery. An accurate evaluation of the dental, skeletal, and soft-
tissue relationships through the normative values of three-dimensional (3D)
cephalometric parameters, specifically palatal and alveolar bone thickness,
mandibular body and maxillary basal curve length, and basal arch form have
been demanded. The normative values of innovative 3D cephalometric
parameters for palatal bone thickness in adolescents versus adults may be
helpful for clinicians to enhance the success in application and use of
temporary skeletal anchorage devices. Characteristics of alveolar bone thick-
ness, basal arch form, and facial asymmetry of both normal occlusion and
Class III malocclusion are also discussed with 3D CBCT images. In this article,
the various 3D analyses and clinical applications including bone thickness,
facial asymmetry, basal curve length, and basal arch form are addressed.
(Semin Orthod 2015; 21:254–262.) & 2015 Elsevier Inc. All rights reserved.

Introduction on evaluation of a limited sample of selected

normal occlusion volunteers.7,8
ephalometric analysis has been a vital
C component in the diagnosis and treatment
planning in orthodontics and orthognathic sur-
However, several inherent drawbacks existed
in all those analyses such as magnification,
superimposition of anatomical structures and
gery. Various cephalometric analyses have been
presentation of a two-dimensional (2D) projec-
reported throughout decades.1–4 Each analysis
tion of a three-dimensional (3D) object. These
was based on identification of several specific
may have led to the main factor affecting the
landmarks to calculate the linear and angular
reliability of cephalometric analyses: error in
relationships among them.
identification of landmarks and their projection
All the analyses aimed to evaluate the devia-
on 2D.9,10 Also, the implication of head ori-
tions in the skeletal and dentoalveolar relation-
entation increased the inquiry about the reli-
ships by comparing these relationships to
ability of these measurements.11,12
normative values. These values, in some analyses,
The introduction of cone-beam computed
have been derived from growth studies such as
tomography (CBCT) has resulted in the
Bolton5 and Burlington6 growth studies.
improvement of the quality of radiographic data
Meanwhile, in other analyses they were based
due to overcoming the disadvantages of the
conventional 2D radiographic techniques. The
The Catholic University of Korea, Seoul, South Korea; Department precision, accuracy, and reliability of landmarks
of Postgraduate Studies, the Universidad Autonoma del Paraguay, identification as well as linear and angular
Asuncion, Paraguay; Arizona School of Dentistry & Oral Health, measurements on CBCT images had been
A.T. Still University, Mesa, AZ; Graduate School of Dentistry, Kyung
comprehensively evaluated.13–16 High intra- and
Hee University, Seoul, South Korea; Department of Orthodontics,
Seoul St. Mary’s Hospital, The Catholic University of Korea, 505 inter-observer reliability of measurements were
Banpo-Dong, Seocho-Gu, Seoul 137-701, South Korea. reported.17,18 In addition, the image distortion
Address correspondence to Yoon-Ah Kook, DDS, PhD, Department and the spatial resolution of a CBCT machine
of Orthodontics, Seoul St. Mary’s Hospital, The Catholic University were assessed.19
of Korea, 505 Banpo-Dong, Seocho-Gu, Seoul 137-701, South Korea.
E-mail: [email protected]
The analysis of CBCT images has been val-
uable in numerous applications such as: (1)
& 2015 Elsevier Inc. All rights reserved.
1073-8746/15/1801-$30.00/0 diagnosis of impacted and supernumerary teeth,
http://dx.doi.org/10.1053/j.sodo.2015.07.003 in which the 3D images were superior to the

Seminars in Orthodontics, Vol 21, No 4, 2015: pp 254–262 254

 3D analysis and clinical applications of CBCT images 255

conventional 2D images,20–22 (2) treatment 3D cephalometric analysis

planning for placement of skeletal anchorage
The 3D analysis might represent a key to over-
devices, in which CBCT scans allowed for the
come all the traditional cephalometric dis-
evaluation of bone thickness and density in dif-
advantages. However, a well-established method
ferent areas in the mandible and maxilla,23–25 (3)
to digitize and analyze 3D radiographic images is,
3D volumetric assessment of the upper airway,26
yet, controversial. Kochel et al.36,37 developed a
(4) 3D analysis for diagnosis and treatment
3D soft-tissue analysis based on the data derived
planning of orthodontic and orthognathic
from 3D stereophotogrammetric images. How-
patients.27,28
ever, all the measurements were taken from the
Several advantages of CBCT were reported,
projections of the digitized points. Moreover,
including ability to assess the image from the
they evaluated correlation of the 3D soft-tissue
three planes, the real-size 3D images, and
data to variables retrieved from 2D lateral
absence of distortion or overlapping structures.29
cephalometric analysis. Also, Farronato et al.38
The fine adjustment of the head position is not
proposed a 10-point 3D analysis of CBCT images
essential during taking the image, because the
directly digitized on the rendered view. They
points keep their spatial relationships in 3D
reported the reliability and the reproducibility of
coordinates unchanged.30 Therefore, the
their method and compared it to 2D data.
reorientation of the images, on the contrary to
However, norms of the variables were not
the lateral cephalometric radiograph, is possible.
reported in their study probably due to the small
Moreover, the ease of landmark identification
sample size and the wide age range. More
and high precision of superimposing images have
recently, Cheung et al.28 reported 3D cephalo-
been reported.31,32
metric norms based on CBCT scans of Chinese
Significant differences were reported between
population. Bayome et al.27 proposed a new 3D
angular measurements performed on 2D post-
cephalometric analysis and evaluated the
eroanterior (PA) cephalograms and those on
relationships among skeletal and dentoalveolar
radiographs constructed from CBCT scans.33,34
variables. Their study has also provided the
In addition, Gribel et al.35 showed significant
norms of the 3D variables of a Korean normal
differences between measurements taken on a
occlusion population.
lateral cephalogram and those taken from a CBCT
scan. Therefore, they suggested a mathemati-
Segmentation of CBCT images
cal formula in an attempt to convert the 2D
cephalometric measurements into a 3D CBCT To achieve a sound method for 3D analysis of a
measurement. CBCT image, volume segmentation and image
Although the radiation dose of the CBCT is reorientation should be considered prior to
lower than the medical spiral CT, it is still higher landmark identification. Volume segmentation is
than that of a 2D cephalogram. However, this the allocation and separation of an anatomical
depends on the CBCT scanner’s specifications, structure or region of interest from the 3D vol-
the time of scanning, and the field of view. umes so that it can be viewed individually. The
Therefore, it is recommended to apply the 3D difficulty of segmentation is mainly due to the
cephalometric analysis to the cases that require variability and complexity of the biological tis-
comprehensive treatment. Clinicians should sues. Besides, the large size of the datasets and
always keep in mind that the radiation exposure the limitations of imaging techniques, such as low
to a human being should be kept As Low As contrast, motion and noise, may result in indis-
Reasonably Achievable (ALARA). tinct boundaries of the adjacent structures.
In addition, no well-established digitization Thresholding approach is one of the simplest
techniques of the 3D images has been stand- stochastic segmentation techniques in which one
ardized either on the multi-planar reconstruction or more values “thresholds” are used to create
(MPR) slices or on the rendered view. The recently partitions according to voxel intensities. For
proposed analyses could be technique sensitive. example, a threshold could be set to separate air
Moreover, custom-made analysis by each user from soft tissue and another threshold to sepa-
requires good knowledge of spatial geometry and rate soft tissue from bone while a third one can
experience in working with 3D models. separate bone from teeth. However, it is highly
256 Bayome et al

sensitive to noise and dependent on the transverse line to guide the construction of the
threshold values. horizontal plane in 3D coordinate systems.

Reorientation of head position 3D analysis procedures

The reorientation process depends on placing Several software programs has been developed to
the image of the head into a known repeatable view, digitize, measure, and analyze CBCT data.
position in the coordinate system through Ludlow et al.29 recommended the identification
defining the origin point and the X, Y, and Z of landmarks on the MPR slices due to its high
planes. These definitions should be based on accuracy. Another study showed landmarks
landmarks that are least susceptible to asymmetry digitized on the rendered view due to its ease
and least affected by treatment procedures to and shorter analysis time.27 Nguyen et al.43 found
strengthen the reliability and validity of the high correlation between measurements from
required planes. Nasion (N) and anterior nasal each of the slice section and volume render views
spines (ANS) tend to fall on or very close to the of Invivo software (Anatomage Inc., San Jose,
midsagittal plane in 90% of the population.39 CA) and the physical measurements. Also,
Therefore, Bayome et al.27 selected N as an several studies reported high accuracy of linear
origin of the 3D coordinate system. The and angular measurements in 3D volume
horizontal plane (X) was defined through the render CBCT images compared to physical
right and left orbitales (Or) and the left porion measurements.44–46
(Po) while the midsagittal plane (Y) was defined With the advent of 3D cephalometric analysis,
as the perpendicular plane passing through N new landmarks, reference planes, and measure-
and ANS. The vertical plane (Z) was the ments were made possible. The ability to 3D
perpendicular to both X and Y (Fig. 1). visualize the head and the possibility of taking
Swennen et al.40 proposed a reorientation sections into the 3D volume allowed practitioners
method “the anatomic Cartesian 3D cepha- to place landmarks accurately on structures that
lometric reference system” with the origin at were not available on the 2D cephalograms. In
Sella (S). However, this system is complicated turn, the 3D Cartesian system facilitated the
and time consuming. Kook and Kim41 proposed creation of new reference planes and the eval-
a clinical method to easily reorient head using uation of curvatures, besides the linear and
frontal facial and intraoral photographs. Park angular relationships. The following are exam-
et al.42 suggested the use of the right and left ples of the newly proposed landmarks, planes,
zygomatic suture points or the Or as a stable and measurements.

Figure 1. Reorientation of head and coordinate system. N, nasion; X, the horizontal plane; Y, the midsagittal
plane; Z, the vertical plane.
 3D analysis and clinical applications of CBCT images 257

Figure 2. Mandibular body variables. Me, menton; MBC, mandibular body curve; Go, gonion; 1, menton angle; 2,
mandibular body length; 3, anterior mandibular body length; 4, posterior mandibular body length; 5, MBC angle.

New landmarks the effect of the roll, yaw, and pitch on the
47 measurements that is defined as follows:
Lee et al. suggested that the 2D definition of
the mandibular body might not be able to a:b
cos θ ¼ ; ð1Þ
represent it on 3D image. Therefore, they jajjbj
proposed the mandibular body curve (MBC)
where a and b are the vectors of each line.
points, which lie on the most convex point on the
For example, a change in the roll of one line
curvature of the mandibular body midway
may change the value of the angle between this
between the inner and outer borders (Fig. 2).
line and another when measured on 3D.
They reported a significant difference between
However, this change will not affect the read-
the asymmetry and normal occlusion groups in
ings if it is measured on 2D. This, subsequently,
the posterior mandibular body length (MBC-
changes the interpretation of the line-to-line
Go), but this difference was not significant in the
angle measurements in 3D. The same is true for
mandibular body length (Me–Go).
measuring 3D line-to-plane or plane-to-plane
angles as they are assessed through determi-
New planes nation of the plane’s normal vector and then
Cheung et al.28 proposed a new reference plane, using Eq. (1).
the supraorbital margin plane, to overcome In addition, 3D cephalometric analysis
the limitations of 2D analysis in assessment of allowed volumetric and curvature evaluations.
paranasal and infraorbital areas. However, Recently, Bayome et al.27 suggested mea-
further research applying new reference planes suring the length of the mandibular body
or implementing alternative methods, such as through calculating the length of the curve
volumetric analysis might be required to enhance passing through menton (Me), MBC, and
the evaluation of the midfacial complex gonion (Go) to achieve a more accurate
configuration. representation of the mandibular body. The
coordinates of these points were entered into
MATLABs 7.5 (R2007b) (The MathWorks Inc.,
New measurements
Natick, MA). The 4th degree polynomial
The measurement of angles and distances in 3D equation f ðx Þ of the best fitting curve that
is considerably different from that in 2D due to pass through the five points was generated as an
258 Bayome et al

approximation of the curvature of the measurement was from N to gnathion (Gn) while
mandibular body. the former’s was to the pogonion (Pg). This
suggested an even larger difference between the
f ðx Þ¼ p1 x 4 þ p2 x 3 þp3 x 2 þ p4 x þ p5 ð2Þ
2 different Asian populations. Even though the
It was found that polynomial of 4th order Southern Chinese sample had about 7 mm
approximated the curvature of the mandibular longer upper facial height than Koreans, this
body with tolerable, or even negligible, mean can be attributed to the extension of the Chinese
square error. Mathematically, the length of a measurement to the A point while the Korean
path from point a to point b on a curve repre- was till ANS. Meanwhile, the Korean sample had
sented by the function f ðx Þ is given by about 2 mm longer lower facial height than
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 ffi
Z b Chinese although their measurement was from
df ðx Þ 2 ANS to Me while ours was from ANS to Pg, which
Length ¼ 1þ dx ð3Þ
a dx implies even longer lower facial height for
Koreans. Therefore, the difference in total
Then, the equation was entered into Maple™
11.0 (Waterloo Maple Inc., Waterloo, ON, Can- facial height might be attributed mainly to the
lower facial third.
ada) to calculate the differentiation of the
Several studies have evaluated condylar vari-
function f ðx Þ:
ables.49–52 You et al.49 suggested that the condylar
df ðx Þ unit (condyle, condylar neck, and part of the
¼ 4p1 x 3 þ3p2 x 2 þ 2p3 x þp4 ð4Þ
dx ramus) plays a central role in the mandibular
Then, the length of the curve from Go to Me asymmetry. However, Huntjens et al.50 found
was found by solving the integration: condylar asymmetries did not correlate well with
Z b qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi facial asymmetry. Also, Sanders et al.51 reported
 2
Length ¼ 1 þ 4p1 x 3 þ3p2 x 2 þ2p3 x þ p4 dx no significant asymmetries among condylar
a measurements in Class I or Class II subdivision
ð5Þ groups. Nevertheless, it was assumed that the
where a and b are the values of X coordinates of subclinical condylar and mandibular asymmetry
Me and Go, respectively (Fig. 3). could be natural in juvenile patients, but the
The same procedures were followed to cal- extend of this asymmetry is still indistinct.52
culate the length of the curve of the basal arch of Also, Bayome et al.27 reported correlation
the maxilla by incorporating the A point, and between the condylar and mandibular variables
right and left canine eminence, and tuberosity, which might be attributed to the adaptive
where a and b in the equation are the values of capacity of the condyle as suggested by Enlow
X coordinates of A point and tuberosity, res- and Hans53; for example, the negative correlation
pectively (Fig. 4). between the condylar anteroposterior inclination
and the gonial angle tends to preserve a
proportion between the height of the mandible
3D analysis and variable relationships
and its sagittal position in normal occlusion
Bayome et al.27 showed significant differences population.
between males and females in several vertical and Recently, the difference in ramus length
transverse measurements. Similarly, Thilander between both sides was reported as a charac-
et al.48 reported that the linear craniofacial teristic of both mandibular retrusion and prog-
measurements on lateral cephalograms were nathism groups.54 Bayome et al.27 reported that
larger in males than in females, while angular the ramus length demonstrated a significant
measurements showed no statistical differences. moderate negative correlation with the gonial
This might suggest that the dimensions of angle (r ¼ 0.62). This might suggest that the
the face played a major role in the gender longer the ramus the smaller its angle with the
dimorphism. mandibular body. This configuration can be a
The total facial height was about 5 mm lar- mechanism to prevent elongation of the facial
ger in the study of Bayome et al.27 than in height. Also, deviation from this relationship on
that of Cheung et al.,28 although the latter’s one side may result in facial asymmetry.
 3D analysis and clinical applications of CBCT images 259

Figure 3. Mandibular body curve length. Go, gonion; Me, menton; MBC, mandibular body curve.

Soft-tissue 3D analysis Several facial analyses have been suggested

based on soft-tissue landmarks and parameters
The facial appearance has a strong influence on registered either on a tracing of a radiograph or
the self-esteem and social acceptance.55,56 photograph.59–67 However, besides the criticism
Therefore, it is imperative that special atten- of the reliability and creditability of several
tion should be paid for the facial esthetics during proposed variables, all these analyses suffered the
orthodontic and orthognathic treatment plan- same disadvantages related to the 2D repre-
ning, especially, since the perception of facial sentation of a 3D object. Moreover, the response
attractiveness is subjective, and dependent on of soft-tissue features to changes in skeletal and
ethnicity, age, gender, culture, and person- dentoalveolar relationships was overlooked.
ality.57,58 In addition, lack of agreement on Several authors have endeavored to develop a
standards, and personal and cultural bias 3D soft-tissue analysis based on various 3D rec-
heightened the controversy between the ideal- ording techniques such as laser scanning, hol-
ization and individualization of treatment. ography, and stereophotogrammetry.36,37,68–70

Figure 4. Maxillary basal curve length. Max. T, maxillary tuberosity; C E, canine eminence; A, A point.
260 Bayome et al

Plooij et al.71 defined new bone-related soft-tissue third of the face, moderate to weak correlations
landmarks on 3D stereophotogrammetric images were found between the lower facial height and
and reported high reproducibility and reliability nasal and mouth widths. In addition, the upper
of identification. They suggested that a 3D soft- facial width had strong to moderate correlation
tissue analysis can be accurately produced with- with the maxillary height and length and
out the need to obtain hard tissue records. mandibular body curve length.72
Kochel et al.36,37 developed a 3D soft-tissue
analysis on 3D stereophotogrammetric images,
but all the measurements were taken from pro-
Conclusion
jections of the digitized points.
A previous study on CBCT images of a normal 3D cephalometric analysis is becoming a vital tool
occlusion population showed that males had a to evaluate the relationships among skeletal and
significantly greater intercanthal distance, nasal dentoalveolar cephalometric variables. New
and mouth widths, and posterior facial width landmarks and variables were suggested to assess
than females.72 Also, it demonstrated that there anatomical structures that were not recognizable
were no significant differences between the right on 2D radiographs especially in the transverse
and left sides.72 However, other studies reported dimension and the midfacial area. The develop-
asymmetry in normal occlusion population with ment of 3D hard- and soft-tissue cephalometric
pleasing facial features.70,73 This disagreement analyses may produce a new understanding of the
might be due to differences in the evaluation relationships between soft-tissue, skeletal and
methods and errors in landmark identification. dentoalveolar cephalometric variables. These
Nevertheless, up to date, no 3D cephalometric analyses can be useful for accurate diagnosis and
analysis has been widely used in clinical practice, treatment planning and for evaluation of treat-
and validated for practicality, reliability, repro- ment outcomes of orthodontics or orthognathic
ducibility, and clinical relevance. surgery.

Relationships between hard and soft

tissues References
1. Jacobson A. The “Wits” appraisal of jaw disharmony. 1975.
Several articles have evaluated the relationship Am J Orthod Dentofacial Orthop. 2003;124:470–479.
between hard- and soft-tissue cephalometric 2. McNamara JA Jr. A method of cephalometric evaluation.
variables. However, a consensus has not been Am J Orthod. 1984;86:449–469.
reached. Ferrario and Sforza74 reported that the 3. Steiner CC. The use of cephalometrics as an aid to
planning and assessing orthodontic treatment. Am J
facial soft tissue was influenced by skeletal
Orthod. 1960;46:721–735.
variables and McNamara et al.75 demonstrated 4. Sassouni V. A roentgenographic cephalometric analysis
a significant relationship between incisor position of cephalo-facial-dental relationships. Am J Orthod. 1955;
and the thickness of upper lip. Nonetheless, 41:735–764.
other authors found that correlations between 5. Broadbent BH Sr, Broadbent BH Jr, Golden WH. Bolton
Standards of Dentofacial Developmental Growth. St. Louis:
hard- and soft-tissue parameters were not
C.V. Mosby; 1975.
common.59,76 6. Popovich F, Thompson GW. Craniofacial templates for
Recently, a 3D soft-tissue analysis showed weak orthodontic case analysis. Am J Orthod. 1977;71:406–420.
to moderate correlations between the SNA and 7. Bjork A. The nature of facial prognathism and its relation
Tragus-N-subspinale angle, and between the SNB to normal occlusion of the teeth. Am J Orthod. 1951;
and Tragus-N-sublabiale angle, sagittally.36 Also, 37:106–124.
8. Björk A. The Face in Profile. Svensk Tandl Tidskr. 1947;40
correlations were reported between the lower (suppl):56.
gonial angle and N-gonion (Go)–gnathion (Gn) 9. Baumrind S, Frantz RC. The reliability of head film
angle and between the basal plane angle and the measurements. 1. Landmark identification. Am J Orthod.
subnasale (Sn)–Pg–Go–Gn angle vertically.37 1971;60:111–127.
10. Ahlqvist J, Eliasson S, Welander U. The cephalometric
However, another study reported no significant
projection. Part II. Principles of image distortion in
correlations between hard- and soft-tissue varia- cephalography. Dentomaxillofac Radiol. 1983;12:101–108.
bles except for the posterior facial width with the 11. Solow B, Tallgren A. Head posture and craniofacial
maxillary length and facial height.72 In the lower morphology. Am J Phys Anthropol. 1976;44:417–435.
 3D analysis and clinical applications of CBCT images 261

12. Major PW, Johnson DE, Hesse KL, et al. Effect of head I pattern and normal occlusion. Korean J Orthod. 2013;
orientation on posterior anterior cephalometric land- 43:62–73.
mark identification. Angle Orthod. 1996;66:51–60. 28. Cheung LK, Chan YM, Jayaratne YS, et al. Three-
13. Berco M, Rigali PH Jr., Miner RM, et al. Accuracy and dimensional cephalometric norms of Chinese adults in
reliability of linear cephalometric measurements from Hong Kong with balanced facial profile. Oral Surg Oral
cone-beam computed tomography scans of a dry human Med Oral Pathol Oral Radiol Endod. 2011;112:e56–e73.
skull. Am J Orthod Dentofacial Orthop. 2009;136(17):e1–e9. 29. Ludlow JB, Gubler M, Cevidanes L, et al. Precision of
14. Gribel BF, Gribel MN, Frazao DC, et al. Accuracy and cephalometric landmark identification: cone-beam com-
reliability of craniometric measurements on lateral puted tomography vs conventional cephalometric views.
cephalometry and 3D measurements on CBCT scans. Am J Orthod Dentofacial Orthop. 2009;136(312):e1–10.
Angle Orthod. 2011;81:26–35. 30. Ludlow JB, Laster WS, See M, et al. Accuracy of
15. Lamichane M, Anderson NK, Rigali PH, et al. Accuracy of measurements of mandibular anatomy in cone beam
reconstructed images from cone-beam computed tomog- computed tomography images. Oral Surg Oral Med Oral
raphy scans. Am J Orthod Dentofacial Orthop. 2009;136(156): Pathol Oral Radiol Endod. 2007;103:534–542.
e1–e6. 31. Cevidanes LH, Bailey LJ, Tucker GR Jr., et al. Super-
16. Moerenhout BA, Gelaude F, Swennen GR, et al. Accuracy imposition of 3D cone-beam CT models of orthognathic
and repeatability of cone-beam computed tomography surgery patients. Dentomaxillofac Radiol. 2005;34:369–375.
(CBCT) measurements used in the determination of 32. Cevidanes LH, Bailey LJ, Tucker SF, et al. Three-
facial indices in the laboratory setup. J Craniomaxillofac dimensional cone-beam computed tomography for
Surg. 2009;37:18–23. assessment of mandibular changes after orthognathic
17. Lagravere MO, Gordon JM, Guedes IH, et al. Reliability of surgery. Am J Orthod Dentofacial Orthop. 2007;131:44–50.
traditional cephalometric landmarks as seen in three- 33. van Vlijmen OJ, Maal TJ, Berge SJ, et al. A comparison
dimensional analysis in maxillary expansion treatments. between two-dimensional and three-dimensional ceph-
Angle Orthod. 2009;79:1047–1056. alometry on frontal radiographs and on cone beam
18. de Oliveira AE, Cevidanes LH, Phillips C, et al. Observer computed tomography scans of human skulls. Eur J Oral
reliability of three-dimensional cephalometric landmark Sci. 2009;117:300–305.
identification on cone-beam computerized tomography. 34. van Vlijmen OJ, Berge SJ, Bronkhorst EM, et al. A
Oral Surg Oral Med Oral Pathol Oral Radiol Endod. comparison of frontal radiographs obtained from cone
2009;107:256–265. beam CT scans and conventional frontal radiographs
19. Ballrick JW, Palomo JM, Ruch E, et al. Image distortion of human skulls. Int J Oral Maxillofac Surg. 2009;38:
and spatial resolution of a commercially available cone- 773–778.
beam computed tomography machine. Am J Orthod 35. Gribel BF, Gribel MN, Manzi FR, et al. From 2D to 3D: an
Dentofacial Orthop. 2008;134:573–582. algorithm to derive normal values for 3-dimensional
20. Haney E, Gansky SA, Lee JS, et al. Comparative analysis of computerized assessment. Angle Orthod. 2011;81:3–10.
traditional radiographs and cone-beam computed tomog- 36. Kochel J, Meyer-Marcotty P, Strnad F, et al. 3D soft tissue
raphy volumetric images in the diagnosis and treatment analysis—part 1: sagittal parameters. J Orofac Orthop.
planning of maxillary impacted canines. Am J Orthod 2010;71:40–52.
Dentofacial Orthop. 2010;137:590–597. 37. Kochel J, Meyer-Marcotty P, Kochel M, et al. 3D soft tissue
21. Botticelli S, Verna C, Cattaneo PM, et al. Two- versus analysis—part 2: vertical parameters. J Orofac Orthop.
three-dimensional imaging in subjects with unerupted 2010;71:207–220.
maxillary canines. Eur J Orthod. 2011;33:344–349. 38. Farronato G, Garagiola U, Dominici A, et al. “Ten-point”
22. Katheria BC, Kau CH, Tate R, et al. Effectiveness of 3D cephalometric analysis using low-dosage cone beam
impacted and supernumerary tooth diagnosis from computed tomography. Prog Orthod. 2010;11:2–12.
traditional radiography versus cone beam computed 39. Harvold E. Cleft lip and palate: morphologic studies of
tomography. Pediatr Dent. 2010;32:304–309. the facial skeleton. Am J Orthod. 1954;40:493–506.
23. Han S, Bayome M, Lee J, et al. Evaluation of palatal bone 40. Swennen GR, Schutyser F, Barth EL, et al. A new method
density in adults and adolescents for application of of 3-D cephalometry Part I: the anatomic Cartesian 3-D
skeletal anchorage devices. Angle Orthod. 2012;82: reference system. J Craniofac Surg. 2006;17:314–325.
625–631. 41. Kook YA, Kim Y. Evaluation of facial asymmetry with
24. Ryu JH, Park JH, Vu Thi Thu T, et al. Palatal bone three-dimensional cone-beam computed tomography. J
thickness compared with cone-beam computed tomog- Clin Orthod. 2011;45:112–115; quiz 92.
raphy in adolescents and adults for mini-implant place- 42. Park JU, Kook YA, Kim Y. Assessment of asymmetry in a
ment. Am J Orthod Dentofacial Orthop. 2012;142:207–212. normal occlusion sample and asymmetric patients with
25. Farnsworth D, Rossouw PE, Ceen RF, et al. Cortical bone three-dimensional cone beam computed tomography: a
thickness at common miniscrew implant placement sites. study for a transverse reference plane. Angle Orthod.
Am J Orthod Dentofacial Orthop. 2011;139:495–503. 2012;82:860–867.
26. Alsufyani NA, Flores-Mir C, Major PW. Three- 43. Nguyen E, Boychuk D, Orellana M. Accuracy of cone-
dimensional segmentation of the upper airway using beam computed tomography in predicting the diameter
cone beam CT: a systematic review. Dentomaxillofac Radiol. of unerupted teeth. Am J Orthod Dentofacial Orthop. 2011;
2012;41:276–284. 140:e59–e66.
27. Bayome M, Park JH, Kook YA. New three-dimensional 44. Lagravere MO, Carey J, Toogood RW, et al. Three-
cephalometric analyses among adults with a skeletal Class dimensional accuracy of measurements made with
262 Bayome et al

software on cone-beam computed tomography images. descriptors and traditional cephalometric methods. Am J
Am J Orthod Dentofacial Orthop. 2008;134:112–116. Phys Anthropol. 2003;121:172–180.
45. Stratemann SA, Huang JC, Maki K, et al. Comparison of 60. Bergman RT. Cephalometric soft tissue facial analysis. Am
cone beam computed tomography imaging with physical J Orthod Dentofacial Orthop. 1999;116:373–389.
measures. Dentomaxillofac Radiol. 2008;37:80–93. 61. Legan HL, Burstone CJ. Soft tissue cephalometric analysis
46. Periago DR, Scarfe WC, Moshiri M, et al. Linear accuracy for orthognathic surgery. J Oral Surg. 1980;38:744–751.
and reliability of cone beam CT derived 3-dimensional 62. Holdaway RA. A soft-tissue cephalometric analysis and its
images constructed using an orthodontic volumetric use in orthodontic treatment planning. Part I. Am J
rendering program. Angle Orthod. 2008;78:387–395. Orthod. 1983;84:1–28.
47. Lee H, Bayome M, Kim SH, et al. Mandibular dimensions 63. Holdaway RA. A soft-tissue cephalometric analysis and its
of subjects with asymmetric skeletal class III malocclusion use in orthodontic treatment planning. Part II. Am J
and normal occlusion compared with cone-beam com- Orthod. 1984;85:279–293.
puted tomography. Am J Orthod Dentofacial Orthop. 64. Peck H, Peck S. A concept of facial esthetics. Angle orthod.
2012;142:179–185. 1970;40:284–318.
48. Thilander B, Persson M, Adolfsson U. Roentgen- 65. Ricketts RM. Esthetics, environment, and the law of lip
cephalometric standards for a Swedish population. A relation. Am J Orthod. 1968;54:272–289.
longitudinal study between the ages of 5 and 31 years. Eur 66. Powell N, Humphreys B. Proportions of the esthetic face. New
J Orthod. 2005;27:370–389. York: Thieme–Stratton; 1984.
49. You KH, Lee KJ, Lee SH, et al. Three-dimensional 67. Arnett GW, Jelic JS, Kim J, et al. Soft tissue cephalometric
computed tomography analysis of mandibular morphol- analysis: diagnosis and treatment planning of dentofacial
ogy in patients with facial asymmetry and mandibular deformity. Am J Orthod Dentofacial Orthop. 1999;116:
prognathism. Am J Orthod Dentofacial Orthop. 2010;138 239–253.
(540):e1–e8. 68. Ferrario VF, Sforza C, Poggio CE, et al. Distance from
50. Huntjens E, Kiss G, Wouters C, et al. Condylar asymmetry symmetry: a three-dimensional evaluation of facial asym-
in children with juvenile idiopathic arthritis assessed by
metry. J Oral Maxillofac Surg. 1994;52:1126–1132.
cone-beam computed tomography. Eur J Orthod.
69. Stauber I, Vairaktaris E, Holst A, et al. Three-dimensional
2008;30:545–551.
analysis of facial symmetry in cleft lip and palate patients
51. Sanders DA, Rigali PH, Neace WP, et al. Skeletal and
using optical surface data. J Orofac Orthop. 2008;69:
dental asymmetries in Class II subdivision malocclusions
268–282.
using cone-beam computed tomography. Am J Orthod
70. Hwang HS, Yuan D, Jeong KH, et al. Three-dimensional
Dentofacial Orthop. 2010;138(542):e1–20.
soft tissue analysis for the evaluation of facial asymmetry
52. Liukkonen M, Sillanmaki L, Peltomaki T. Mandibular
in normal occlusion individuals. Korean J Orthod.
asymmetry in healthy children. Acta Odontol Scand.
2012;42:56–63.
2005;63:168–172.
71. Plooij JM, Swennen GR, Rangel FA, et al. Evaluation of
53. Enlow DH, Hans MG. Essentials of Facial Growth. Phila-
delphia: W B Saunders; 1996. reproducibility and reliability of 3D soft tissue analysis
54. Kim EJ, Palomo JM, Kim SS, et al. Maxillofacial character- using 3D stereophotogrammetry. Int J Oral Maxillofac Surg.
istics affecting chin deviation between mandibular retru- 2009;38:267–273.
sion and prognathism patients. Angle Orthod. 2011;81: 72. Bayome M. Evaluation of Three-Dimensional Relationships
988–993. Among Skeletal, Dentoalveolar, and Soft Tissue Variables in
55. Hershon LE, Giddon DB. Determinants of facial profile Normal Occlusion Sample. Seoul: The Catholic University of
self-perception. Am J Orthod. 1980;78:279–295. Korea; 2011, [Doctoral dissertation].
56. Langlois JH, Kalakanis L, Rubenstein AJ, et al. Maxims or 73. Shah SM, Joshi MR. An assessment of asymmetry in the
myths of beauty? A meta-analytic and theoretical review normal craniofacial complex. Angle Orthod. 1978;48:
Psychol Bull. 2000;126:390–423. 141–148.
57. Mandall NA, McCord JF, Blinkhorn AS, et al. Perceived 74. Ferrario VF, Sforza C. Size and shape of soft-tissue facial
aesthetic impact of malocclusion and oral self- profile: effects of age, gender, and skeletal class. Cleft
perceptions in 14-15-year-old Asian and Caucasian chil- Palate Craniofac J. 1997;34:498–504.
dren in greater Manchester. Eur J Orthod. 2000;22: 75. McNamara L, McNamara JA Jr, Ackerman MB, et al.
175–183. Hard- and soft-tissue contributions to the esthetics of the
58. Sahin Saglam AM, Gazilerli U. Analysis of Holdaway soft- posed smile in growing patients seeking orthodontic
tissue measurements in children between 9 and 12 years treatment. Am J Orthod Dentofacial Orthop. 2008;133:
of age. Eur J Orthod. 2001;23:287–294. 491–499.
59. Rose AD, Woods MG, Clement JG, et al. Lateral facial soft- 76. Saxby PJ, Freer TJ. Dentoskeletal determinants of soft
tissue prediction model: analysis using Fourier shape tissue morphology. Angle Orthod. 1985;55:147–154.
3-Dimensional cone-beam computed
tomography superimposition: A review
Jae Hyun Park, DMD, MSD, MS, PhD (Professor and chair),
Kiyoshi Tai, DDS, PhD (Visiting adjunct professor), and
Payam Owtad, DDS, MS, DHEd (Orthodontist)

Traditionally, the superimposition of serial, two-dimensional cephalometric

radiographs has been used for the evaluation of growth and treatment
effects. Nowadays, the superimposition of cone-beam computed tomogra-
phy (CBCT) images has become an important tool for three-dimensional (3D)
assessment of changes with growth or treatment. However, the assessment
of changes with 3D image superimposition poses many challenges such as
accuracy and reproducibility. Various methods for the reconstruction of 3D
CBCT images have been used in diagnosis, treatment planning, and
simulation. In this article, the various 3D CBCT superimposition techniques
and relevant evidences are discussed. Furthermore, their clinical applications,
benefits, and limitations are addressed. (Semin Orthod 2015; 21:263–273.) &
2015 Elsevier Inc. All rights reserved.

Introduction been used in orthodontic diagnosis, treatment

planning, and simulation to overcome common
one-beam computed tomography (CBCT)
C is considered as a supplemental imaging
technique when two-dimensional (2D) imaging
2D cephalometric errors.6–8
In dentistry and orthodontics, high resolution
CBCT is used to acquire a low distortion digital
does not provide sufficient information in ortho-
image of the hard and soft tissues of the cra-
dontic clinical assessments.1,2 Common errors with
niofacial structures. CBCT machines have a cone
2D cephalometric radiographs include patient and
shape x-ray beam that is the source of the cone-
hardware orientation errors and geometric error
beam name, which is unlike conventional CT.
which is due to differential magnification and
Conventional CT uses a fan-shaped beam to
association error, which makes it difficult to identify
create multiple thin slices. In CBCT systems, the
an identical point seen from different angles on
resolution is measured in voxels instead of pixels,
separate projections.3
which is often sharper than a conventional CT.
CBCT scanners were introduced in the late
Panoramic and cephalometric projections that
1990s as an evolutionary process, resulting from
are produced by CBCT are transformed into a 3D
the demand for three-dimensional (3D) infor-
format after the data has been reformatted in a
mation obtained by conventional computerized
volume by computer software.9
tomography (CT) scans.4,5 3D CBCT images have
3D CBCT can be used to gather diagnostic
information on temporomandibular joints, ana-
tomic features of the mandible, and the mor-
Arizona School of Dentistry & Oral Health, A.T. Still University,
Mesa, AZ; Graduate School of Dentistry, Kyung Hee University,
phology of the palate, and to measure the width
Seoul, South Korea; Private Practice of Orthodontics, Okayama, of alveolar bones, to determine the position of the
Japan. dental roots and supernumerary and impacted
Address correspondence to Jae Hyun Park, DMD, MSD, MS, teeth, and to identify sites for implant placement
PhD, Arizona School of Dentistry & Oral Health, A.T. Still
or osteotomies.10 CBCT imaging is also used to
University, 5835 East Still Circle, Mesa, AZ 85206. E-mail:
[email protected]
plan for orthodontic and orthognathic surgery
& 2015 Elsevier Inc. All rights reserved.
treatment, to assess skeletal displacements after
1073-8746/15/1801-$30.00/0 osteotomies, to verify treatment outcomes, and to
http://dx.doi.org/10.1053/j.sodo.2015.07.007 determine stability.11

Seminars in Orthodontics, Vol 21, No 4, 2015: pp 263–273 263

264 Park et al

Figure 1. Multi-planar reconstruction (MPR) and volume rendered images. (A) Pre- and post-treatment
superimposed MPR images and soft tissue volume rendered image. (B) Volume rendered hard tissue. (C) Volume
rendered hard tissue overlaid by translucent soft tissue.

Multi-planar reconstruction (MPR) and vol- 3D CBCT superimposition in

ume rendered are the most commonly used 3D orthodontics
planes for measuring CBCT-derived cephalo-
Two common 2D superimposition methods are
grams (Fig. 1).12–15 When comparing physical
the best-fit method and the structural method.
skull measurements with 3D image measure-
The usual approaches for 3D superimposition are
ments, point-to-point MPR measurements have
registration points or mathematical algorithms.3 In
been found to be highly accurate, while minor
the registration point approach, certain landmarks
error is seen with volume rendered mode
are registered on two volumetric images, which will
measurements, most likely due to surface con-
coincide when the superimposition is made. After
tour estimations.16,17 Therefore, identification
initial superimposition of the two images, the
and targeting of anatomical landmarks should be
operator can manually move the superimposed
done in the MPR display mode.15
images for further 3D fine-tuning. In mathematical
Superimposition of orthodontic 3D CBCT
algorithms, the initial 3D CBCT scan is considered
images is usually used for evaluating a patient’s
to be the volume of interest (VOI) or the reference
craniofacial growth or for comparing craniofacial
volume. Software, based on probability and
structures before and after treatment.11,18 The
information theory, then superimposes the second
superimposition can be performed by registering
scan over the VOI in its best-fit position. The fusion
stable anatomic landmarks or by software-assisted
process of the two images occurs automatically. It is
best-fit registration of stable anatomic structures
not dependent on an operator’s skill and is faster
such as the anterior cranial base.18 A variety of
than manual methods.3
software tools have been designed for 3D CBCT
Various 3D superimposition methods are used
superimposition, which align the registered
for clinical diagnosis and treatment evaluation
landmarks or cranial base structures.19 For
purposes in orthodontic treatment and cranio-
example, one of the anatomical structures that
facial surgeries. There are valuable benefits and
is commonly used for CBCT registration is the
some limitations with each method. They are
anterior cranial base.11,18 It should be remem-
discussed in this article. A summary of 3D CBCT
bered that the accuracy and reproducibility of
superimposition methods, applications, advan-
these superimposition methods are directly
tages, and limitations is presented in Table 1.
dependent on the accuracy of the landmarks and
anatomic structures.
3D CBCT superimposition methods
Superimposition of CBCT images has become
an important tool for 3D assessment of changes With most software programs, a clinician does
with growth or treatment. 3D CBCT super- the initial alignment of the landmarks or
imposition methods and clinical applications, anatomic structures of the two images to be
benefits and limitations are discussed in this superimposed, then computer software
review. measures the changes in other anatomic
 3D CBCT superimposition 265

Table 1. 3D CBCT superimposition

Methods Clinical applications Advantages Limitations

Registration points: soft- Evaluating patients’ 3D assessments by the Time consuming and
ware-assisted best-fit craniofacial fusing of two 3D computing
registration of arbi- growth and/or images in growing intensive
trary selected points, craniofacial patients and/or dur- Measurement
also called point- structures before ing treatment process outcomes of a
based and after ortho- Observation of CBCT images can
Information theory and dontics and bilateral structures differ depending
mathematical algorithm: orthognathic sur- such as condylion, in on window set-
software-assisted gery treatment real size 3D images is tings, plane defi-
superimposition of Quantifying and more precise and nition, number
the registered visually assessing more reproducible and thickness of
structures: hard and soft tis- than conventional 2D sections
Surface-based: sue changes cephalometric images Slightly inferior to
iterative closest between the two Provides a more realistic reality when
point (ICP) scans relative to way to evaluate the threshold-based
method the registered changes measurements
Voxel-based: voxel landmarks or A comprehensive visual are used
by voxel best-fit structures; and quantitative Clinical
registration i.e., evaluating analysis measurements
method mid-facial soft tis- Software-assisted 3D may be impacted
sue changes in CBCT analysis by segmentation
the condyles, Human error is process
rami, chin, max- eliminated during the Landmark
illa, and denti- automated superim- identification and
tion, quantifying position processes registration errors
transverse dimen- Final manual
sions and axial fine-tuning in some
inclinations of of the techniques
dentition, and Creates accurate 3D
ramus displace- surface models and
ment and airway accurate measure-
assessments in 3D ments with multi-pla-
nar reconstruction
and color-coded
imagesa
Great repeatability and
reproducibilitya
Minimal errorsa
Creates an unbiased
scientific set up for
the superimposition
processa
a
Subjects of controversy in the literature.

structures relative to the registered points or Choi and Mah23 reported a 3D superimposition
structures. The final superimposed image method that is performed without prior landmark
shows changes that have resulted from growth designation. This method is based on a matching
or treatment. 15,19–22 criterion that uses the information theory and
266 Park et al

mutual information or relative entropy that was other software programs, it takes time to become
originally used for conventional CT images and proficient. After mastering the use of the software,
magnetic resonance imaging.24 By applying this image measurements can be made with great
theory, an operator is able to acquire geometrical repeatability.26 With the ICP 3D superimposition
information from a software program and use it in technique, registration of the scans over the cranial
another one for a pre-segmented surface model base is reported to be an accurate method for
comparison. In this way, it is possible to achieve a superimposition.29–31 This method can be used for
clear and precise registration plus subvoxel accu- a valid and reproducible assessment of treatment
racy.25 Moreover, with this system, it is possible to outcomes for growing subjects. ICP is also consid-
superimpose images and scans from different ered to be clinically valuable because of the man-
CBCT devices and software programs.23 ageability and 3D accuracy of data comparison with
With the iterative closest point (ICP) method, MPR images (Fig. 3).30
a more accurate measurement can be made by The other method is voxel-based image reg-
using the same points on the same surface with istration, which is an accurate and reproducible
fusion at different time points.26,27 The accuracy semi-automated technique for 3D CBCT super-
of linear measurements in 2D cephalograms and imposition.32,33 For example, when a patient is
3D scans are not the same because of a difference fully grown and developed, registration of the
in the size and location of the objects in the two superimposed CBCT images of the zygomatic
imaging systems.28 Unlike conventional cephalo- arches can be considered as an alternative to the
metric radiographic analysis, the ICP method anterior cranial base.34 After superimposition,
allows for the precise fusing of two 3D images the differences between the two surfaces are
from growing patients.26,27 In the ICP technique, mapped with about 600,000 color-coded surface
an operator manually defines a certain domain distances in millimeters, which helps a practi-
on the surface of the CBCT scans such as the tioner quantify and visually assess the hard and
outline of the anterior cranial base from the soft tissue changes between the two scans relative
superior view (Fig. 2). Then the software to the cranial base.19,35
automatically matches and registers the identical Gianquinto et al.36 introduced a reproducible
landmarks of the selected domains on the two scans CBCT superimposition method based on the
and completes the superimposition process. The posterior cranial base in a single software
operator can evaluate and measure the changes package using a step-by-step manual technique.
relative to the registered surfaces. Learning to take With this method, the craniofacial volume for
image measurements is not difficult, but as with each of their patients is imported to their 3D

Figure 2. The iterative closest point (ICP) method. (A) Cranial base superimposition performed on all areas of the
cranial base except the peripheral growing zone. (B) Merged image of pre-(T0) and post-treatment (T1) CBCT
scans, superimposed at the cranial base.
 3D CBCT superimposition 267

Figure 3. Superimposition using the iterative closest point (ICP) method. (A) The combined images, pre- (gray)
and post-treatment (light blue) images, with illustration of an arbitrary coronal plane (green). (B) To facilitate
measurement, the 3D sections were converted to 2D data. (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of this article.)

CBCT superimposition software. The software craniofacial models (Fig. 4).37–39 In addition,
resamples the scans to a 0.5 mm voxel size, and CBCT images can be used to evaluate mid-facial
then superimposes the images with a mutual soft tissue changes,21,40 to assess changes in the
information algorithm. At this stage, the operator condyles, rami, chin, maxilla, and dentition,11,18
uses a semi-automatic technique to extract pos- and to measure the transverse dimensions and
terior cranial base surface data, which results in a axial inclinations of dentition.26
colored map based on the distance between the CBCT imaging has been proven to be an
two volumes relative to the cranial base.37 essential tool for assessing growth and development,
Triple voxel-based CBCT superimposition is which provides the most comprehensive visual and
another method, which is used to build an quantitative analysis.11,18 While some researchers
augmented 3D skull model using rigid registra- consider 3D superimposition systems to be accu-
tion points from three separate CBCT scans rate,15,41,42 there are some scientific reports that
made at a certain time point; the first scan with question the accuracy of these techniques.34,43
wax bite wafer in place and the second low-dose The anterior cranial base has been used as a
scan with a triple tray bite registration in place. reference structure for CBCT superimposition
Then the third CBCT scan is taken from the after orthognathic surgery in cases of skeletal
triple tray bite registration. With this method, the Class III malocclusion to evaluate the post-
three scans are fused to create detailed occlusal operative mid-facial soft-tissue changes (Fig. 4B
and intercuspation data without the use of dental and C).26,44 Anterior cranial bases are super-
models.33 imposed using the best fit of certain anatomic
landmarks of the anterior cranial base followed
by an automated voxel-based superimposition by
Clinical applications of 3D CBCT
the software. The anatomic landmarks of the
superimpositions
anterior cranial base might be the inner cortical
3D CBCT superimposition can be used to eval- layer of the frontal bones, superior surfaces of the
uate facial growth, craniofacial anomalies, and orbital roofs, superior aspect of the ethmoid and
skeletal changes after orthodontics and orthog- cribriform plate and cortical ridges on the
nathic surgery (i.e., nasal morphologic and medial, lesser wings of the sphenoid, planum
maxillary or mandibular skeletal changes after bi- sphenoidale, anterior clinoid processes, or the
maxillary surgery), and to create 3D virtual anterior wall of sella.19 Cevidanes et al.19 stated that
268 Park et al

Figure 4. Different ways of visualization of the treatment outcome using overlay of registered 3D models. (A) The
displacement magnitude of interface distances of pre- and post-treatment, registered by voxel-based method, is
expressed by the different colors using quantitative color mapping by Morpheus3D CT Viewer (Morpheus). (B)
Pre-treatment scan shown as a dotted red mesh and post-treatment scan shown in light blue, registered by ICP
method, by Micro AVS (Cybernet). (C) The displacement magnitude of interface distances of pre- and post-
treatment, registered by ICP method; quantitative color mapping by 3D-Rugle (Medic Engineering). (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)

a precise registration on the anterior cranial base difference when comparing linear and angular
provides a reliable 3D assessment of craniofacial measurements of traditional cephalograms and
changes during growth and development by CBCT-derived cephalograms.15,18,41,42 However,
evaluating displacement of facial structures rela- the accuracy is still a subject of some controversy
tive to the cranial base. In addition, clinicians can in the literature because of the errors that could
evaluate and describe treatment outcomes and be introduced in CBCT scans.26,32–34,43
changes due to future remodeling after treatment Observing full-scale 3D superimposed images
in a more comprehensive manner. of bilateral structures offers a more realistic way
The anterior cranial base is reported to be one to evaluate the changes.18,19,44 Comparison of 3D
of the most accurate registration structures for surfaces is also more reliable and precise than
3D superimposition. However, with the anterior traditional 2D cephalometric evaluations.18 By
cranial base superimposition method, clinicians visualizing 3D model superimpositions, the
cannot definitively describe mandibular move- location, magnitude, and direction of the hard
ment relative to the maxilla.45,46 In order to tissue changes can easily be determined. This
evaluate mandibular changes relative to the allows quantifying the ramal displacement in
maxilla, practitioners can register the two scans three dimensions.47
on a maxillary structure and perform the The availability of automated software for
superimposition. This process is conventionally analyzing 3D CBCT scans is a big advantage of 3D
used in 2D structural superimposition methods. digital systems over conventional 2D analyzing
For this purpose, the registration points or methods. In addition, using software for cranial
structures can arbitrarily be selected by the base superimposition is better than conventional
operator, i.e., registering CBCTs over zygomatic superimposition methods, since errors relative to
structures in order to evaluate the changes in landmark identification are eliminated and it is
craniofacial structures relative to the zygomatic possible to determine anatomic structures and
bones as a mid-face structure.34 their contours. The software creates an unbiased
scientific method for using the superimposition
process.18,42
Advantages of 3D CBCT superimposition 3D CBCT superimposition software offers
The accuracy of CBCT-derived cephalograms has operators the ability to superimpose 3D CBCT
been noted in numerous peer-reviewed articles, scanned surfaces, but it also provides a platform for
and it is reported that there is no statistical performing quantitative analysis, which can pro-
 3D CBCT superimposition 269

vide valuable information.48 Some software allows areas of the condyle and the lingual side of the
the operator to open two CBCT scans at the same mandible, as it is reported that the risk of
time to register common landmarks more measurement errors is higher in these areas.51
precisely. This software creates a cross-sectional Even though 3D CBCT superimposition soft-
visualization of the superimposed data set with ware can provide a significant amount of quan-
different colors, which makes it possible to eval- titative data, this high amount of information
uate the differences between the two CBCT scans sometimes makes it difficult for the practitioner
relative to the registered landmarks (Fig. 4).42 to formulate a straight and concise conclusion.18
This problem can be overcome by limiting the
quantitative analysis to particular landmarks and
Limitations of 3D CBCT superimposition
very particular surfaces or points at each
Even though CBCT is now widely used in intervention and quantitative analysis.
orthodontics,47–49 orthognathic surgery treat- In order to apply 3D CBCT superimposition
ments18,25,47 and airway assessments,1,9,10,50 only techniques in routine orthodontic treatment
limited evidence is currently available in the lit- planning, a faster, simpler, and more user-
erature regarding the accuracy of 3D CBCT friendly method must be implemented. It is
superimposition techniques and the options for imperative that scientific studies on the current
registration points or structures.34 Certain errors 3D superimposition techniques be conducted in
have been associated with 3D superimpositions orthodontic clinical practices and graduate
such as errors during visualization and improper programs, to further the developmental process
identification or location of landmarks. It is also of these techniques.11
challenging to use the anterior cranial base as a
registration structure.3 In addition, reliable
Photographs/digital models and CBCT
directional tendencies relative to growth are
superimposition
hard to determine since different structures grow
and move in different directions. In the future, the More recently, CBCT has been used with the
development of vectorial analysis tools may help to registration of skin surface images,29 so clinicians
clarify the displacement directions.10,18,19 can quantitatively assess 3D maxillofacial
Moreover, while the use of CBCT super- morphology, and evaluate linear and angular
imposition has significantly increased for assess- changes in facial soft and hard tissues in clinical
ing changes between serial CBCT scans, the procedures. Standard normative 3D values for
accuracy of CBCT scan superimposition techni- the craniofacial hard and soft tissues of normal
ques and a standardized method of selecting women were calculated by Terajima et al.,52 and
anatomical structures for 3D superimposition has were then compared with 3D CT measurements
not yet been established.34 The measurement before and after patients had orthognathic
outcomes from CBCT images can differ depen- surgery. It is reported that with this method,
ding on window settings, plane definition, and the they were able to quantitatively assess craniofacial
number and thickness of sections. Therefore, structure deviation from the norm before surgery
universal software should be used for CBCT and the changes in the hard and soft tissues after
image measurement as well as software with an surgery. While this method is reliable and clinically
ability to control variables.43 sound, they reported that more investigation is
One of the other limitations of 3D CBCT needed with before this method is used in other
superimposition is the fact that even though diagnostic and treatment planning applications.52
CBCT images can be used to create very accurate Cevidanes et al.19 also reported that because 3D
3D surface models, they are somewhat inferior to surface models superimposition is currently time
reality when threshold-based methods are used. consuming and computing intensive, its use in
Differences in the segmentation process of CBCT routine clinical practice is not very practical.
analysis can result in clinically significant var- Therefore, more simplified analysis techniques
iances between measurements that can affect are required for 3D superimposition techniques to
treatment. Therefore, when making a decision be viable in routine daily practice.
based on CBCT measurement on a 3D surface Clinicians, scientists, and engineers have
model, special attention must be given in the developed techniques for superimposing facial
270 Park et al

Figure 5. 2D photograph and 3D stereophotograph superimposition over CBCT scan. (A) Limited view of a 2D
photographs superimposed over volume rendered maxillary dentition from a CBCT scan with InVivoDental
software by Anatomage. (B) The matching process of CBCT soft tissue, and 3D stereophotograph, captured by 3D
MDfaceTM stereophotograhic system; initial registration surfaces are shown in green. (C) Merged image of the 3D
soft tissue stereophotograph and the 3D CBCT hard tissue image. (For interpretation of the references to color in
this figure legend, the reader is referred to the web version of this article.)

2D photographs48 and 3D photographs,29,37,39,53 Conclusion

and digital models40,42,54,55 over CBCT scans. It is
Superimposition of 3D CBCT scans in ortho-
reported that the integration of 3D photographs
dontics and orthognathic surgery is usually used
and CBCT images has shown minimal errors in the
for craniofacial growth evaluations or for post-
assessment of bone and soft tissue.56 Therefore,
this process can be used as an objective tool for treatment assessment of changes in the cra-
diagnosis and treatment planning in orthodontics niofacial structures such as a 3D evaluation of
and orthognathic surgery (Figs. 5 and 6).56 changes in the mid-facial soft tissue, the con-
Multi-sensor image fusion techniques have dyles, rami, chin, maxilla, and dentition. Per-
allowed for more advanced diagnosis and treat- forming quantitative analysis on superimposed
ment strategies in the field of 3D imaging.57 In images and visualization of the craniofacial
addition, image fusion is considered to be a changes at different time points by color-coded
dependable and precise method which is not fused images are two of the main advantages of
affected by surgical and longitudinal changes.19,58 3D superimpositions.
There has also been an increase in use of image Various 3D superimposition methods have
fusion techniques in orthodontics and max- been developed including the use of registration
illofacial surgery, which has provided more accu- points, information theory such as the ICP
rate and reliable information for surgical method and voxel-based image registration, a
outcomes and treatment progress.10,25,59 However, single software package using a step-by-step
3D CBCT image superimposition via semi- manual technique, and triple voxel-based
automatic registration methods is a relatively superimposition. These methods are mainly
new concept on which little research has been based on registration of anatomic structures and
conducted yet relative to its accuracy.58 The automated fusion of the images using compu-
superimposition of certain landmarks is a more terized mathematical algorithms. More recently,
precise and reproducible method than conven- superimposition of CBCT scans with 2D and 3D
tional 2D cephalometric images, especially with skin surface photographs have been used to
the bilateral landmarks such as condylion, gonion, assess 3D maxillofacial hard and soft tissue
and orbitale in 3D images. On the other hand, it is morphology, in addition to the evaluation of
more difficult to accurately define certain the linear and angular changes in clinical
landmarks in the third dimension of the medio- procedures.
lateral direction.60 A summary of photographs/ Even though the 3D superimposition meth-
digital models and CBCT superimposition ods are clinically more accurate and reprodu-
methods, applications, advantages, and limita- cible than conventional 2D methods, the
tions is presented in Table 2. precision of these methods is still a subject of
 3D CBCT superimposition 271

Figure. 6. 3D photograph superimposition over CBCT scan with Morpheus3D CT software (Morpheus). The 3D
photographs superimposed over (A) volume rendered mandible, (B) mid-sagittal section of the CBCT (frontal
view), (C) mid-sagittal section of the CBCT (lateral view), (D) full CBCT (¾ view), (E) superior view with anterior
coronal section of the CBCT, and (F) inferior ¾ view with sectioned out mandible.

controversy in the literature. Most of the settings are time consuming. Therefore,
limitations of 3D superimposition techni- more research-based technological develop-
ques are related to imaging and landmark ments are required in this field to create
identification errors and software/hardware more efficient and faster 3D superimposition
related errors. In addition, most of the techniques with higher accuracy and higher
methods that are currently being used in clinical reproducibility.

Table 2. Photographs/digital models and CBCT superimposition

Methods Clinical applications Advantages Limitations

Registration and superimposition Quantitatively assess 3D maxillofacial Precise and Expensive

of 3D digital models, and/or morphology, and evaluate the linear dependable Limited evidence
skin surface of facial 2D and 3D and angular changes in facial soft method Time consuming
photographs over CBCT scans and hard tissues in clinical procedures Technique sensitive
Assessment of bone and soft tissue
Advanced diagnosis and treatment
planning in orthodontics and
orthognathic surgery
Provides a more accurate and reliable
information for craniofacial growth,
surgical outcomes, and treatment
progress than conventional methods
272 Park et al

Acknowledgment 17. Stratemann SA, Huang JC, Maki K, Miller AJ, Hatcher
DC. Comparison of cone beam computed tomography
The authors would like to thank Ms. Jennifer Huynh imaging with physical measures. Dentomaxillofac Radiol.
for her help with literature review. 2008;37:80–93.
18. Motta AT, de Assis Ribeiro Carvalho F, Cevidanes LH, de
Oliveira Almeida MA. Assessment of mandibular advance-
ment surgery with 3D CBCT models superimposition.
References Dental Press J Orthod. 2010;15:45e1–45e12.
1. Mah JK, Huang JC, Choo H. Practical applications of 19. Cevidanes LH, Heymann G, Cornelis MA, DeClerck HJ,
cone-beam computed tomography in orthodontics. J Am Tulloch JF. Superimposition of 3-dimensional cone-beam
Dent Assoc. 2010;141(suppl 3):7S–13S. computed tomography models of growing patients. Am J
2. Smith BR, Park JH, Cederberg RA. An evaluation of cone- Orthod Dentofacial Orthop. 2009;136:94–99.
beam computed tomography use in postgraduate ortho- 20. Mah JK, Yi L, Huang RC, Choo H. Advanced applications
dontic programs in the United States and Canada. J Dent of cone beam computed tomography in orthodontics.
Educ. 2011;75:98–106. Semin Orthod. 2011;17:57–71.
3. Heon JC. Three-dimensional superimposition. PCSO Bull. 21. Cevidanes LH, Oliveira AE, Grauer D, Styner M, Proffit
2010;82:23–26. WR. Clinical application of 3D imaging for assessment of
4. Danforth RA, Peck J, Hall P. Cone beam volume treatment outcomes. Semin Orthod. 2011;17:72–80.
tomography: an imaging option for diagnosis of complex 22. Nguyen T, Cevidanes L, Paniagua B, Zhu H, Koerich L,
mandibular third molar anatomical relationships. J Calif De Clerck H. Use of shape correspondence analysis to
Dent Assoc. 2003;31:847–852. quantify skeletal changes associated with bone-anchored
5. Vishwanath AE, Sandesh SP, Nandini SN, Neelima A. Class III correction. Angle Orthod. 2014;84:329–336.
CBCT—an advanced diagnostic tool for orthodontics and 23. Choi JH, Mah J. A new method for superimposition of
other specialties in dentistry—a systematic review. Orthod CBCT volumes. J Clin Orthod. 2010;44:303–312.
Cyber J. 2012. 24. Maes F, Collignon A, Vandermeulen D, Marchal G,
6. Hatcher DC, Aboudara CL. Diagnosis goes digital. Am J Suetens P. Multimodality image registration by max-
Orthod Dentofacial Orthop. 2004;125:512–515. imization of mutual information. IEEE Trans Med Imaging.
7. Harrell WE Jr, Hatcher DC, Bolt RL. In search of 1997;16:187–198.
anatomic truth: 3-dimensional digital modeling and the 25. Cevidanes LH, Bailey LJ, Tucker GR Jr., et al. Super-
future of orthodontics. Am J Orthod Dentofacial Orthop. imposition of 3D cone-beam CT models of orthognathic
2002;122:325–330. surgery patients. Dentomaxillofac Radiol. 2005;34:369–375.
8. Miller AJ, Maki K, Hatcher DC. New diagnostic tools in 26. Tai K, Park JH, Mishima K, Shin JW. 3-Dimensional cone-
orthodontics. Am J Orthod Dentofacial Orthop. 2004;126: beam computed tomography analysis of transverse
395–396. changes with Schwarz appliances on both jaws. Angle
9. Agrawal JM, Agrawal MS, Nanjannawar LG, Parushetti Orthod. 2011;81:670–677.
AD. CBCT in orthodontics: the wave of future. J Contemp 27. Tai K, Park JH, Superimposition of 3-dimensional cone-
Dent Pract. 2013;14:153–157. beam computed tomography for 2-dimensional image
10. Cevidanes LH, Styner MA, Proffit WR. Image analysis and analysis,. In Park JH: Computed Tomography: New
superimposition of 3-dimensional cone-beam computed Research, ed Nova Science Publishers, Inc., Hauppauge,
tomography models. Am J Orthod Dentofacial Orthop. NY, 2013, 457–475.
2006;129:611–618. 28. Rayapudi N, Padmalatha C, Gandikopta CS, Yudhistar PV,
11. da Motta AT, de Assis Ribeiro Carvalho F, Oliveira AE, Tircoveluri S. A comparative study of linear measure-
Cevidanes LH, de Oliveira Almeida MA. Superimposition ments of facial skeleton using computed tomography and
of 3D cone-beam CT models in orthognathic surgery. traditional cephalometry. APOS Trends Orthod. 2013;3:7.
Dent Press J Orthod. 2010;15:39–41. 29. Naudi KB, Benramadan R, Brocklebank L, Ju X, Khambay
12. Scarfe WC, Farman AG, Sukovic P. Clinical applications of B, Ayoub A. The virtual human face: superimposing the
cone-beam computed tomography in dental practice. simultaneously captured 3D photorealistic skin surface of
J Can Dent Assoc. 2006;72:75–80. the face on the untextured skin image of the CBCT scan.
13. Periago DR, Scarfe WC, Moshiri M, Scheetz JP, Silveira Int J Oral Maxillofac Surg. 2013;42:393–400.
AM, Farman AG. Linear accuracy and reliability of cone 30. Tai K, Park JH, Mishima K, Hotokezaka H. Using
beam CT derived 3-dimensional images constructed superimposition of 3-dimensional cone-beam computed
using an orthodontic volumetric rendering program. tomography images with surface-based registration in
Angle Orthod. 2008;78:387–395. growing patients. J Clin Pediatr Dent. 2010;34:361–367.
14. Farman AG, Scarfe WC. The basics of maxillofacial cone 31. Tai K, Hotokezaka H, Park JH, et al. Preliminary cone-
beam computed tomography. Semin Orthod. 2009;15:11. beam computed tomography study evaluating dental and
15. Kapila S, Conley RS, Harrell WE Jr. The current status of skeletal changes after treatment with a mandibular
cone beam computed tomography imaging in orthodon- Schwarz appliance. Am J Orthod Dentofacial Orthop.
tics. Dentomaxillofac Radiol. 2011;40:24–34. 2010;138:262.e1–262.e11.
16. Stratemann SA, Huang JC, Maki K, Hatcher DC, Miller 32. Heymann GC, Cevidanes L, Cornelis M, De Clerck HJ,
AJ. Evaluating the mandible with cone-beam computed Tulloch JF. Three-dimensional analysis of maxillary
tomography. Am J Orthod Dentofacial Orthop. 2010;137: protraction with intermaxillary elastics to miniplates.
S58–S70. Am J Orthod Dentofacial Orthop. 2010;137:274–284.
 3D CBCT superimposition 273

33. Swennen GR, Mollemans W, De Clercq C, et al. A cone- 47. Cevidanes LH, Bailey LJ, Tucker SF, et al. Three-
beam computed tomography triple scan procedure to dimensional cone-beam computed tomography for
obtain a three-dimensional augmented virtual skull assessment of mandibular changes after orthognathic
model appropriate for orthognathic surgery planning. surgery. Am J Orthod Dentofacial Orthop. 2007;131:44–50.
J Craniofac Surg. 2009;20:297–307. 48. Grauer D, Cevidanes LS, Proffit WR. Working with
34. RM, Maal TJ, Breuning KH, Berge SJ, Mostafa YA, DICOM craniofacial images. Am J Orthod Dentofacial
Kuijpers-Jagtman AM. Accuracy and reproducibility of Orthop. 2009;136:460–470.
voxel based superimposition of cone beam computed 49. Cho HJ. A three-dimensional cephalometric analysis.
tomography models on the anterior cranial base and the J Clin Orthod. 2009;43:235–252.
zygomatic arches. PLoS One. 2011;6:e16520. 50. Park SB, Kim YI, Son WS, Hwang DS, Cho BH. Cone-
35. Cevidanes LH, Motta A, Proffit WR, Ackerman JL, Styner M. beam computed tomography evaluation of short- and
Cranial base superimposition for 3-dimensional evaluation long-term airway change and stability after orthognathic
of soft-tissue changes. Am J Orthod Dentofacial Orthop. surgery in patients with Class III skeletal deformities:
2010;137:S120–S129. bimaxillary surgery and mandibular setback surgery. Int J
36. Gianquinto JR, Tuncay OC, Sciote JJ, Yang J. A method of Oral Maxillofac Surg. 2012;41:87–93.
superimposition of CBCT volumes in the posterior cranial 51. Engelbrecht WP, Fourie Z, Damstra J, Gerrits PO, Ren Y.
base. The Temple University Digital Library. The Temple The influence of the segmentation process on 3D
University; 2011. measurements from cone beam computed tomography-
37. Kau CH, Olim S, Nguyen JT. The future of orthodontic derived surface models. Clin Oral Investig. 2013;17:1919-27.
diagnostic records. Semin Orthod. 2011;17:6. 52. Terajima M, Yanagita N, Ozeki K, et al. Three-
38. Sun Y, Luebbers HT, Politis C. Three-dimensional virtual dimensional analysis system for orthognathic surgery
model surgery to fabricate the digital intermediate splint.
patients with jaw deformities. Am J Orthod Dentofacial
J Craniofac Surg. 2013;24:563–565.
Orthop. 2008;134:100–111.
39. Kau CH. Creation of the virtual patient for the study of
53. Jayaratne YS, McGrath CP, Zwahlen RA. How accurate are
facial morphology. Facial Plast Surg Clin North Am.
the fusion of cone-beam CT and 3-D stereophotographic
2011;19:615–622.
images?PLoS One. 2012;7:e49585.
40. Park TJ, Lee SH, Lee KS. A method for mandibular dental
54. Lin HH, Chiang WC, Lo LJ, Sheng-Pin Hsu S, Wang CH,
arch superimposition using 3D cone beam CT and
Wan SY. Artifact-resistant superimposition of digital
orthodontic 3D digital model. Korean J Orthod. 2012;42:
dental models and cone-beam computed tomography
169–181.
images. J Oral Maxillofac Surg. 2013;71:1933–1947.
41. Farman AG, Scarfe WC. Development of imaging
55. Cevidanes LH, Tucker S, Styner M, et al. Three-
selection criteria and procedures should precede ceph-
dimensional surgical simulation. Am J Orthod Dentofacial
alometric assessment with cone-beam computed tomog-
Orthop. 2010;138:361–371.
raphy. Am J Orthod Dentofacial Orthop. 2006;130:257–265.
42. Chenin DL, Chenin DA, Chenin ST, Choi J. Dynamic 56. Jayaratne YSN, McGrath CPJ, Zwahlen RA. How accurate
cone-beam computed tomography in orthodontic treat- are the fusion of cone-beam CT and 3-D stereophoto-
ment. J Clin Orthod. 2009;43:507–512. graphic images? PLoS One. 2012;7.
43. Spin-Neto R, Marcantonio E Jr, Gotfredsen E, Wenzel A. 57. Tsao J. Interpolation artifacts in multimodality image
Exploring CBCT-based DICOM files. A systematic review registration based on maximization of mutual informa-
on the properties of images used to evaluate maxillofacial tion. IEEE Trans Med Imaging. 2003;22:854–864.
bone grafts. J Digit Imaging. 2011;24:959–966. 58. Lee JH, Kim MJ, Kim SM, Kwon OH, Kim YK. The 3D CT
44. Lee JY, Kim YI, Hwang DS, Park SB. Effect of setback Le superimposition method using image fusion based on the
Fort I osteotomy on midfacial soft-tissue changes as maximum mutual information algorithm for the assess-
evaluated by cone-beam computed tomography super- ment of oral and maxillofacial surgery treatment results.
imposition for cases of skeletal Class III malocclusion. Int J Oral Surg Oral Med Oral Pathol Oral Radiol. 2012;114:
Oral Maxillofac Surg. 2013;42:790–795. 167–174.
45. Efstratiadis S, Baumrind S, Shofer F, Jacobsson-Hunt U, 59. Lagravere MO, Major PW, Carey J. Sensitivity analysis for
Laster L, Ghafari J. Evaluation of Class II treatment by plane orientation in three-dimensional cephalometric
cephalometric regional superpositions versus conven- analysis based on superimposition of serial cone beam
tional measurements. Am J Orthod Dentofacial Orthop. computed tomography images. Dentomaxillofac Radiol.
2005;128:607–618. 2010;39:400–408.
46. Ghafari J, Engel FE, Laster LL. Cephalometric super- 60. Ludlow JB, Gubler M, Cevidanes L, Mol A. Precision of
imposition on the cranial base: a review and a comparison cephalometric landmark identification: cone-beam com-
of four methods. Am J Orthod Dentofacial Orthop. puted tomography vs conventional cephalometric views.
1987;91:403–413. Am J Orthod Dentofacial Orthop. 2009;136:312.e1–312.e10.
Diagnosis and evaluation of skeletal Class III
patients with facial asymmetry for orthognathic
surgery using three-dimensional computed
tomography
Sung-Jin Kim, Hyoung-Seon Baik, Chung-Ju Hwang, and Hyung-Seog Yu

Patients with facial asymmetry, especially skeletal Class III, have skeletal
asymmetry in the cranial base, maxilla, mandible as well as asymmetry of
soft tissues including the masticatory muscles. Three-dimensional (3D)
computed tomography is considered as an effective tool to analyze these
patients, since it enables accurate measurement without distortion regard-
less of head orientation using real anatomic surface landmarks and 3D
representation of complex morphology including volumetric measurements.
In this article, morphologic characteristics of the cranial base, mandible, and
masticatory muscle in skeletal Class III patients with facial asymmetry are
addressed, and 3D analysis for these patients is suggested. Furthermore, 3D
evaluation of hard- and soft-tissue changes after orthognathic surgery is
discussed, which can contribute to treatment planning for predictable
treatment results. (Semin Orthod 2015; 21:274–282.) & 2015 Elsevier Inc.
All rights reserved.

Introduction been used for the evaluation of facial asymmetry.

However, such two-dimensional (2D) images
acial asymmetry is a relatively common
F phenomenon. The prevalence of clinically
apparent asymmetry has been reported as 23% in
have inherent sources of error including internal
orientation error, external orientation error,
geometric error, and association error, mainly
the orthodontic population1 and 34–38.6% in
due to visual phenomena related to a 2D rep-
patients with dentofacial deformities.2,3 Facial
resentation of a three-dimensional (3D) object.9
asymmetry is more frequently found in patients
The usefulness of 2D cephalometry is seriously
of skeletal Class III, with an incidence of 40–
limited for the evaluation of facial asymmetry
80%,1–4 possibly due to excessive mandibular
because most 2D measurements are distorted in
growth in the case of mandibular prognathism.3
facial asymmetry patients, and 2D cephalometry
Therefore, thorough evaluation for facial
does not provide enough information on the
asymmetry is particularly important in skeletal
complex morphology of asymmetric facial units
Class III patients.
such as size, shape, position, and orientation.10,11
Traditionally, frontal cephalograms,5,6 sub-
3D CT has recently been introduced for
mentovertex views,7 and panoramic views8 have
orthodontic diagnosis and is considered to be an
effective tool for analyzing facial asymmetry
Department of Orthodontics, School of Dentistry, Yonsei
because it enables accurate measurement of
University, 134 Shinchon-dong, Seodaemun-gu, Seoul 120-752, anatomical structures in 3 dimensions and pro-
South Korea. vides a 3D representation of complex morphol-
Address correspondence to Hyung-Seog Yu, DDS, MS, PhD, ogy, including volume.12,13 Thus, non-growing
Department of Orthodontics, School of Dentistry, Yonsei University,
skeletal Class III patients with severe facial asym-
134 Shinchon-dong, Seodaemun-gu, Seoul 120-752, South Korea.
E-mail: [email protected]
metry, who are typically treated with a combina-
& 2015 Elsevier Inc. All rights reserved.
tion of orthognathic surgery and orthodontic
1073-8746/15/1801-$30.00/0 therapy,14–16 can benefit from 3D CT for diagnosis
http://dx.doi.org/10.1053/j.sodo.2015.07.004 and the planning of surgical procedures.

Seminars in Orthodontics, Vol 21, No 4, 2015: pp 274–282 274

 3D CT analysis for skeletal Class III patients with facial asymmetry 275

Morphologic characteristics of skeletal this type of asymmetry is asymmetric growth of the

Class III patients with facial asymmetry condyle or mandible.26,27 Rolling of the maxillary
occlusal plane toward the deviated side usually
Mandible
accompanies these malformations since the dif-
Mandibular asymmetry is one of the most ference in ramal heights results in compensatory
prominent features in patients with facial asym- asymmetric vertical growth of the maxilla. These 2
metry. Severt and Proffit2 reported that 75% of groups include 44.2–48.6% of the patients with
patients with facial asymmetry showed deviation facial asymmetry.24,25
of the chin, while 36% had middle-third asym- The third subgroup is characterized by menton
metry; asymmetry of the upper face was observed deviation without differences between ramal
in only 5%. Haraguchi et al.4 and Maeda et al.17 heights or rolling of the maxillary occlusal plane.
also reported that asymmetry was observed most Asymmetric ramal inclination on the coronal plane
often in the lower part of the face. Thus, most contributes to the menton deviation. This type of
previous 3D CT studies of facial asymmetry have asymmetry has been described as displacement
focused on mandibular asymmetry. asymmetry28 or functional asymmetry,29 which
The mandible is usually divided into deviated results from a functional shift of the mandible
and non-deviated sides according to menton caused by occlusal interference. Previous studies
deviation from the midsagittal reference line, have indicated that posterior unilateral crossbite
since deviation of the chin has a great impact on can result in asymmetric condylar position,30,31 and
the perception of facial asymmetry.4,18 In typical adaptive remodeling of the temporomandibular
facial asymmetry, the mandibular body, ramus, joint to the mandibular displacement might lead to
and condyle are elongated on the non-deviated skeletal asymmetry.32,33 Of the patients with facial
sides.13,19–22 The ramal and condylar volumes are asymmetry, 22.2–44.2% belong to this group.24,25
also increased on the non-deviated sides,13,20 Atypical facial asymmetry, the fourth subgroup,
while the mandibular body is not increased in accounts for 11.6% (22.2%) of the patients with
volume despite its elongation on the non- facial asymmetry.24,25 The most prominent feature
deviated sides.13 In addition to being of this type is reverse canting of the maxillary
elongated, the ramus shows mesial and medial occlusal plane toward the non-deviated side. This
inclination on the non-deviated sides, con- canting is not caused by a shortened ramus or
tributing to menton deviation.22,23 Patients with condyle on the non-deviated side. In fact, the ramal
asymmetry show more superior, posterior, and height, measured as the distance between the
lateral positioning of the gonion19,23 and a larger gonion and condylion in 3D images, is still longer
gonial angle on the deviated sides.23 Finally, the on the non-deviated side,25 although the gonion
menton deviation, which usually determines the and antegonion are positioned more superiorly
degree of facial asymmetry, correlates with and laterally on the non-deviated side.24,25 The
differences in the hemi-mandibular volume, etiology of this type of asymmetry is not clear. A
mandibular body length, ramal height, ramal unilateral masticatory habit and aberrant growth of
volume, condylar length, condylar volume, and the mandible have been proposed as the possible
ramus inclination between the non-deviated and cause,24 but this hypothesis requires further
deviated sides.13,22,23 evidence to be confirmed.
The morphology of the mandible in individual
patients is so variable that facial asymmetry cannot
be identified merely on the basis of typical features. Cranial base
Based on this premise, Hwang et al.24 and Baek Derived from the Greek word for “twisted head,”
et al.25 classified facial asymmetry patients into 4 plagiocephaly describes a unilateral flattening of an
subgroups based on morphologic characteristics of anterior or posterior quarter of the skull34 that
maxillary and mandibular asymmetry using 2D results in asymmetry of the cranial base. Depending
frontal cephalograms and 3D CT images, on the cause, plagiocephaly is generally divided into
respectively. The first and second subgroups 2 types: synostotic plagiocephaly, caused by uni-
include the aforementioned typical facial lateral craniosynostosis with premature fusion of the
symmetry, with the only difference between them coronal or lambdoid suture, and deformational
being the magnitude of discrepancy. The cause of plagiocephaly, resulting from external molding
276 Kim et al

forces applied to the pliable infant head.35 Most not significantly different between the non-
studies regarding cranial base asymmetry have deviated and deviated sides, or between
focused on plagiocephaly in infancy in response patients with and without asymmetry, concurring
to a large increase in the incidence of deformational with Kwon et al.19 and Baek et al.23 Regarding the
plagiocephaly after initiation of the “back to sleep” relationship to mandibular asymmetry, the
campaign by the American Academy of Pediatrics in menton deviation and ramal volume correlated
199236–38 to prevent sudden infant death syn- with the difference in cranial base volume
drome.39 Although synostotic plagiocephaly is between the non-deviated and deviated sides.
reported to occur in approximately 1 out of 10,000 These findings suggest that cranial base
live births,40 the prevalence of deformational asymmetry can affect mandibular asymmetry,
plagiocephaly in infants at 4 months of age is which is the most prominent feature of facial
reported to be 19.7–48%.41,42 However, observation asymmetry.46 The temporomandibular joint
of the natural history of deformational plagiocephaly position and mandibular shape are reported to
without treatment shows that the prevalence dimin- be affected in plagiocephalic patients with an
ishes over time, with 3.3% at 2 years and 2.4% at 3 asymmetric cranial base.47–49 Kwon et al.19 found
years of age41,43; only 0.5–1% of children are esti- that the position of the mandibular condyle
mated to enter school with notable plagiocephaly.35 correlated with the position of the mastoid
Varying degrees of cranial base asymmetry can process and the petrous ridge angle in adults.
be observed in adult patients with no notable In addition, Endo et al.50 reported that facial
facial asymmetry44,45 as well as in those with facial asymmetry patients had larger condylar fossae
asymmetry.7 However, only a few studies address and longer condylar processes on the non-
cranial base asymmetry in adults. Kwon et al.19 deviated side. Therefore, the cranial base and
and Baek et al.23 investigated the morphology of mandible are not only structurally related but
the cranial base in adult patients with and also overgrow together on the non-deviated sides,
without facial asymmetry using 3-D CT images. contributing to the facial asymmetry.
These studies reported no difference in the
anterior, middle, and posterior cranial base angle
between the non-deviated and deviated sides, or Masticatory muscle
between patients with and without asymmetry. The size and orientation of masticatory muscles
Based on these findings, the authors concluded vary with craniofacial morphology. Long-faced
that although facial asymmetry was accompanied adults have thinner masticatory muscles with a
by various degrees of cranial base asymmetry, the smaller cross-sectional area51,52; patients with
degree of facial asymmetry was not related to the mandibular prognathism also have thinner and
morphology of the cranial base. smaller masseter muscles that are more anteriorly
A recent study by Kim et al.46 found that tilted compared to those of normal subjects.53
cranial base asymmetry was related to With regard to facial asymmetry, masticatory
mandibular asymmetry in patients of skeletal muscles in hemifacial microsomia have been
Class III. In this study, 60 adults with mandibular observed to have a smaller volume on the
prognathism were separated into groups based affected side than on the unaffected side.54,55
on the presence (MD 4 4 mm; n ¼ 30) or However, unlike hemifacial microsomia, which is
absence (MD o 2 mm; n ¼ 30) of mandibular caused by unilateral hypoplasia, facial asymmetry
asymmetry, and linear, angular, and volumetric in mandibular prognathism results from unilat-
measurements of the cranial base and mandible eral or bilateral hyperplasia; thus, the morphology
were performed using 3-D CT images. The results of the masticatory muscle is also different.
showed that in the patients with mandibular In patients with mandibular prognathism and
asymmetry, the hemi-base, anterior cranial base, facial asymmetry, there are no significant dif-
and middle cranial base volumes were larger, and ferences in the volume of the masticatory muscles
the crista galli to sphenoid, sphenoid to petrous except in the medial pterygoid muscle, which is
ridge, anterior clinoid process to petrous ridge, smaller on the non-deviated side.56–58 Further,
and vomer to petrous ridge lengths were greater no significant differences has been found in the
on the non-deviated sides. However, the anterior, cross-sectional area, thickness, length, or width of
middle, and posterior cranial base angles were masseter muscles between the deviated and non-
 3D CT analysis for skeletal Class III patients with facial asymmetry 277

deviated sides.57–59 Only the angle between the (Cg–ANS) plane, which have also been used for
FH plane and the anterior border of the masseter 3D CT analysis.23,63,64 However, these internal
muscle (ABM) differs significantly, which is more reference planes are highly unreliable for anal-
vertical on the non-deviated side.57,58 These ysis of facial asymmetry because cranial base
findings contrast those of Goto et al.,60 who asymmetry often accompanies and is related to
reported that the masseter muscle was facial asymmetry,46 resulting in distortion of the
significantly shorter and of lower volume on reference planes. Therefore, external reference
the deviated side, possibly because their subjects planes, which are defined by natural head
did not have mandibular prognathism. The more posture, are recommended for the analysis of
vertical orientation of the masseter muscle on the facial asymmetry.11 Further studies are needed to
non-deviated side can be explained by the mesial determine the most reproducible and
inclination of the ramus on the non-deviated side convenient way of capturing the natural head
and posterior positioning of the gonion on the posture and to investigate postural change of the
deviated side,19,23 since its superficial head, head after correction of asymmetry.
which is a major part of the muscle, arises from Park et al.65 have proposed an analytical method
the zygomatic arch, passing downward and for evaluating craniofacial morphology using 3D
backward to be inserted into the angle of the CT, and reported mean values of the zygomatic,
mandible. Although the sizes of masticatory maxillary, and mandibular measurements in 30
muscles generally do not differ between the non- subjects who had normal occlusion and a balanced
deviated and deviated sides in facial asymmetry face. A 14-year-old boy who visited the Orthodontic
patients with mandibular prognathism, they are Department of Yonsei University with the chief
significantly smaller than in normal subjects.57,58 complaint of facial asymmetry was evaluated using
Orthognathic surgery combined with ortho- this method (Fig., Table 1). He demonstrated
dontic therapy has been used to treat these cranial base asymmetry and a parallelogram-shaped
patients, not only correcting skeletal discrep- cranium resulting from deformational plagioce-
ancies but also improving the balance of soft phaly with flattening of the left occipital bone. The
tissues and masticatory function. The size of the menton was deviated to the right side because the
masseter muscle reportedly decreases tempora- maxilla and mandible showed clockwise rolling
rily but reverts back to preoperative status within (canting) and yawing (rotation) to the right side
one year61,62 and continues to increase, reaching due to the longer mandibular body and increased
the values of normal subjects in cross-sectional ramal height on the left side and the differences in
area, thickness, and width within 4 years of follow- ramal inclinations between the 2 sides.
up.58 The differences in the masseter muscle
angle between the non-deviated and deviated
Hard- and soft-tissue changes following
sides and between patients with and without
orthognathic surgery
asymmetry also resolve within 1 year.57,58
Orthognathic surgery combined with orthodontics
is often required to correct severe skeletal dis-
3D analysis for skeletal Class III patients
crepancies in skeletal Class III patients with facial
with facial asymmetry
asymmetry. However, ideal repositioning of the
To analyze craniofacial structure in 3 dimen- underlying skeleton may not lead to ideal soft
sions, a proper reference system needs to be tissue appearance because the soft tissue response
established. Reference planes are divided into does not follow the exact changes of the hard
internal reference planes, which are determined tissue.66 Therefore, the soft-tissue response, espe-
by internal elements such as landmarks of cra- cially the ratio of soft-tissue changes to hard-tissue
niofacial structure, and external reference changes, should be considered in the diagnosis
planes, which are determined by external ele- and treatment planning for these patients.
ments such as the natural head posture.11 Cranial Traditionally, the soft-tissue response following
base landmarks define most traditional reference orthognathic surgery in skeletal Class III patients
planes for 2D cephalometry, including the has been investigated primarily in the sagittal
Frankfort Horizontal (FH) plane, Sella–Nasion and vertical direction using 2D lateral cephalo-
(SN) plane, and Crista galli–Anterior nasal spine grams.67–72 However, these studies have the
278 Kim et al

Figure. 3D CT of a patient with facial asymmetry and 3D chart of this patient.

inherent limitations of 2D analysis. More recently, but just used a summary based on all subjects. Lee
3D laser scanners have been introduced and used et al.77 used CBCT to investigate 18 skeletal Class
for the evaluation of facial soft tissue, but they III patients who were divided into groups based
cannot examine hard and soft tissues simulta- on the presence (MD from the facial midline
neously; thus, additional lateral cephalograms are
needed in studies evaluating underlying hard Table 1. 3D analysis65 of a patient
tissues,66,73,74 which can cause inaccuracy in the
Patient
measurements. Furthermore, when facial asym-
metry is present in skeletal Class III patients, the Zygoma
hard- and soft-tissue changes in the transverse Facial index 93.6
Midface angle (1)
direction, which cannot be measured in lateral Rt 131.8
cephalograms or 3D laser scanners should also be Lt 140.7
considered. 3D cone beam CT (CBCT) imaging is Maxilla
suitable for analysis of these patients since it can Canting (1) 2.6
Rotation (1) 2.6
produce accurate images of both soft and hard Divergence (1) 13.4
tissues simultaneously and with patients posi- Mandible
toned in an upright position, giving a natural Canting (1) 1.3
shape to the facial soft tissue.75 Rotation (1) 6.3
Divergence (1) 42.8
Several studies have investigated the ratio of
Mandibular measurements
soft-tissue changes to hard-tissue changes fol- Body length (mm)
lowing orthognathic surgery in skeletal Class III Rt 88.3
patients.66–74 Two of these studies used 3D CT for Lt 94.2
Ramal height (mm)
their analysis of skeletal Class III patients with Rt 58.0
facial asymmetry (Table 2).76,77 Jung et al.76 used Lt 61.6
multi-slice CT (MSCT) to investigate 3D changes Internal ramal inclination (1)
Rt 81.6
of hard and soft tissues in 17 skeletal Class III Lt 89.1
patients who were grouped based on the pres- External ramal inclination (1)
ence or absence of facial asymmetry according to Rt 82.4
Lt 85.8
the amount of MD from the facial midline. Lateral ramal inclination (1)
However, the study failed to calculate the ratio Rt 80.9
between hard and soft tissues not for each group, Lt 74.0
 3D CT analysis for skeletal Class III patients with facial asymmetry
Table 2. Proportions of soft-tissue changes to hard-tissue change
ΔLi/ΔL1E ΔB0 /ΔB ΔPog0 /ΔPog ΔMe0 /ΔMe
A–P Vertical Transverse A–P Vertical Transverse A–P Vertical Transverse A–P Vertical Transverse
ratio ratio ratio ratio ratio ratio ratio ratio ratio ratio ratio ratio

2D Lateral Cephalogram
Marsan et al.67 0.55 NS NS NS NS NS
Marsan et al.68 0.28 0.55 0.30 0.58 0.36 0.61
Chew69 0.84 0.88 1.01 0.78 0.85 1.08
Lin and Kerr70 0.99 0.88 1.02 0.82 0.99 0.89 0.99 0.96
Becker et al.71 0.74 0.88 0.87
Jokic et al.72 1.07 1.05 1.02
3D Laser scan & 2D Lateral
Cephalogram
Soncul and Bamber66 0.64 0.90 0.97
Baik and Kim74 0.72 0.85 0.89 0.98
3D CT
Jung et al.76 0.78 NS 0.66 0.78 NS 0.62 0.86 NS 0.58 0.56 NS NS
Lee et al.77 Symmetry group 0.88 NS NS 0.94 NS 0.53 0.84 NS 0.67 0.96 NS 0.90
Asymmetry group 0.88 NS 0.52 0.95 NS 0.77 0.91 NS 0.88 1.03 NS NS
NS: statistically no significant correlation.

279
280 Kim et al

44 mm) or absence (MD from the facial midline 8. Van Elslande DC, Russett SJ, Major PW, Flores-Mir C.
o2 mm) of facial asymmetry and found that the Mandibular asymmetry diagnosis with panoramic imag-
ing. Am J Orthod Dentofacial Orthop. 2008;134:183–192.
ratio of ΔB0 to ΔB and ΔPog0 to ΔPog on the 9. Quintero JC, Trosien A, Hatcher D, Kapila S. Craniofacial
horizontal axis were higher in patients with facial imaging in orthodontics: historical perspective, current
asymmetry than in those without. This result may status, and future developments. Angle Orthod. 1999;69:
be due to the greater deviation in the mandible 491–506.
on the horizontal axis in patients with facial 10. Gateno J, Xia JJ, Teichgraeber JF. Effect of facial
asymmetry on 2-dimensional and 3-dimensional cepha-
asymmetry, resulting in more stretching of the lometric measurements. J Oral Maxillofac Surg. 2011;69:
soft tissues. These differences should be taken 655–662.
into account when planning orthognathic 11. Gateno J, Xia JJ, Teichgraeber JF. New 3-dimensional
surgery in skeletal Class III patients with facial cephalometric analysis for orthognathic surgery. J Oral
Maxillofac Surg. 2011;69:606–622.
asymmetry.
12. Yanez-Vico RM, Iglesias-Linares A, Torres-Lagares D,
Gutierrez-Perez JL, Solano-Reina E. Diagnostic of cranio-
facial asymmetry. Literature review. Med Oral Patol Oral Cir
Bucal. 2010;15:e494–e498.
Conclusion
13. You KH, Lee KJ, Lee SH, Baik HS. Three-dimensional
Facial asymmetry is frequently found in skeletal computed tomography analysis of mandibular morphol-
Class III patients, often requiring orthognathic ogy in patients with facial asymmetry and mandibular
prognathism. Am J Orthod Dentofacial Orthop. 2010;138:540.
surgery in those who are no longer growing. e1–540.e8; discussion 1.
Mandibular asymmetry is the primary patho- 14. Miyatake E, Miyawaki S, Morishige Y, Nishiyama A, Sasaki A,
logical feature in these patients, but asymmetries Takano-Yamamoto T. Class III malocclusion with severe
of the cranial base and masticatory muscle are facial asymmetry, unilateral posterior crossbite, and tem-
also found in the majority of patients. Traditional poromandibular disorders. Am J Orthod Dentofacial Orthop.
2003;124:435–445.
2D analysis cannot reveal the complex mor- 15. Fukui T, Tsuruta M, Yong-Bong C, Kuwahara Y. Multi-
phology involved in these asymmetric structures lingual bracket treatment combined with orthognathic
in individual patients; however, 3D CT analysis surgery in a skeletal Class III patient with facial asymme-
can provide valuable information for the diag- try. Am J Orthod Dentofacial Orthop. 1999;115:654–659.
nosis and surgical evaluation of skeletal Class III 16. Collins SM, Poulton DR. Orthodontic and orthognathic
surgical correction of Class III malocclusion. Am J Orthod
patients with facial asymmetry. Dentofacial Orthop. 1996;109:111–115.
17. Maeda M, Katsumata A, Ariji Y, et al. 3D-CT evaluation of
facial asymmetry in patients with maxillofacial deform-
ities. Oral Surg Oral Med Oral Pathol Oral Radiol Endod.
References 2006;102:382–390.
1. Willems G, De Bruyne I, Verdonck A, Fieuws S, Carels C. 18. Ahn JS, Hwang HS. Relationship between Perception of
Prevalence of dentofacial characteristics in a belgian Facial Asymmetry and Posteroanterior Cephalometric
orthodontic population. Clin Oral Investig. 2001;5: Measurements. Korean J Orhthod. 2001;31:489–498.
220–226. 19. Kwon TG, Park HS, Ryoo HM, Lee SH. A comparison of
2. Severt TR, Proffit WR. The prevalence of facial asymmetry craniofacial morphology in patients with and without facial
in the dentofacial deformities population at the Univer- asymmetry—a three-dimensional analysis with computed
sity of North Carolina. Int J Adult Orthodon Orthognath Surg. tomography. Int J Oral Maxillofac Surg. 2006;35:43–48.
1997;12:171–176. 20. Lee B, Kang D, Son W, et al. The relationship between
3. Chew MT. Spectrum and management of dentofacial condyle position, morphology and chin deviation in
deformities in a multiethnic Asian population. Angle skeletal Class III patients with facial asymmetry using
Orthod. 2006;76:806–809. cone-beam CT. Korean J Orthod. 2011;41:87–97.
4. Haraguchi S, Takada K, Yasuda Y. Facial asymmetry in 21. Park KR, Park HS, Piao Z, et al. Three-dimensional vector
subjects with skeletal Class III deformity. Angle Orthod. analysis of mandibular structural asymmetry. J Craniomax-
2002;72:28–35. illofac Surg. 2013;41:338–344.
5. Grummons DC, Kappeyne van de Coppello MA. A frontal 22. Lee H, Bayome M, Kim SH, Kim KB, Behrents RG, Kook
asymmetry analysis. J Clin Orthod. 1987;21:448–465. YA. Mandibular dimensions of subjects with asymmetric
6. Letzer GM, Kronman JH. A posteroanterior cephalomet- skeletal Class III malocclusion and normal occlusion
ric evaluation of craniofacial asymmetry. Angle Orthod. compared with cone-beam computed tomography. Am J
1967;37:205–211. Orthod Dentofacial Orthop. 2012;142:179–185.
7. Forsberg CT, Burstone CJ, Hanley KJ. Diagnosis and 23. Baek SH, Cho IS, Chang YI, Kim MJ. Skeletodental factors
treatment planning of skeletal asymmetry with the affecting chin point deviation in female patients with
submental-vertical radiograph. Am J Orthod. 1984;85: Class III malocclusion and facial asymmetry: a three-
224–237. dimensional analysis using computed tomography. Oral
 3D CT analysis for skeletal Class III patients with facial asymmetry 281

Surg Oral Med Oral Pathol Oral Radiol Endod. 2007;104: 42. Saeed NR, Wall SA, Dhariwal DK. Management of
628–639. positional plagiocephaly. Arch Dis Child. 2008;93:82–84.
24. Hwang HS, Youn IS, Lee KH, Lim HJ. Classification of 43. Boere-Boonekamp MM, van der Linden-Kuiper LL.
facial asymmetry by cluster analysis. Am J Orthod Dentofacial Positional preference: prevalence in infants and follow-
Orthop. 2007;132(279):e1–e6. up after two years. Pediatrics. 2001;107:339–343.
25. Baek C, Paeng JY, Lee JS, Hong J. Morphologic evaluation 44. Arnold TG, Anderson GC, Liljemark WF. Cephalometric
and classification of facial asymmetry using 3-dimensional norms for craniofacial asymmetry using submental-
computed tomography. J Oral Maxillofac Surg. 2012;70: vertical radiographs. Am J Orthod Dentofacial Orthop. 1994;
1161–1169. 106:250–256.
26. Bruce RA, Hayward JR. Condylar hyperplasia and 45. Kim YH, Sato K, Mitani H, Shimizu Y, Kikuchi M.
mandibular asymmetry: a review. J Oral Surg. 1968;26: Asymmetry of the sphenoid bone and its suitability as a
281–290. reference for analyzing craniofacial asymmetry. Am J
27. Obwegeser HL, Makek MS. Hemimandibular hyperplasia Orthod Dentofacial Orthop. 2003;124:656–662.
—hemimandibular elongation. J Maxillofac Surg. 1986;14: 46. Kim SJ, Lee KJ, Lee SH, Baik HS. Morphologic relation-
183–208. ship between the cranial base and the mandible in
28. Schmid W, Mongini F, Felisio A. A computer-based patients with facial asymmetry and mandibular prognath-
assessment of structural and displacement asymmetries ism. Am J Orthod Dentofacial Orthop. 2013;144:330–340.
of the mandible. Am J Orthod Dentofacial Orthop. 1991;100: 47. St John D, Mulliken JB, Kaban LB, Padwa BL. Anthro-
19–34. pometric analysis of mandibular asymmetry in infants
29. Joondeph DR. Mysteries of asymmetries. Am J Orthod with deformational posterior plagiocephaly. J Oral Max-
Dentofacial Orthop. 2000;117:577–579. illofac Surg. 2002;60:873–877.
30. Myers DR, Barenie JT, Bell RA, Williamson EH. Condylar 48. Sakurai A, Hirabayashi S, Sugawara Y, Harii K. Skeletal
position in children with functional posterior crossbites: analysis of craniofacial asymmetries in plagiocephaly
before and after crossbite correction. Pediatr Dent. (unilateral coronal synostosis). Scand J Plast Reconstr Surg
1980;2:190–194. Hand Surg. 1998;32:81–89.
31. Hesse KL, Artun J, Joondeph DR, Kennedy DB. Changes 49. Kane AA, Lo LJ, Vannier MW, Marsh JL. Mandibular
in condylar postition and occlusion associated with dysmorphology in unicoronal synostosis and plagioce-
maxillary expansion for correction of functional unilat- phaly without synostosis. Cleft Palate Craniofac J. 1996;33:
eral posterior crossbite. Am J Orthod Dentofacial Orthop. 418–423.
1997;111:410–418. 50. Endo M, Terajima M, Goto TK, Tokumori K, Takahashi I.
32. Pirttiniemi P, Kantomaa T, Lahtela P. Relationship Three-dimensional analysis of the temporomandibular
between craniofacial and condyle path asymmetry in joint and fossa-condyle relationship. Orthodontics (Chic.).
unilateral cross-bite patients. Eur J Orthod. 1990;12: 2011;12:210–221.
408–413. 51. Kiliaridis S, Kalebo P. Masseter muscle thickness meas-
33. O‘Byrn BL, Sadowsky C, Schneider B, BeGole EA. An ured by ultrasonography and its relation to facial
evaluation of mandibular asymmetry in adults with morphology. J Dent Res. 1991;70:1262–1265.
unilateral posterior crossbite. Am J Orthod Dentofacial 52. van Spronsen PH, Weijs WA, Valk J, Prahl-Andersen B, van
Orthop. 1995;107:394–400. Ginkel FC. A comparison of jaw muscle cross-sections of
34. Besson A, Pellerin P, Doual A. Study of asymmetries of the long-face and normal adults. J Dent Res. 1992;71:1279–1285.
cranial vault in plagiocephaly. J Craniofac Surg. 2002;13: 53. Ariji Y, Kawamata A, Yoshida K, et al. Three-dimensional
664–669. morphology of the masseter muscle in patients with
35. Robinson S, Proctor M. Diagnosis and management of mandibular prognathism. Dentomaxillofac Radiol. 2000;29:
deformational plagiocephaly. J Neurosurg Pediatr. 2009;3: 113–118.
284–295. 54. Takashima M, Kitai N, Murakami S, Furukawa S, Kreiborg
36. Argenta LC, David LR, Wilson JA, Bell WO. An increase in S, Takada K. Volume and shape of masticatory muscles in
infant cranial deformity with supine sleeping position. patients with hemifacial microsomia. Cleft Palate Craniofac
J Craniofac Surg. 1996;7:5–11. J. 2003;40:6–12.
37. Littlefield TR, Kelly KM, Pomatto JK, Beals SP. Multiple- 55. Huisinga-Fischer CE, Vaandrager JM, Prahl-Andersen B,
birth infants at higher risk for development of deforma- van Ginkel FC. Masticatory muscle right-left differences in
tional plagiocephaly. Pediatrics. 1999;103:565–569. controls and hemifacial microsomia patients. J Craniofac
38. Turk AE, McCarthy JG, Thorne CH, Wisoff JH. The “back Surg. 2004;15:42–46.
to sleep campaign” and deformational plagiocephaly: is 56. Kwon TG, Lee KH, Park HS, Ryoo HM, Kim HJ, Lee SH.
there cause for concern?J Craniofac Surg 1996;7:12–18. Relationship between the masticatory muscles and man-
39. AAP Task Force on Infant Positioning and SIDS. Position- dibular skeleton in mandibular prognathism with and
ing and SIDS. Pediatrics. 1992;89:1120–1126. without asymmetry. J Oral Maxillofac Surg. 2007;65:
40. Cohen M Jr, MacLean R. Craniosynostosis. Diagnosis, 1538–1543.
Evaluation, and Management. New York: Oxford University 57. Seo SA, Baik HS, Hwang CJ, Yu HS. Analysis of masseter
Press; 2000. muscle in facial asymmetry before and after orthognathic
41. Hutchison BL, Hutchison LA, Thompson JM, Mitchell surgery using 3-dimensional computed tomography.
EA. Plagiocephaly and brachycephaly in the first two years Korean J Orthod. 2009;39:18–27.
of life: a prospective cohort study. Pediatrics. 2004;114: 58. Lee DH, Yu HS. Masseter muscle changes following
970–980. orthognathic surgery: a long-term three-dimensional
282 Kim et al

computed tomography follow-up. Angle Orthod. 2012;82: 68. Marsan G, Cura N, Emekli U. Soft and hard tissue
792–798. changes after bimaxillary surgery in Turkish female Class
59. Ueki K, Takazakura D, Marukawa K, Shimada M, III patients. J Craniomaxillofac Surg. 2009;37:8–17.
Nakagawa K, Yamamoto E. Relationship between the 69. Chew MT. Soft and hard tissue changes after bimaxillary
morphologies of the masseter muscle and the ramus and surgery in Chinese Class III patients. Angle Orthod. 2005;
occlusal force in patients with mandibular prognathism. 75:959–963.
J Oral Maxillofac Surg. 2006;64:1480–1486. 70. Lin SS, Kerr WJ. Soft and hard tissue changes in Class III
60. Goto TK, Nishida S, Yahagi M, et al. Size and orientation patients treated by bimaxillary surgery. Eur J Orthod.
of masticatory muscles in patients with mandibular 1998;20:25–33.
laterognathism. J Dent Res. 2006;85:552–556. 71. Becker OE, Avelar RL, Dolzan Ado N, Haas OL Jr., Scolari N,
61. Katsumata A, Fujishita M, Ariji Y, Ariji E, Langlais RP. 3D Oliveira RB. Soft and hard tissue changes in skeletal Class III
CT evaluation of masseter muscle morphology after patients treated with double-jaw orthognathic surgery-
setback osteotomy for mandibular prognathism. Oral maxillary advancement and mandibular setback. Int J Oral
Surg Oral Med Oral Pathol Oral Radiol Endod. 2004;98: Maxillofac Surg. 2014;43:204–212.
461–470. 72. Jokic D, Jokic D, Uglesic V, Macan D, Knezevic P. Soft tissue
62. Ueki K, Okabe K, Mukozawa A, et al. Assessment of changes after mandibular setback and bimaxillary surgery in
ramus, condyle, masseter muscle, and occlusal force Class III patients. Angle Orthod. 2013;83:817–823.
73. Nkenke E, Vairaktaris E, Kramer M, et al. Three-
before and after sagittal split ramus osteotomy in patients
dimensional analysis of changes of the malar-midfacial
with mandibular prognathism. Oral Surg Oral Med Oral
region after LeFort I osteotomy and maxillary advance-
Pathol Oral Radiol Endod. 2009;108:679–686.
ment. Oral Maxillofac Surg. 2008;12:5–12.
63. Damstra J, Fourie Z, De Wit M, Ren Y. A three-
74. Baik HS, Kim SY. Facial soft-tissue changes in skeletal
dimensional comparison of a morphometric and conven-
Class III orthognathic surgery patients analyzed with 3-
tional cephalometric midsagittal planes for craniofacial
dimensional laser scanning. Am J Orthod Dentofacial Orthop.
asymmetry. Clin Oral Investig. 2012;16:285–294.
2010;138:167–178.
64. Kim EJ, Palomo JM, Kim SS, Lim HJ, Lee KM, Hwang HS.
75. Plooij JM, Maal TJ, Haers P, Borstlap WA, Kuijpers-
Maxillofacial characteristics affecting chin deviation Jagtman AM, Berge SJ. Digital three-dimensional image
between mandibular retrusion and prognathism patients. fusion processes for planning and evaluating orthodon-
Angle Orthod. 2011;81:988–993. tics and orthognathic surgery. A systematic review. Int J
65. Park SH, Yu HS, Kim KD, Lee KJ, Baik HS. A proposal for Oral Maxillofac Surg. 2011;40:341–352.
a new analysis of craniofacial morphology by 3- 76. Jung YJ, Kim MJ, Baek SH. Hard and soft tissue changes
dimensional computed tomography. Am J Orthod Dento- after correction of mandibular prognathism and facial
facial Orthop. 2006;129:600.e23–600.e34. asymmetry by mandibular setback surgery: three-
66. Soncul M, Bamber MA. Evaluation of facial soft tissue dimensional analysis using computerized tomography.
changes with optical surface scan after surgical correction Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2009;
of Class III deformities. J Oral Maxillofac Surg. 2004;62: 107763.e8–771.e8.
1331–1340. 77. Lee TY, Kim KH, Yu HS, Kim KD, Jung YS, Baik HS.
67. Marsan G, Oztas E, Kuvat SV, Cura N, Emekli U. Changes Correlation analysis of three-dimensional changes of hard
in soft tissue profile after mandibular setback surgery in and soft tissues in class III orthognathic surgery patients
Class III subjects. Int J Oral Maxillofac Surg. 2009;38: using cone-beam computed tomography. J Craniofac Surg.
236–240. 2014;25:1530–1540.
Comparison of the anterior alveolar bony
changes of moderately crowded cases treated
either with extraction or non-extraction
orthodontic treatment
Tanner Cook, Fräns Currier, Onur Kadioglu, and Thomas Griffin

Extraction vs. non-extraction treatment in orthodontics has been debated for

many decades. The alveolar bone surrounding the anterior permanent teeth
should be a factor in making this decision. This article examines two separate
studies that measured bony changes following orthodontic treatment in
moderately crowded cases. A total of 59 cases (39 non-extraction and 20 four
premolar extraction) were examined using CBCT, 2-D extraoral radiographs,
and orthodontic models. The heights and thicknesses of the anterior bony
support were measured on both the facial and lingual surfaces in extraction and
non-extraction cases with the changes compared between the groups. Results
showed that both treatment modalities experienced statistically significant
bone loss with the lingual surfaces, which received the greatest and most
consistent changes from orthodontic treatment. There were no statistically
significant locations that showed a positive bony change. The pattern of bone
loss from treatment was very similar for extraction and non-extraction cases
with the non-extraction cases, on average, having a greater amount of
bone loss. (Semin Orthod 2015; 21:283–290.) Published by Elsevier Inc.

Introduction often depends upon the orthodontistsʼ past

experiences or the available literature. Other
here has been a long-standing debate in
T orthodontics over which treatment is
“better,” extraction or non-extraction treatment.
considerations for whether to extract or not,
include the original bony structure. If the roots
are moved outside of the original bony cortical
There are many different factors that are
plate, problems can arise. The roots can be
involved in making this decision, but often many
blunted and the supporting alveolar bone and
patients present in the gray area of extractions
gingival tissues can be diminished.1 The
in treatment planning. The decision to treat
periodontal response to teeth moving outside
these patients either with or without extractions
of the cortical plate have shown that when teeth
are held outside of the original cortical plate
Private Practice, Amarillo, TX; Department of Orthodontics, dimensions, no bone is present after 1 month,
College of Dentistry, University of Oklahoma Health Sciences Center, but in 3 months osteoblasts are re-forming
1201 N Stonewall Ave, Oklahoma City, OK 73117; Private Practice,
Broken Arrow, OK.
the cortical plate.2 This could provide some
Address correspondence to Onur Kadioglu, DDS, MS, Department explanation for why teeth can be moved to the
of Orthodontics, College of Dentistry, University of Oklahoma Health cortical plate, and possibly beyond. Nevertheless,
Sciences Center, 1201 N. Stonewall Ave, Oklahoma City, OK 73117. knowingly encroaching on the cortical plate can
E-mail: [email protected] result in unintended adverse consequences. If
All research was performed at the University of Oklahoma teeth are moved to the cortical plate, they must
Graduate Orthodontics clinic in Oklahoma City, OK.
be held there for at least 3 months to begin to see
All funding for the project was provided by the University of
Oklahoma Graduate Orthodontics Department.
a bony response.
Among the multitude of reasons to extract or
Published by Elsevier Inc.
1073-8746/15/1801-$30.00/0 not extract in a case is the predicted amount of
http://dx.doi.org/10.1053/j.sodo.2015.07.005 bony support following treatment as compared to

Seminars in Orthodontics, Vol 21, No 4, 2015: pp 283–290 283

284 Cook et al

the initial levels. Will oneʼs treatment lead to an statistical analysis. Both studies were approved by
increase, or at least less of a decrease, in alveolar the University of Oklahoma Health Sciences
bone over the other? Centerʼs Institutional Review Board.
It is known that after non-extraction ortho- In the extraction group, 15 of the extraction
dontic treatment, there is a decrease in the cases had all first four premolars extracted, four
overall density of the anterior bone.3 CBCT had all four second premolars extracted, and one
allows for a 3-dimensional view of the bony had maxillary first premolars and mandibular
structure with accuracy in vivo of 0.2–0.4 mm and second premolars extracted. The patients were
is detailed enough to find dehiscences and fen- evenly distributed among males and females
estrations in the alveolar bone.4,5 across both groups. There were 21 males and 18
A CBCT study of four premolar extraction females in the non-extraction group and 10
cases have shown that 84% of the lingual surfaces males and 10 females in the extraction group.
of the mandibular incisors showed a 2 mm or Mean age at T0 was 15 years 1 month and 13 years
more loss of alveolar bone height. However, this 6 months for non-extraction and extraction
study took the second CBCT image at debond cases, respectively. Average treatment time was 20
without allowing enough time for the bone to months for non-extraction and 26.5 months for
mineralize and seen radiographically. extraction cases.
Other studies that have allowed the bone to re- Records used for this study were part of the
mineralize for at least 6 months have also found a standard records taken for patients treated in the
decrease in bone on the lingual surfaces of the Graduate Orthodontic Clinic and were as follows.
incisors.6,7 There have been no studies that have
compared the bony changes that have occurred T0 records
in the anterior arches of extraction vs. non-
Pre-treatment records, which included ortho-
extraction treatment of moderately crowded
dontic study models, a lateral cephalometric
cases.
radiograph, and a CBCT scan were taken prior to
The aim of this study was to compare the
placement of orthodontic appliances.
anterior alveolar in extraction and non-
extraction orthodontic treatment of moderately
T1 records
crowded cases.
Post-treatment records were taken no less than 6
months after orthodontics with a mean time of 10
Materials and methods
months after debond for non-extraction cases
This study included 59 patients treated at the and 9 months for extraction cases. These records
Graduate Orthodontic Clinic at the University of included orthodontic study models, a lateral
Oklahoma from two separate thesis projects. A cephalometric radiograph, and a CBCT scan. T1
total of 39 patients were treated non-extraction time point was chosen to be at least 6 months
from a thesis by Cook8 and 20 with four premolar after debond due to the alveolar bone needing at
extraction treatment from a thesis by Griffin.9 least that much time to mineralize to be properly
Inclusion criteria was the same for both projects: visualized on radiographs.10
(1) at least 4.0 mm of clinical crowding in either
arch; (2) Class I, Class II, or end/end malocclu- Cast analysis
sion; (3) permanent dentition; and (4) treated
All orthodontic casts were digitally scanned into a
with four quadrant premolar extractions or non-
2-D image and calibrated. Littleʼs Irregularity
extraction treatment (excluding third molars).
Index11 and the intercanine widths were taken for
Exclusion criteria was also the same for both
both the mandibular and the maxillary arches.
theses and included (1) no history of periodontal
Measurements were made at both time points.
disease, (2) no previous orthodontic treatment,
(3) no expansion auxiliaries, (4) no clefts or
Cephalometric analysis
craniofacial anomalies, and (5) no orthognathic
surgery. Patients could be included in the study The cephalometric analysis was performed at both
with initial unerupted permanent canines, as time points using Dolphin Imaging Premium
those teeth were excluded from the study and version 11.0 (Dolphin Imaging, 9200 Eton Ave,
 Comparison of the anterior alveolar bony changes of moderately crowded cases 285

Chatsworth, CA 91311) with images being 5 mm below the alveolar crest bone height on
imported from a Planmeca Proline XC Dimax3 both the lingual and facial surfaces (Fac3, Fac5,
Ceph machine (Planmeca Oy, Helsinki, Ling3, and Ling5). The alveolar crest height
Finland). The measurements included were measurements were made along the reference
U1–SN1, U1–NA1, U1–NA mm, U1–PP mm, IMPA, line, and the thickness measurements were made
L1–NB1, L1–NB mm, L1–MP mm, and FMA. perpendicular to this line at 3 mm and 5 mm
below the crest of the alveolar bone on the facial
and the lingual. Fig. 1 illustrates the measure-
CBCT analysis
ments on a sample CBCT image. These values
All CBCT scans were taken at both time points were made for all teeth at T0; the lines for these
using the Iluma Ultra Cone Beam CT Scanner measurements were then copied and pasted onto
(IMTEC, Ardmore, OK). The spatial resolution that toothʼs image at T1. The lines and tooth
was measured at 0.19 mm and was defined as how outline were rotated to best fit the tooth at T1 to
closely two lines can be positioned relative to allow for adjustment for any root resorption or
each other and still be perceived as separate.12 enameloplasty that occurred during treatment.
Thus, this study allowed the ability to differ- Once the T0 lines were matched up to the T1
entiate differences greater than 0.19 mm. The image, the alveolar crest height and bone thickness
method used for image capture and measure- lines could be adjusted to allow the reference lines
ment is described by Romero et al.,13 using a and the 3 mm and 5 mm lines to remain at the
tooth superimposition method for T0 and T1 exact spot for the tooth as they were for T0. All
and six measurements per tooth. The images values were compared at T0–T1, so that a positive
used for measurements were sagittal slices taken value of change showed an increase in bone and a
down the long axis of each tooth. The tooth negative value for a decrease in alveolar bone.
was then outlined, and a reference line was
constructed down the long axis of the tooth. The
Statistics
measurements included facial and lingual
alveolar bone heights (FacH and LingH) as The minimum, maximum, mean, median, and
well as thickness measurements at 3 mm and standard deviations were calculated for all

Figure 1. Magnified maxillary right permanent central incisor to illustrate facial and lingual alveolar crest heights
and thicknesses.
286 Cook et al

Table 1. Cast measurements for the non-extraction cases

Measurement n T0 (SD) T1 (SD) Change p Value

Irregularity index (mm) 38 8.3 (4.4) 0.8 (0.9) 7.5 (4.0) o0.0001
Maxillary 3–3 distance (mm) 36 32.9 (3.8) 34.8 (2.1) 1.9 (3.6) 0.038
Mand. 3–3 distance(mm) 39 25.5 (2.7) 26.5 (2.0) 1.0 (1.5) 0.0014

Significant of bold values is p o 0.05.

measurements. Paired t-tests were used to com- changes were negative demonstrating loss of
pare the mean changes between T0 and T1 for all alveolar bone. There were no locations that showed
model, lateral cephalometric, and CBCT meas- a positive change, or increase. Fig. 2 shows the
urements. Correlation analyses were used to locations of the statistically significant bone changes
determine the relationship between bone changes that occurred in the maxilla for non-extraction vs.
in the maxillary and mandibular central incisors extraction cases and Fig. 3 shows the mandibular
and selected cephalometric changes as well as the bony changes. Fig. 4 shows the magnitude of the
permanent canines and their intercanine dis- bone loss in the maxilla, which combines all six
tances. A two-sided 0.05 alpha level was used to anterior teeth in the maxilla to show a range of the
define the statistical significance. statistically significant changes at each location.
Fig. 5 shows the statistically significant changes
combining all six anterior teeth at each location
Results for the mandible. If no bar marker is present at a
Cast measurements location, there were no statistically significant
changes at that location for that treatment.
The cast measurements for non-extraction cases are
presented in Table 1 and extraction cases in Table 2.
The irregularity index for both non-extraction and Discussion
extraction show a statistically significant decrease,
or improvement, while both the maxillary and Both modalities used in this study successfully
mandibular intercanine widths increased. resolved crowding within clinically acceptable
treatment results. There were no statistical dif-
ferences in the crowding or changes in irregu-
Cephalometric analysis larity between the treatments at T0 or T1.
The cephalometric changes with treatment The patients who were treated with non-
showed that, in general, the permanent incisors extraction treatment were treated using a wide
in the non-extraction cases were proclined and variety of prescriptions and wire sizes and
protracted, while for the extraction cases they sequences. There were 10 patients treated with
were reclined and retracted (Table 3). conventional ligating brackets, 14 treated with
active self-ligating brackets, and 15 passive self-
ligating brackets. The different types of treat-
CBCT analysis
ment and tooth movements were analyzed to
The changes seen in the anterior teeth often determine if bracket type had any effect on the
demonstrated a few statistically significant changes; amount of bone loss. The 15 cases treated with
therefore, only the statistically significant values are passive self-ligating brackets did have an
presented. All instances with statistically significant increased bone loss as compared to conventional

Table 2. Cast measurements for the extraction cases

Measurement n T1 (SD) T2 (SD) Difference p Value

Irregularity index (mm) 20 10.9 (3.38) 3.2 (2.05) 7.7 (3.80) o0.0001
Maxillary 3–3 distance (mm) 18 34.6 (2.63) 36.00 (1.40) 1.4 (1.93) 0.0064
Mand. 3–3 distance (mm) 20 25.9 (2.84) 27.6 (1.53) 1.7 (2.10) 0.0073

Significant of bold values is p o 0.05.

 Comparison of the anterior alveolar bony changes of moderately crowded cases 287

Table 3. Cephalometric changes comparing extraction to non-extraction treatment

Non-Extraction Extraction
Measurement T0 T1 Difference p Value T0 T1 Difference p Value

U1 to SN1 104.0 106.9 2.9 0.0059 107.1 102.0 5.1 0.0004

IMPA1 90.8 95.9 5.1 o0.0001 92.4 88.6 3.8 0.0009
U1 to NA1 23.3 26.1 2.8 0.0120 26.5 22.6 3.9 0.0067
U1 to NA (mm) 5.8 6.4 0.6 0.0900 5.73 4.6 1.1 0.0360
L1 to NB1 22.5 27.8 5.3 o0.0001 27.5 24.2 3.3 0.0785
L1 to NB (mm) 4.5 6.2 1.7 o0.0001 6.4 4.4 2.0 0.0010
FMA1 24.3 25.1 0.8 0.3281 27.0 26.0 1.0 0.1172
U1 to PP (mm) 26.8 27.8 1.0 0.0007 27.3 28.4 1.1 0.0002
L1 to MP (mm) 38.0 39.0 1.0 0.0001 38.6 39.2 0.6 0.0833

Significant of bold values is p o 0.05.

and active self-ligating brackets. However, the reclined and retracted during treatment, which
groups as a whole behaved very similarly. would be expected in extraction treatment. All of
The bracket types for the extraction cases were the non-extraction treatment cases showed that
also varied. The cases were treated with variable the incisors proclined and protracted. However,
methods of retraction and had different ortho- the cast measurements showed statistically
dontic residents under the supervision of varying insignificant differences for changes in inter-
faculty treating the cases. canine widths changes.
The cephalometric values for the extraction The CBCT demonstrated that the bone was
cases showed that the permanent incisors were lost in both non-extraction and extraction cases

Figure 2. Areas of significant alveolar bone loss for extraction vs. non-extraction cases in the maxilla. Each NX/X
represents an area of statistically significant bone loss.14
288 Cook et al

Figure 3. Areas of significant bone loss for extraction vs. non-extraction cases in the mandible. Each NX/X
represents an area of statistically significant bone loss.14

Figure 4. All six maxillary teeth were combined to create this graph with the ranges of statistically significant
alveolar bone changes at each level.
 Comparison of the anterior alveolar bony changes of moderately crowded cases 289

Figure 5. All six mandibular teeth were combined to create this graph with the ranges of statistically significant
alveolar bone changes at each level.

on the facial and the lingual. The facial surfaces the mandible. There were no instances of a
showed the least amount of bone loss for all positive gain in alveolar bone during treat-
patients. When comparing the magnitude of ment. Based on reports in the literature, bone
bone loss, the numbers were similar but show changes of 0.5 mm or more are considered to be
that at almost every location the non-extraction clinically significant changes in either height or
cases lost more bone. There were locations thickness.15
where extraction cases lost more bone, i.e., the Why non-extraction treatment lost more
lingual alveolar bone heights in the maxilla and alveolar bone is difficult to explain. One
the lingual thicknesses at 3 mm and 5 mm in explanation could be that the expanding forces

Figure 6. Initial CBCT image (T0) for an UR1 on the left with an arrow showing the rotation that can occur from
T0–T1. Follow-up CBCT image (T1) of same UR1 on the right. The yellow lines show the T1 lingual thicknesses
and the orange lines showing the change that occurred in both lingual bone heights and thicknesses. (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)
290 Cook et al

with non-extraction treatment are more likely treatment at almost every location, but bone
to move teeth out of their biological limits loss was greater with non-extraction treatment.
than extraction treatment. Another consid-
eration could be that the incisors in non-
extraction cases showed a higher magnitude References
of flaring than the non-extraction cases. This
1. Handelman CS. The anterior alveolus: its importance in
larger scale of movement could have certainly limiting orthodontic treatment and its influence on the
contributed to why non-extraction cases lost occurrence of iatrogenic sequelae. Angle Orthod. 1996;66
more bone. (2):95–110.
Non-extraction cases would be expected, at 2. Wainwright WM. Faciolingual tooth movement: its influ-
first glance, to loosen more bone on the facial ence on the root and cortical plate. Am J Orthod. 1973;64
(3):278–302.
areas, as that is the general direction of the tooth 3. Hsu JT, Chang HW, Huang HL, et al. Bone density
movement. However, a greater amount of changes around teeth during orthodontic treatment. Clin
alveolar bone loss was seen on the lingual in both Oral Invest. 2011;15:511–519.
cases. Through CBCT, the lingual surfaces seem 4. Vroome KM. Anterior alveolar bone changes following pre-
to be more susceptible reduction in bony sup- molar extractions: a cone beam computed tomography eva-
luation [Masterʼs Thesis]. Oklahoma City: University of
port. A possible reason for this lingual bone loss is
Oklahoma; 2009.
the flaring of the incisors creating a larger area 5. Ahn HW, Moon SC, Baek SH. Morphometric evaluation
for bone to fill in on the lingual as compared to of changes in the alveolar bone and roots of the maxillary
the facial. This can be visualized in Fig. 6. From anterior teeth before and after en masse retraction using
this image, the tooth appears to have rotated with cone beam computed tomography. Angle Orthod.
lingual root torque moving the root toward the 2013;83:212–221.
6. Lund H, Grondahl K, Grondahl HG. Cone beam
lingual cortical plate. This might make the computed tomography for assessment of root length
lingual plate thinner and create a decrease in and marginal bone level during orthodontic treatment.
bone height, which may provide a possible Angle Orthod. 2010;80:466–473.
explanation for why the alveolar bone loss 7. Garib DG, Yatabe MS, Ozawa TO, et al. Alveolar bone
pattern in non-extraction treatment was pri- morphology under the perspective of the computed
tomography: defining the biological limits of tooth
marily on the lingual and not as much on the
movement. Dental Press J Orthod. 2010;15(5):192–205.
facial. 8. Cook T. CBCT assessment of anterior alveolar bone levels after
non-extraction orthodontic treatment: conventional vs. self-
ligating brackets [Thesis]. Oklahoma City: University of
Summary and conclusions Oklahoma; 2014.
9. Griffin T. Anterior arch changes, including the alveolar bone
A total of 39 non-extraction patients and 20 four
housing, in extraction orthodontic therapy [Thesis]. Oklahoma
premolar extraction cases with moderate City: University of Oklahoma; 2014.
crowding were studied to determine if any dif- 10. Starnbach H, Bayne D, Cleall J, et al. Facioskeletal and
ferences existed in alveolar bone loss during dental changes resulting from rapid maxillary expansion.
treatment in the anterior segment. The results Angle Orthod. 1966;36:152–164.
showed that all the cases lost alveolar bone. 11. Little Robert M. The irregularity index: a quantitative
score of mandibular anterior alignment. Am J Orthod.
There were no locations that had a statistically 1975;68(5):554–563.
significant location that had an increase in bone. 12. Ballrick JW, Palomo JM, Ruch E, et al. Image distortion
The conclusions are as follows: and spatial resolution of a commercially available cone-
beam computed tomography machine. Am J Orthod
(1) Orthodontic treatment of patients with mod- Dentofacial Orthop. 2008;134(4):573–582.
13. Romero Delmastro A, Currier GF, Kadioglu O, et al.
erate crowding resulted in anterior alveolar
Digital tooth based superimposition method for assess-
bone loss, regardless of whether they were ment of alveolar bone levels on cone-beam computed
treated with extractions or not. tomography images. Am J Orthod Dentofacial Orthop.
(2) Non-extraction and extraction treatment 2014;146(2):255–263.
presented similar patterns of alveolar bone 14. Dendram, Graphics and Printing, Monrovia, CA. Perio
Screening Exam (Morgulis 11) - Item No. M-11. Web.
loss with most of the significant changes seen
25 July 2014.
on the lingual surfaces. 15. Zachrisson BU. Cause and prevention of injuries to teeth
(3) The magnitude of alveolar bone loss was and supporting structures during orthodontic treatment.
similar for extraction and non-extraction Am J Orthod. 1976;69(3):285–300.
 Seminars in Orthodontics
Future Issues
Vol 22 No 1 (March 2016)
INTERACTIONS BETWEEN ORTHODONTICS AND ORAL MAXILLOFACIAL SURGERY
Jae Hyun Park, DMD, MSD, MS, PhD, Guest Editor

Recent Issues
Vol 21 No 3 (September 2015)
ACCELERATED ORTHODONTICS
Mani Alikhani, DDS, MS, PhD, Guest Editor
Vol 21 No 2 (June 2015)
JUVENILE IDIOPATHIC ARTHRITIS AND TEMPOROMANDIBULAR JOINT INVOLVEMENT: AN INTERDISCIPLINARY APPROACH
Bjørn Øgaard, DDS, Dr Odont, Guest Editor
Vol 21 No 1 (March 2015)
INTERDISCIPLINARY MANAGEMENT OF THE ORTHODONTIC PATIENT
Pratik Kumar Sharma, BDS(Hons), MFDS, MSc, MOrth, FDSOrth, Guest Editor
Vol 20 No 4 (December 2014)
ALL ROADS LEAD TO ROME: NEW DIRECTIONS FOR CLASS II
S. Jay Bowman, DMD, MSD, FACD, FICD, Guest Editor
Vol 20 No 3 (September 2014)
PERIODONTAL-ORTHODONTIC INTERACTIONS
Ramzi V. Abou-Arraj, DDS, MS, Guest Editor
Vol 20 No 2 (June 2014)
AGE-APPROPRIATE ORTHODONTIC TREATMENT, PART II
Gerry Samson, DDS, and Elliott M. Moskowitz, DDS, MSd, Guest Editors
Vol 20 No 1 (March 2014)
AGE-APPROPRIATE ORTHODONTIC TREATMENT, PART I
Elliott M. Moskowitz, DDS, MSd, and Gerry Samson, DDS, Guest Editors
Vol 19 No 4 (December 2013)
THE VERTICAL DIMENSION IN ORTHODONTICS
Nada M. Souccar, DDS, MS, Guest Editor
Vol 19 No 3 (September 2013)
EVIDENCE-BASED ORTHODONTICS
Katherine Vig, BDS, MS, FDS, DOrth, and Greg Huang, DMD, MSD, MPH, Guest Editors
Vol 19 No 2 (June 2013)
PROGRESSIVE CONDYLAR RESORPTION AND DENTOFACIAL DEFORMITIES
Chester S. Handelman, DMD, and Charles S. Greene, DDS, Guest Editors
Vol 19 No 1 (March 2013)
INTERDISCIPLINARY TREATMENT OF ADOLESCENTS WITH MISSING ANTERIOR TEETH
Mark R. Yanosky, DMD, MS, Guest Editor
Vol 18 No 4 (December 2012)
UPDATES ON THE BIOLOGICAL FOUNDATIONS OF ORTHODONTIC TOOTH MOVEMENT
Vinod Krishnan, BDS, MDS, M Orth RCS D, PhD, and Ze’ev Davidovitch, DMD, Cert Ortho, Guest Editors
Vol 18 No 3 (September 2012)
AN OVERVIEW OF FACIAL ATTRACTIVENESS FOR ORTHODONTISTS
Margaret Collins, BDS, FDSRCPS, DOrth, MSc, MOrthRCS, MA, Guest Editor
Vol 18 No 2 (June 2012)
MAXILLARY EXPANSION AND MANDIBULAR WIDENING: TREATMENT METHODS AND STABILITY
Haluk İ şeri, DDS, PhD, Guest Editor
Vol 18 No 1 (March 2012)
FUNCTION AND DYSFUNCTION OF THE TEMPOROMANDIBULAR JOINT
Rakesh Koul, MDS (Orthodontics), Guest Editor