SECTION 4: SIMULATION GROWTH AND TREATMENT SIMULATIONS, AND VIRTUAL

REALITY

K Maki
N Inou
A Takanishi
AJ Miller

Computer-assisted
simulations in orthodontic
diagnosis and the application
of a new cone beam X-ray
computed tomography

Authors' affiliations:
K. Maki, Department of Orthodontics,
Showa University, Tokyo, Japan
N. Inou, Graduate School of Science and
Engineering, Tokyo Institute of Technology,
Tokyo, Japan
A. Takanishi, Department of Engineering,
Waseda University, Tokyo, Japan
A.J. Miller, Department of Growth and
Development, University of California,
San Francisco, CA, USA

Abstract

Correspondence to:
Koutaro Maki
Department of Orthodontics
Showa University
145-8515, 2-1-1 Kitasenzoku, Ohta-Ku
Tokyo, Japan
Tel.: +81 3 3787 1151
Fax: +81 3 3784 6641
E-mail: makihome@cd.mbn.or.jp

Authors Maki K, Inou N, Takanishi A, Miller AJ

Computational simulations which include three-dimensional
(3-D) image processing and biomechanical calculations
should provide useful information to our research and orthodontic clinic as a clinical tool defined as thinking. In this
review, 1) biomechanical simulations applied to predict the
mandibular growth; 2) mathematical models of virtual bone
cells and 3) 3-D images and solid model simulations for surgical planning are introduced. In biomechanical simulation,
biting force, electromyographic (EMG) activity and cephalograms of 32 subjects were applied. Computational results of
mathematical model were compared with actual bone growth in
a rat. Three-dimensional image and solid model of 14 patients
were utilized for their treatment planning. From the results,
several concepts of our simulations were confirmed: 1) reaction
forces generated by masticatory muscles at the condyle control the direction of mandibular growth; 2) some mathematical
models have the possibility to describe the process of bone
growth; 3) 3-D image processing software and solid models are
necessary for diagnosis and planning of orthognathic surgery.
We also believe that the orthodontists can more accurately

To cite this article:

Orthod Craniofacial Res 6 (Suppl. 1), 2003; 95101
Maki K, Inou N, Takanishi A, Miller AJ:
Computer-assisted simulations in orthodontic
diagnosis and the application of a new cone beam
X-ray computed tomography
Copyright  Blackwell Munksgaard 2003
ISSN 1741-2420

predict the affects of surgical procedures and orthodontic tooth

movement using the new cone beam X-ray computed tomography (CT) (CB MercuRay
; Hitachi Medico Technology,
Tokyo, Japan) and its advanced application software.
Key words: biomechanics; condylar growth; cone beam CT;
simulation; surgery

Maki et al. Computational simulations and new cone beam CT

Introduction
The progress of imaging technology and the development of numerically controlled mathematical models
makes it possible to create three-dimensional (3-D)
observations of the maxillofacial structures. Computerassisted simulations in medicine are closely related to
the progress in industrial technology. Computerassisted manufacturing (CAM), computer-assisted
design (CAD), and finite element methods (FEM) were
created to assist the manufacturing industry after
World War II. In medicine, the utilization of digital
modalities began when the company, EMI, developed
X-ray computed tomography (CT) starting in 1972.
Both technologies dealing with digital information have
complemented each other, and in 1980, computerassisted simulations were invented. The publics
interest in high quality medicine and the support from
venture capital have accelerated this progress.
The advantages of simulation are visualization,
repetition, feedback and prediction. Simulated
results based on collected data and logical calculations
support diagnosis, treatment planning, and assist the
actual operation. In the dental field, implant treatment
has become a standard. Dental CAD/CAM programs for
prosthetic treatment, and orthodontic simulators such
as ALIGN or Orametrix are produced on a commercial
basis. Recently, dental-CT has also been developed to
provide 3-D evaluations of patients. Now is the time to
establish new simulation methods that will be utilized
for more accurate orthodontic diagnosis and treatment.
In this review, we will introduce: 1) biomechanical
simulations (prediction of condylar growth); 2)
mathematical models of bone remodeling and 3) 3-D
image simulations and solid modeling for surgical
orthodontics. In addition, the application of the
newly developed cone beam X-ray CT for the dentomaxillofacial region is shown.
Simulation of condylar growth (study 1)

In our previous studies, we showed that the bone

mineral density and morphological changes of the
human mandible are related to muscle loading (1, 2).
However, the biomechanical factors in controlling the
human condylar cartilage are still unknown. From the
biomechanical point of view, the mandible is regarded

96

Orthod Craniofacial Res 6 (Suppl. 1), 2003/95101

as a class III lever system. Occlusal contacts are estimated as points of reaction, muscle insertions as a
point of force applied, and the condyle is the fulcrum.
Direction and magnitude of reaction forces generated
by mastication at the condyle can be calculated
from the vectors of muscle traction and loads on the
dentition. From our observations of over 162 CT studies
in our orthodontic clinic, subjects with large masseter
muscles and square faces have large condyles, and
subjects with weak muscles and a narrow face have a
smaller condyle.
What affects the growth direction and size of
condyle?
In the temporomandibular joint, above and below
the disc are small sac-like compartments called synovial cavities. Part of the tissue lining these cavities is
an epithelium that secretes synovial fluid. Stress generated in synovial cavities is reduced by the hydrostatic effect of this fluid. Reaction forces are dispersed
to maintain optimum pressure for cartilage and
smooth movement of the condyle. Subjects with large
magnitudes of pressure (reaction force) need larger
surface areas of the condyle, and small pressure
requires small areas. The extremely high or low levels
of pressure might present a physical impediment to
cartilage and bone formation. Therefore, our hypothesis is as follows: muscle traction and occlusal biting
force generate reaction forces and shear stress at the
condyle, and
increase the intra-articulator pressure;
the hydrostatic pressure enhances cartilage maintenance;
the condylar cartilage and endochondral ossification
work to stabilize the optimum biomechanical conditions.
To verify this hypothesis, 3-D simulations were performed. Reaction forces on the human condyle were
calculated under various loading conditions in a
standard model. Occlusal points, direction of biting
forces and direction of muscle tractions corresponding
to the insertion areas on mandibular surface were
determined. Only the magnitudes of muscle loading
forces were changed. From these calculations, if the
vector of temporalis muscle force exceeded that of the
masseter, the vector of the condylar reaction force was
inclined backward. Figure 1 shows this result.

Maki et al. Computational simulations and new cone beam CT

Fig. 1. Three-dimensional simulation of reaction forces at condyle.

Arrows at condylar head indicating vectors of reaction force. (A)
Temporalis develops more force than the masseter. (B) Masseter
develops more force than the temporalis.

In order to confirm this result with large number of

living subjects, two-dimensional (2D) simulations
from standard orthodontic cephalograms were used.
In this study, simultaneous measurements of the biting force were determined for each patient using
pressure sensor sheets. The electromyographic (EMG)

activity of the masseter muscle, and the anterior and

posterior parts of the temporal muscles were calculated using surface EMG. Integrated values of each
muscles EMG activity during clenching were measured as traction forces. As the measurement of the
pterygoid muscles were impossible, the ratio of the
anatomical muscle cross-section areas was used as
follows masseter:medial pterygoid:lateral pterygoid
5:3:2. The direction of the biting force was regarded
as the direction of the tooth axis. The direction of the
muscle traction force was determined at the point of
the anatomical muscle insertions. By using these
vectors, the moment at the condyle was calculated as
a reaction force. Using the longitudinal data of the
cephalograms, the direction of this reaction force was
calculated at the initial tracing, and then compared
with the actual growth shown in the second later
cephalogram as the child grew. Figure 2 shows the
calculated reaction forces and the superimposition of
the mandibular growth in two different patients. From
the results, the direction of the reaction force coincides with the direction of mandibular growth on the

Fig. 2. Two-dimensional simulation of

reaction forces on the cephalogram.
(A) Vectors of reaction force at initial stage in
two different subjects. (B) Superimposition
of initial and 2 years later on mandibular
plane at menton. Vectors of reaction forces
coincide with the direction of condylar
growth.
Orthod Craniofacial Res 6 (Suppl. 1), 2003/95101

97

Maki et al. Computational simulations and new cone beam CT

Fig. 3. Virtual cell model (two-dimensional

simulation of bone formation). (A) Actual
femur of rat. 6 weeks of age (upper) and
29 weeks (lower). (B) Process of mathematical simulation using virtual cells. Arrow
indicates applied loading in initial structure.

Fig. 4. Surgical simulation. (A) Threedimensional (3-D) image simulation.

Simulated result (middle) and actual 3-D
CT image after surgical operation (left).
(B) Solid model. Milling model (left) and
laser lithograph model (right). (C) Simulation
of sagittal split of ramus osteotomy in
milling model.

98

Orthod Craniofacial Res 6 (Suppl. 1), 2003/95101

Maki et al. Computational simulations and new cone beam CT

superimposed image. These findings suggest that

condylar growth is affected by the occlusal points, the
magnitude of biting strength, the muscle loading
direction, and the ratio of their muscle forces. We are
now testing this simulation, but this technique seems
to be extremely useful for orthodontic diagnosis.
Mathematical simulation of bone formation (study 2)

To replicate the interaction between bone morphogenesis and loading history, a mathematical simulation
for the growth and functional adaptation of bone was
created.
In this simulation, a structure consisting of some
cells is developed, and the rule that only the cells in the
regions that resist the weight applied to the surface are
kept is applied, and unnecessary cells are eliminated.
The following processes one through five are repeated until convergence.

Simulation of orthognathic surgery (study 3)

Three-dimensional image processing and individual

solid modeling provide useful information for preoperative planning of surgeries to the human mandible.
Many studies have attempted to produce a better
prediction of what a certain treatment will complete.
Since 1987, we have been applying these simulations to
predict the effects of osteotomy as shown in Fig. 4 (3).
Image processing and transformation software was
used to translate from CT data format to STL (stereolithography interface; CAD/CAM) file format.
Sagittal split ramus osteotomy of the mandible is
widely used as the correction of mandibular deformities. However, condylar displacement after sagittal
splitting of the mandible has been reported as an
important factor in post-surgical relapse. Without
proper setting of the sectioned rami, skeletal displacement can occur. The magnitude of rotation and

1 A given structure consisting of small cells is assumed

as a prototype.
2 Boundary conditions for weight bearing and its
restrictions are determined.
3 The stress and strain that develop in the prototype
are calculated.
4 The quantities of distortion corresponding to the
stress (Von Meses) values at junctions (or cells) in the
structure are determined.
5 Youngs modulus changes, the structure grows
automatically. Unnecessary cells are eliminated in
order to maintain minimum stiffness and optimum
strength.
As a result of this repeated calculation, a trabecular like
structure is generated. Figure 3 compares actual X-ray
images of the rat and the simulated result. With this
method, the differentiation of cells into osteoblasts and
osteoclasts is expressed mathematically. Extensions
and refinements will provide additional support for the
fundamental concepts that are embodied in this model.
These studies have just begun, however, such types
of computer simulations might become a powerful tool
for understanding biological reactions. In the future,
comparisons of experimental results and interpretations based on computational simulation will contribute to defining new aspects of bone formation. The
accuracy of the information will strongly determine the
strength of this approach.

Fig. 5. New cone beam X-ray CT. The X-ray tube and the image
intensifier (II) are placed in opposite positions, and the conical beam
rotates 360 around the patient. The X-ray is detected by the II,
transmitted by a high-resolution charge-coupled device camera, and
re-constituted with a three-dimensional processor used in the voxel
transmission method.

Orthod Craniofacial Res 6 (Suppl. 1), 2003/95101

99

Maki et al. Computational simulations and new cone beam CT

displacement of the condylar proximal segments are

affected by the split direction and physical interferences of each bony segment. Solid modeling has an
advantage for detection of these interferences. Preoperative evaluation of the grinding area on each segment could be helpful in reducing skeletal relapse. The
influence of the fixation of segments, and the newly
acquired biomechanical circumstances after the surgical operation are problems which cannot be avoided.
Description of the occlusal condition in a simulation
model, and translation from simulated results to the
actual operation should be improved. Solutions of
these problems are needed with development of
biomechanical modeling (4), application of robotic
technology, and more accurate resolution in CT images
and reconstruction.
Application of new cone beam X-ray CT (study 4)

Conventional CT scanning has been limited in its use

in general dentistry, because of two critical issues, a
low vertical resolution and the high radiation dose.
We have been working with Hitachi Medical Technology (Tokyo, Japan) to develop a new X-ray CT
using conical X-ray beams and a plane detector to

enhance the usefulness of CT in dental diagnosis and

treatment (5). Cone beam X-ray CT (CB MercuRayTM)
was designed to drastically increase the number of
layers of detectors by employing 2-D detectors and
changing the X-ray beam from the fan-shaped beam
to the conical beam. Cone beam CT has several
advantages compared with conventional CT a shorter
scan time, better vertical resolution, and lower exposure dose.
Our objectives to develop a new CT system included
the following factors.
to limit the scanning area to the maxillofacial region;
to restrict the area occupied by the system to a
square of 2.5 m 2.5 m;
to improve the vertical resolution;
to provide images similar to conventional panoramic
images or cephalograms;
to add application software that supports orthodontic, surgical and implant operations.
The apparatus developed was an open gantry type, in
which the X-ray irradiator and the detector rotated
around the patient. Figure 5 shows an overview of this
system. The X-ray tube and the image intensifier (II)
are placed in opposite positions in the gantry, and the

Fig. 6. Images derived from cone beam

X-ray CT. (A) Frontal section at first molars.
Geometrical relation of root and cortical
bone is presented clearly. (B) Threedimensional (3-D) position and actual size of
permanent teeth could be measured
before eruption. (C) 3-D facial image.
(D) Simulation of SSRO.

100

Orthod Craniofacial Res 6 (Suppl. 1), 2003/95101

Maki et al. Computational simulations and new cone beam CT

conical beam rotates 360 around the patient. The II of

the detector can be increased in size from 4.5, 7.0, 9.0 to
12.0 inches (changeable), and 512 projection images of
a 512  512 matrix are obtained in each field of view by
a single rotation. The scanning condition is as follows
tube voltage, 120 kV; tube current, 15 mA; voxel size,
100340 lm (x, y, z); scan time, 9.6 s/scan. Data format
conforms to DICOM 3. A detailed description of
the performance characteristics was introduced by
Yamamoto et al. (6).
As the system has a resolution of 512 pixels along all
x, y and z-axes, clear images are obtained in any crosssection. Observation at this fine resolution is useful not
only to visualize skeletal shapes but also to obtain
actual measurements of the individual teeth. Images
that have never been obtained were readily available.
Figure 6 shows 3-D images of dentition, non-erupted
teeth, and the geometrical relationship of the molars
and cortical bone. Panoramic and cephalometric views
were also reproduced using image-processing software.
Adjustment of the segmentation values (CT numbers), allowed obtaining the morphological soft tissue
data as seen in a facial photograph, a panoramic X-ray
image, a cephalometric image, and a dental cast model.
All the required orthodontics records are obtained from
only one scanning of a patient. Various surgical simulations can also be performed as shown in Fig. 6. An
additional software program (V Works, CyberMed
Corp., Seoul, Korea) was applied to visualize the data. It
may also be possible to manufacture a patient orthodontic device, using 3-D laser lithograph modeling,
from the above records. In the near future, virtual set-up
modeling of digital dentition, which includes the actual
position and inclination of root, will be developed.
Presently, we are attempting to visualize jaw movement
and to complete automated finite element modeling
by using this CT system. Orthodontics will certainly
change with these new approaches.

Conclusion

Question 1. How will we predict condylar growth?

Question 2. Which new techniques in imaging, genetics, and molecular biology will affect orthodontics?
Question 3. What does the practicing orthodontist
need for more accurate planning to treat a patient with
less relapse and more sustainable outcomes?
Presently, computer science has accelerated the rapid progress in orthodontics beyond comparison with
the pace of progress in the past. In the near future,
computational simulation will provide various advantages for orthodontic treatment and research. To
achieve a more scientific approach in orthodontics, the
following strategies are required:
unification of various data formats;
decreasing expenses for software development;
spread and wider application of technology and
knowledge;
collaboration with different fields of study.

References
1. Maki K, Miller AJ, Okano T, Shibasaki Y. Changes in cortical bone
mineralization in the developing mandible: a three-dimensional
quantitative computed tomography study. J Bone Min Res
2000;15:7009.
2. Maki K, Miller AJ, Okano T, Shibasaki Y. A three-dimensional,
quantitative computed tomographic study of changes in
distribution of bone mineralization in the developing human
mandible. Arch Oral Biol 2001;46:66778.
3. Maki K, Inou N, Mikawa M, Tanaka N, Usui T, Toki Y et al. Computer-aided biomechanical simulations for the diagnosis of maxillofacial functions. In: Lemke HU, Vannier MW, Inamura K,
Farman AG, editors. Computer Assisted Radiology and Surgery.
Amsterdam: Elsevier; 1998. pp. 81923.
4. Inou N, Maki K, Motojima S, Koseki M, Ujihashi S. Biomechanical
analysis of the human mandible in surgical operations. In:
Power H, Hart RT, editors. Computer Simulations in Biomedicine.
Boston: Computational Mechanics Publications; 1995. pp. 37784.
5. Maki K, Usui T, Kubota M, Nakano H, Shibasaki Y, Araki K et al.
Application of cone-beam X-ray CT in dento-maxillofacial region.
In: Lemke HU, Vannier MW, Inamura K, Farman AG, Doi K, Reiber
JHC, editors. Computer Assisted Radiology and Surgery. Amsterdam: Elsevier; 2002. pp. 10038.
6. Yamamoto K, Ueno K, Seo K, Shinohara D. Development of dentomaxillofacial cone beam X-ray computed tomography system.
Ortho Craniofac Res 2003;6(Suppl. 1):160162.

There are several questions that will become important

in the future of orthodontic diagnosis and treatment.

Orthod Craniofacial Res 6 (Suppl. 1), 2003/95101

101