University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

AN ASSESSMENT OF ORTHOGNATHIC SURGERY OUTCOMES UTILIZING
VIRTUAL SURGICAL PLANNING AND A PATENTED FULL-COVERAGE 3D-
PRINTED ORTHOGNATHIC SPLINT

By
Robert Hann

A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
in Partial Fulfillment of the Requirements for the Degree of
MASTER OF SCIENCE (CRANIOFACIAL BIOLOGY)

May 2016
2
TABLE OF CONTENTS

Abstract 4
List of Figures and Tables 7
Chapter 1: Background 8
Traditional Surgical Planning and Surgical Splints 8
Traditional Plaster Model Surgery Protocol 9
Virtual Surgical Planning and 3D Printed Splints 10
3D Imaging 11
Virtual Surgical Planning 12
3D Printed Splints 15
Virtual Model Surgery Protocol 18
CBCT Scan 18
Intraoral Optical Scan 18
Image Data Processing and Software Usage 19
Obstacles to Adoption of the Protocol 20
3D Image Superimpositions 22
Chapter 2: Our Current Study 25
General Overview 25
Treatment Considerations 25

Chapter 3: Research Question 27
Chapter 4: Materials and Methods 28
Sample 28
Pre-surgical preparation 30
VSP 33
Post-surgical image acquisition 35
Superimposition and registration 35
Overlay and color mapping 36

3
Chapter 5: Results 39
Patient One 44
Patient Two 45
Patient Three 46
Patient Four 47
Patient Five 48
Patient Six 48
Patient Seven 49

Chapter 6: Discussion 51
Maxillary trends 51
Mandibular trends 51
Other trends 52
Rami 52
Proximity to Dentition 52
Class II and Class III Corrections 53

Chapter 7: Conclusions 54
Acknowledgements 55
Bibliography 56

4
ABSTRACT

INTRODUCTION
To assess the accuracy of a patented 3D-printed tray-type splint in orthognathic surgery
by comparing postoperative CBCT scans to virtual surgical plans.

METHODS
Seven adult patients underwent pre-surgical orthodontic decompensation. Immediately
prior to surgical workup, orthodontic appliances or attachments adjunct to clear aligner
therapy were removed to ensure a smooth, scannable dentition. The arches were
scanned optically, maxillofacial structures were scanned via CBCT, and a composite
image of these scans was generated. Virtual surgery was performed, and the 3D-printed
tray-type splint was fabricated according to the virtually planned postoperative position.
Orthognathic surgery (LeFort I and BSSO in for all cases) was completed with the splint,
and intermaxillary fixation was obtained by means of miniscrews. A postoperative CBCT
was taken immediately following surgery. The virtual plan and postoperative CBCT were
superimposed by registration of the images at the anterior cranial base. These
superimpositions were analyzed and measured by semitransparent overlay and color
maps of surface positional differences. Postoperative orthodontic treatment was
completed with clear aligner therapy.

5
RESULTS
Qualitative analysis of semi-transparent overlays and color maps for each patient
reveals accuracy of position for each of the seven included patients, with the possible
exception of patient 3. The mean directional differences between VSP and post-
operative CBCT positions for all measured tooth positions were 0.15 ± 0.57mm posterior
in the sagittal plane, 0.22 ± 0.88mm right in the horizontal plane, and 0.11 ± 0.56mm
downward in the vertical plane.

DISCUSSION
The left lateral maxilla tended to be slightly more accurate than the right, and in
cases for which the left lateral maxilla was notably more accurate, the anterior maxilla
was typically of intermediate accuracy. The anterior maxilla tended by be more accurate
than the posterior, with the maxillary tuberosities being of least accuracy. Within these
observed maxillary trends, the overwhelming majority of surfaces were within 2.0mm of
accuracy, and many were within 0.4mm or accuracy.
Mandibular positioning tended to be slightly more accurate overall than
maxillary positioning. And as with the maxilla, the overwhelming majority of surfaces
were within 2.0mm of accuracy, and many were within 0.4mm or accuracy. Also as
observed in the maxilla, a slight increase in left-sided accuracy was noted. Unlike the
maxilla, however, no anteroposterior trends were noted in the mandible.
Other observations include: the rami presenting with more positional variability
than any other segment, maxillary and mandibular areas closer to the dentition tending
6
to be more accurate than those areas farther away, and an apparent absence of trends
for relative differences in accuracy between class II and class III corrections. Ramus
presentation is of particular note because the rami, as bony segments proximal to the
surgical osteotomies, are not under direct positional control of the tray-type splint.

CONCLUSIONS
The combined use of pre-operative CBCT scans, VSPs, and 3D printing to create a
full-coverage orthognathic splint results in predictable, accurate post-operative
positions. A future prospective study could include pre-surgical and post-surgical scans
of equal resolution. Future research could also incorporate software-driven registration
of superimpositions as computerized surface-matching algorithms become more robust.

7
LIST OF FIGURES AND TABLES

Figure 1. Principle of Rapid Prototyping 16
Figure 2. Diagram of SLA 17
Figure 3. Diagram of laser surface digitizer 19
Figure 4. Intra-oral scanning 32
Figure 5. 3D printed splint, occlusal view 35
Figure 6. 3D printed splint, labial view 36
Figure 7. Overlay registration example 37
Figure 8. Color map example 38
Figure 9. Semi-transparent overlay and color map of patient 1 39
Figure 10. Semi-transparent overlay and color map of patient 2 40
Figure 11. Semi-transparent overlay and color map of patient 3 40
Figure 12. Semi-transparent overlay and color map of patient 4 41
Figure 13. Semi-transparent overlay and color map of patient 5 41
Figure 14. Semi-transparent overlay and color map of patient 6 42
Figure 15. Semi-transparent overlay and color map of patient 7 42
Table 1. Differences between VSP and post-operative positions at lower 43
incisor edge
Table 2. Differences between VSP and post-operative positions at all 44
measured tooth positions

8
CHAPTER 1: BACKGROUND

An accurate preoperative plan is the primary requirement for efficacious
orthognathic surgery. Currently, there are two standard planning methods: one,
traditional planning, including model surgery performed on plaster and splinted with an
acrylic wafer; two, virtual planning, including 3D model surgery performed in software
and splinted with a CAD/CAM wafer of various biocompatible materials. The splint
evaluated in this paper is part of a method that modifies and improves on the second
standard, utilizing virtual surgical planning to generate a 3D printed full-coverage splint.
To understand this splint and its accuracy, these standard methods must be reviewed.

TRADITIONAL SURGICAL PLANNING AND SURGICAL SPLINTS

Plaster model surgery is performed in order to generate an acrylic wafer splint
that transfers the surgical treatment plan to the operating room. While plaster model
surgery has been the gold standard for hundreds of thousands of patients and
generations of orthognathic surgeons, the protocol includes several potential sources of
error, and can be time consuming as well. (Sun et al., 2013b) These sources of error
include: distortion of the impression material, distortion of the plaster casts, loss of
precision positioning as the bite and facebow are registered and transferred to the
articulator, incorrect repositioning and/ or measurements of the casts relative to their
planned post-surgical positions, and distortion of the acrylic splint. (Sun et al., 2013b)
9
Plaster models also lack any information concerning the position and
morphology of the surrounding bone tissue. (Xia et al., 2005) This means that model
surgery precludes the surgeon from making cuts that are informed by bony
presentation, a great irony considering the critical role of maxillofacial hard tissues in
orthognathic surgery. (Xia et al., 2005)
To fully understand virtual model surgery and the potential for improved
accuracy that this method includes and which is discussed in this paper, a review of
plaster model surgery is necessary.

TRADITIONAL PLASTER MODEL SURGERY PROTOCOL

In traditional plaster model surgery, the first step is to create plaster models that
duplicate the maxillary and mandibular dental arches accurately and with as much detail
as possible. These arches are then transferred via bite registration and facebow to an
articulator, in which they are oriented to the Frankfort Plane such that the upper
mounting plate equivocates to the patient's inferior orbital margin. (Anwar and Harris,
1990)
Reference points and vertical and horizontal reference lines are then marked on
the plaster that is used to join the casts and articulator mounting plates, allowing for
proper recording of midlines as well as the magnitude and direction of movements
desired during surgery. (Anwar and Harris, 1990; Santoro et al., 2003) After these
references have been marked, lines are drawn to guide the location of planned
10
osteotomies. The surgeon cuts the plaster along these guides and repositions the
resulting model segments to post-operative positions. Within the constraints of
anatomy and patient presentation, these segments are presented with a goal of
maximum intercuspation between the maxillary and mandibular arches with a class I
canine orientation. Once the segments have provided a satisfactory postsurgical
presentation, they are joined in place with firm wax. (Anwar and Harris, 1990; Olszewski
and Reychler, 2004) The acrylic splint is then fabricated from these segmented and
repositioned arches. If both arches are repositioned, an intermediate splint is made to
first align the maxilla to the non-operated mandible along with a final splint that is used
after the maxilla is fixed to reposition the operated mandible to the maxilla. (Mavili et
al., 2007)
In contrast to traditional planning and model surgery as described above, virtual
planning and model surgery require include less opportunity for material distortion and
error. Once the method is studied and practiced, time savings can also be considerable.
But to acquire the necessary technical knowledge, a review of the technology that
makes the method possible is required.

VIRTUAL SURGICAL PLANNING AND 3D PRINTED SPLINTS

Virtual surgical planning is based on manipulation of accurate 3D images. These
images are obtained as follows.

11
3D IMAGING
Beginning approximately 35 years ago, digital images acquired by computed
tomography and magnetic resonance began to be coded in a standardized manner,
courtesy of the efforts of the American College of Radiology and the National Electrical
Manufacturers Association. These efforts resulted in the DICOM, or Digital Imaging and
COmmunications in Medicine, standard. Complete DICOM records consist of a sequence
of coded axial image slices and a DICOMDIR file. The combined axial slices form the 3D
image, and the DICOMDIR file provides information regarding the patient, image
acquisition, and the image itself. (Grauer et al., 2009)
Once a CBCT scan has been performed, various proprietary software packages
are able to convert the raw scan data to a 3D image. CBCT imaging software also allows
for the conversion of these images into DICOM files. Research opportunities remain for
evaluating the conversion process from proprietary formats to DICOM for accuracy, loss
of fidelity, image distortion, etc. (Grauer et al., 2009)
The 3D CBCT is constructed by layering two dimensional image slices, and the
building blocks of 3D images are the voxels rendered from the pixels in the stacked
image slices. Each voxel is assigned a gray-level value determined by the amount of
radiation captured by the imaging machines charge-coupled device (CCD), and voxels
are filtered for visualization based on a threshold value. A user-defined threshold value
is compared to the gray-level value of each voxel. Those voxels with a gray-level value
greater than the threshold are rendered visible. All visible voxels contribute to the
composite 3D image displayed on screen. Viewing software allows the user to review
12
the component image layers in axial, coronal, and sagittal directions. Users may also
turn, change viewing perspective with fluid transitions, and reposition the 3D image.
Because the user determines visualization threshold values as well as image
orientation and perspective, it is important to note that the final rendered image is
highly user-dependent. The user's perception dictates the final image presentation and
can be influenced by image contrast and noise, as well as user bias regarding structural
landmarks and visual skills. (Grauer et al., 2009) Despite this possibility for user
influence, studies have reported the reliability of measurements and landmark
identification in CBCT more accurate when compared to 2D images. (Adams et al., 2004)
3D virtual surface models, called segmentations, are constructed from
volumetric data that can come from either virtual surfaces or rendered images.
Segmentation is a key part of rendering 3D images used in virtual surgical planning for
two reasons: first, it allows for user exportation of 3D models in nonproprietary formats;
second, it provides a way to import those models to imaging software. These factors
allow for the multimodal combination of digital models of a patient's dentition obtained
from optical scanning with digital models of a patient's hard and soft tissue obtained
from CBCT data. The resulting composite images are the basis of virtual dentistry,
including the virtual surgical planning described below. (Grauer et al., 2009)

VIRTUAL SURGICAL PLANNING
Virtual surgical planning (VSP), also known as computer-aided surgical simulation
(CASS), is a relatively new process for transferring the pre-surgical workup of a case to
13
the operating room. (Aboul-Hosn Centenero and Hernandez-Alfaro, 2012; Xia et al.,
2005) VSP, when combined with CAD/CAM technology to fabricate wafer-type surgical
splints, allows surgeons to avoid plaster model surgery during planning. (Aboul-Hosn
Centenero and Hernandez-Alfaro, 2012) Virtual surgery gives surgeons a way to
generate and view virtual outcomes that include visualizations of both dental and bony
anatomy.
The process reduces surgical planning time, decreases the amount of manual
labor required, and can also reduce time spent in the operating room. (Nadjmi et al.,
2010; Xia et al., 2005) Plans and outcomes can also be shared worldwide with
colleagues, collaborators, and researchers. (Aboul-Hosn Centenero and Hernandez-
Alfaro, 2012) Perhaps most importantly, post-surgical outcomes incorporating VSP and
CAD/CAM wafer splints appear to be more accurate than outcomes including traditional
methodologies. (Farronato et al., 2015; Olszewski et al., 2010; Xia et al., 2005)
When comparing traditionally fabricated acrylic wafer splints and CAD/CAM
fabricated biocompatible wafer splints, the high degree of similarity has established the
VSP and CAD/CAM method as a valid, reliable surgical protocol. (Aboul-Hosn Centenero
and Hernandez-Alfaro, 2012) Additionally, some planning procedures, including 3D
model creation and digitizing landmarks, can be carried out by a trained technician
instead of a surgeon. (Gateno et al., 2007) This reduces the amount of preparatory time
for surgeons, and, if trained technicians are in a centralized location, introduces
efficiency to the field with one VSP center serving multiple surgeons. (Gateno et al.,
2007)
14
Prior to the advent of intraoral optical scanners, reliable VSP was not as viable
because the dental portion of the CBCT scan could not provide accurate occlusal detail.
Initially, this problem was addressed with a double CBCT procedure in which separate
scans were made of the patient and of plaster casts of the dentition. While representing
an improvement, the process introduced streak artifacts and resulted in lower
resolution that was exacerbated upon later determination of intermaxillary occlusion.
(Swennen et al., 2007) The current method of optically scanning the patient's dentition
and superimposing this image of anatomy and occlusion onto the CBCT for virtual
surgery incorporates highly accurate digital detail as part of the CAD data and decreases
the overall amount of radiation exposure necessary. (Swennen et al., 2007)
VSP is gradually becoming more common, and is beginning to replace the
traditional protocol. (Xia et al., 2006) VSP addresses one of the greatest drawbacks of
traditional planning, namely the absence of any representation of boney structures. (Xia
et al., 2006) VSP preparation procedures, such as scanning, uploading, image
verification, and data entry, can be carried out by a trained technician. Diagnoses can be
shared with patients, providing a useful tool for explanation and interpretation
analogous to demonstration of physical models. (Aboul-Hosn Centenero and
Hernandez-Alfaro, 2012) Because the plan is digital, preparatory steps can also be
completed in one location with the surgeon generating the VSP remotely. (Xia et al.,
2006) Other surgeons, orthodontists and specialists can also remotely contribute
expertise as needed. (Aboul-Hosn Centenero and Hernandez-Alfaro, 2012)
15
The digital basis of the process provides another valuable benefit: a complete
absence of physical materials that can introduce error and distortion, however minimal,
and that depend on technique and skill which may vary between professionals. 3D
printed splints, described presently, are fabricated from information based solely on the
desired post-surgical position that is digitally determined by the surgeon in virtual
planning.

3D PRINTED SPLINTS
Once the post-surgical position has been determined in virtual planning, the 3D
splint is fabricated. Several methods of CAD/CAM rapid prototyping exist, but all can be
categorized as either subtractive or additive technologies and both categories require
an initial 3D virtual model usually obtained form CBCT or MRI scans.
Subtractive methods begin with a block of material, and much like sculpture,
remove areas of that material to reveal the desired shape within the block. There are
two primary limitations to the subtractive method as it pertains to virtual dentistry.
First, the machines that mill the block of material down to the desired 3D object have
restricted motion capabilities that can inhibit the creation of complex geometries
(Potamianos et al., 1998) Second, the material that is milled must be hard enough to
resist distortion of the desired final shape as the subtracted material is removed. (Klein
et al., 1992)
Additive methods do not have these limitations. They can provide more
anatomical detail and internal features inclusive of complex geometries, and to a much
16
greater degree, they do not rely on material hardness. Common additive technologies
(from this point on referred to as 3D printing technologies) include selective laser
sintering, fused deposition modeling, inkjet-based systems, and stereolithography (SLA).
(Liu et al., 2006)
The patented splint used in this study was fabricated via SLA. SLA is a kind of
rapid prototyping based on layered manufacturing, in which a complex 3D printing
problem is changed via computer algorithm to a simpler 2D layer-building problem. The
virtual 3D shape is converted to a stack of discrete layers, each of which is deposited
sequentially to form the desired 3D print model, often with supporting attachments
printed to provide structural integrity to overhanging or undercut features while the
print is in process. See Figure 1 for a diagram of the principles of discretization and
sequential stacking. (Hungate, 2015)

Figure 1. Principle of Rapid Prototyping. (Liu et al., 2006)
17
The SLA technology of Formlabs' Form 1+ hardware used in this study prints full-
coverage splints one photopolymerized layer at a time by tracing a laser on the bottom
surface of a vat of unpolymerized liquid photopolymer as in Figure 2. The first layer self-
adheres to a build platform suspended in the vat, and subsequent layers cure to the
layer preceding them until splint fabrication is complete. Once fabrication is complete,
the build platform is raised out of the liquid vat, the splint is removed from the
platform, and any supports are trimmed from the splint. Processing steps after removal
from the vat include removal of residual uncured photopolymer and a brief heat cure.
(Hungate, 2015) Advantages of this process include accuracy, clinically acceptable
smoothness of finished surfaces, and choice of material properties. Disadvantages
include initial equipment costs, liquid materials storage and handling, and processing
prints after photopolymerization. (Cooper, 2001; Hilton, 2000; Jacobs, 1992)

Figure 2. Diagram of SLA (Liu et al., 2006)
18
The full-coverage splint thus fabricated is the end product of the virtual model
surgery protocol described in the following paragraphs, a notable contrast to the
traditional plaster model surgery protocol previously reviewed.

VIRTUAL MODEL SURGERY PROTOCOL

CBCT SCAN
Virtual model surgery is what makes the tray-type orthognathic splint possible. It
begins with a CBCT scan of the patient's maxillofacial tissues and an intraoral optical
scan of the patient's maxillary and mandibular dentition and intermaxillary occlusion.
CBCT scan resolution is directly proportional to total scanning time, partly because it
involves external and internal volumetric data. Though desirable resolution must be
balanced against total patient exposure to radiation, equipment and staff time, and
patient comfort, it is worth noting that every generation of CBCT scanning technology
has resulted in smaller slice intervals as well as faster scans. (Liu et al., 2006)

INTRAORAL OPTICAL SCAN
Intraoral optical scan resolution is dependent not on total scanning time but on
laser and CCD fidelity. The image is produced by shining a planar beam of diode-based
laser light on the surface of the teeth. The surface profile thus illuminated is recorded by
a CCD, which sends this data to a processor that computes coordinates for the surface in
3D space (See Figure 3 for a diagram of a laser surface digitizer). A composite model
19
emerges from a series of these coordinated surface profiles. The patient is not irradiated
and, because the data captured by optical scan concerns only external surfaces and not
internal volumes, the resultant data file is relative small when compared to a CBCT scan.
This can reduce computer processing time. (Hungate, 2015)

Figure 3. Diagram of laser surface digitizer.

IMAGE DATA PROCESSING AND SOFTWARE USAGE
Once the scan data is obtained, it must be processed to generate a composite 3D
image. The output image data is converted to a format used by virtual surgery planning
software, which is the orthognathic surgery analog of manufacturing software used in
industrial processes. These CAD software suites allow for editing and modification. (Kai,
20
1994; Kai et al., 1998) Within the context of this paper, that editing translates to the
surgeon performing virtual surgery.
The virtual surgery software not only allows for simulation of osteotomies and
segment repositioning but also for the determination of the degree of occlusion. The
desired occlusion is achieved virtually with a rigid motion simulation module that
prevents overlap of the virtual models, and with a virtual grinding function that can
eliminate undesired interferences. Any virtual removal of interferences, just like all
planned osteotomies and segment repositioning, must be replicated in the operating
room for the 3D printed splint to fit properly. Like most digital applications, the software
does require a time investment in order to learn how to use each function well. (Nadjmi
et al., 2010)
The 3D splint is created by introducing a virtual solid, shaped by the surgeon, to
the planned post-surgical occlusion. The software treats this object as a virtual bite
registration, recording incisal and cuspal positions of the planned intercuspated bite.
This registration is printed in 3D as described above and serves as the surgical splint.

OBSTACLES TO ADOPTION OF THE PROTOCOL
Two obstacles to greater adoption of VSP and 3D splint fabrication have included
time and money. (Mavili et al., 2007) Given how rapidly the technology involved in both
3D imaging and 3D printing has pervaded many industries, software has become much
more user friendly and hardware has become much cheaper, to the degree that these
obstacles may be based more on perception than reality, however. VSP has been found
21
to be less costly than traditional planning for complex cranio-maxillofacial surgery, with
respective average costs totaling $1,900 and $3,510. (Xia et al., 2006) Not only can it be
less costly, it can also be less time consuming, with an average of 5.25 hours for VSP
cases compared to 9.75 hours for traditional planning. (Xia et al., 2006)
It must be stated that this protocol does not include physical models at any step,
and there may be no current substitute for physical models for a number of reasons.
One reason is that representation of a 3D model on a 2D screen gives surgeons
geometric information that must be interpreted with practice to produce a clear mental
map of the patient's anatomy. Although recent technological developments to provide a
greater sense of virtual depth include head-mounted displays, stereoscopic eyewear,
and holographic projections, there is not yet a way to simulate the physical feel of the
3D image. (Chelule et al., 2000) Another reason is that physical models provide a
measure of comfort for surgeons. This is partly because handling casts has been an
integral part of surgical training and practice for decades, and partly because physical
model surgery involves surgical instruments with which the surgeon has vast
experience. A third reason is that physical models can be a vital part of a surgeon's
communication with patients, serving as a centerpiece of constructive dialogue that
ensures patient education via a tangible 'prop'. (Petzold et al., 1999) However this
reason must be tempered with what was noted previously concerning 3D models as
useful tools for explanation and interpretation analogous to demonstration of physical
models. (Aboul-Hosn Centenero and Hernandez-Alfaro, 2012)

22
3D IMAGE SUPERIMPOSITIONS

Reliable image superimposition is necessary for interpretation of serial 3D
images. The traditional quantitative assessment of changes over time in orthodontic
treatment is superimposition of lateral cephalogram. These cephalograms are traced,
and orientation landmarks are recorded. Stable structural landmarks are used to
register the superimposition, and changes are assessed relative to these structures of
reference. This registration allows for two 2D images with unique coordinate systems to
be combined according to a shared coordinate system. (Grauer et al., 2009) The
analogous three-dimensional quantitative assessment of change involves registered
superimposition of CBCT records, providing a volumetric tool for analysis of changes due
to growth, orthodontic treatment, and of particular relevance to this study,
orthognathic surgery.
One group of researchers found that 3D cepahlometrics can yield more precise
landmark identification than 2D cephalometric analysis. (Adams et al., 2004) Though
research was not found on this topic, the reported increase in precision suggests a
resultant increase in precision of 3D cephalometric superimpositions, because the
superimpositions are predicated in part on the landmarks identified.
VSP models and post-surgical CBCTs can be superimposed by registration at the
surface of the anterior cranial base via automated surface matching. Anterior cranial
base is often selected for surface registration of superimpositions because this area is
not repositioned during surgery and can serve as a reliable, reproducible reference.
23
(Cevidanes et al., 2005; Cevidanes et al., 2007; Cevidanes et al., 2011; de Paula et al.,
2013)
There are several methods of registration. One study evaluating computer-
generated guides in bilateral sagittal split osteotomies registered images at the condyles
and recorded separate 2D image slices in each of three dimensions (anteroposterior,
mediolateral and superoinferior). These images were then used to make separate
assessments of changes in each direction. (Abdel-Moniem Barakat et al., 2014)
A series of illustrative comparison studies at the University of North Carolina
include software developed by the university. The software masks portions of the
patient's bony anatomy that are affected by growth or surgical change or orthodontic
appliance therapy, and registers 3D images to one another at unchanged surfaces by a
digitally automated process. The automation is a voxelwise, rigid registration, meaning
both superimposed images are of the same resolution and of the same coordinate
system. The automated superimposition is then verified or improved by the software
user, and quality is supported by the independent verification of an additional user.
(Cevidanes et al., 2005; Cevidanes et al., 2007; Cevidanes et al., 2009; Cevidanes et al.,
2011; Cevidanes et al., 2010; de Paula et al., 2013; Heymann et al., 2010; Hino et al.,
2013; Nguyen et al., 2011)
A study of VSP use with an intraoral marker for positioning, referred to as the
Charlotte method, included a process of registration similar to that described in the
'Materials and Methods' section of this paper, but with different reference surfaces.
Instead of using anterior cranial base, post-operative anatomy was superimposed on the
24
VSP according to best fit at different surfaces depending on the alignment. For maxillary
alignment, the aligning surface used was the midface; for mandibular alignment, the
aligning surface used included the condylar head. (Bobek et al., 2015)

25
CHAPTER 2: OUR CURRENT STUDY

GENERAL OVERVIEW

The purpose of the current study was to evaluate the accuracy of a recently
patented 3D printed splint for orthognathic surgery. Several studies have evaluated the
accuracy of CAD/CAM wafer-type surgical splints, our study evaluates the accuracy of a
new, full-coverage, tray-type 3D printed splint. This splint holds promise as a surgical
tool that includes not just incisal and cuspal positioning information, but tip and torque
information as well. And because occlusal contacts are not obstructed by the splint,
much of the patient's post-surgical occlusion is what provides post-surgical stability.
Multi-segment maxillary treatment modalities experience greater segment control as
well. Also, because it is an integral part of the surgeon's virtual surgical plan, it
eliminates several physical laboratory steps that have the potential to induce error.

TREATMENT CONSIDERATIONS

Unique to the usage of the splint evaluated in this paper is that prior to surgery,
all fixed appliances or clear aligner therapy attachments must be removed. Because the
splint includes crown coverage, the dentition must present with smooth, easily scanned
crown surfaces. Additionally, patients typically experience post-surgical finishing
orthodontics by means of clear aligner therapy. Accordingly, patients must be educated
26
regarding the timeline of their overall treatment, the need for a unique pre-operative
orthodontic appointment in addition to the pre-operative orthognathic appointment, as
well as the costs, benefits, risks, advantages, and compliance requirements inherent in
clear aligner therapy.

27
CHAPTER 3: RESEARCH QUESTION

For patients with skeletal discrepancies that require orthodontic-orthognathic
correction, does the combined use of Virtual Surgical Planning (VSP) and a recently
patented, full-coverage 3D-printed splint yield a predictable, accurate post-surgical
position?

28
CHAPTER 4: MATERIALS & METHODS

SAMPLE

The study sample comprises the first seven patients to be treated utilizing the
full-coverage splint. The patients include four adult women and three adult men. Of the
seven, four presented with class III discrepancies, and three presented with class II
discrepancies. All patients experienced two-jaw surgery involving LeFort I maxillary
osteotomy and bilateral sagittal split osteotomy (BSSO) of the mandibe. One LeFort I
osteotomy involved no segments, two involved two segments, and the remaining four
involved three segments.
Patient one was an adult Caucasian female diagnosed with excess lower face
height, concave profile, horizontal and transverse maxillary deficiency, horizontal
mandibular excess, mandibular asymmetry, upper occlusal plane cant, and class III
dental relationship. She was treated with LeFort I osteotomy in three segments to
advance, widen, and correct midline in the maxilla and level the occlusal plane, and with
BSSO to setback, correct midline, and correct yaw in the mandible.
Patient two was an adult Caucasian female diagnosed with dolichofacial pattern,
excess lower face height, excess convexity, labial insufficiency, vertical maxillary
hyperplasia, horizontal mandibular hypoplasia, mandibular asymmetry, short ascending
mandibular ramus, high angle class II malocclusion, anterior open bite, reverse lower
Curve of Spee, right temporomandibular (TMJ) articular disc disorder. She was treated
29
with LeFort I osteotomy in three segments to impact the maxilla and flatten the occlusal
plane, and with BSSO to advance and correct asymmetry in the mandible.
Patient three was an adult Caucasian female diagnosed with excess lower face
height, labial insufficiency, labial strain, anterior vertical maxillary hyperplasia,
horizontal maxillary hypoplasia, horizontal mandibular hypoplasia, and class II occlusion.
She was treated with LeFort osteotomy in three segments to advance and widen the
maxilla and to flatten the sagittal plane, and with BSSO to advance the mandible.
Patient four was an adult Caucasian male diagnosed with mesiofacial pattern,
horizontal maxillary hypoplasia, horizontal mandibular hyperplasia and asymmetry, class
III malocclusion, and anterior open bite. He was treated with LeFort I osteotomy in three
segments to advance and down-graft the posterior maxilla, and with BSSO to set back
the mandible.
Patient five was an adult Asian male diagnosed with horizontal maxillary
hypoplasia, mild vertical maxillary hyperplasia, short upper lip, horizontal mandibular
hyperplasia and asymmetry, nasal deviation, and class III malocclusion. He was treated
with LeFort I osteotomy without segmentation to advance and impact the maxilla, and
with BSSO to setback and correct asymmetry in the mandible.
Patient six was an adult Hispanic male diagnosed with horizontal and transverse
maxillary hypoplasia, horizontal mandibular hyperplasia, class III malocclusion and
anterior open bite. He was treated with LeFort I osteotomy in two segments to advance
and widen the maxilla and flatten the plane of occlusion, and with BSSO to set back the
mandible.
30
Patient seven was an adult Caucasian female diagnosed with excess lower face
height, labial insufficiency, high smile line, convex profile, vertical maxillary excess,
mandibular deficiency, mandibular occlustal plane cant, lower dental midline
discrepancy, and high angle class II dental relationship. She was treated with LeFort I
osteotomy in three segments to impact and advance the maxilla, and with BSSO to
advance the mandible.

PRE-SURGICAL PREPARATION

Prior to surgery, patients were orthodontically decompensated. For every
patient in this study, traditional appliances were used. However, clear aligner therapy is
a viable option as well, particularly because aligners are an integral part of the
treatment modality (see below). (Hungate, 2015)
When the necessary tooth movements had been achieved and clinical
approximation of the corrected jaw relationship was possible, progress records were
taken and study casts were made. The interocclusal relationship of the post-surgical
dentition was analyzed for occlusal function, stability and esthetics in the
anteroposterior, transverse, and vertical dimensions. If this analysis showed that
additional orthodontic corrections were required, then fixed appliance therapy
continued and additional study casts were made. (Hungate, 2015)
Once study cast analysis showed that all desired pre-surgical orthodontic
correction was adequate and immediately prior to surgical workup, fixed appliances
31
were removed. If clear aligners therapy was utilized and included attachments, then all
attachments would have been removed. Lingual retainers were then bonded to the
anterior dentition, from canine to canine in the mandible, and to and from the teeth
mesial to the planned osteotomy cuts in the maxilla. (Hungate, 2015)
The patient's dentition was scanned with an iTero 3D scanner. See Figure 4.
These scans generated virtual 3D models of the patient's dental arches that were
imported into Medical modeling VSP software. In the software, the virtual models were
segmented according to the osteotomies of the surgical plan. The virtual arch segments
were oriented to planned post-surgical positions that included correct tip, torque and
rotation. These virtually repositions models (VRMs) were occluded with one another in
virtual space. Any premature contacts were identified by and equilibrated in the
software, with corresponding notes for equilibration added to the surgical plan. The
following separate 3D files were then exported to Medical Modeling VSP software:
VRMs in occlusion, maxillary arch alone, and mandibular arch alone. (Hungate, 2015)

32

Figure 4. Intra-oral scanning. Optical scanning of a patient's dentition using an iTero 3D
scanner. (Hungate, 2015)

Cone-beam computed tomography (CBCT) scans of the patients craniofacial
tissues were obtained from an iCAT FLX machine. Pre-surgical scans were performed
with a total patient exposure time of 17.9 seconds and saved as digital imaging and
communications in medicine files.
VRMs and CBCTs were combined in software, and the surgeon created a virtual
surgical plan utilizing the composite skeletal and dental 3D image in the software.
Skeletal corrections were made according to hard and soft tissue presentations and
desired results. The VRM-determined interocclusal relationship remained intact by
registering the maxillary model to the same coordinate system as the patient's
segmented skull. The interocclusal relationship is maintained through any positional
changes in the maxilla relative to the skull. (Hungate, 2015)
33
VSP

An online video conference for each patient was hosted by Medical Modeling,
Inc., whose VSP software was interfaced with by a trained software technician, at the
direction of the maxillofacial surgeon and with input from the orthodontist. The
conference is held with the technician, the surgeon, and the orthodontist because VSP
software simulates surgical repositioning but it requires a level of software expertise
currently beyond the training of most dental professionals.
Once the desired post-surgical position was finalized, the arches were exported
to Orchestrate software as one merged file. In the software, a virtual solid block larger
than the merged arches was made to occupy the volume of the merged file plus an
additional surrounding space at least 5mm in depth. The merged arches were then
volumetrically subtracted from the virtual solid, yielding a 3D negative space within the
virtual solid of the exact size and morphology as the dental arches. This negative space
was then 'shelled' from the virtual solid. That is, all but a veneer of the virtual solid was
discarded, leaving a shell around the arch-defined space with a depth of 1.0mm at
buccal surfaces and 0.1mm at lingual surfaces, and with exposed occlusal contacts. The
resulting 3D object was exported to Formlabs PreForm software and 3D printed via
photopolymerization of Formlabs Resin 1, a composite photopolymer. This 3D print
served as the surgical splint. The splint thus created provided requisite clinical rigidity,
ensured that occlusal contact was the final determinant of interocclusal positioning, and
34
allowed for easy removal. (Hungate, 2015) Figures 5 and 6 provide views of a print
created using this method.

Figure 5. 3D printed splint, occlusal view. (Hungate, 2015)

35

Figure 6. 3D printed splint, labial view. (Hungate, 2015)

POST-SURGICAL IMAGE ACQUISITION

Immediately following surgery, and per ALARA (As Low As Reasonably
Achievable) principles, to minimize patients' exposure to radiation because higher
resolution is not required for post-operative purposes, post-surgical CBCT scans were
performed with a total patient exposure time of 8.9 seconds and saved as DICOM files.
Post-surgical care included clear aligner therapy to detail and finish orthodontic
treatment, beginning as soon as adequate opening and tray tolerance were achieved.

SUPERIMPOSITION AND REGISTRATION

First, skull base anatomy was digitally segmented from the post-operative CBCT
scan by a software engineer at Medical Modeling (E. Koenig). This anatomical segment
36
was saved in .stl format. The .stl information was then imported to the patient's pre-
operative CBCT scan image. Next, post-operative anterior cranial base contour was
aligned by the software engineer according to best fit with the pre-operative cranial
base, using Mimics software, version 13.1.0.70. A second software engineer then
independently reviewed the alignment to ensure best fit.

OVERLAY AND COLOR MAPPING

Once the alignment was established, color overlays were generated in Simplant
software, version 13.0.0.66. Figure 4 shows an example of the end result of this part of
the method. In the figure, the post-operative skull base segment is color-coded in red
and the pre-operative CBCT scan image is green. Theoretically, the red and green
surfaces should match perfectly, as the skull base was not altered during surgery and
patients experienced no growth between pre- and post-operative scans. However, as
mentioned earlier, scan resolutions are not identical as ALARA principles dictated
reduced post-operative exposure time.

37

Figure 7. Overlay registration example. Post-operative skull base (red) registered to pre-
operative CBCT (green) according to best fit at anterior cranial base surface contour.

Lastly, numerical differences between planned and post-operative anatomical
positions were generated using Dolpin Imaging software, version 11.7. The software was
then used to plot these numerical differences on a virtual image composite of pre- and
post-surgical anatomy with a color-coded gradient to elucidate various areas with
differing degrees of absolute numerical difference. Figure 5 provides an example of the
kind of image created using this method.

38

Figure 8. Color map example. Composite image of pre- and post-operative bony
anatomy, with color-coded gradient showing absolute positional differences at surface
contours, from 0.0 to 2.0mm.

39
CHAPTER 5: RESULTS

The semi-transparent overlays and color maps for each of the seven patients
included in this study are presented in Figures 9-15.

Figure 9. Semi-transparent overlay and color map of patient 1. Overlay: red areas
represent post-operative anatomy, gray areas are VSP surfaces; color map: Greener
areas represent areas in which the two images are most similar, ranging through purple
for areas differing by up to 2.0mm

40

Figure 10. Semi-transparent overlay and color map of patient 2. Overlay: red areas
represent post-operative anatomy, gray areas are VSP surfaces; color map: Greener
areas represent areas in which the two images are most similar, ranging through purple
for areas differing by up to 2.0mm

Figure 11. Semi-transparent overlay and color map of patient 3. Overlay: red areas
represent post-operative anatomy, gray areas are VSP surfaces; color map: Greener
areas represent areas in which the two images are most similar, ranging through purple
for areas differing by up to 2.0mm
41

Figure 12. Semi-transparent overlay and color map of patient 4. Overlay: red areas
represent post-operative anatomy, gray areas are VSP surfaces; color map: Greener
areas represent areas in which the two images are most similar, ranging through purple
for areas differing by up to 2.0mm

Figure 13. Semi-transparent overlay and color map of patient 5. Overlay: red areas
represent post-operative anatomy, gray areas are VSP surfaces; color map: Greener
areas represent areas in which the two images are most similar, ranging through purple
for areas differing by up to 2.0mm

42

Figure 14. Semi-transparent overlay and color map of patient 6. Overlay: red areas
represent post-operative anatomy, gray areas are VSP surfaces; color map: Greener
areas represent areas in which the two images are most similar, ranging through purple
for areas differing by up to 2.0mm

Figure 14. Semi-transparent overlay and color map of patient 7. Overlay: red areas
represent post-operative anatomy, gray areas are VSP surfaces; color map: Greener
areas represent areas in which the two images are most similar, ranging through purple
for areas differing by up to 2.0mm
43
Ten total tooth positions were measured. These include: upper central incisor tip
midpoint, lower central incisor tip midpoint, upper right and left canine tips, lower right
and left canine tips, upper right and left first molar mesiobuccal cusp tip, and lower right
and left first molar mesiobuccal cusp tip. For example, the mean difference between
VSP positions and post-operative CBCT positions at the lower incisor tip midpoint was
0.66 ± 0.49mm in the sagittal dimension, 0.70 ± 0.63 in the horizontal dimension, and
0.42 ± 0.29 in the vertical dimension. See Table 1 for absolute differences for each
patient.

Patient Sagittal Horizontal Vertical
1 0.03 0.36 0.72
2 1.24 0.83 0.4
3 0.22 1.53 0.06
4 0.49 0.5 0.27
5 1.07 1.53 0.22
6 1.17 0.07 0.41
7 0.37 0.07 0.89
Table 1. Differences between VSP and post-operative positions at lower incisor edge. All
measurements in mm.

The mean difference between VSP and post-operative CBCT positions for all
measured positions was 0.15 ± 0.57mm posterior, 0.22 ± 0.88mm right, and 0.11 ±
0.56mm downward in the sagittal, horizontal, and vertical dimensions, respectively. See
Table 2 for the differences for each patient.

44
Patient Sagittal Horizontal Vertical
1 0.097 0.171 0.746
2 0.859 0.046 0.377
3 0.406 1.663 1.039
4 0.516 0.782 0.032
5 0.718 1.195 0.257
6 0.481 0.399 0.191
7 0.407 0.004 0.127
Table 2. Differences between VSP and post-operative positions at all measured tooth
positions. All measurements in mm.

The greatest advantage of 3D comparisons is in relation to 2D comparisons is the
ability to look at surfaces and areas as opposed to points. For example, a qualitative
analysis of the area of the lateral maxilla may give a better indication of accurate
maxillary segment repositioning than pinpoint landmarks such as A-point. Much of the
following case summary content employs this advantage.

PATIENT ONE

Recall that patient one was a class III case treated with LeFort I osteotomy in
three segments to advance, widen, and correct midline in the maxilla and level the
occlusal plane, and with BSSO to set back, correct midline, and correct yaw in the
mandible.
Patient one is one of the three most accurate cases overall. The color map
displays a preponderance of surfaces as accurate to within 0.4mm, i.e. green,
throughout the lateral and anterior maxilla. Several red pinpoints, indicating areas of up
45
to 1.6mm positional difference, are present at the anterior surfaces. However, given
their isolated state, these points may be artifacts of the resolution differences between
the pre- and post-scans (as described in Materials and Methods, above). The overall
impression is profoundly green, with a tendency toward yellow, or differences up to
0.8mm, beginning over the root surfaces of the first molars and continuing to maxillary
tuberosities.
The body of the mandible is almost entirely accurate to within 0.8mm. The
exception is a small, distinct red area at the anterior portion of the right inferior border.
The ramus, by contrast, is noteworthy for increasing inaccuracy from the anterior to
posterior borders. As these proximal segments were not altered in surgery, it may be
that patient positioning, patient movement, or patient muscular posturing during the
post-surgical scan may have influenced the image.

PATIENT TWO

Patient two was a class II case treated with LeFort I osteotomy in three segments
to impact the maxilla and flatten the occlusal plane, and with BSSO to advance and
correct asymmetry in the mandible.
Maxillary surfaces are generally accurate to within 0.8mm, with the exception of
the right lateral maxilla. Here, beginning approximately 3mm superior to interdental
bone levels, accuracy gradually decreases approaching the osteotomy to up to 1.6mm.
This right-sided decrease in accuracy is mirrored in the mandible.
46
Left mandible body positioning is accurate to within 0.4mm beginning from just
anterior to just distal to pogonion. From this area through the right osteomoty, accuracy
reduces to 1.6mm. An area approximating and inferior to the oblique line of the right
mandible approaches positional differences of up to 2.0mm. As with patient one, the
rami, proximal to the BSSO cuts and therefore not under positional influence of the
splint, present with varying degrees of accuracy.

PATIENT THREE

Patient three was a class II case treated with LeFort osteotomy in three
segments to advance and widen the maxilla and to flatten the sagittal plane, and with
BSSO to advance the mandible.
Accuracy in the maxillary and mandibular surfaces is less accurate for this patient
than any other. This is due to the fact that final surgical positions were altered in the
operating theater. Bony presentation and effects on soft tissue, when visualized
clinically, dictated alteration to the surgical plan. This alteration included left yaw
relative to the VSP, and this is reflected in the maxilla and mandible. In the maxilla,
positional accuracy ranges up to 2.0mm on the right and goes beyond the scale of the
color map on the left. The mandible also has a left-to-right skew ranging from very
accurate to inaccuracy outside the scale of the color map.

47
PATIENT FOUR

Patient four was a class III case treated with LeFort I osteotomy in three
segments to advance and down-graft the posterior maxilla, and with BSSO to set back
the mandible.
Patient four is the second of the three most accurate cases overall. The left
lateral maxilla is generally accurate to within 0.4mm, with a tendency toward
differences up to 1.2mm, and small areas of up to 1.6mm, closer to the maxillary
tuberosity. A majority of the anterior maxilla is accurate to within 0.8mm, with notable
portions accurate to 1.6mm. The right lateral maxilla is highly accurate, i.e. within
0.4mm, at the superior-most portion of the junction with the anterior segment,
decreasing in accuracy steadily in an inferior-posterior direction to areas of up to 2.0mm
or more of discrepancy. Notably, the right maxillary dentition appears to range in
discrepancy, anterior to posterior, from 0.4mm to more than 2.0mm. This range may be
due in part to a change in the patient's surgical timeline, which prevented an ideal pre-
surgical scan that would have included removal of all clear aligner therapy attachments
and may have resulted in a 3D printed splint with a less than ideal fit.
At mandible body surfaces, positional accuracy is remarkably uniform to within
1.2mm over nearly all surfaces distal to the BSSO cuts. The exception is an area accurate
to within 1.6mm at, and to the right of, the mental protuberance. Proximal mandibular
segments include areas with a similar presentation to patient two.

48
PATIENT FIVE

Patient five was a class III case treated with LeFort I osteotomy without
segmentation to advance and impact the maxilla, and with BSSO to setback and correct
asymmetry in the mandible.
A majority of the maxilla is accurate to within 0.8mm, with superior portions of
the right lateral maxilla and posterior portions of the left lateral maxilla approaching
accuracy to within 2.0mm.
The body of the mandible is also largely accurate to within 0.8mm. A highly
notable exception to this accuracy concerns the patient's mental protuberance. The
patient's surgery also included genioplasty. As chin positioning is not under direct
positional influence of the 3D splint, this area is not considered in this discussion.

PATIENT SIX

Patient six was a class III case treated with LeFort I osteotomy in two segments
to advance and widen the maxilla and flatten the plane of occlusion, and with BSSO to
set back the mandible.
The left lateral maxilla is almost uniformly accurate to within 0.8mm, with a
majority of the surface accurate to within 0.4mm. The anterior maxilla is accurate to
within 1.6mm, with most surfaces accurate to either 1.2mm or 0.8mm. Similar to
patient four, the right lateral maxilla is highly accurate, i.e. within 0.4mm, in the
49
superior-most portion of the junction with the anterior segment, decreasing in accuracy
steadily in an inferior-posterior direction to areas of up to 2.0mm of discrepancy.
The body of the mandible shows positional accuracy within 0.8mm beginning at
the left osteotomy and continuing past the chin midpoint. The right mandibular body,
however, is accurate to 1.6mm, with an area approximating the oblique line and
traversing the right osteotomy accurate to 2.0mm or more. Proximal mandibular
segments include areas with a similar presentation to patient two.
This case is also notable for the least amount of green surface area at skull base,
the area used to register the 3D superimposition. This may be because of patient
movement during the post-surgical CBCT, and may have influenced imaging and
positional comparisons at the distal segments.

PATIENT SEVEN

Patient seven was a class II case treated with LeFort I osteotomy in three
segments to impact and advance the maxilla, and with BSSO to advance the mandible.
Patient seven is the third of the three most accurate cases overall. The maxilla
presents across all segments as a relatively even gradient of accuracy, with
discrepancies closest to the dentition of 0.4mm or less and those nearer the LeFort
osteotomy approaching up to 1.6mm.
The mandible presents as almost uniformly accurate to within 0.8mm, with a
majority of surface areas accurate to within 0.4mm or less. The exception is a small,
50
distinct area at the anterior portion of the left inferior border. This area is accurate to
within 2.0mm, and is part of the structure included in genioplastic movement not under
direct positional influence of the 3D splint. Proximal mandibular segments include areas
with a similar presentation to patient one.

51
CHAPTER 6: DISCUSSION

MAXILLARY TRENDS

The left lateral maxilla tended to be slightly more accurate than the right. This is
most clearly evidenced in patients two through six. In cases for which the left lateral
maxilla was notably more accurate than the right the anterior maxilla was typically of
intermediate accuracy. With the exception of patient three, both left and right accuracy
was to within 2.0mm.
The anterior maxilla tended to be more accurate than the posterior. The
maxillary areas with the least amount of accuracy tended to be at or near the
tuberosities, but even these areas were accurate to within 2.0mm, with exceptions as
noted for individual patients. The exceptions include patients three and six.

MANDIBULAR TRENDS

Mandibular positioning appeared slightly more accurate overall than maxillary
positioning. This was true for cases that did not include genioplasty as well as those that
did include surgical chin modification. It was also true for both cases that underwent
mandibular setback and cases that experienced mandibular advancement.
The left side of the body of the mandible tended to be slightly more accurate
than the right. This slight increase left-sided accuracy mirrors what is seen in the maxilla.
52
And as with the maxilla, nearly all mandibular surfaces were accurate to within 2.0mm,
with exceptions as noted for individual patients. Unlike the maxilla, however, is that no
anteroposterior trends were noted in the mandible.

OTHER TRENDS

RAMI
The rami, which were proximal to sagittal osteotomies and therefore
theoretically unaltered by surgery, showed some of the greatest variations in positional
accuracy. Given that the rami are not under direct positional control of the splint
described in this paper, this observation may be beyond the scope of this discussion.
Regardless, it is worth mentioning that positional variability of the rami did not appear
to show bilateral symmetry. Also, despite a lack of obvious anteroposterior patterns, the
color maps and overlays do portray a tendency toward the post-surgical position of both
rami being relatively lateral to the VSP position.

PROXIMITY TO DENTITION
Among all cases, maxillary and mandibular areas closer to the dentition tended
to be more accurate than those areas farther away. However, it must be noted that with
the upper arch, areas of the anterior maxilla farther from the dentition were positionally
fixed with titanium plates. As these plates are not accounted for in the VSP, it is possible
that the plates themselves skewed accuracy readings. In other words, maxillary anterior
53
areas farther from the dention and closer to the osteotomy that are closer to 2.0mm
than 0.4mm of discrepancy between pre- and post-surgical positions may be related to
the post-surgical CBCT's inclusion of internal surgical fixation plates rather than actual
post-surgical anterior maxillary bony presentation.

CLASS II AND CLASS III CORRECTIONS
No trend was noticed for relative differences in accuracy between class II and
class III corrections: the trends already discussed were noted across Class II and Class III
corrections, and of the three cases with the most overall accuracy (patients one, four
and seven), two were Class III cases and one was a Class II case. Though all cases were
surgically treated with LeFort I and BSSO, the anteroposterior corrective movement of
both maxillary and mandibular distal segments was not uniform within classes. Perhaps
normalizing for directional correction within classes would provide additional data.

54
CHAPTER 7: CONCLUSIONS

The combined use of pre-operative CBCT scans, VSPs, and 3D printing to create a
full-coverage orthognathic splint results in predictable, accurate post-operative
positions to within 2.0mm and often to within less than 0.4mm. The tray-type splint can
contribute to VSP outcome reproducibility by providing surgical positioning accuracy
and predictability.
There is ample room for future study. A prospective study could include pre-
surgical and post-surgical scans of equal resolution, and/ or software-driven registration
of superimpositions as computerized surface-matching algorithms become more robust.
Single-jaw procedures that utilize the splint could contribute valuable data. It may be
beneficial to determine if the accuracy of the technique contributes to reduced post-
surgical orthodontic treatment time. Lastly, the ability of the splint to control tooth tip
and torque relative to wafer-type splints, manufactured traditionally or with
comparable digital protocols, could be analyzed.

55
ACKNOWLEDGEMENTS

Particular gratitude goes to Dr. Ryan Hungate, DDS, MS, for the original idea--
and the tremendous amount of follow-through effort-- that resulted in the invention of
the tray-type splint discussed herein and which has proven to be a valuable part of the
orthodontic-orthognathic armamentarium.
Thank you also to the Los Angeles Center for Oral and Maxillofacial Surgery, and
to Dr. Robert Relle, DDS, in particular, for pioneering the use of the tray-type splint, for
delivering outstanding patient care, and for maintaining impeccable records, all of which
made this project possible.
Lastly, a special note of appreciation to Medical Modeling, and to Evan Koenig in
particular, for providing the software, the technical knowledge, and the time required to
bring such compelling data to life through impressively precise imaging.

(Abdel-Moniem Barakat et al., 2014; Aboul-Hosn Centenero and Hernandez-Alfaro, 2012; Adams et al., 2004; Adolphs et al., 2014;
Baker et al., 2012; Bobek et al., 2015; Cevidanes et al., 2005; Cevidanes et al., 2007; Cevidanes et al., 2009; Cevidanes et al., 2011;
Cevidanes et al., 2010; Choi et al., 2012; Choi et al., 2009; Dammous et al., 2015; de Paula et al., 2013; Farronato et al., 2015; Gateno
et al., 2007; Grauer et al., 2009; Hernandez-Alfaro and Guijarro-Martinez, 2013; Heymann et al., 2010; Hino et al., 2013; Hsu et al.,
2013; Kamiishi et al., 2007; Levine et al., 2012; Li et al., 2013; Liebregts et al., 2015; Mavili et al., 2007; Metzger et al., 2008; Nadjmi
et al., 2010; Nguyen et al., 2011; Noguchi et al., 2007; Olszewski et al., 2010; Olszewski et al., 2007; Schouman et al., 2015; Song and
Baek, 2009; Sun et al., 2013a; Sun et al., 2013b; Swennen et al., 2007)
(Tucker et al., 2010; Uribe et al., 2013; Xia et al., 2005; Xia et al., 2006; Zinser et al., 2012)

56
BIBLIOGRAPHY

Abdel-Moniem Barakat, A., Abou-ElFetouh, A., Hakam, M.M., El-Hawary, H., and Abdel-
Ghany, K.M. (2014). Clinical and radiographic evaluation of a computer-generated
guiding device in bilateral sagittal split osteotomies. J Craniomaxillofac Surg 42, e195-
203.

Aboul-Hosn Centenero, S., and Hernandez-Alfaro, F. (2012). 3D planning in orthognathic
surgery: CAD/CAM surgical splints and prediction of the soft and hard tissues results -
our experience in 16 cases. J Craniomaxillofac Surg 40, 162-168.

Adams, G.L., Gansky, S.A., Miller, A.J., Harrell, W.E., Jr., and Hatcher, D.C. (2004).
Comparison between traditional 2-dimensional cephalometry and a 3-dimensional
approach on human dry skulls. Am J Orthod Dentofacial Orthop 126, 397-409.

Adolphs, N., Liu, W., Keeve, E., and Hoffmeister, B. (2014). RapidSplint: virtual splint
generation for orthognathic surgery - results of a pilot series. Comput Aided Surg 19, 20-
28.

Baker, S.B., Goldstein, J.A., and Seruya, M. (2012). Outcomes in computer-assisted
surgical simulation for orthognathic surgery. J Craniofac Surg 23, 509-513.

Bobek, S., Farrell, B., Choi, C., Weimer, K., and Tucker, M. (2015). Virtual surgical
planning for orthognathic surgery using digital data transfer and an intraoral fiducial
marker: the charlotte method. J Oral Maxillofac Surg 73, 1143-1158.

Cevidanes, L.H., Bailey, L.J., Tucker, G.R., Jr., Styner, M.A., Mol, A., Phillips, C.L., Proffit,
W.R., and Turvey, T. (2005). Superimposition of 3D cone-beam CT models of
orthognathic surgery patients. Dentomaxillofac Radiol 34, 369-375.

Cevidanes, L.H., Bailey, L.J., Tucker, S.F., Styner, M.A., Mol, A., Phillips, C.L., Proffit, W.R.,
and Turvey, T. (2007). Three-dimensional cone-beam computed tomography for
assessment of mandibular changes after orthognathic surgery. Am J Orthod Dentofacial
Orthop 131, 44-50.

Cevidanes, L.H., Heymann, G., Cornelis, M.A., DeClerck, H.J., and Tulloch, J.F. (2009).
Superimposition of 3-dimensional cone-beam computed tomography models of growing
patients. Am J Orthod Dentofacial Orthop 136, 94-99.

Cevidanes, L.H., Oliveira, A.E., Grauer, D., Styner, M., and Proffit, W.R. (2011). Clinical
application of 3D imaging for assessment of treatment outcomes. Semin Orthod 17, 72-
80.

57
Cevidanes, L.H., Tucker, S., Styner, M., Kim, H., Chapuis, J., Reyes, M., Proffit, W., Turvey,
T., and Jaskolka, M. (2010). Three-dimensional surgical simulation. Am J Orthod
Dentofacial Orthop 138, 361-371.

Chelule, K., Coole, T., and Cheshire, D. (2000). Fabrication of medical models from scan
data via rapid prototyping techniques. Paper presented at: Proceedings of the 2000
Conference on Time Compression Technologies.

Choi, J.Y., Hwang, J.M., and Baek, S.H. (2012). Virtual model surgery and wafer
fabrication using 2-dimensional cephalograms, 3-dimensional virtual dental models, and
stereolithographic technology. Oral Surg Oral Med Oral Pathol Oral Radiol 113, 193-200.

Choi, J.Y., Song, K.G., and Baek, S.H. (2009). Virtual model surgery and wafer fabrication
for orthognathic surgery. Int J Oral Maxillofac Surg 38, 1306-1310.

Cooper, K. (2001). Rapid prototyping technology: selection and application (CRC press).

Dammous, S., Dupont, Q., and Gilles, R. (2015). Three-dimensional computed
tomographic evaluation of bilateral sagittal split osteotomy lingual fracture line and le
fort I pterygomaxillary separation in orthognathic surgery using cadaver heads:
ultrasonic osteotome versus conventional saw. J Oral Maxillofac Surg 73, 1169-1180.

de Paula, L.K., Ruellas, A.C., Paniagua, B., Styner, M., Turvey, T., Zhu, H., Wang, J., and
Cevidanes, L.H. (2013). One-year assessment of surgical outcomes in Class III patients
using cone beam computed tomography. Int J Oral Maxillofac Surg 42, 780-789.

Farronato, G., Giannini, L., Galbiati, G., Mortellaro, C., and Maspero, C. (2015).
Presurgical virtual three-dimensional treatment planning. J Craniofac Surg 26, 820-823.

Gateno, J., Xia, J.J., Teichgraeber, J.F., Christensen, A.M., Lemoine, J.J., Liebschner, M.A.,
Gliddon, M.J., and Briggs, M.E. (2007). Clinical feasibility of computer-aided surgical
simulation (CASS) in the treatment of complex cranio-maxillofacial deformities. J Oral
Maxillofac Surg 65, 728-734.

Grauer, D., Cevidanes, L.S., and Proffit, W.R. (2009). Working with DICOM craniofacial
images. Am J Orthod Dentofacial Orthop 136, 460-470.

Hernandez-Alfaro, F., and Guijarro-Martinez, R. (2013). New protocol for three-
dimensional surgical planning and CAD/CAM splint generation in orthognathic surgery:
an in vitro and in vivo study. Int J Oral Maxillofac Surg 42, 1547-1556.

Heymann, G.C., Cevidanes, L., Cornelis, M., De Clerck, H.J., and Tulloch, J.F. (2010).
Three-dimensional analysis of maxillary protraction with intermaxillary elastics to
miniplates. Am J Orthod Dentofacial Orthop 137, 274-284.
58

Hilton, P. (2000). Rapid tooling: technologies and industrial applications (CRC Press).

Hino, C.T., Cevidanes, L.H., Nguyen, T.T., De Clerck, H.J., Franchi, L., and McNamara, J.A.,
Jr. (2013). Three-dimensional analysis of maxillary changes associated with facemask
and rapid maxillary expansion compared with bone anchored maxillary protraction. Am
J Orthod Dentofacial Orthop 144, 705-714.

Hsu, S.S., Gateno, J., Bell, R.B., Hirsch, D.L., Markiewicz, M.R., Teichgraeber, J.F., Zhou,
X., and Xia, J.J. (2013). Accuracy of a computer-aided surgical simulation protocol for
orthognathic surgery: a prospective multicenter study. J Oral Maxillofac Surg 71, 128-
142.

Hungate, R. (2015). The creation of a novel full-coverage orthognathic surgical splint
utilizing 3D printing & virtual surgical planning. MS Thesis. University of Southern
California Graduate School.

Kamiishi, H., Miyasato, Y., and Kosaka, M. (2007). Development of the 3D-cephalogram:
a technical note. J Craniomaxillofac Surg 35, 258-260.

Jacobs, P.F. (1992). Rapid prototyping & manufacturing: fundamentals of
stereolithography (Society of Manufacturing Engineers).

Kai, C.C. (1994). Three-dimensional rapid prototyping technologies and key
development areas. Computing & Control Engineering Journal 5, 200-206.

Kai, C.C., Meng, C.S., Ching, L.S., Hoe, E.K., and Fah, L.K. (1998). Rapid prototyping
assisted surgery planning. Int J Adv Manuf Technol 14, 624-630.

Klein, H.M., Schneider, W., Alzen, G., Voy, E.D., and Gunther, R.W. (1992). Pediatric
craniofacial surgery: comparison of milling and stereolithography for 3D model
manufacturing. Pediatr Radiol 22, 458-460.

Levine, J.P., Patel, A., Saadeh, P.B., and Hirsch, D.L. (2012). Computer-aided design and
manufacturing in craniomaxillofacial surgery: the new state of the art. J Craniofac Surg
23, 288-293.

Li, B., Zhang, L., Sun, H., Yuan, J., Shen, S.G., and Wang, X. (2013). A novel method of
computer aided orthognathic surgery using individual CAD/CAM templates: a
combination of osteotomy and repositioning guides. Br J Oral Maxillofac Surg 51, e239-
244.

59
Liebregts, J.H., Timmermans, M., De Koning, M.J., Berge, S.J., and Maal, T.J. (2015).
Three-dimensional facial simulation in bilateral sagittal split osteotomy: a validation
study of 100 patients. J Oral Maxillofac Surg 73, 961-970.

Mavili, M.E., Canter, H.I., Saglam-Aydinatay, B., Kamaci, S., and Kocadereli, I. (2007). Use
of three-dimensional medical modeling methods for precise planning of orthognathic
surgery. J Craniofac Surg 18, 740-747.

Metzger, M.C., Hohlweg-Majert, B., Schwarz, U., Teschner, M., Hammer, B., and
Schmelzeisen, R. (2008). Manufacturing splints for orthognathic surgery using a three-
dimensional printer. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 105, e1-7.

Nadjmi, N., Mollemans, W., Daelemans, A., Van Hemelen, G., Schutyser, F., and Berge, S.
(2010). Virtual occlusion in planning orthognathic surgical procedures. Int J Oral
Maxillofac Surg 39, 457-462.

Nguyen, T., Cevidanes, L., Cornelis, M.A., Heymann, G., de Paula, L.K., and De Clerck, H.
(2011). Three-dimensional assessment of maxillary changes associated with bone
anchored maxillary protraction. Am J Orthod Dentofacial Orthop 140, 790-798.

Noguchi, N., Tsuji, M., Shigematsu, M., and Goto, M. (2007). An orthognathic simulation
system integrating teeth, jaw and face data using 3D cephalometry. Int J Oral Maxillofac
Surg 36, 640-645.

Olszewski, R., Tanesy, O., Cosnard, G., Zech, F., and Reychler, H. (2010). Reproducibility
of osseous landmarks used for computed tomography based three-dimensional
cephalometric analyses. J Craniomaxillofac Surg 38, 214-221.

Olszewski, R., Zech, F., Cosnard, G., Nicolas, V., Macq, B., and Reychler, H. (2007). Three-
dimensional computed tomography cephalometric craniofacial analysis: experimental
validation in vitro. Int J Oral Maxillofac Surg 36, 828-833.

Petzold, R., Zeilhofer, H.F., and Kalender, W.A. (1999). Rapid protyping technology in
medicine--basics and applications. Comput Med Imaging Graph 23, 277-284.

Potamianos, P., Amis, A.A., Forester, A.J., McGurk, M., and Bircher, M. (1998). Rapid
prototyping for orthopaedic surgery. Proc Inst Mech Eng H 212, 383-393.

Schouman, T., Rouch, P., Imholz, B., Fasel, J., Courvoisier, D., and Scolozzi, P. (2015).
Accuracy evaluation of CAD/CAM generated splints in orthognathic surgery: a cadaveric
study. Head Face Med 11, 24.

Song, K.G., and Baek, S.H. (2009). Comparison of the accuracy of the three-dimensional
virtual method and the conventional manual method for model surgery and
60
intermediate wafer fabrication. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 107,
13-21.

Sun, Y., Luebbers, H.T., Agbaje, J.O., Schepers, S., Vrielinck, L., Lambrichts, I., and Politis,
C. (2013a). Accuracy of upper jaw positioning with intermediate splint fabrication after
virtual planning in bimaxillary orthognathic surgery. J Craniofac Surg 24, 1871-1876.

Sun, Y., Luebbers, H.T., and Politis, C. (2013b). Three-dimensional virtual model surgery
to fabricate the digital intermediate splint. J Craniofac Surg 24, 563-565.

Swennen, G.R., Barth, E.L., Eulzer, C., and Schutyser, F. (2007). The use of a new 3D
splint and double CT scan procedure to obtain an accurate anatomic virtual augmented
model of the skull. Int J Oral Maxillofac Surg 36, 146-152.

Tucker, S., Cevidanes, L.H., Styner, M., Kim, H., Reyes, M., Proffit, W., and Turvey, T.
(2010). Comparison of actual surgical outcomes and 3-dimensional surgical simulations.
J Oral Maxillofac Surg 68, 2412-2421.

Uribe, F., Janakiraman, N., Shafer, D., and Nanda, R. (2013). Three-dimensional cone-
beam computed tomography-based virtual treatment planning and fabrication of a
surgical splint for asymmetric patients: surgery first approach. Am J Orthod Dentofacial
Orthop 144, 748-758.

Xia, J.J., Gateno, J., and Teichgraeber, J.F. (2005). Three-dimensional computer-aided
surgical simulation for maxillofacial surgery. Atlas Oral Maxillofac Surg Clin North Am 13,
25-39.

Xia, J.J., Phillips, C.V., Gateno, J., Teichgraeber, J.F., Christensen, A.M., Gliddon, M.J.,
Lemoine, J.J., and Liebschner, M.A. (2006). Cost-effectiveness analysis for computer-
aided surgical simulation in complex cranio-maxillofacial surgery. J Oral Maxillofac Surg
64, 1780-1784.

Zinser, M.J., Mischkowski, R.A., Sailer, H.F., and Zoller, J.E. (2012). Computer-assisted
orthognathic surgery: feasibility study using multiple CAD/CAM surgical splints. Oral
Surg Oral Med Oral Pathol Oral Radiol 113, 673-687.

Abstract (if available)

Abstract INTRODUCTION: To assess the accuracy of a patented 3D-printed tray-type splint in orthognathic surgery by comparing postoperative CBCT scans to virtual surgical plans. ❧ METHODS: Seven adult patients underwent pre-surgical orthodontic decompensation. Immediately prior to surgical workup, orthodontic appliances or attachments adjunct to clear aligner therapy were removed to ensure a smooth, scannable dentition. The arches were scanned optically, maxillofacial structures were scanned via CBCT, and a composite image of these scans was generated. Virtual surgery was performed, and the 3D-printed tray-type splint was fabricated according to the virtually planned postoperative position. Orthognathic surgery (LeFort I and BSSO in for all cases) was completed with the splint, and intermaxillary fixation was obtained by means of miniscrews. A postoperative CBCT was taken immediately following surgery. The virtual plan and postoperative CBCT were superimposed by registration of the images at the anterior cranial base. These superimpositions were analyzed and measured by semitransparent overlay and color maps of surface positional differences. Postoperative orthodontic treatment was completed with clear aligner therapy. ❧ RESULTS: Qualitative analysis of semi-transparent overlays and color maps for each patient reveals accuracy of position for each of the seven included patients, with the possible exception of patient 3. The mean directional differences between VSP and post-operative CBCT positions for all measured tooth positions were 0.15 ± 0.57mm posterior in the sagittal plane, 0.22 ± 0.88mm right in the horizontal plane, and 0.11 ± 0.56mm downward in the vertical plane. ❧ DISCUSSION: The left lateral maxilla tended to be slightly more accurate than the right, and in cases for which the left lateral maxilla was notably more accurate, the anterior maxilla was typically of intermediate accuracy. The anterior maxilla tended by be more accurate than the posterior, with the maxillary tuberosities being of least accuracy. Within these observed maxillary trends, the overwhelming majority of surfaces were within 2.0mm of accuracy, and many were within 0.4mm or accuracy. ❧ Mandibular positioning tended to be slightly more accurate overall than maxillary positioning. And as with the maxilla, the overwhelming majority of surfaces were within 2.0mm of accuracy, and many were within 0.4mm or accuracy. Also as observed in the maxilla, a slight increase in left-sided accuracy was noted. Unlike the maxilla, however, no anteroposterior trends were noted in the mandible. ❧ Other observations include: the rami presenting with more positional variability than any other segment, maxillary and mandibular areas closer to the dentition tending to be more accurate than those areas farther away, and an apparent absence of trends for relative differences in accuracy between class II and class III corrections. Ramus presentation is of particular note because the rami, as bony segments proximal to the surgical osteotomies, are not under direct positional control of the tray-type splint. ❧ CONCLUSIONS: The combined use of pre-operative CBCT scans, VSPs, and 3D printing to create a full-coverage orthognathic splint results in predictable, accurate post-operative positions. A future prospective study could include pre-surgical and post-surgical scans of equal resolution. Future research could also incorporate software-driven registration of superimpositions as computerized surface-matching algorithms become more robust.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

PDF

Three-dimensional quantification of post-surgical condylar displacement

PDF

Comparison of facial midline landmark and condylar position changes following orthognathic surgery

PDF

Utilizing voxel based superimposition to asses orthognathic surgical treatment

PDF

Assessment of 3D surface changes following virtual bracket removal

PDF

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

PDF

Classification of 3D maxillary incisor root shape

PDF

Three-dimensional immediate post-surgery condylar displacement

PDF

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

PDF

Monitering of typodont root movement via crown superimposition of single CBCT and consecutive iTero scans

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

Prevalence of TMJ morphological changes and scoring system based on CBCT imaging

PDF

Root shape frequency and direction of dilaceration: a CBCT study

PDF

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

PDF

The long-term cephalometric and respiratory outcomes of mandibular distraction osteogenesis in infants with Pierre Robin sequence

PDF

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

PDF

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

PDF

Use of digital models to assess orthodontic treatment progress and identify deficiencies

PDF

Detrimental effects of dental encroachment on secondary alveolar bone graft outcomes in the treatment of patients with cleft lip and palate: a cone-beam computed tomography study

Asset Metadata

Creator Hann, Robert (author)

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

School School of Dentistry

Degree Master of Science

Degree Program Craniofacial Biology

Publication Date 04/20/2016

Defense Date 03/17/2016

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

Tag 3D printing,OAI-PMH Harvest,orthognathic splint,orthognathic surgery,virtual surgical planning

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

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

Creator Email hannr@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-234402

Unique identifier UC11279098

Identifier etd-HannRobert-4311.pdf (filename),usctheses-c40-234402 (legacy record id)

Legacy Identifier etd-HannRobert-4311.pdf

Dmrecord 234402

Document Type Thesis

Format application/pdf (imt)

Rights Hann, Robert

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