University of Southern California Dissertations and Theses

3D ssessment of bracket position accuracy for lingual appliances using CAD/CAM technology: a pilot study

(USC Thesis Other)

3D ssessment of bracket position accuracy for lingual appliances using CAD/CAM technology: a pilot study

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

3D Assessment of
Bracket Position Accuracy
for Lingual Appliances
Using CAD/CAM Technology:
A Pilot Study

By Tarim Song DDS
May 2018

A Thesis Presented to
the Faculty of The USC Graduate School
University of Southern California
in Partial Fulfillment of
the Requirements for the Degree
Master of Science
(Craniofacial Biology)
2

Acknowledgements

This project would not have been possible without both Dr. Hongsheng Tong and Dr.
Andre Weissheimer’s gentle guidance, mentorship, brilliance, and enthusiasm; the truly magical
engineering team of INBRACE that made ideas come to life, especially Allen Huynh, Ishan
Shinde, Joseph Hernandez, Frederique Norpetlian, and Robert Lee; my handy statisticians Dr.
Glenn Sameshima and Dr. Kevin Kaifeng Yin; and the warm fuzzy moral support and
overwhelming unconditional love from my dear Tofflemire.

3

Table of Contents

I. Abstract 4
II. Literature Review 5
a. A Brief History of Lingual Orthodontics 5
b. Bracket Failure: A Never-Ending Universal Problem 10
c. Current Rebond Techniques 12
d. Assessment of Bracket Position Accuracy 20
III. Purpose 21
IV. Materials and Methods 22
a. Sample 22
b. Jig Fabrication 23
c. Rebond Methods 23
i. Direct Bond 24
ii. Flexible Tray Indirect Bond 25
iii. Rigid Jig Indirect Bond 27
d. Scanning, Superimposition, & Error Measurement 27
e. Statistical Analysis 32
V. Results 33
VI. Discussion 41
a. Limitations & Future Suggestions 43
b. Other Future Suggestions 46
VII. Conclusion 46
VIII. References 47

4

I. ABSTRACT

Introduction: While lingual orthodontics have evolved and matured to deliver excellent treatment
outcomes comparable to labial appliances, one problem has yet to be completely resolved for any
bonded appliance. Despite the meticulous digital setup and transfer of lingual appliances, bond
failure of individual brackets is inevitable throughout the course of treatment. An average of 2-3
bond failures within the first year of orthodontic treatment can be expected (Ziebura et al, 2014).

Objectives: The overall goal of this investigation was to determine the accuracy of and to compare
three different single-tooth rebond methods for a CAD/CAM lingual orthodontic appliance. The
three rebond methods to be assessed and compared were direct bond, flexible tray indirect bond,
and 3D-printed rigid jig indirect bond. The specific goals of this pilot study were: (1) To quantify
and compare the accuracy for rebonding brackets with a flexible tray, 3D printed rigid jig, and
direct bonding, (2) to compare the accuracy of the three methods for different tooth morphologies
(molar vs. premolar vs. anterior tooth), and (3) to compare the accuracy of each rebond method in
each translational dimension. The null hypotheses were that there would be no significant
difference in the precision of bracket position for each of the rebond methods; no difference
between incisor, premolar, and molar bracket position accuracies for any of the rebond methods;
and no difference in the position accuracy between the translational dimensions. The significance
level was set at α = 0.05, and the clinically acceptable limit was set at 0.5mm.

Materials & Methods: 30 identical malocclusion models were 3D-printed with the MoonRay 3D
printer (SprintRay, Inc., Los Angeles, CA). 10 were assigned to each of the 3 rebond methods.
After brackets were bonded, they were scanned with an iTero scanner (Align Technologies, San
Jose, CA) for a superimposition study of the deviations which were evaluated by a proprietary
imaging software (Swift Health Systems, Irvine, CA). Statistical analysis was with descriptives
and a two-way ANOVA using SPSS (Armonk, NY).

Results: The largest deviations in bracket placement were observed for the direct bond and rigid
jig methods, between which there was no significant difference, but the flexible tray method had
significantly less error than either. There was a tendency for lingual displacement using the rigid
jig. The tendency for gingival deviation was observed when using the direct bond method for all
teeth and the flexible tray method for the incisor and premolar.

Conclusions: The overall most consistently accurate rebond method was the flexible tray.
placement onto the molar was the most accurate across all rebond methods, compared to bracket
placement onto the premolar or incisor. Brackets were positioned most accurately in the
mesiodistal translational dimension for all rebond methods. The flexible tray method resulted in
the overall most accurate bracket placement in each of the translational planes. Brackets were
positioned most accurately in the mesiodistal translational dimension for all rebond methods.

5

II. LITERATURE REVIEW

II.a. A Brief History of Lingual Orthodontics
The age of adult orthodontics and patient-driven care in the late 20
th
century have paralleled
the growing demand for lingual appliances (Mizrahi, 2006; Pauls et al, 2017). The rise in esthetic
expectations and adult clientele led to the development of the first plastic brackets in the early
1970s and the first “invisible” lingual appliances in the late 1970s (Mavreas et al, 2017; Russell,
2005). Adults have been seeking an improved smile more than ever, whether to be treated for the
first time or retreated, many of whom are reluctant or refuse to face the prospect of wearing
conventional labial appliances that are immediately visible during speech or upon smiling
(Schubert et al, 2013). Esthetic labial brackets are an acceptable compromise, but many
orthodontists are also discouraged by their perceived unfavorable characteristics, such as frictional
resistance, reduced torque stability, staining tendency, and brittleness (Chatoo, 2013; Mavreas et
al, 2017).
(a) (b)
Figure 1: (a) Dr. Kinya Fujita of Kanagawa, Japan. (b) Dr. Craven Kurz of Beverly
Hills, CA, USA.

The advent of modern lingual appliances is characterized by innovative and pioneering
individuals that challenged the status quo (Chatoo, 2013). Dr. Kinya Fujita (Assistant Professor
of the Orthodontic Department at Kanagawa Dental University in Kanagawa, Japan) and Dr.
Craven Kurz (Beverly Hills, CA; Assistant Professor of Occlusion and Gnathology at UCLA
6

School of Dentistry in Los Angeles, CA) are recognized for having independently developed the
concept of a full, multi-bracket appliance that can be placed on the lingual surfaces of teeth (Figure
1). Interestingly, Fujita was encouraged to develop his lingual appliance not only by an esthetic
demand, but also to meet the orthodontic needs of those practicing high-impact sports like football
and judo who wanted to protect their lips and cheeks from trauma (Mizrahi, 2006). Thus the
discipline of lingual orthodontics was spawned. He first reported his concepts on a lingual
appliance in 1967, and began production and use of his appliance in patients in 1975 (Fujita, 1979).
He published on the development and use of the mushroom-shaped archwire (Figure 2) in 1979
and soon after on extraction cases in both adults and children (Fujita, 1982).
(a) (b)

(c)
Figure 2: The lingual appliance first developed by Fujita (Fujita, 1979). (a) The
mushroom-shaped archwires, due to the buccolingual thickness change from canine
to premolar. (b) The bracket slots were designed for archwire insertion from the
occlusal, with pins and elastic ties for engaging the wire. (c) Fujita’s lingual
appliance for an extraction case, 6 months into treatment (Fujita, 1982).

Kurz too was independently inspired to develop a lingual appliance by a Playboy Bunny
Club employee with crowding who refused a “tin grin” or even esthetic labial appliances (Echarri;
2006). Kurz collaborated with Ormco (Sybron Dental Specialties, Orange, California) to design
and develop a lingual bracket in 1976, with the production of the first preliminary edgewise
7

prototype in 1979. In December 1980, the Ormco Lingual Task Force was formed to further test,
develop, and refine the lingual appliance and technique that would soon be a viable esthetic
solution for patients and clinically convenient tool for orthodontists (Figure 3) (Alexander et al,
1982).

Figure 3: Lingual arch form templates designed by the Ormco Lingual Task Force,
also with the mushroom shaped arch wire due to the construction in arch width as
the wire proceeds distally from canine to bicuspid. (Alexander et al, 1982)

It became apparent that lingual appliances had to be customized for each patient, as the
lingual morphology varies significantly not only between different patients but also between the
individual teeth within the same patient (Alexander et al, 1982). It was also determined that
indirect bonding (IDB) was absolutely necessary, as precise bracket placement was far more
critical than for labial braces due to a lack of adequate visualization. Bending lingual archwires
and accurately positioning lingual brackets freehandedly was significantly more difficult also due
to a reduced arch perimeter and inter-bracket distance. Furthermore, an ideal treatment result can
be attained only if the bracket position on the setup cast is also successfully transferred to the
patient’s mouth, despite the accuracy of the setup (Schubert et al, 2013).
The clinical debut of lingual appliances was met with an initial period of enthusiasm and
euphoria, followed by a long period of discouragement and frustration in the USA by 1987 as
shown in Figure 4 (Chatoo, 2013). The decline in popularity was due to difficulties achieving the
same high standard of results as with labial appliances, resulting in waned commercial interest in
8

developing the technique and growing reservations among clinicians about using the appliance
(Chatoo, 2013; Echarri, 2006).

Figure 4: Graph showing the rise and fall in popularity of lingual appliances in the
USA, while the number of cases treated in Europe and Japan increased slowly from
1980 to 2001 (Chatoo, 2013)

Nevertheless, devoted enthusiasts continued to develop the lingual appliance and find
solutions to those pitfalls and complications commonly faced by clinicians, leading to its eventual
redemption in 2004 back into the mainstream of orthodontics (Chatoo, 2013). Advanced
laboratory techniques, sophisticated wires and brackets, ever-improving bonding materials, and
digital technology eventually led to the successful treatment of all malocclusions using lingual
appliances and regaining clinicians’ confidence. Digital setups from therapeutic models have
allowed for more definitive treatment planning, so that tooth movements are more precise and
predictable (Larson et al, 2013). Computer-aided design and computer-aided manufacture
(CAD/CAM) of appliances led the way towards fully customized lingual bracket systems. With a
fully customized appliance, the orthodontist is given more three-dimensional (3D) control over the
entire treatment, and the most difficult part of conventional lingual appliances is nearly eliminated,
which is the orthodontic finishing of each case (Kothari et al, 2016).
9

The first fully customized lingual bracket system, Incognito (3M Unitek, Monrovia, CA),
was introduced by Dr. Dirk Wiechmann in 2002, using CAD/CAM technology to make
individually customized brackets perfectly adapted to the lingual anatomy as in Figure 5a, which
was one of its winning attributes against previous lingual bracket designs that were customized by
only modifying the composite base (Wiechmann et al, 2003). A robot bends the series of wires to
be used over the whole treatment based on the virtual target position for each individual tooth
(Grauer & Proffit, 2011). Several other fully customizable lingual appliances have since been
developed around the world and in popular use among orthodontists today, such as Suresmile
(Orametrix, Inc., Richardson, TX), STb (Scuzzo-Takemoto bracket; Ormco, Orange, CA), in-
Ovation L (Dentsply Sirona GAC, York, PA), and even self-ligating systems such as Harmony
(American Orthodontics Corporation, Sheboygan, WI) and Alias (Ormco).
(a) (b)
Figure 5: (a) Incognito appliance with broad, fully-adapted bracket pads and
robot-bent archwire (Grauer & Proffit, 2011). (b) INBRACE pre-fabricated
brackets with modified composite bases and a looped round archwire for torque
control.

The newest fully-customized lingual appliance today, INBRACE (Swift Health Systems,
Inc., Irvine, CA), is a unique and innovative lingual orthodontic solution consisting of looped
round archwires that engage into a springboard within pre-fabricated brackets that have a
modifiable composite base (Figure 5b). The loops at the interproximal extend gingivally, allowing
patients to floss freely and eliminating the tedious chore of threading underneath each interbracket
10

area, as is still the disadvantage with all other lingual appliances. The looped form also allows for
torque control on gentle round NiTi wires. Single-tooth adjustments can be made freely with the
extra lengths of wire available interproximally.

II.b. Bracket Failure: A Never-Ending Universal Problem
While lingual orthodontics have evolved and matured to deliver excellent treatment
outcomes comparable to labial appliances, one problem has yet to be completely resolved for all
bonded appliances. The common challenge that still flusters all orthodontists today is the bond
failure of individual brackets throughout the course of treatment, whether by poor bond technique
or isolation, a compromised tooth surface, or the patient’s lack of prudence in food choice. An
average of 2-3 bond failures within the first year of orthodontic treatment can be expected for both
appliances, with no significance difference in failure rate between labial versus lingual appliances
(a mean of 2.61 versus 2.63 bond failures, respectively) (Ziebura et al, 2014). Advanced bonding
systems with improved retention have helped to reduce but does not eliminate bracket failure, and
there are no significant differences in failure rates between modern adhesive systems (Mavreas et
al, 2017; Sunna & Rock, 1998).
For labial appliances, a failed bracket can be rebonded directly with great ease and speed;
however, for lingual appliances, the limited access, lack of direct visibility, difficult isolation, and
the wide variation in lingual tooth anatomy make IDB a crucial procedure (Mavreas et al, 2017).
IDB of a failed bracket is two-fold more accurate for all teeth in both labial and lingual systems
(Shpack et al, 2007). Direct bond (DB) of lingual brackets remains an option, but is not
recommended (Bittner, 2015). Although IDB may entail more extensive laboratory preparation,
this is counterbalanced by the simpler and faster chairside bonding (Lombardo et al, 2011).
11

Different tooth types—incisors, canines, premolars, and molars—have significant
variations in etch pattern after acid etching as well as in bond strength, which may have an
influence on the differences in failure rates between them (Linklater & Gordon, 2001). In-vivo
studies showed that posterior brackets fail more frequently than anterior brackets, possibly due to
increased masticatory loading. Molar brackets have a significantly higher bond failure rate (21%)
than premolars (12%), and premolars are more prone to bone failure than canines and incisors
(5%) (Ziebura et al, 2014). For labial appliances, mandibular brackets have a significantly higher
failure rate (9%) than for the maxillary (4%), perhaps due to the buccal overjet of the maxillary
teeth over the mandibular teeth (Linklater & Gordon, 2003). In lingual appliances, the opposite
would be suspected to be the case theoretically, but there is no significant difference between bond
failure of maxillary versus mandibular brackets (Ziebura et al, 2014). In this study, three tooth
types – molars, premolars, and anterior teeth – were independently evaluated for which rebond
method may best be suited for that type.
Individual bracket replacement after failure must be transferred as precisely as had been
initially planned for a successful treatment outcome. Given that one of the main goals of
orthodontic treatment is to reach a functional and esthetic balance between dental, skeletal, and
facial structures, optimal bracket placement is critical. The afflictions of poor bracket placement
are plenty. Even a minor deviation in bracket placement from the target position can result in
significant tooth misalignment and inadequate finish (Schubert et al, 2013). For example, the
wrong vertical position can result in unlevel teeth, torque alterations, misalignment, and arch
length alterations, compromising the functional integrity of the teeth (Joiner, 2010). A drift in the
mesiodistal position of the bracket would result in tooth rotation, and rotation about a buccolingual
axis would cause unwanted tipping. Bracket placement on posterior teeth is especially critical as
12

they tend to extrude readily as archwires level out, potentially opening the bite, from which
recovery is challenging.
The penalties to the orthodontist for poor bracket placement include the consequent need
for time-consuming compensatory bends, complicated treatment, longer chair time, and
shortcomings in final treatment outcomes, since the bracket positions determine the final teeth
positions (Birdsall et al, 2012; Mota Júnior et al, 2015). To bend the archwire for precise
adjustment of angulation, inclination, and torque is extremely challenging due to the reduced
lingual interbracket distance between anterior teeth (Schubert et al, 2013). To avoid such errors
and to lessen the clinician’s potential distress, a reliably accurate rebond method must be utilized
for optimal effectiveness and efficiency of treatment (Brown et al, 2015).

II.c. Current Rebond Techniques
The most common method today for rebonding single-tooth lingual bracket is by sectioning
the original full-arch transfer PVS or silicone tray that was used to initially deliver the appliance.
This indirect method is described in some clinical manuals as seen in Figure 8. The interface
between the tooth and bracket must be thoroughly cleaned of any residual adhesive. The failed
bracket is then carefully replaced into the flexible tray, and after proper preparation of primer and
adhesive, the tray is then carefully positioned over the tooth for bonding.
Direct bonding (DB) is another modality to replace failed brackets. DB may be necessary
for incorporating previously unerupted teeth, for which a positioning jig or tray would not have
yet been able to be fabricated in the initial treatment planning phase (Bittner, 2015). Sufficient
accuracy can be assumed for the customized brackets that are fully adapted to the lingual surface,
such as Incognito (Figure 5a) and Harmony. Because of the adaptation of the pad to the individual
13

tooth, the bracket is able to be decently stabilized over the contours of the lingual surface. Yet
there is no such scientific literature today that has evaluated the accuracy of directly rebonding
lingual brackets, whether with fully adapted pads or by eyeballing.
Another notable indirect bracket-transfer system is the Hiro technique in Figure 6 that was
created by Toshiaki Hiro in 2008 and improved by Kyoto Takemoto and Giuseppe Scuzzo (Buso-
Frost & Fillion, 2006; Hiro et al, 2008; Schubert et al, 2013). The advantages of this technique
are its simplicity and low cost. The transfer trays are made individually with Fermit or any
provisional resin material for each bracket on a setup model. The trays can be transferred directly
from the setup model to the mouth, but chair time is longer as each tray has to be transferred,
positioned, and bonded one at a time. Another disadvantage is that the transfer trays cannot be
reused, as they get destroyed during its removal after bonding. Modifications of this technique to
be made reusable have been attempted, such as the Quick Modul System in Figure 7 (QMS,
Halbich Lingualtechnik, Berlin, German) (Schubert et al, 2013).
The objective of IDB protocols is to correctly position brackets on a handheld or virtual
model and then accurately transfer the ideally-placed brackets to the patient’s teeth. The accuracy
of laboratory-fabricated transfer techniques for labial appliances have had varying success across
several investigations (Castilla et al, 2014; Nichols et al, 2013; Wendl et al, 2008). One study
found minimal improvements that were not statistically significant with laboratory-fabricated IDB
techniques when compared with DB of labial brackets (Koo et al, 1999). Nevertheless, there are
very limited studies available about the accuracy of single-tooth bonding techniques for lingual
appliances (Schubert et al, 2013).
CAD/CAM technology has become a growing focus in orthodontics to minimize such
human errors. The ideal lingual indirect rebond jig would be accurate, easy to fabricate, low-cost,
14

simple to use, quick to deliver, and reusable for future bond failures. The goals of incorporating
CAD/CAM into orthodontics are to improve reproducibility, efficiency, and quality of orthodontic
treatment (Brown et al, 2015). Materials used today for indirect bracket transfer include vacuum-
formed thermoplastic foils, single- and multi-layer silicone trays, prosthetic putties,
polyvinylsiloxane (PVS), and silicone gels (Brown et al, 2015). With CAD/CAM technology,
bracket loss can be seen as a less complex issue, as precisely-milled IDB jigs can be looked to as
a routine and easily manageable solution to bond failure in the foreseeable future. Being able to
consistently rebond brackets in the correct position is more cost-effective in terms of both time
and materials. For this study, a novel indirect rebond method was designed using a rigid 3D-
printed jig, which will be described later in the Materials and Methods section.

Figure 6: The non-reusable Hiro system for individual transfer of lingual brackets
is made of a light-curable resin. (Buso-Frost & Fillion, 2006)

(a) (b)
Figure 7: The reusable QMS jig for transferring lingual brackets. To the left (a) is
the jig fabricated upon the setup model, and demonstrated to the right (b) is the
intraoral transfer (Schubert et al, 2013)
15

(a)
16

(b)
17

(c)
18

(d)
19

(e) (Harmony, 2012)

20

(f)

Figure 8: Rebond protocols available online for (a) indirect rebond of Incognito
by 3M Unitek, (b) direct rebond of Incognito, (c) indirect rebond of Suresmile
Fusion by OraMetrix, (e) direct rebond of Suresmile Fusion which is not
recommended, (e) Harmony by AO, with individual “Positioning Jigs” provided,
and (f) INBRACE by Swift Health Systems.

II.d. Assessment of Bracket Position Accuracy
The precision of bonding techniques have been evaluated in various ways. Two-
dimensional evaluations of bracket precision include traced superimpositions and photographic
superimpositions. Traced superimpositions entail enlargening photographed images of the
experimental and control groups for tracing the bracket upon acetate paper, and then
superimposing the tracings over the outline of the tooth (Koo et al, 1999). Photographic
superimpositions are overlapping high-quality two-dimensional scaled images taken of the control
and experiment that are assessed at marked reference points (Aguirre et al, 1982; Koo et al, 1999;
Wendl et al, 2008).
3D evaluations are possible with virtual superimposition of teeth scanned with a handheld
intraoral optical scanner or cone-beam computed tomography (Larson et al, 2013). Several 3D
imaging software programs are available to calculate the changes or deviations between two
21

superimposed 3D images, such as the GeoAnalyzer (Orametrix), eModel Compare (GeoDigm),
Geomagic Design X (3D Systems), and Cimatron (3D Systems) (Kim et al, 2018; Larson et al,
2013; Schubert et al, 2013). Deviations can also be visualized by color maps where different colors
represent varying degrees of devation (Grauer & Proffit, 2011).

III. PURPOSE
The overall goal of this investigation was to determine the accuracy of and to compare
three different single-tooth rebond methods for a CAD/CAM customized lingual appliance. The
three rebond methods to be assessed and compared were DB, flexible tray IDB, and 3D-printed
rigid jig IDB. The specific goals of this pilot study were:
1. To quantify and compare the accuracy for rebonding brackets with a flexible tray,
3D printed rigid jig, and DB.
2. To compare the accuracy of the three methods for different tooth morphologies
(molar vs. premolar vs. anterior tooth), and
3. To compare the accuracy of each rebond method in each translational dimension.
The null hypotheses were that there would be no significant difference in the precision of
bracket position for each of the rebond methods; no difference between incisor, premolar, and
molar bracket position accuracies for any of the rebond methods; and no difference in the position
accuracy between the translational dimensions. The significance level was set at α = 0.05, and the
clinically acceptable limit was set at 0.5mm.

22

IV. MATERIALS & METHODS
The preparation, setup and execution of this study are described in great details as this
study is intended to contribute towards the development of a single-tooth IDB methodology
applicable to all fully customizable lingual appliances.
IV.a. Sample
A digitized malocclusion model of a mandibular arch with mild crowding, rotations and
tip (Figure 9) was used to fabricate the samples for this study. The CAD/CAM bracket position
setup was designed for the INBRACE lingual appliance. The sample comprised of 30 identical
3D-printed copies using the MoonRay 3D printer (SprintRay, Inc., Los Angeles, CA). As a pilot
study, it was deemed sufficient to have 10 samples for each of the three rebond methods. The
material used to 3D-print the models was a gray photopolymer resin (SprintRay, Inc.). It was
determined that only three teeth of different morphologies in one quadrant would be used to
evaluate bracket positions for this study: LL1, LL4, and LL6. One orthodontic resident of the
University of Southern California with prior experience of delivering the INBRACE appliance
performed this investigation. The samples were divided into three groups of 10 and labeled from
1 to 10 for each rebond method category: FT for the flexible tray group; RJ for the rigid jig group;
and DB for the direct bond group. Each sample was labeled by the category then number, e.g.
FT01 for flexible tray sample 1.

Figure 9: Labeled 3D printed models composed of a gray photopolymer resin
(Sprintray, Inc.)
23

IV.b. Jig Fabrication
The flexible trays were fabricated with Lumaloc (Opal Orthodontics, South Jordan, UT), a
light blue translucent PVS, by a single laboratory technician trained to make the trays with uniform
dimensions. The dimensions of the single-tooth IDB trays were made to extend to the mesial and
distal edges of the tooth; cover the entire bracket and lingual surface of the tooth extending 1.5
mm from the cervical edge of the bracket; and cover two-thirds to three-fourths of the buccal or
labial surface. The flexible trays were for single use only, hence a total of 30 trays for each
individual trial were produced to perform the flexible tray rebond procedure to eliminate
confounding error (Figure 11).

Figure 10: 3D printed model with bracket wells for fabricating flexible bonding
trays with Lumaloc, a light blue transparent PVS.

Figure 11: Individual Lumaloc flexible trays fabricated and loaded with
corresponding brackets. There are apparent differences in the dimensional
thicknesses of the trays for each tooth type. The top row of trays are for LL6; the
middle row for LL4; and the bottom row for LL1.

24

The rigid jigs used were designed using a Swift Health Systems (SHS) proprietary software
and 3D-printed by the MoonRay 3D printer with a white translucent photopolymer resin fabricated
so that the brackets could be light-cured (Figure 12). Only one set for each tooth was printed to
be used repeatedly. The jigs were checked for adaptation to the model teeth so that it can be held
consistently in a stable position for each trial.

(a) (b)

(c) (d)
Figure 12: (a) Freshly 3D printed rigid jigs prior to finalized curing process. (b)
Rigid jigs after the curing process. (c) Rigid jig for LL6 after detachment from 3D-
printed base, showing the void in the well gingival to the bracket for clearance of
removing the jig without accidentally debonding the bracket. (d) Rigid jigs placed
over respective model teeth.

IV.c. Rebond Methods
The accuracy of three methods of rebonding brackets was to be tested: one DB method and
two IDB methods.

25

IV.c.i. Direct Bond
The DB method was to freehandedly position the bracket as guided by the Virtual Bracket
Positioning Card provided in the original INBRACE package (Figure 13). Assure Plus and Light
Bond Paste were the primer and adhesive, respectively (Reliance Orthodontic Products Inc., Itasca,
IL) used for placing every bracket in this study. The Bracket Positioning Guide was used to
determine the bracket position upon the lingual surface of the tooth, as well as the thickness of
composite required which is indicated in red.

Figure 13: Virtual Bracket Positioning Guide showing the position of the bracket and
the amount of composite required in the bracket base in red.

IV.c.ii. Flexible Tray IDB
The IDB method using a flexible tray was performed as follows. First, the tray was placed
upon the corresponding tooth to check that it does not rock and can be positioned without
interference in the proper mesiodistal position. Then the bracket was placed into the well, making
sure it was fully seated and engaged. Then a minimal layer of Assure Plus primer was applied,
with air from a tooth dryer applied for 3 seconds to evaporate the solvent in the primer. Then a
26

minimal sesame-seed sized amount of adhesive was applied, making sure all four corners of the
bracket base were covered (Figure 14a). The jig was carefully positioned over the tooth from the
lingual to the buccal, so not to distort the load of adhesive. The two-finger stabilization technique
was utilized, with one finger on the bucco-occlusal portion of the jig, and one other finger on the
lingual surface (Figure 14b). Gentle pressure was applied so not to distort the form of the PVS.
Then it was checked again that the tray does not rock. The curing light was applied for 10 seconds
with 2-finger pressure while holding tray. The light was applied again for an additional 10 seconds
without holding the tray. Then the tray was released towards the lingual to minimize tension upon
the bracket so not to potentially break it off the model (Figures 14c-d).

(a) (b)

(c) (d)
Figure 14: Proper removal of the flexible tray from buccal (a) to lingual (b) so not
to inadvertently break the bracket off the model.

27

IV.c.iii. Rigid Jig IDB
The IDB method using a rigid jig required meticulous preparation. The rigid jigs were
prepared by applying a thin layer of Vaseline petroleum jelly to be used as a separating agent with
a microbrush to the bracket well and areas surrounding the well to prevent the adhesive from
bonding to the resin jig itself. The bracket inserted into the well using a bracket holding plier,
avoiding any contact between the Vaseline jelly and the base to ensure bonding of the bracket to
the adhesive. White wax was packed into the gingival void in the well to hold the bracket and
also to prevent overflow or excess adhesive into the springboard area.

Figure 15: 3D printed models with indirectly bonded brackets using rigid jigs on
LL1, LL4, and LL6.

IV.d. Scanning, Superimposition, & Error Measurement
Each 3D-printed malocclusion model was scanned thoroughly once by the same high-
resolution optical intraoral scanner after all brackets were bonded, as shown in Figure 15 (iTero
version 5.2, Align Technology, San Jose, California). The brackets were scanned especially
carefully from several angles to ensure that the dimensions and details of the brackets were
captured accurately. The scans were exported as stereolithography (STL) files.
28

The control was a digital model of the same malocclusion but with the virtually-planned
bracket positions (Figure 16a). The digital control model was used as the baseline from which to
observe the deviations of the brackets from each bonding trial. The flexible trays, rigid jigs, and
virtual positioning card were all fabricated based on the control model.
To assess the bracket position accuracy, the STL files of the control malocclusion model
with the digital bracket positions and the sample models after the bondings were imported into the
SHS software. A surface-based superimposition was performed after importing the STL files of
both models, in which the unchanged surfaces of the scanned and control teeth were registered and
superimposed (Figure 16b). The deviations of the experimentally bonded brackets were then able
to be visualized and quantified.
The deviations were quantified using the linear measurement tool in the SHS software in
three translational dimensions: buccal-lingual, mesial-distal, and occlusal-cervical. The buccal,
mesial, and occlusal directions were considered numerically “positive” while the latter lingual,
distal, and cervical directions were recorded as numerically “negative.” The reference points to
measure the amount of bracket position error were determined both visually and digitally to be at
the area of the brackets with the greatest displacement from the control (Figures 16c-f).
Measurements were made to the nearest hundredth of a millimeter and recorded as positive or
negative depending on the direction of the discrepancy for each bracket (Table 2).

Figure 16 (below): (a) In white, the control model with digitally-planned bracket
positions. In purple, the digitally scanned sample model after bonding brackets. (b)
Superimposed views from the occlusal and lingual. (c) Enlarged view of the visible
deviations of the brackets on LL6, LL4, and LL1, respectively, after
superimposition. (d-i) Measurements of the deviations along the three translational
planes: buccolingual, mesiodistal, occlusocervical. (d) LL6 superimposition
showing measurements at reference points of greatest deviation between the
brackets. (e-f) A closeup of B-L measurements for LL6. (g) LL4 superimposition
29

showing measurements. (h) A closeup of LL4 measurements. (i) View of LL1
superimposition with measurements in all three planes.

(a)

(b)

(c)

(d)
30

(e)

(f)

31

(g)

(h)

32

(i)

IV.e. Statistical Analysis
The independent variables were (1) the three rebond methods and (2) the individual teeth
(LL1, LL4, LL6) that were bonded. The dependent variables were the deviations in millimeters
of the rebonded brackets from the control brackets in the three translational dimensions
(buccolingual, mesiodistal, occlusocervical).
Our objectives were (1) to analyze the overall accuracy of each rebond method; (2) to
compare the accuracy of each rebond method for each tooth; and (3) to compare the accuracy of
each rebond method in each dimension. The significance level was set at α = 0.05, and the
acceptable measurement error for the clinical limit was set at 0.5 mm in the M-D, B-L, and O-G
dimensions, as in the superimposition study design for bracket placement accuracy by Kim et al,
2018. Descriptive data and two-way ANOVA were run using SPSS (Armonk, NY) to evaluate
the data.

33

V. RESULTS
In Table 1 below is the raw data collected. Table 2 simply clarifies the meaning of the
negative and positive values.

LL6 LL4 LL1
Sample B/L O/G M/D B/L O/G M/D B/L O/G M/D
FT01 -0.46 -0.04 0 -0.41 -0.42 0.20 0.14 0.05 0.16
FT02 -0.58 -0.22 0.14 -0.51 -0.47 -0.22 -0.11 -0.22 0.08
FT03 -0.20 -0.13 -0.04 -0.23 -0.35 -0.10 -0.30 -0.39 -0.10
FT04 -0.20 0.29 0.2 0.20 -0.68 -0.54 0.32 -0.31 -0.50
FT05 -0.22 0 0.23 -0.18 -0.37 -0.13 -0.23 -0.35 0
FT06 -0.15 0.12 0.21 0.19 -0.19 -0.12 0.14 -0.24 0.09
FT07 -0.40 -0.08 0.16 -0.18 -0.27 -0.40 -0.15 -0.23 -0.08
FT08 -0.14 -0.17 0.24 -0.20 -0.22 -0.20 -0.26 -0.28 -0.04
FT09 -0.24 0 0.24 -0.18 -0.18 -0.07 -0.24 -0.43 -0.19
FT10 -0.54 -0.08 0.45 -0.25 -0.24 -0.14 -0.13 -0.52 -0.42
DB01 -0.15 -0.12 0.25 0.05 -0.50 0.09 -0.14 -0.71 -0.62
DB02 -0.34 0.19 0.17 -0.64 -0.74 -0.41 0.47 -0.79 -0.69
DB03 -0.14 0.25 -0.10 -0.25 -0.61 -0.50 0.28 -0.31 -0.09
DB04 -0.2 0.30 0.34 0.19 -0.65 -0.62 -0.35 -0.28 -0.48
DB05 -0.31 -0.24 -0.10 0.20 -0.79 -0.57 -0.26 -0.81 -0.37
DB06 -0.45 -0.34 0.25 0.14 -0.60 -0.29 0.16 -0.46 -0.41
DB07 -0.23 -0.28 -0.25 0.18 -0.55 -0.36 0.21 -0.40 -0.19
DB08 -0.19 -0.41 -0.22 0.25 -0.64 -0.41 -0.28 -0.73 0.23
DB09 -0.22 -0.41 0.10 0.21 -0.80 -0.66 0.25 -0.63 -0.45
DB10 -0.31 -0.57 0.11 -0.25 -0.71 -0.27 -0.32 -0.82 -0.28
RJ01 -0.62 0.38 0.12 -0.31 0.13 -0.10 -0.46 -0.53 0.53
RJ02 -0.95 0.18 0.39 -0.36 -0.09 0.07 -0.36 -0.17 0.09
RJ03 -0.13 0.28 -0.10 -0.55 -0.32 -0.18 -0.98 -0.25 0.22
RJ04 -0.14 0.49 -0.07 -0.44 0.26 -0.25 -0.79 -0.09 -0.12
RJ05 -0.37 0.34 0.26 -0.70 0.26 -0.13 -0.98 -0.23 0.11
RJ06 -0.43 0.24 0.14 -0.31 0.03 -0.11 -0.38 0.08 -0.47
RJ07 -0.69 0.27 0.19 -0.39 0.28 0.20 -0.40 -0.44 -0.12
RJ08 -0.45 0.23 0.07 -0.26 0.11 -0.12 -0.36 -0.50 -0.24
RJ09 -0.48 0.29 -0.20 -0.36 0.24 -0.10 -0.33 -0.28 0.16
RJ10 -0.59 0.25 0.12 -0.41 0.15 -0.10 -0.39 -0.44 0.09
Table 1. Measurements from the corner of each bracket with the most
displacement in the three translational planes: B/L (buccolingual), O/G
(occlusogingival), M/D (mesiodistal).

Negative (-) Positive (+)
Buccal/lingual lingual buccal
Occlusal/gingival gingival occlusal
Mesio-distal distal mesial
Table 2. Clarification of the meaning of positive and negative values.

34

Table 3. Maximum Absolute Error (MAE)
Sample/Tooth
LL6 LL4 LL1
B/L O/G M/D MAE B/L O/G M/D MAE B/L O/G M/D MAE
FT FLEXIBLE TRAY - Total mean: 0.34
FT01 -0.46

0.46

-0.42

0.42

0.16 0.16
FT02 -0.58

Mx 0.58 -0.51

0.51

-0.22

0.22
FT03 -0.20

0.2

-0.35

0.35

-0.39

0.39
FT04

0.29

0.29

-0.68 ME 0.68

-0.50 0.5
FT05

0.23 0.23

-0.37

0.37

-0.35

0.35
FT06

0.21 0.21 0.19

0.19

-0.24

0.24
FT07 -0.40

0.4

-0.4 0.4

-0.23

0.23
FT08

0.24 0.24

-0.22

0.22

-0.28

0.28
FT09

0.24 0.24 -0.18

0.18

-0.43

0.43
FT10 -0.54

0.54 -0.25

0.25

-0.52 Mx 0.52

Mean 0.34

Mean 0.36

Mean 0.33
Dir 50% L 10% O 40%M

30% L 50% G 10% D

80% G 10% D

10% B

10% M

DB DIRECT BOND - Total mean: 0.54
DB01

0.25 0.25

-0.50

0.50

-0.71

0.71
DB02 -0.34

0.34

-0.74

0.74

-0.79

0.79
DB03

0.25

0.25

-0.61

0.61

-0.31

0.31
DB04

0.34 0.34

-0.65

0.65

-0.48 0.48
DB05 -0.31

0.31

-0.79

0.79

-0.81

0.81
DB06 -0.45

0.45

-0.60

0.60

-0.46

0.46
DB07

-0.28

0.28

-0.55

0.55

-0.40

0.40
DB08

-0.41

0.41

-0.64

0.64

-0.73

0.73
DB09

-0.41

0.41

-0.80 Mx 0.80

-0.63

0.63
DB10

-0.57 Mx 0.57

-0.71

0.71

-0.82 Mx 0.82

Mean 0.36

Mean 0.66

Mean 0.61
Dir 30% L 40% G 20% M

100% G

90% G 10% D

10% O

RJ RIGID JIG - Total mean: 0.51
RJ01 -0.62

0.62 -0.31

0.31

-0.53

0.53
RJ02 -0.95

0.95 -0.36

0.36 -0.36

0.36
RJ03

0.28

0.28 -0.55

0.55 -0.98

0.98
RJ04

0.49

0.49 -0.44

0.44 -0.79

0.79
RJ05 -0.37

0.37 -0.70

Mx 0.7 -0.98

Mx 0.98
RJ06 -0.43

0.43 -0.31

0.31

-0.47 0.47
RJ07 -0.69

Mx 0.69 -0.39

0.39

-0.44

0.44
RJ08 -0.45

0.45 -0.26

0.26

-0.50

0.50
RJ09 -0.48

0.48 -0.36

0.36 -0.33

0.33
RJ10 -0.59

0.59 -0.41

0.41

-0.44

0.44

Mean 0.54

Mean 0.41

Mean 0.58
Dir 80% L 20% O

100% L

50% L 40% G 10% D

Table 3. All values besides the greatest absolute value for each tooth in each sample
were eliminated for visualization of a pattern in the translational planes. The
proportions of observations of the greatest error for each plane were recorded for
each tooth type and rebond method. Mx = the highlighted value of the greatest
absolute error for that tooth for each rebond method.

35

In Table 3, only the values of the maximum absolute error (MAE) for each sample tooth
were retained. One reason for this was to assess only the values of the greatest deviation for each
individual bracket. The other value to organizing the table this way was to be able to visualize any
patterns in the translational dimensions for each tooth type and rebond method.
The MAE means of each tooth type for the FT method were very similar to each other
while they were also the lowest for each tooth type, even considering that the trays were flexible.
The MAE between LL6 and LL1 was similar even though the LL6 bulk of the Lumaloc material
had more extensions and would be expected to be more stable than for LL1. The other interesting
finding with LL1 using the FT method was that most of the greatest absolute errors vertical were
in the O-G direction – 80% of the maximum bracket deviations were in the gingival direction. The
flexibility of the LL1 tray was apparently not important for the M-D or B-L planes but was
significant for the O-G vertical dimension.
When assessing the MAE values from the DB method, there is no apparent pattern of error
for LL6. The most apparent pattern of errors was seen for LL4 and LL1. In 100% of the MAEs,
the bracket was positioned too gingivally for LL4 using DB method. For LL1, 90% of the MAEs
were towards the gingival direction, as well. The DB method has less accuracy than other methods
except with the LL6, which had decent accuracy.
In assessing the accuracy of the RJ method, the mean errors of the MAE values for each
tooth were bigger than for the FT method. For LL6, 80% of bracket positions were too lingual;
100% for LL4; and 50% for LL1. MAE values did not occur in the M-D dimension for LL6 or
LL4, and only 10% for LL1, meaning the rigid jig was excellent for M-D positional control with
this rebond method.

36

A tooth-by-tooth comparison using Table 3 shows that for LL6, both FT and DB methods
had the lowest mean MAE and were also similar in mean MAE (0.34mm and 0.36 mm,
respectively), while the RJ method had the greatest mean MAE (0.54mm deviation). For both LL4
and LL1, the MAE means using the FT method were lower than that of the DB and RJ methods.

Statistical Analysis
Statistical analyses tested if the aforementioned inferences made from Table 3 were valid
or invalid. The descriptive data in Table 4 was generated using the MAE values from Table 3.

Table 4. Descriptives
Measurement Method Statistic Tooth Statistic
Mean
FT
0.34
LL6
0.41
St Err 0.03 0.03
Median 0.32 0.41
Mode 0.24 0.24
SD 0.14 0.17
Variance 0.02 0.03
Kurtosis -0.47 2.02
Skewness 0.65 1.17
Range 0.52 0.75
Minimum 0.16 0.20
Maximum 0.68 0.95
Sum 10.28 12.35
Count 30 30
Cnf Lvl (95%) 0.05 0.06
Mean
DB
0.54
LL4
0.48
St Err 0.03 0.03
Median 0.56 0.43
Mode 0.25 0.55
SD 0.19 0.18
Variance 0.03 0.03
Kurtosis -1.34 -1.10
Skewness -0.06 0.17
Range 0.57 0.62
Minimum 0.25 0.18
Maximum 0.82 0.80
Sum 16.34 14.25
Count 30 30
Cnf Lvl (95%) 0.07 0.07
Mean
RJ
0.51
LL1
0.51
St Err 0.04 0.04
37

Median 0.45 0.47
Mode 0.36 0.50
SD 0.20 0.22
Variance 0.04 0.05
Kurtosis 0.79 -0.48
Skewness 1.22 0.60
Range 0.72 0.82
Minimum 0.26 0.16
Maximum 0.98 0.98
Sum 15.26 15.28
Count 30 30
Cnf Lvl (95%) 0.07 0.08

(a) (b)

(c) (d)
Figures 17a-d above shows the histographic and boxplot distributions of the deviations
for each rebond method. (a) The FT method has the most normal distribution, as well
more bracket deviations that were within clinical limits (set at 0.5mm). (b) The DB
method is apparently more prone to greater bracket position deviations. (c) The RJ
method histogram has a right-skewed distribution.

4
17
8
1
0 0
0
5
10
15
20
0.2 0.4 0.6 0.8 1 More
Frequency
Error (mm)
FT Method Histogram
0
8
9
11
2
0
0
5
10
15
0.2 0.4 0.6 0.8 1 More
Frequency
Error (mm)
DB Method Histogram
0
10
13
4
3
0
0
2
4
6
8
10
12
14
0.2 0.4 0.6 0.8 1 More
Frequency
Error (mm)
RJ Method Histogram
0
0.2
0.4
0.6
0.8
1
1.2
FT DB RJ
Methods Box Plot
38

(a) (b)
(c) (d)
Figures 18a-d above shows the histographic and boxplot distributions of the
deviations for each tooth type: LL1, LL4, and LL6.

Table 5. Two-way ANOVA results from SPSS.

1
14
12
2
1
0
0
5
10
15
0.2 0.4 0.6 0.8 1 More
Frequency
Error (mm)
LL6 Histogram
2
11
8
9
0 0
0
5
10
15
0.2 0.4 0.6 0.8 1 More
Frequency
Error (mm)
LL4 Histogram
1
10 10
5
4
0
0
5
10
15
0.2 0.4 0.6 0.8 1 More
Frequency
Error (mm)
LL1 Histogram
0
0.2
0.4
0.6
0.8
1
1.2
LL6 LL4 LL1
Tooth Box Plot
39

Table 6. Tukey HSD post-hoc test results for rebond methods.

Table 7. Pairwise comparisons – the tooth types for each method were isolated to
make individual comparisons between method and tooth type.

The results of a two-way ANOVA are provided in Table 5. There was no statistically
significant difference in deviations between the tooth types (p=.06). A very highly significant
40

difference existed between the three rebond methods (p<.01). The interaction between tooth type
and rebond method were highly statistically different (p=.001), meaning that there were
statistically significant differences of the deviations of particular teeth bonded using particular
methods. A pairwise comparisons (Table 7) isolated each tooth type for each method.
Table 6 shows that there was a very highly statistically significant difference between FT
and both DB and RJ methods (p<.01), but there is no significant difference between DB and RJ
(p=0.66).
In Table 7, pairwise comparisons show that for the FT method, there were no significant
differences between any of the tooth types (p=0.73, p=0.92, p=0.80 between LL1-4, LL1-6, and
LL4-6 respectively). For the DB method, there was no significant difference in error between LL1
and LL4 (p=0.53), while there was a highly statistically significant difference between LL1 and
LL6 (p=0.001). For the RJ method, there was a statistically significant difference only between
LL1 and LL4 (p=0.02), but not between LL6 and both LL1 and LL4 (p=0.51, p=0.08 respectively).

Figure 19. Estimated marginal means of error.
41

Figure 19 provides a visualization of the mean MAE for each rebond method relative to
each other. For the FT method, it is clear that the mean errors for each tooth type were not
significantly different from one another, and thus similarly low. For the DB method, there was no
significant difference between LL1 and LL4 mean errors, but they were significantly worse
compared with LL6. For the RJ method, there was a significant difference between the mean
errors of LL1 and LL4, but not between LL1 and LL6 or between LL4 and LL6.

VI. DISCUSSION
Across the mean errors for each rebond method, the FT method had a statistically
significantly low mean error relative to that of the DB and RJ methods, from which it can be
inferred that this rebond method is the most accurate. There was also no significant difference
between any of the teeth for the FT method, which while the mean errors were low, confirms the
inference from Table 3 and with reference from Table 7 that that the FT method was consistently
the most accurate method for all tooth types.
The interesting finding with the FT method is that 80% of the greatest bracket deviations
on LL1 and 50% on LL4 were in the gingival direction. The flexibility of the LL1 tray was
apparently not a problem for either the M-D or B-L planes, but was significant for the vertical O-
G dimension. This can be explained by the likelihood that too much gingival pressure was
consistently applied, since there is only a small incisal surface area to provide sufficient vertical
stability of holding the tray. The tray size is also significantly smaller, making the tray more
flexible than for the other teeth. The vertical deviations were not as prominent for LL6, perhaps
due to the broad occlusal surface area that stabilizes the vertical position of the tray, but in turn
there was a consistent lingual deviation in bracket position for LL6. Therefore it is recommended
42

that very gentle incisal pressure be applied for rebonding LL1 and LL4 with a flexible tray, and
that more lingual pressure to be applied when rebonding LL6. The consequence of a bracket
positioned too gingivally is extrusion, which can in turn result in traumatic hyperocclusion,
premature contacts in function, and the need for another reposition of the bracket or detailing
bends.
An explanation for the inaccuracy of the DB method is that perhaps the Virtual Positioning
Card was proportionally scaled too small for a clinician to accurately gauge and identify reference
points to the correct dimension when trying to position the bracket. Furthermore, it seemed that
the teeth printed on the provided Virtual Bracket Position Card were not oriented properly,
specifically with excessive lingual crown torque. Using a tooth printed with excessive lingual
torque may result in a bracket positioned too gingivally, and that with excessive buccal torque may
result in an occlusally deviated bracket placement. The LL6 and LL4 printed on the card used in
this study (Figure 13) appear to be torqued lingually because of the visibility of the occlusal
surface, while it cannot be assumed of LL1 as the incisal edge would appear thin in any orientation,
hence perhaps this is what led to the very consistent gingival deviations in bracket placement (40%
for LL6, 100% for LL4, and 90% for LL1, in Table 3). Furthermore, since the cards are not printed
to size, it is difficult to gauge the accurate thickness of composite suggested in red.
In assessing the accuracy of the RJ IDB method, the mean errors of the maximum absolute
values for each tooth were bigger than for the FT IDB method. For LL6, 80% of the greatest
bracket deviations were towards the lingual; 100% for LL4; and 50% for LL1. This significant
tendency for lingual displacement points to an error in the design of the rigid jig. The depth of the
bracket well may have been too deep relative to the digital design, or perhaps the overall dimension
of the jig was too long buccolingually and the stable position of the jig during bonding was tended
43

towards the lingual. Regardless, the greatest deviations did not occur in the M-D dimension for
LL6 or LL4, and only 10% for LL1, meaning that the rigid jig is excellent for M-D positional
control with the RJ method. Overall, the 3D printed rigid jigs are not ready to be used as routine
rebond method.
Another clinical significance of these results is that we can assume the error in an intraoral
in vivo setting to be greater due to limited access to the individual teeth. This study was on
freestanding typodonts with easy access to each tooth, whereas in the mouth, the molar is the most
difficult to access, while the incisor is the easiest to access. While the DB results in our study had
large errors, this error can be expected to be greater in a clinical situation, and thus DB should be
avoided. The patient should rather be rescanned for the fabrication of an individual tooth tray or
jig for IDB.

VI.a. Limitations & Future Suggestions
There were many limitations of this study. Many confounding variables were identified
throughout the course of this study that possibly contributed to the errors in our results. One such
variable to explain the misfit of the rigid jig upon the model teeth was that the different colored
resins used to 3D-print the models and the jigs were suspected to have unpredictable amounts of
shrinkage. Even though the differences in shrinkage were accounted for each color of resin by the
manufacturer, there may have been unpredictable shrinkage in different dimensions for various
thicknesses and forms during the hour-long curing process that follows the printing, even if the
printing may be accurate. In one trial during the research and development of this study design,
the rigid jigs were printed from the same exact digital design by the same MoonRay 3D printer in
the white resin and again in a different orange surgical-guide resin manufactured by the same
44

company, with the varying shrinkages accounted for in the design program; however, the jigs fit
differently on the model teeth. It might have been best if the same color resin were used for both
the model and the jig to eliminate this error, but the curing light cannot be shined through the
opaque grey resin to bond the brackets, and visualization is difficult if the models were to printed
with the white resin, which is actually transparent. A future superimposition study can compare a
3D scan of the rigid jigs with the digital design to identify the areas of shrinkage, to either confirm
or eliminate this source of error and possibly point to another.
The possibility that there might be error between 3D-printed models of an identical design
was also suspected to contribute to the degree of error in our study. This could have possibly
occurred due to the potential unpredictable shrinkage during the curing process. This suspicion
arose during printing of multiple rigid jigs that were digitally identical copies made of the same
exact resin but had varying fit onto the models. This suspicion could be relieved or verified with
a future study to assess accurate reproducibility of 3D printed designs, by superimposing digital
scans of the 3D-printed models or jigs themselves to one each other to evaluate the degree of
identicalness between them.
Another potential source of error with the rigid jig method is that the high friction between
the resin materials of the rigid jig and 3D-printed models made it slightly challenging to gauge the
truly most stable position over its respective tooth. The visible and rough printed layers of the
models and jigs can be assumed to have additionally contributed to a higher degree of friction.
The inconsistencies in bracket position from using the flexible jig method, though they
were minimal, can potentially be explained by the lack of exact uniformity Lumaloc trays, as seen
in Figure 11. If the trays were perfectly uniform, that would have at least eliminated another
45

potential confounding variable contributing to the error in the variation of the bracket positions
observed using this indirect rebond method.
It was also apparent Accuracy of iTero 5.2 3D scanner did not fully capture all details of
the brackets or models as seen in the superimpositions in Figures 16d-i, preventing the accurate
measurement of the true greatest bracket deviations. A future study can repeat this experiment
using a more precise, perhaps newer, model of a digital scanner in capturing the minute details and
corners of the brackets.
Another limitation to this study is that an “ex vivo” setting was not simulated. The
positions of each tooth type in the intraoral cavity might contribute to errors in transferring
individual brackets. A future study can incorporate a typodont affixed to a manikin. A “real life”
setting may reveal errors in bracket placement contributed by restricted head movement of the
patient and the obstructive presence of the tongue and cheek.
While the proprietary SHS software used was an important tool, it did not allow for the
quantification of the bracket rotations in 3D, but instead only the larger linear displacements. In a
future study using another imaging software, each tooth can be assigned a coordinate system to be
able to assess axial deviations in rotation, inclination, or angulation along with the linear deviations
along translational planes (Grauer & Proffit, 2011; Schubert et al, 2013).
A limitation to the reliability of our measurements in our study was that we did not repeat
measurements. Time was limited, and data collection for this experiment was extremely tedious.
An intrarater agreement and intraclass correlation coefficient would have been useful calculations
for reliability of the experimenter. The iTero 5.2 scanner had difficulty finding, focusing, and
capturing images of the grey 3D printed models with the bonded brackets. A total of 3 hours was
spent scanning the 30 models with this particular scanner. The source of this problem is unknown.
46

Furthermore, 7.5 hours was spent on the proprietary 3D imaging software to individually select
reference points and to measure the errors between the superimposed images. In the future, a 3D
imaging software that can automatically and more accurately measure the greatest deviations in
each translational direction can be implemented.

VI.b. Other Future Suggestions
- The handle of the 3D printed rigid jig can be redesigned for more stability when positioned
over a tooth intraoral in the clinic.
- For unerupted teeth, perhaps CBCT scans can be utilized to fabricate single-tooth jigs to
be able to accurately incorporate them into lingual appliances without necessitating a new
scan when they fully erupt in order to indirectly bond a bracket.

VII. CONCLUSION
The conclusions to the aims of our pilot study were:
(1) The overall most consistently and significantly accurate rebond method was the flexible
tray. Its accuracy was within the set clinical limits.
(2) Bracket placement onto the molar (LL6) was the most accurate across all rebond
methods, compared to bracket placement onto the premolar or incisor.
(3) The flexible tray method resulted in the overall most accurate bracket placement in
each of the translational planes. Brackets were positioned most accurately in the
mesiodistal translational dimension for all rebond methods.
Despite the meticulous digital setup of teeth, patients are constantly breaking brackets for
all bonded orthodontic appliances. Provided that lingual braces are looked to by both orthodontists
47

and patients alike as the ultimate esthetic appliance capable of providing full 3D control, it is
crucial for failed brackets to be rebonded accurately in order to reach the intended CAD/CAM-
designed outcome. Whether the clinician uses the simplest in-house laboratory setup or the most
sophisticated CAD setup, the results of this study is applicable to all customized lingual appliances.
The results of our study confirmed that the DB method is neither accurate nor recommended, and
the design and/or printing errors for the 3D-printed IDB rigid jig from this study need to be
identified and corrected. Nevertheless, individual variations in tooth anatomy are what give these
investigated IDB methods tremendous potential for improvement in accuracy and can still be
looked forward to as ideal solutions to bracket position problems in the future.

48

VIII. REFERENCES

3M Oral Care (2017). Incognito Step-by-Step Bonding Protocols. Monrovia, CA: 3M Unitek.
Available at: http://multimedia.3m.com/mws/media/1141484O/incognito-bonding
protocols-70-2021-5281-8.pdf [Accessed 22 Dec. 2017].

Aguirre, M. J., King, G. J. & Waldron, J. M. (1982) Assessment of bracket placement and bond
strength when comparing direct bonding to indirect bonding techniques. Am J Orthod,
82(4), 269-76.

Alexander, C. M., Alexander, R. G., Gorman, J. C., Hilgers, J. J., Kurz, C., Scholz, R. P. & Smith,
J. R. (1982) Lingual orthodontics. A status report. J Clin Orthod, 16(4), 255-62.

American Orthodontics. (2012). Harmony Protocol Handbook. [PDF] Sheboygan, WI: American
Orthodontics Corporation. Available at: http://www.americanortho.com/downloads/
brochures/LITSL-52_Harmony_Protocol.pdf [Accessed 22 Dec. 2017].

Birdsall, J., Hunt, N. P., Sabbah, W. & Moseley, H. C. (2012) Accuracy of positioning three types
of self-ligating brackets compared with a conventionally ligating bracket. J Orthod, 39(1),
34-42.

Bittner, W. (2015). Recommended Protocol: Rebonding Fusion Brackets. [PDF] Richardson, TX:
OraMetrix, Inc. Available at: https://www.suresmile.com/wp-content/uploads
/2015/02/DOC-500426-2-Protocol_Rebonding.pdf [Accessed 22 Dec. 2017].

Brown, M. W., Koroluk, L., Ko, C. C., Zhang, K., Chen, M. & Nguyen, T. (2015) Effectiveness
and efficiency of a CAD/CAM orthodontic bracket system. Am J Orthod Dentofacial
Orthop, 148(6), 1067-74.

Buso-Frost, L. & Fillion, D. (2006) An Overall View of the Different Laboratory Procedures Used
in Conjunction with Lingual Orthodontics. Seminars in Orthodontics, 12(3), 203-210.

Castilla, A. E., Crowe, J. J., Moses, J. R., Wang, M., Ferracane, J. L. & Covell, D. A., Jr. (2014)
Measurement and comparison of bracket transfer accuracy of five indirect bonding
techniques. Angle Orthod, 84(4), 607-14.

Chatoo, A. (2013) A view from behind: a history of lingual orthodontics. J Orthod, 40 Suppl 1,
S2-7.

Echarri, P. Revisiting the History of Lingual Orthodontics: A Basis for the Future. Seminars in
Orthodontics, 12(3), 153-159.

Echarri, P. (2006) Revisiting the History of Lingual Orthodontics: A Basis for the Future. Seminars
in Orthodontics, 12(3), 153-159.

49

Fujita, K. (1979) New orthodontic treatment with lingual bracket mushroom arch wire appliance.
American Journal of Orthodontics, 76(6), 657-675.

Fujita, K. (1982) Multilingual-bracket and mushroom arch wire technique. A clinical report. Am J
Orthod, 82(2), 120-40.

Grauer, D. & Proffit, W. R. (2011) Accuracy in tooth positioning with a fully customized lingual
orthodontic appliance. American Journal of Orthodontics and Dentofacial Orthopedics,
140(3), 433-443.

Hiro, T., Iglesia, F. & Andreu, P. (2008) Indirect bonding technique in lingual orthodontics: the
HIRO system. Prog Orthod, 9(2), 34-45.

INBRACE (2017). INBRACE Clinical Guide. [PDF] Irvine, CA: Swift Health Systems. Available
at: http://provider.myinbrace.com/wp-content/uploads/2017/06/SM-001-Rev-G-UserGui
de.pdf [Accessed 28 Dec. 2017].

Joiner, M. (2010) In-house precision bracket placement with the indirect bonding technique. Am J
Orthod Dentofacial Orthop, 137(6), 850-4.

Kothari, J (2016). Digital and customized lingual orthodontics: The next step. The Journal of
Indian Orthodontic Society, 50(5), 33-43.

Kim, J., Chun, Y.-S. & Kim, M. (2018) Accuracy of bracket positions with a CAD/CAM indirect
bonding system in posterior teeth with different cusp heights. American Journal of
Orthodontics and Dentofacial Orthopedics, 153(2), 298-307.

Koo, B. C., Chung, C. H. & Vanarsdall, R. L. (1999) Comparison of the accuracy of bracket
placement between direct and indirect bonding techniques. Am J Orthod Dentofacial
Orthop, 116(3), 346-51.

Larson, B. E., Vaubel, C. J. & Grunheid, T. (2013) Effectiveness of computer-assisted orthodontic
treatment technology to achieve predicted outcomes. Angle Orthod, 83(4), 557-62.

Linklater, R. A. & Gordon, P. H. (2001) An Ex Vivo Study to Investigate Bond Strengths of
Different Tooth Types. Journal of Orthodontics, 28(1), 59-65.

Linklater, R. A. & Gordon, P. H. (2003) Bond failure patterns in vivo. American Journal of
Orthodontics and Dentofacial Orthopedics, 123(5), 534-539.

Lombardo, L., Kaplan, A., Lapenta, R., Bratti, E., Pera, C., Scuzzo, G. & Siciliani, G. (2011) A
comparative study of lingual bracket bond strength. Orthodontics (Chic.), 12(3), 178-87.

Mavreas, D., Cuzin, J. F., Boonen, G. & Vande Vannet, B. (2017) The effect of various adhesives,
enamel etching, and base treatment on the failure frequency of customized lingual brackets:
a randomized clinical trial. Eur J Orthod.
50

Mizrahi, E. (2006) Introduction. Seminars in Orthodontics, 12(3), 151-152.

Mota Júnior, S. L., de Andrade Vitral, J., Schmitberger, C. A., Machado, D. B., Avelar, J. C.,
Fraga, M. R., da Silva Campos, M. J. & Vitral, R. W. F. (2015) Evaluation of the vertical
accuracy of bracket placement with the Boone gauge. American Journal of Orthodontics
and Dentofacial Orthopedics, 148(5), 821-826.

Nichols, D. A., Gardner, G. & Carballeyra, A. D. (2013) Reproducibility of bracket positioning in
the indirect bonding technique. Am J Orthod Dentofacial Orthop, 144(5), 770-6.

Park, J. M., Choi, S. A., Myung, J. Y., Chun, Y. S. & Kim, M. (2016) Impact of Orthodontic
Brackets on the Intraoral Scan Data Accuracy. Biomed Res Int, 2016, 5075182.

Pauls, A., Nienkemper, M., Schwestka-Polly, R. & Wiechmann, D. (2017) Therapeutic accuracy
of the completely customized lingual appliance WIN : A retrospective cohort study. J
Orofac Orthop, 78(1), 52-61.

Pauls, A. H. (2010) Therapeutic accuracy of individualized brackets in lingual orthodontics. J
Orofac Orthop, 71(5), 348-61.

Russell, J. S. (2005) Aesthetic orthodontic brackets. J Orthod, 32(2), 146-63.

Schubert, K., Halbich, T., Jost-Brinkmann, P. G. & Müller-Hartwich, R. (2013) Precision of
indirect bonding of lingual brackets using the Quick Modul System (QMS)®. Journal of
Orofacial Orthopedics / Fortschritte der Kieferorthopädie, 74(1), 6-17.

Shpack, N., Geron, S., Floris, I., Davidovitch, M., Brosh, T. & Vardimon, A. D. (2007) Bracket
Placement in Lingual vs Labial Systems and Direct vs Indirect Bonding. The Angle
Orthodontist, 77(3), 509-517.

Sunna, S. & Rock, W. P. (1998) Clinical performance of orthodontic brackets and adhesive
systems: a randomized clinical trial. British Journal of Orthodontics, 25(4), 283-287.

Wendl, B., Droschl, H. & Muchitsch, P. (2008) Indirect bonding--a new transfer method. Eur J
Orthod, 30(1), 100-7.

Ziebura, T., Hohoff, A., Flieger, S. & Stamm, T. (2014) Accidental debondings: Buccal vs fully
individualized lingual multibracket appliances. Am J Orthod Dentofacial Orthop, 145(5),
649-54.

Abstract (if available)

Abstract Introduction: While lingual orthodontics have evolved and matured to deliver excellent treatment outcomes comparable to labial appliances, one problem has yet to be completely resolved for any bonded appliance. Despite the meticulous digital setup and transfer of lingual appliances, bond failure of individual brackets is inevitable throughout the course of treatment. An average of 2-3 bond failures within the first year of orthodontic treatment can be expected (Ziebura et al, 2014). ❧ Objectives: The overall goal of this investigation was to determine the accuracy of and to compare three different single-tooth rebond methods for a CAD/CAM lingual orthodontic appliance. The three rebond methods to be assessed and compared were direct bond, flexible tray indirect bond, and 3D-printed rigid jig indirect bond. The specific goals of this pilot study were: (1) To quantify and compare the accuracy for rebonding brackets with a flexible tray, 3D printed rigid jig, and direct bonding, (2) to compare the accuracy of the three methods for different tooth morphologies (molar vs. premolar vs. anterior tooth), and (3) to compare the accuracy of each rebond method in each translational dimension. The null hypotheses were that there would be no significant difference in the precision of bracket position for each of the rebond methods

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Bonding accuracy of a novel lingual customized orthodontic appliance (INBRACE™): an in-vivo study

PDF

An evaluation of bond strength using sham lingual brackets with differences in base morphology and preparation

PDF

3D assessment of virtual bracket removal for modern orthodontic retainers: a prospective clinical study

PDF

Shear bond strength comparison of mesh, sandblasted and laser-etched orthodontic brackets

PDF

An assessment of orthognathic surgery outcomes utilizing virtual surgical planning and a patented full-coverage 3D-printed orthognathic splint

PDF

Orthodontic rotational relapse: prevalence and prevention

PDF

Use of digital models to assess orthodontic treatment progress

PDF

Assessment of 3D surface changes following virtual bracket removal

PDF

Root shape frequency and direction of dilaceration: a CBCT study

PDF

Overbite correction with fully customized lingual appliances

PDF

Three-dimensional quantification of post-surgical condylar displacement

PDF

Adhesive performance of hybrid CAD/CAM materials. Chapter I, Influence of surface treatment on the shear bond strength of hybrid CAD/CAM materials. Chapter II, Luting protocol for novel CAD/CAM m...

PDF

Evaluating the treatment efficiency of Carriere Distalizer: a cephalometric and study model comparison of Class II appliances

PDF

Classification of 3D maxillary incisor root shape

PDF

Operator-determined and reoriented natural head position in three-dimensional imaging

PDF

Cirtual 3D placement of temporary orthodontic anchorage implants

PDF

Three-dimensional assessment of tooth root shape and root movement after orthodontic treatment: a retrospective cone-beam computed tomography study

PDF

The creation of a novel full-coverage orthognathic surgical splint utilizing 3D printing & virtual surgical planning

PDF

Three dimensional analysis of maxillary retromolar alveolar bone before and after en‐masse distalization

PDF

Mandibular plane angle changes with or without premolar extraction treatment in adult orthodontics measured using 3-D cone beam technology

Asset Metadata

Creator Song, Tarim S. (author)

Core Title 3D ssessment of bracket position accuracy for lingual appliances using CAD/CAM technology: a pilot study

School School of Dentistry

Degree Master of Science

Degree Program Craniofacial Biology

Publication Date 03/28/2018

Defense Date 02/27/2018

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag 3D,3D jig,bond failure,bracket,bracket rebond,CAD/CAM,INBRACE,indirect,indirect bonding,OAI-PMH Harvest,rebond

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Sameshima, Glenn (committee chair), Grauer, Dan (committee member), Paine, Michael (committee member)

Creator Email tarimson@usc.edu,tarimsong@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-487890

Unique identifier UC11268101

Identifier etd-SongTarimS-6129.pdf (filename),usctheses-c40-487890 (legacy record id)

Legacy Identifier etd-SongTarimS-6129.pdf

Dmrecord 487890

Document Type Thesis

Rights Song, Tarim S.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA