Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
The accuracy of intraoral scans of interproximal spaces with and without saliva: a prospective clinical study
(USC Thesis Other)
The accuracy of intraoral scans of interproximal spaces with and without saliva: a prospective clinical study
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
The Accuracy of Intraoral Scans of Interproximal
Spaces With and Without Saliva: A Prospective
Clinical Study
By
Katherine Schwartz, DDS
A Thesis Presented to the
Faculty of the USC Herman Ostrow School of Dentistry
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
Master of Science
Cranio-Facial Biology
School of Dentistry
May 2020
2
Acknowledgements:
I would like to thank Dr. Andre Weissheimer for his mentorship and guidance throughout this
research project. From the initial brainstorming to the data collection and review of thesis drafts,
Dr. Weissheimer has been an invaluable resource. I would also like to thank Dr. Kevin Yin for
his help with the statistics used to analyze the results of the project. Dr. Yin was an invaluable
resource throughout the project.
I would also like to offer special thanks to Dr. Hany Youssef, Dr. Cory Nasoff and Dr. Kimberlin
Low for their contributions to this project
-
3
Table of Contents:
1. Abstract……………………………………………………………………………………4
2. Introduction………………………………………………………………………………..6
3. Materials and Methods…………………………………………………………………...10
a. Sample
b. Protocols
c. Measurements
d. Statistical Analysis
4. Results……………………………………………………………………………………16
5. Discussion………………………………………………………………………………..27
6. Conclusions…………………………………………………………………………........33
7. References………………………………………………………………………………..34
8. Appendix…………………………………………………………………………………36
4
Abstract:
Introduction:
Digital workflow in orthodontic treatment planning relies on the accuracy of scanners to generate
3D models from patients’ malocclusion. At the end of treatment, undiagnosed diastemas lead to
spaces that are both unhygienic and unaesthetic, often requiring further treatment. The purpose
of this clinical study is to evaluate whether the presence of saliva affects the accuracy of
interproximal spaces of digital models made from intraoral scans.
Methods:
The sample consisted of 15 individuals ranging from 12 to 60 years old with a total of 17 arches
and 29 interproximal spaces. Inclusion criteria was that subjects had at least one space in a dental
arch measuring 0.2 mm or greater. The spaces were measured intraorally using IPR gauges as the
gold standard. The arches were scanned with saliva and without saliva using the 3Shape
TRIOS3
®
Intraoral Scanner. Stereolithography files of digital models were generated from the
intraoral scans. The 3D models, including those with and without saliva, were uploaded to
Meshmixer™. An orthodontist and two orthodontic residents were asked to identify, measure
and record spaces on each model a total of two times generating two different trials of
measurements (single blind). Wilcoxon signed rank and Mann Whitney U tests were performed
to determine whether saliva significantly affected the visibility of intraoral spaces on the digital
models.
Results:
Measurements of spaces on models generated with and without saliva were not significantly
different (P value > 0.05) from the intraoral measurements of the same spaces. Models with and
without saliva did not accurately represent spaces less than 0.3 mm. Model measurements for
5
spaces 0.4 mm and above, regardless of the presence of saliva, were significantly different than
the intraoral measurements. The absence of saliva did not lead to more accurate measurements
by orthodontists, although digital models in Meshmixer™ appeared visually clearer when saliva
was absent.
Conclusions:
Regardless the presence of saliva, small spaces ( ≦ 0.3 mm) were not diagnosed on the intraoral
scans while spaces 0.4 mm and above, could be recognized but not accurately measured. For
small interproximal spaces, diagnosis still remains a challenge in the orthodontic digital
workflow when intra-oral scans obtained with the 3Shape TRIOS3® Intraoral Scanner are used.
6
Introduction:
The field of orthodontics is rapidly changing with the introduction of digital workflow including
the use of intraoral scans for diagnosis, treatment planning, virtual setups of the dentition and the
manufacture of custom appliances such as clear aligners and lingual orthodontic appliances.
According to an informal survey by the Journal of Clinical Orthodontics in 2016, sixty-two
percent of subscribers to the journal reported using intraoral scanners in their private practice
(Sinclair, 2016). With improvements in technology and digital treatment planning, it is likely that
this percentage has increased in the years since this publication.
Digital workflow in the field of orthodontics is a multi-step process to streamline the formation
of a diagnosis, treatment plan and customized appliance for each patient. Each patient is scanned
with an intraoral scanner in order to create a digital 3D model in the form of a stereolithography
(STL) file. STL files are a standard format used by most intraoral scanners and 3D modeling
software programs (Taneva et al., 2015). The STL file may be used as a digital study model for
diagnosis and treatment planning, to place brackets digitally for indirect bonding and for the
manufacture of orthodontic appliances such as retainers and expanders. The STL file may also be
uploaded to a digital software where the boundaries of teeth are identified such as their cusps,
marginal ridges and occlusal surfaces in order to reposition them (Kumar et al., 2012).
Repositioning teeth allows for the creation of a digital setup of the proposed treatment plan. The
treatment plan is then used to manufacture a customized appliance for the patient (Taneva et al.,
2015).
7
There are many benefits to intraoral scanning when compared with conventional techniques.
Conventional impression taking and the pouring of stone models requires many steps that have
the possibility of introducing errors, leading to the production of inaccurate stone models (Ender
et al., 2013; Renne et al., 2017). The ability to integrate digital impressions into the diagnostic
data collected allows for more predictable, customizable treatment for patients (Zimmermann et
al., 2015). The use of intraoral scanners in dentistry and orthodontics has made impressions
more comfortable for patients, especially those patients with strong gag reflexes. The ease of
digitally transferring files to dental technicians allows for a quicker turnaround of appliances and
it is easier to transfer digital records between practitioners. Digital impressions allow for
magnification of the dentition and the ability to edit defective areas of the model (Renne et al.,
2017). Digital impressions also allow for more powerful communication and marketing with
patients since they feel more involved in their treatment by visualizing their treatment digitally
(Mangano et al., 2017). Finally, the integration of digital models into one’s practice allows
orthodontists to no longer store stone models and save digital records indefinitely (Kihara et al.,
2019).
Digital workflow in orthodontic treatment planning relies on the accuracy of scanners to generate
3D models from patients’ malocclusions. Every rotation, overlap and space between teeth must
be captured accurately in order for an accurate diagnosis to be made and an effective treatment
plan to be generated. It is accepted in orthodontics that measurement accuracy up to 0.1 mm is
necessary for a study model to be diagnostically accurate. The literature has shown that digital
models and traditional stone models have clinically insignificant differences of orthodontic
measurements (Taneva et al., 2015).
8
In order for computer-aided customized treatment to be manufactured, the digital model must be
segmented so that each tooth is free to be moved within the arch. In addition to aiding in the
fabrication of customized appliances, segmentation of teeth is necessary for computing arch
length and performing a Bolton analysis. The surface curvature, long axis, width and height of
each tooth is determined in order to properly segment it. It is often difficult to identify the
interstices between teeth when malocclusion exists. Additionally, the shapes of teeth and artifacts
from scanning make identifying each tooth a challenge (Liao et al., 2015). Errors in the diagnosis
of the malocclusion can be introduced if the teeth are not segmented correctly.
According to Dr. Andrews’ Six Keys to Normal Occlusion article, a dentition with normal
occlusion has tight contacts between teeth in which no spaces exist (Andrews, 1972). Contact
surfaces, usually located in the upper middle third of teeth, allow for the patient to clean properly
and avoid food impaction. Contact surfaces also allow for the transmission of occlusal forces and
the maintenance of the dental arch length (Keogh, 2001). At some forums it has been brought up
by orthodontists that residual spaces are sometimes found at the end of digital customized
treatment. These undiagnosed spaces are both unhygienic and unaesthetic. Undiagnosed spaces
at the end of the prescribed treatment require further treatment such as refinement trays or wires
(Papadimitriou et al., 2018), costing the companies more resources and the patient increased time
in treatment. It seems this is understood and incorporated by the digital software programs as
Invisalign aligner prescription forms provide the option to include overcorrection trays to close
spaces at the end of the prescribed treatment (Align Technology, 2020). Additionally,
INBRACE, the new generation of customized lingual appliance, programs 0.2 mm of digital
9
interproximal collisions between teeth into the appliance in order to compensate for possible
dental segmentation errors (Tong et al., 2019). These errors have the possibility of resulting in
untreated spaces at the end of treatment. This overcorrection is known as a “virtual power
chain” prescribed for all cases in order to prevent spaces from opening clinically (Align
Technology, 2020).
Intraoral scanners have been reviewed in current literature by comparing different types of
scanners and assessing the accuracy of the scans of the dentition compared to stone models
digitized with desktop scanners (Tomita et al., 2018). One study postulated that when intraoral
scanning, patient movement and saliva may lead to inaccuracy of the digital impressions,
although these were not tested (Kihara et al., 2019). Current research has not studied the effects
of saliva, blood or patient movement on the accuracy of intraoral scanners (Kihara et al., 2019).
While intraoral environment simulators have been tested, an in vivo study assessing the accuracy
of the 3Shape TRIOS3
®
scanner with saliva has yet to be completed (Park et al., 2018). We
hypothesized that if saliva is present in the interproximal area, then spaces up to of 0.3 mm will
not be diagnosed, leading to inaccuracy of the orthodontic diagnosis, digital setup process and
manufacturing of the digital customized appliance, especially if multiple small spaces are
present.
10
Methods:
Sample
This study qualified as Human Subjects Research and required approval by the University Park
Institutional Review Board (UPIRB). This prospective clinical study was conducted in the
Advanced Orthodontics clinic of the University of Southern California (USC). Subjects’
dentitions were scanned with the 3Shape TRIOS3
®
Intraoral Scanner (3Shape A/S, Copenhagen,
Denmark) in the USC orthodontics clinic (Figures 1 and 2). Stereolithography (STL) files were
exported and uploaded to the desktop application Meshmixer™ to be analyzed by the
orthodontist measurers, one practicing orthodontist and two orthodontic residents (Figure 3).
Fig 1. TRIOS3
®
Intraoral Scanner Fig 2. 3Shape TRIOS
®
MOVE
(3Shape A/S, Copenhagen, Denmark). (3Shape A/S, Copenhagen, Denmark).
Image from: https://www.3shape.com Image from: https://www.3shape.com
/en/products/trios/intraoral-scanners /en/products/trios/intraoral-scanners
11
Fig 3. Meshmixer™ Desktop Application (version 3.5.474; ©Autodesk, Inc. 2017 San Rafael,
CA) Image taken from desktop Meshmixer™ application
In this prospective clinical trial, the sample consisted of 15 individuals ranging from 12 to 60 years
old with a total of 17 arches and 29 spaces. All subjects were either patients of the University of
Southern California orthodontics clinic or orthodontic residents. The inclusion criteria included
at least one space in a dental arch measuring 0.2 mm or greater. Calibrated interproximal
reduction gauges were used to measure spaces to the hundredth decimal point intraorally.
Fig 4. Interproximal reduction gauges to the nearest hundredth decimal were used to measure
spaces intraorally prior to scanning with the 3Shape TRIOS
®
scanner.
12
Fig 5. The 3Shape TRIOS3
®
Intraoral scanner was used to create digital models of the dentition.
The teeth were scanned with saliva and without saliva.
The 3Shape TRIOS3
®
scanner was calibrated before use on each patient. Without drying, the
dental arches containing spaces with saliva in the interproximal areas were scanned with the
3Shape TRIOS
®
scanner. The dentition was then dried thoroughly and scanned again with the
3Shape TRIOS
®
scanner. Two separate digital models were created of each arch: one taken with
saliva present and one after drying the dentition. The digital models of the dental arches
including those with and without saliva were blinded, numbered, randomized, and uploaded to
the application Meshmixer™. Meshmixer™ is a commonly used, free software program used to
analyze and edit scans, segment teeth and print 3D models (Autodesk, 2019). Two orthodontic
residents and one orthodontist were each blinded, regarding the purpose of the research, and
asked to identify and measure spaces they observed on each model. These individuals were not
informed that the study would compare arches with and without saliva and were instead told the
purpose of the study was to assess the accuracy of the scanned digital models and to locate and
measure spaces between teeth. Spaces were to be measured at the narrowest point between the
greatest convexities of the two teeth and recorded. This process was completed twice for each
measurer, one week apart. There were fifty-eight total spaces in all thirty-four dental arches
including those with and without saliva.
13
Protocols
Meshmixer™ was used as it is a free computer-aided design software available for Windows and
MacOS and is widely used by many orthodontists to work with STL files in the orthodontic
digital workflow. Meshmixer™ allows for quick 3D measurements on the digital models
(Autodesk, 2019). Digital models were uploaded to Meshmixer™ using the Import tool. Using
two fingers to control the models, they could be moved around to view the models from all
angles and determine where spaces exist. It was also possible to zoom in and out of the models in
order to view the model clearer. Measurers were told to create a measurement line using
Meshmixer™ and to evaluate this line from different views to determine if it was between the
greatest convexities of adjacent teeth. Next the analysis setting was selected followed by the
measure tool. Within the measure tool there are options to choose the type and direction of the
measurement (Fig 6.). Measurers were instructed to choose the second option within the type of
measurement that most closely resembles the measurement of space between two teeth (Fig 7.).
Measurers were then instructed to choose the first option within the direction of measurement
that most closely resembles a measurement from the most external surface of one tooth to the
next (Fig 8.).
14
Fig 6. Measurement tool within Meshmixer™. The type of measurement and the desired
direction of measurement are chosen.
Fig 7. Measurers were instructed to choose the second option within the type of measurement
that most closely resembles the measurement of space between two teeth.
Fig 8. Measurers were instructed to choose the first option within the direction of measurement
that most closely resembles a measurement from the most external surface of one tooth to the
next.
15
Fig 9. A space between two teeth measuring 0.44 mm at the narrowest point between the two
convexities of the adjacent teeth.
Measurements
The study consisted of 15 individuals with 17 arches and 29 total spaces. Each arch was scanned
with and without saliva, leading to a total of 34 stereolithography files. It took each measurer
approximately one and a half hours to complete all 34 measurements. Measurements were made
by one orthodontist and two orthodontic residents. The 34 stereolithography files of models with
and without saliva were randomized so that the models with and without saliva of each patient
were not consecutive. Each measurer was instructed to measure spaces on the models twice to
produce two trials. Measurements between trials one and two were made at least one week apart
for the intra-observer reliability test. Measurers were not told where the spaces were on each
model and were free to assume some models may not have spaces. Each measurer was trained
how to manipulate and perform digital measurements on the models using the Meshmixer™
software and were instructed to record their measurements to the nearest hundredth decimal
point on an excel spreadsheet.
16
Statistical Analysis
The two sets of measurements recorded in Meshmixer™ by the measurers, measured one week
apart, were compared to the intraoral measurements taken on each patient. Inter-observer and
intra-observer reliability were measured using Chronbach’s alpha. Error was calculated for the
mean measurement of spaces with and without saliva compared to the intraoral measurements.
The Wilcoxon signed rank test was performed to detect the potential difference in space
measurement between the digital scans and gold standard intraoral measurements. The Mann-
Whitney U test was used to determine if there is a significant difference between the errors with
and without saliva. Type I error rate was 0.05.
17
Results:
The results of Chronbach’s alpha test for intra-operator and inter-operator reliability were 0.94
and above, indicating high reliability among the measurers in their ability to measure the spaces.
Due to the limited sample size used in the study, nonparametric statistical analysis was used.
The descriptives of all measurements in the study including the intraoral measurements (gold
standard) and mean measurements with and without saliva were determined (Table 1). The
median for the gold standard was 0.45 mm, the median for the saliva group was 0.55 mm and the
median for the group without saliva was 0.6 mm. It is evident that the mean measurements for
models with and without saliva were not far from the intraoral (gold standard) measurements,
although the standard deviation of 0.45 for the saliva group and 0.43 for the group without saliva
were larger than the standard deviation of 0.21 for the gold standard group.
Table 1. Summary of measurements (mm): Intraoral measurements (Gold Standard) compared
to measurements with saliva (Saliva Mean) and measurements without saliva (No Saliva Mean).
Descriptives Median Min Max Mean SD
Gold Standard 0.45 0.2 0.9 0.4815 0.21
Saliva mean 0.55 0 1.41 0.4817 0.45
No Saliva mean 0.6 0 1.26 0.52 0.43
The errors made with and without saliva compared to the intraoral measurements were calculated
(Table 2). Because the data was not normally distributed, we rely on the median to further
evaluate the data set. For the models with saliva, the median error was 0.017 mm while the
median error without saliva was 0.057 mm. These numbers are very small indicating that the
errors made when measuring models with and without saliva were not significant compared to
18
the intraoral measurements.
Table 2. Summary of errors (mm) for measurements on models with and without saliva.
Descriptives Median Min Max Mean SD
Saliva_error 0.017 -0.4 0.51 0.0032 0.28
No_saliva_err
or
0.057 -0.55 0.5 0.035 0.28
The Wilcoxon Signed-Rank test was used to compare the measurements with saliva to the
intraoral measurements (gold standard) as well as compare the measurements without saliva to
the intraoral measurements. For the models with saliva compared to the intraoral measurements,
the P-value was calculated to be 0.866. For the models without saliva compared to the intraoral
measurements the P value was 0.508.
The Mann-Whitney U test was calculated to be 0.714, which indicates that there is no significant
difference between the two groups: measurements of spaces with saliva and measurements of
spaces without saliva.
Table 3. Report of Wilcoxon Signed Rank test and Mann Whitney U test for all measurements.
Statistical analysis P values
Wilcoxon Signed Rank- Saliva 0.866
Wilcoxon Signed Rank- No Saliva 0.508
Mann Whitney U Test 0.714
19
Table 4. Intraoral measurements and the mean measurements of spaces with and without saliva.
Space Number Intraoral Measurement Mean Saliva Mean No Saliva
Space #1 0.4 0 0.476
Space #2 0.5 0.737 0.721
Space #3 0.7 0.717 1.06
Space #4 0.55 0.255 0
Space #5 0.9 1.41 1.262
Space #6 0.7 1.088 0.888
Space #7 0.55 0.613 0.702
Space #8 0.65 0.761 0.707
Space #9 0.6 0.83 1.005
Space #10 0.8 0.977 0.83
Space #11 0.3 0 0
Space #12 0.3 0 0
Space #13 0.3 0 0
Space #14 0.3 0 0
Space #15 0.3 0 0.208
Space #16 0.3 0.48 0.598
Space #17 0.7 1.12 1.2
Space #18 0.6 0.935 0.972
Space #19 0.3 0 0
Space #20 0.2 0 0
Space #21 0.3 0 0
Space #22 0.6 0.997 0.86
Space #23 0.8 1.008 1.073
Space #24 0.9 0.912 0.917
Space #25 0.2 0 0.183
Space #26 0.5 0.768 0.737
Space #27 0.3 0.262 0.282
Space #28 0.4 0.548 0.597
Space #29 0.45 0.607 0.603
20
Fig 10. Graph summarizing mean measurements taken on digital models of spaces with saliva
(red) and without saliva (yellow) compared to the intraoral measurements (blue). Spaces
measuring 0.3 mm or less were less accurately identified, regardless of the presence of saliva.
The spaces equal to or less than 0.3 mm were rarely identified by the orthodontists on the digital
models with and without saliva (Fig 10). Additionally, the measured spaces on the digital models
were consistently higher than the measurements taken intraorally. Figure 10 shows the
measurements with and without saliva (yellow and red) averaged higher values than the intraoral
measurement.
The descriptives of the error with saliva and without saliva for measurements 0.4 mm and greater
were calculated (Table 5). When the error for the saliva group and the no saliva group for
measurements 0.4 mm and greater were compared to the errors made for the entire data set
(Table 2.), the errors for these greater measurements were larger. The mean error for
measurements of spaces equal to or greater than 0.4 mm with saliva was 0.15 mm, while the
21
mean error of all spaces with saliva was 0.0032 mm. The mean error for measurements of spaces
equal to or greater than 0.4 mm without saliva was 0.17, while the mean error of all spaces
without saliva was 0.035 mm.
Table 5. Descriptives of the data for measurements 0.4 mm and above.
Median Min Max Mean Standard
Deviation
Saliva Error 0.17 -0.4 0.51 0.15 0.24
No Saliva
Error
0.19 -0.55 0.5 0.17 0.24
When measurements 0.3 mm and below were removed from the data set, the Wilcoxon Signed
Rank test for models with saliva was 0.023 with a P value of less than 0.05, indicating a
significant difference in measurements between these models and the gold standard. The
Wilcoxon Signed Rank test for models without saliva compared to the intraoral measurements
measured 0.0007 with a P value of less than 0.05, also indicating a significant difference between
these models and the gold standard. The Mann-Whitney U test was calculated to be 0.897, which
indicates that there is no significant difference between the two groups (Table 6.).
Table 6. Report with Measurements of 0.4 mm and above.
Statistical analysis P values
Wilcoxon Signed Rank- Saliva 0.023
Wilcoxon Signed Rank- No Saliva 0.0007
Mann Whitney U Test 0.897
Another hypothesis was tested that the distribution is the same across the three groups: intraoral
measurements (gold standard), saliva group and group without saliva. A Kruskal-Wallis test was
22
completed to test whether the samples had the same distribution. The type I error rate was 0.05.
A Post-Hoc test was also completed for this new hypothesis. Refer to the Appendix for results of
this test and the Post Hoc test.
One patient presented with an intraoral space measuring 0.3 mm between the lower left central
and lateral incisors (Fig 11.). All three measurers did not find a space between the lower left
central and lateral on the digital model generated in Meshmixer with no saliva present (Fig 12.).
Similarly the measurers did not find a space between the lower left central and lateral on the
digital model when saliva was present (Fig 13.).
Fig 11. Intraoral photograph of Patient #1 with spacing between lower left central and lateral
incisors after drying the dentition. Intraoral measurement of 0.3 mm.
Fig 12. Digital model of Patient #1 from 3Shape TRIOS
®
scan with no saliva present. All three
orthodontists did not find a space between the lower left central and lateral.
23
Intraoral
Measurement
Orthodontist A
Trial #1
Orthodontist A
Trial #2
Orthodontist B
Trial #1
Orthodontist B
Trial #2
Orthodontist C
Trial #1
Orthodontist C
Trial #2
0.3 mm 0 mm 0 mm 0 mm 0 mm 0 mm 0 mm
Table 8. Intraoral vs. model measurements after drying the dentition for patient #1 with space
between lower left central and lateral.
Fig 13. Digital model of Patient #1 from 3Shape TRIOS
®
scan with saliva present. All three
orthodontists did not find a space on this model between the lower left central and lateral.
Intraoral
Measurement
Orthodontist A
Trial #1
Orthodontist A
Trial #2
Orthodontist B
Trial #1
Orthodontist B
Trial #2
Orthodontist C
Trial #1
Orthodontist C
Trial #2
0.3 mm 0 mm 0 mm 0 mm 0 mm 0 mm 0 mm
Table 9. Intraoral vs. model measurements with saliva present for patient #1 with space between
lower left central and lateral.
Another patient presented with an intraoral space measuring 0.4 mm between the lower right
lateral incisor and canine (Fig 14.). When saliva was not present (Fig 15.), the measurers
measured the space on the digital model ranging from 0.23 mm to 0.69 mm (Table 10.). When
saliva was present (Fig 16.), the measurers did not recognize a space present on the digital model
uploaded to Meshmixer™ (Table 11.).
24
Fig 14. Patient #2 with spacing between the lower right lateral and canine, measuring 0.4 mm
intraorally.
Fig 15. Digital model generated of Patient #2 from a 3Shape TRIOS3
®
scan with no saliva
present. The three orthodontists measured space between the lower right lateral and canine
ranging from 0.23 mm to 0.69 mm.
Intraoral
Measurement
Orthodontist A
Trial #1
Orthodontist A
Trial #2
Orthodontist B
Trial #1
Orthodontist B
Trial #2
Orthodontist C
Trial #1
Orthodontist C
Trial #2
0.4 mm 0.398 mm 0.45 mm 0.40 mm 0.23 mm 0.69 mm 0.69 mm
Table 10. Intraoral vs. model measurements with no saliva present for patient #2 with space
between lower right lateral and canine.
25
Fig 16. Meshmixer™ model generated of Patient #2 from 3Shape TRIOS
®
scan with saliva
present. All orthodontists were not able to identify a space between the lower right lateral and
canine.
Intraoral
Measurement
Orthodontist A
Trial #1
Orthodontist A
Trial #2
Orthodontist B
Trial #1
Orthodontist B
Trial #2
Orthodontist C
Trial #1
Orthodontist C
Trial #2
0.4 mm 0 mm 0 mm 0 mm 0 mm 0 mm 0 mm
Table 11. Intraoral vs. model measurements with saliva present for patient #2 with space
between lower right lateral and canine.
Another patient presented with a space of 0.3 mm between the upper central incisors (Fig 17.).
The teeth were photographed with and without saliva present (Fig 17.). The next figure (Fig 18.)
When no saliva was present (Fig 18.), the measurers measured this space in Meshmixer™
ranging from 0 mm to 0.61 mm (Table 12.). When saliva was present (Fig 19.), the measurers
measured this space in Meshmixer™ ranging from 0 to 0.55 mm (Table 13).
A. B.
Fig 17. Photos of Patient #3 with saliva between the upper central incisors (A) and after drying
the space between the upper central incisors (B). This space measured 0.3 mm intraorally.
26
Fig 18. Meshmixer™ digital model of Patient #3 with no saliva present between the upper
central incisors. This space measured 0.3 mm intraorally.
Intraoral
Measurement
Orthodontist A
Trial #1
Orthodontist A
Trial #2
Orthodontist B
Trial #1
Orthodontist B
Trial #2
Orthodontist C
Trial #1
Orthodontist C
Trial #2
0.3 mm 0 mm 0.61 mm 0 mm 0 mm 0.55 mm 0.53 mm
Table 12. Intraoral vs. model measurements with no saliva present for patient #3 with space
between upper central incisors.
Fig 19. Meshmixer™ digital model of Patient #3 with saliva present between the upper central
incisors. This space measured 0.3 mm intraorally.
Intraoral
Measurement
Orthodontist A
Trial #1
Orthodontist A
Trial #2
Orthodontist B
Trial #1
Orthodontist B
Trial #2
Orthodontist C
Trial #1
Orthodontist C
Trial #2
0.3 mm 0.55 mm 0 mm 0 mm 0 mm 0.48 mm 0.54 mm
Table 13. Intraoral vs. model measurements with saliva present for patient #3 with space
between upper central incisors.
27
Discussion:
The effect of saliva on the accuracy intraoral scanning has not been investigated in the literature
(Renne et al., 2017). The objective of this prospective clinical study was to test the hypothesis
that if saliva is present in interproximal teeth area, spaces up to of 0.3mm will not be diagnosed
in intraoral scans leading to inaccuracies during the digital orthodontics workflow. Spaces or
diastemas may exist at the end of customized digital orthodontic treatment due several reasons.
One of these can be attributed to the inaccuracy of the diagnosis of small interproximal spaces
and the inaccurate selection of dental borders during the digital setup segmentation process. This
may be seen later in treatment represented by open interproximal spaces. These spaces are
unhygienic and unaesthetic and require further treatment. This leads to longer treatment time for
the patient and orthodontist as well as increased cost to the manufacturer of the treatment
refinement.
All values for Chronbach’s Alpha for the intra-operator and inter-operator reliability tests were
above 0.94, therefore the measurements taken by the measurers should be considered reliable.
This is interesting given the fact that spaces that appear clearly visible on the digital models (ie.
Fig 15) were not measured accurately by all three measurers. Additionally, often measurements
taken on digital models were much greater than the spaces intraorally and many spaces were not
consistently measured. An example of this is illustrated in table 14, space #15 (Appendix) when
measurer #3 measured 0 mm in trial #1 and then 0.63 mm in trial #2 for a space that measured
0.3 mm intraorally.
28
In this study, there was not a significant difference in the ability to diagnose spaces in the
presence or absent of saliva, when digital models were measured by an orthodontist and
orthodontic residents. The orthodontist and residents’ measurements tended to overlook or
overestimate the size of spaces when compared to the intraoral measurements (control group).
However, these differences were not statistically significant (P > 0.05). For the models with
saliva compared to the intraoral measurements (gold standard), the P-value was calculated to be
0.866, indicating that the presence of saliva did not cause the orthodontists to make more errors
when measuring the models. For the models without saliva compared to the intraoral
measurements the P value was 0.508, also indicating that the measurements of spaces on models
without saliva were not significantly different than the intraoral measurements. The presence of
saliva did not cause the orthodontists to make errors when measuring the spaces. Individual
digital models, however when compared side by side, such as with figures 15 and 16, showed
that models created from dry dentition were more defined and clearer than those generated from
teeth with saliva present.
When statistical analysis was completed on the data with only spaces 0.4 mm and greater
included, the spaces with and without saliva were significantly different than the intraoral
measurements (control group). The Wilcoxon Signed Rank test had p values less than 0.05 when
saliva was present and when saliva was absent. Therefore model measurements for spaces 0.4
mm and above, regardless of the presence of saliva, were significantly different than the intraoral
measurements. The Mann-Whitney U test of 0.897 means that there is no significant difference
between the two groups despite them being significantly different from the intraoral
measurements. Therefore it cannot be concluded that saliva was the reason for the disparity in
29
measurements from the digital models to the intraoral measurements but it can be concluded that
measurements made on the digital models were inaccurate from those made intraorally when
spaces were equal to or greater than 0.4 mm.
Many of the measurements taken on the digital models illustrate that the measurers
overestimated the amount of spaces when compared to the control group. Measuring larger
spaces than what exist intraorally can have negative effects on the virtual setup designed as well
as during the manufacturing of the customized appliances. Overestimating the size of dental
spaces during the segmentation of teeth for the construction of the virtual setup has the
possibility of increasing the programmed interproximal collisions between adjacent teeth. This
would lead to tighter interproximal contacts created by the appliance leading to more difficult
dental movements. This process is similar to interproximal reduction (IPR) executed virtually but
not clinically. This may explain the need for clinicians who use digital customized appliances
such as aligners or INBRACE to routinely check the interproximal contacts with dental floss in
order to alleviate the contacts with light IPR. It is understood that performing IPR where
collisions exist enhances tooth movement.
Interestingly, spaces on the digital model less than or equal to 0.3 mm were rarely identified,
regardless of the presence of saliva. For most of these spaces a connection was formed between
the teeth (ie. Fig. 12) rendering these spaces equivalent to teeth with true contact points. These
“contacts” were generally not recorded as spaces by the measurers. This can be attributed to the
inability of the 3Shape TRIOS3
®
scanner to capture and reconstruct small spaces between two
very close surfaces. This is significant for if a dental arch has multiple small spaces that are not
30
accurately illustrated on the STL file, then these spaces will not be closed in the digital treatment
setup. The patient will be left with many open contacts at the end of treatment. In addition, it
must be considered that the human eye may not be able to recognize spaces small than 0.3 mm.
These very small, undiagnosed spaces may be the reason for residual spaces at the end of
treatment when digital customized orthodontic appliances are utilized. To overcome this
problem, 0.2mm to 0.3mm of interproximal collision or “virtual power chain effect” is
recommended to be designed into the digital setups to avoid small spaces at the end of treatment.
With INBRACE, a customized lingual appliance, 0.2mm of interproximal collision between
contacts is programmed into the digital setup to account for these small undiagnosed spaces. This
ensures contacts are formed clinically at the end of treatment. Some clear aligner companies,
such as Invisalign also use this “virtual power chain effect” built into the aligner prescription to
avoid spaces at the end of treatment (Align Technology, 2020).
In this study, the results illustrated that the measurers often overestimated the measurements of
spaces greater than 0.3mm. This can be explained by software limitations of Meshmixer™,
including the difficulty of performing digital model measurements. In the Meshmixer™ software
program, when a small space exists and a false connection is visible between teeth on the digital
model, one must measure facial to the contact point or incisal to it. These distances adjacent and
incisal to the contact points are larger, leading to greater spaces measured. This pattern was
repeated multiple times in the data set. For example, space #2 measured 0.5 mm intraorally and
averaged 0.737 mm when saliva was present and 0.721 mm when saliva was not present (Table
4). Another example is space #17 that had an intraoral measurement of 0.7 mm and a mean
measurement of 1.12 mm when saliva was present and 1.2 mm when saliva was not present
31
(Table 4). This is a possible source of error in the data where model measurements were
drastically larger than those taken intraorally.
Additionally, when using Meshmixer™ to measure spaces, one is not able to click two separate
points from which to measure between. Instead, one must click within a space and a line is
automatically drawn between two objects. The model can be rotated and this measuring line
changed in order to find the most accurate measurement between two teeth (Fig 9.). The clarity
of Meshmixer™ was not compared to that of other software programs capable of dental
segmentation. It is possible that other digital software programs have better methods for space
measurement and tooth segmentation than Meshmixer™. It would be valuable to compare
Meshmixer™ with other segmentation software programs to determine the most accurate
program for manual segmentation.
It was also impossible to standardize the amount of saliva between patients. The amount of saliva
in the patients’ mouths was not quantified. Spaces scanned intraorally with saliva in a patient
who produces minimal saliva may have been equivalent to those spaces scanned after drying the
dentition. This lack of standardization has the possibility of affecting the data, for interproximal
areas believed to be coated with saliva may have been wide open and dry.
Another possible source of error is this clinical trial was the limited sample size. If the scale of
the study were enlarged, it is possible there would be more power to detect a potential difference
in the measurements of spaces with and without saliva. When looked at individually,
measurements for small spaces including those less than 0.3 mm were often recorded as 0 mm.
32
When these numbers were averaged together with larger measurements made above the contact
area, they often yielded larger numbers. It is important to note that any space larger than a
contact point left untreated does not follow Andrews’ Six Keys to Normal Occlusion. The fifth
key is for the dentition to have tight contacts between teeth in which no spaces exist (Andrews,
1972).
33
Conclusions:
In this study, the hypothesis that if saliva is present in the interproximal area between teeth,
spaces up to of 0.3mm will not be diagnosed, was confirmed. Spaces less than or equal to 0.3
mm were often not recognized with and without saliva on the digital models made from intraoral
scans with the 3Shape TRIOS3
®
Intraoral Scanner. Models did appear clearer after the teeth were
dried prior to the scan, although this did not affect the accuracy of measurements. Model
measurements for spaces 0.4 mm and above, regardless of the presence of saliva, could be
recognized but not accurately measured. They were significantly different than the intraoral
measurements. The diagnosis of small interproximal spaces still remains a challenge in the
orthodontic digital workflow when intraoral scans obtained with the 3Shape TRIOS3
®
Intraoral
scanner are used.
34
References:
1. Andrews, Lawrence F. "The six keys to normal occlusion." Am J Orthod 62.3 (1972):
296-309.
2. Autodesk (2019). Meshmixer Overview.
3. Align Technology (2020). Clinical Preference and Prescription Form. Retrieved from
https://learn.invisalign.com/spacing/treatmentplan
4. Ender, A., & Mehl, A. (2013). Influence of scanning strategies on the accuracy of digital
intraoral scanning systems. International journal of computerized dentistry, 16(1), 11-21.
5. Keogh, Thomas P. “Creating Tight, Anatomically Correct Interproximal Contacts.” New
Techniques in Esthetic and Restorative Dentistry, edited by Raymond L Bertolotti, 1st
ed., vol. 45, Elsevier, 2001, pp. 83–86.
6. Kihara, Hidemichi, et al. “Accuracy and Practicality of Intraoral Scanner in Dentistry: A
Literature Review.” Journal of Prosthodontic Research, Aug. 2019,
doi:10.1016/j.jpor.2019.07.010.
7. Kumar, Y., Janardan, R., & Larson, B. (2012). Automatic feature identification in dental
meshes. Computer-Aided Design and Applications, 9(6), 747-769.
8. Kumar, Y., Janardan, R., Larson, B., & Moon, J. (2011). Improved segmentation of teeth
in dental models. Computer-Aided Design and Applications, 8(2), 211-224.
9. Liao, S. H., Liu, S. J., Zou, B. J., Ding, X., Liang, Y., & Huang, J. H. (2015). Automatic
tooth segmentation of dental mesh based on harmonic fields. BioMed research
international, 2015.
10. Mangano, F., Gandolfi, A., Luongo, G., & Logozzo, S. (2017). Intraoral scanners in
dentistry: a review of the current literature. BMC oral health, 17(1), 149.
35
11. Park, Hye-Nan, et al. “A Comparison of the Accuracy of Intraoral Scanners Using an
Intraoral Environment Simulator.” The Journal of Advanced Prosthodontics, vol. 10, no.
1, Feb. 2018, pp. 58–64., doi:10.4047/jap.2018.10.1.58.
12. Papadimitriou, Aikaterini, et al. “Clinical Effectiveness of Invisalign® Orthodontic
Treatment: a Systematic Review.” Progress in Orthodontics, vol. 19, no. 1, 2018, pp. 1–
24., doi:10.1186/s40510-018-0235-z.
13. Renne, W., Ludlow, M., Fryml, J., Schurch, Z., Mennito, A., Kessler, R., & Lauer, A.
(2017). Evaluation of the accuracy of 7 digital scanners: An in vitro analysis based on 3-
dimensional comparisons. The Journal of prosthetic dentistry, 118(1), 36-42.
14. Taneva, E., Kusnoto, B., & Evans, C. A. (2015). 3D scanning, imaging, and printing in
orthodontics. Issues in contemporary orthodontics, 148.
15. Tomita, Y., Uechi, J., Konno, M., Sasamoto, S., Iijima, M., & Mizoguchi, I. (2018).
Accuracy of digital models generated by conventional impression/plaster-model methods
and intraoral scanning. Dental materials journal, 37(4), 628-633.
16. Tong H, Weissheimer A, Pham J, Lee R, Redmond R. Lingual Orthodontics Redefined
with Automation and Friction-Free Mechanics. Journal of Clinical Orthodontics.
2019;53(4):214-24.
17. Wu, K., Chen, L., Li, J., & Zhou, Y. (2014). Tooth segmentation on dental meshes using
morphologic skeleton. Computers & Graphics, 38, 199-211.
36
Appendix A:
Appendix table 1. Summary of all intraoral measurements and the average measurements of
spaces with and without saliva for each of the three measurers.
Space #
Intraoral
Measurem
ent
Measurer
#1: Saliva
Measurer
#1: No
Saliva
Measurer
#2: Saliva
Measurer
#2: No
Saliva
Measurer
#3: Saliva
Measure
r #3: No
Saliva
Space #1 0.4 0 0.42 0 0.32 0 0.69
Space #2 0.5 0.59 0.64 0.85 0.95 0.78 0.58
Space #3 0.7 0.67 0.73 0.95 1.36 0.53 1.1
Space #4 0.55 0 0 0 0 0.77 0
Space #5 0.9 1.01 1.215 1.63 1.37 1.59 1.2
Space #6 0.7 1.08 0.74 0.91 1.04 1.28 0.89
Space #7 0.55 0.48 0.65 0.63 0.79 0.735 0.67
Space #8 0.65 0.29 0.7 1 0.73 1 0.69
Space #9 0.6 0.84 0.77 0.81 1.03 0.85 1.23
Space #10 0.8 0.81 0.88 1.13 0.83 1 0.79
Space #11 0.3 0 0 0 0 0 0
Space #12 0.3 0 0 0 0 0 0
Space #13 0.3 0 0 0 0 0 0
Space #14 0.3 0 0 0 0 0 0
Space #15 0.3 0 0 0 0 0 0.63
Space #16 0.3 0.69 0.5 0 0.57 0.75 0.73
Space #17 0.7 0.87 1.21 0.96 1.17 1.54 1.23
Space #18 0.6 0.92 0.77 0.76 1 1.13 1.15
Space #19 0.3 0 0 0 0 0 0
Space #20 0.2 0 0 0 0 0 0
Space #21 0.3 0 0 0 0 0 0
Space #22 0.6 0.97 0.76 0.95 0.93 1.07 0.89
Space #23 0.8 1.05 0.92 0.88 1.26 1.1 1.04
37
Space #24 0.9 0.88 0.89 1.04 0.92 0.82 0.95
Space #25 0.2 0 0 0 0 0 0.55
Space #26 0.5 0.49 0.55 1.04 0.95 0.78 0.72
Space #27 0.3 0.28 0.31 0 0 0.51 0.54
Space #28 0.4 0.51 0.47 0.51 0.5 0.63 0.82
Space #29 0.45 0.56 0.41 0.495 0.54 0.77 0.86
Appendix table 2. Results from Kruskal-Wallis test.
Null Hypothesis Test Significance Decision
The distribution is the same
across the gold standard,
saliva group and no saliva
group
Independent-Samples
Krukal-Wallis Test
0.014 Reject the null
hypothesis
Appendix table 3. Results from the Post-Hoc test
Sample Adj. Significance
Intraoral vs. saliva group 0.036
Intraoral vs. non-saliva group 0.031
Saliva vs. non-saliva group 1.000
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
3D assessment of virtual bracket removal for modern orthodontic retainers: a prospective clinical study
PDF
Monitering of typodont root movement via crown superimposition of single CBCT and consecutive iTero scans
PDF
Prevalence of arch form types and tooth size discrepancy in a Vietnamese population
PDF
Bonding accuracy of a novel lingual customized orthodontic appliance (INBRACE™): an in-vivo study
PDF
Assessment of 3D surface changes following virtual bracket removal
PDF
Use of digital models to assess orthodontic treatment progress and identify deficiencies
PDF
A comparative study of caucasian and African American mandibular clinical arch forms
PDF
The mesiodistal angulation and faciolingual inclination of each whole tooth in three dimensional space post non-extraction orthodontic treatment
PDF
Use of digital models to assess orthodontic treatment progress
PDF
Premolar extraction influence on third molar angulation and retromolar length
PDF
An overview of the dental and skeletal relationships in individuals of Vietnamese ancestry
PDF
Comparative study of Caucasian and Hispanic mandibular clinical arch forms
PDF
Cephalometric norms for the Vietnamese: a clinical appraisal
PDF
Mandibular plane angle changes with or without premolar extraction treatment in adult orthodontics measured using 3-D cone beam technology
PDF
Three-dimensional assessment of tooth root shape and root movement after orthodontic treatment: a retrospective cone-beam computed tomography study
PDF
An assessment of orthognathic surgery outcomes utilizing virtual surgical planning and a patented full-coverage 3D-printed orthognathic splint
PDF
Evaluating the treatment efficiency of Carriere Distalizer: a cephalometric and study model comparison of Class II appliances
PDF
Orthodontic rotational relapse: prevalence and prevention
PDF
Shear bond strength comparison of mesh, sandblasted and laser-etched orthodontic brackets
PDF
Cirtual 3D placement of temporary orthodontic anchorage implants
Asset Metadata
Creator
Schwartz, Katherine
(author)
Core Title
The accuracy of intraoral scans of interproximal spaces with and without saliva: a prospective clinical study
School
School of Dentistry
Degree
Master of Science
Degree Program
Craniofacial Biology
Publication Date
04/03/2020
Defense Date
02/07/2020
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
accuracy,customized,digital,interproximal,intraoral,measure,Model,OAI-PMH Harvest,orthodontics,Planning,scanner,space,Treatment
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Sameshima, Glenn (
committee chair
)
Creator Email
katienasoff@gmail.com,schwartk@Usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-278219
Unique identifier
UC11675524
Identifier
etd-SchwartzKa-8242.pdf (filename),usctheses-c89-278219 (legacy record id)
Legacy Identifier
etd-SchwartzKa-8242.pdf
Dmrecord
278219
Document Type
Thesis
Rights
Schwartz, Katherine
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
Tags
customized
interproximal
intraoral
measure
orthodontics
scanner
space