Abstract. to the diagnosis of upper airway obstructions.


Abstract. The aim of this study was to
compare the measurement of the upper airway volume and minimum area using
airway analysis function in two software. The sample consisted of 11 cone-beam
computed tomography (CBCT) scans data, evaluated using the Invivo5 (Anatomage)
and Romexis (version 3.8.2.R, Planmeca) software which afforded image
reconstruction, and airway analysis. The measurement was done twice with one
week gap between the two measurements. The measurement obtained was analyzed
with t-tests and intraclass correlation coefficient (ICC), with confidence
intervals (CI) was set at 95%. From the analysis, the mean reading of volume
and minimum area is not significantly different between Invivo5 and Romexis.
Excellent intrarater reliability values were found for the both measurement on
both software, with ICC values ranging from 0.940 to 0.998. The results suggest
that both of these software can be used in further studies to investigate upper
airway, thereby contributing to the diagnosis of upper airway obstructions.

Keywords— cone-beam computed tomography; Invivo5 (Anatomage);
Romexis (version 3.8.2.R, Planmeca); airway analysis; airway volume and minimum



The airway is a system that
consists of tubes that conveys inhaled air from nose and mouth into lungs. The skeletal support for airway is superiorly
provided by the cranial base, posteriorly provided by spine, anterosuperiorly provided
by nasal septum, and anteriorly provided by jaws and hyoid bone. An obstruction of the upper airway will partially or
totally prevent air from getting into lungs. Airway obstructions can be minor or
life-threatening emergencies which require immediate medical attention. Airway
obstructions or encroachments became attentive because they increase airway resistance
that may lead to obstructive sleep-disordered breathing (OSDB). Therefore,
visualization and calculation of the airway dimensions are important. Common
airway obstructions are turbinates, adenoids, long soft palate, large tongue,
and pharyngeal and lingual tonsils. Less commonly found airway encroachments
include polyps and tumors.

Airway obstruction is not diagnosed with
imaging, however, imaging plays a role in the anatomic assessment of the airway
and adjacent structures as imaging can identify the patients with airways who
are at risk for obstruction and other anatomic characteristics that may
contribute to OSDB. Most studies evaluating the
airway have been conducted with 2-dimensional (2D) cephalograms, providing
limited data such as linear and angular measurements, for a complex
3-dimensional (3D) structure 2.The airway extending from the tip of
the nose to the superior end of the trachea can be visualized on conventional
computed tomography (CT) and cone beam CT (CBCT) scans.


CBCT and Image Analysis

CBCT systems have been developed
specifically for the maxillofacial region with the advantage of the reduced
radiation doses compared with conventional CT 3. Accurate and easy evaluation
of the airway anatomy has been possible using those CBCT systems 2. There are
many studies 7, 8, 9, 10, 11, 12 and 13 of the upper airway was analyzed or
assessed using CBCT. The next level up of CBCT
is the advanced software tools involve airway tracing features that give the
user the capability to delineate the airway’s boundaries, measure its volume,
and calculate and locate the Minimum-Cross-Sectional Area (MCA) 1. 

Although numerous
methods have been proposed for upper airway studies, most studies evaluating
the airway have been conducted with 2-dimensional (2D) cephalograms. However,
this method only provides data such as linear and angular measurements, for
3-dimensional (3D) structure. With the introduction of CBCT, the 3D diagnosis
of the patient became more accessible in dentistry. This technology allows the
segmentation and visualization of structures such as the airway in 3

The segmentation of
the airway can be done manually or automatically. Manual segmentation needs the
operator to delineate the airway slice by slice and render the data into a 3D
volume for analysis. Reference 6 has shown that the measurement of the 3D
airway from CBCT data using a semi-assisted software program is accurate,
reliable, and fast. While automatic segmentation can be done by differentiating structures
with different density values as done by 5 which applied a simple gray scale
thresholding based method to segment and measure the upper airway using CBCT.

segmentation of the airway is significantly faster and more practical than  manual segmentation, but the reliability,
reproducibility and the accuracy of the method with commercially available
programs are less be tested. The aim of this current study is to compare the
measurement of upper airway volume and minimum area using airway analysis
function in two software (Invivo5 (Anatomage) and Romexis (version 3.8.2.R,





This study was done
at School of Dental Sciences (PPSG), University Sains Malaysia, Health Campus,
Kubang Kerian Kelantan. 11 CBCT scans data were selected from the dental clinic
database system, School of Dental Sciences (PPSG). The CBCT scan was conducted
using the Promax 3D (Planmeca, Helsinki, Finland). All the 11 CBCT
scans image were analysed with airway analysis function using two software;
Invivo5 (Anatomage) and Romexis (version 3.8.2.R Planmeca).

In the Invivo5 software,
the airway was measured using the airway segmenting tool as in Figure 1. Then
the line was drawn down in the middle of the airway in a sagittal view. After
the line is drawn, the software will automatically detect the airway space
within the soft-tissue boundaries based on the Hounsfield Unit. Once the airway
space has been defined and the boundaries are well established, the volume of
the airway and the minimum area are automatically generated. The setting for
airway analysis can be found in ‘volume render’ menu as in Figure 2.

FIGURE 1. Display of invivo5 software for airway analysis in
‘section’ menu. the airway segmenting tools are shown by the arrow.

FIGURE 2. Display of Invivo5
software for airway analysis in ‘volume render’ menu. The airway segmenting
tools are shown by the arrow.


In Romexis version 3.8.2.R software, the airway was
measured using region growing feature (as in Figure 3). First, a cube was drawn
at the area of airway in a sagittal grayscale view using ‘to draw a cube’
button. Then the ‘3D region growing’ button was used to set parameter to be
used. In ‘3D region growing’ window, the ‘pre-set’ box was set as ‘air cavity’,
the threshold was set at 300, ticked at coloured by areas. Next step was
‘select the seed point’, this step was needed to allow Romexis to know what
type of density to be measured. Click on a space in the airway. Romexis then
rendered up the airway and displayed the air volume and the area of the airway.
In this software, the minimum area is not automatically displayed. To find the
minimum area, the axial view was scrolled until the minimum area was found.

FIGURE 3. Display of Romexis
(version 3.8.2.R) software for airway
analysis using region growing tools. The button of ‘to draw a cube’ and ‘3d
region growing’ are shown by the arrow.


The measurement was
repeated after one week. After all the measurement data was obtained, the data was
analyzed using IBM SPSS software (version 23) with t-test to compare the
measurement between software and ICC intrarater reliability test to assess the
consistency of measurements made by both software in measuring the same
quantity. The confidence interval was set at 95%. For intrarater raliability test, the
‘model’ used was ‘One-Way Random’. Bland & Altman plot was then plotted to
visualize the consistency between measurements.



Table 1 shows the mean, standard deviation (sd)
and the output from t-test analysis for two software. From the data below, the
mean airway volume and mean minimum area measurement from Romexis software is
higher compared to Invivo5 software. However, the standard deviations from
Romexis measurements are lower than Invivo5 software. It also found that the p-value (for volume
and minimum area) are 0.914 which is more than 0.05, therefore,  it can be conclude that the mean reading of
volume and min. area is not significantly different between Invivo5 and


1. T-test
for airway volume and minimum area.


Table 2 shows the
mean, standard deviation and output from Intrarater reliability test. The
correlation value are 0.998 (0.992, 0.999), 0.970 (899, 992), 0.976 (0.918,
0.993) and 0.984 (0.945, 0.996) with 95% CI for measurement of volume in
Invivo5, measurement of minimum area in Invivo5, measurement of volume in
Romexis and measurement of minimum area in Romexis. From the results obtained,
it shows that there is evidence for the repeatability of measurements between
two occasions for the software. A copy of the Bland and Altman plot for this
data is shown in the Figure 4 and Figure 5, which shows good agreement for most
cases. For volume measurement, 7 are nearer to zero, with no outlier for
Invivo5 and 8 are nearer to zero, but with one outlier for Romexis (Figure 4).
For measurement of minimum area, 10 are nearer to zero, but with one outlier
for Invivo5 and 7 are nearer to zero, but with one outlier for Romexis (Figure


Intrarater reliability test (ICC) for airway volume and minimum area.



4. Bland
& Altman plot of 1st and 2nd measurement of volume.



   FIGURE 5. Bland & Altman
plot of 1st and 2nd measurement of min. area.



There are currently
more than 15 third-party DICOM viewers mainly for orthodontics, implantology,
and oral and maxillofacial surgery are available commercially. Although the
reliability, repeatability and accuracy of CBCT machines have been evaluated,
testing the reliability of CBCT-related software has not gone further as they
differ in terms of the statistical test used.

In this study, 2
commercially available CBCT-software programs that use automatic segmentation
to calculate airway volumes were tested. From the t-test analysis, the p-value
is equal to 0.914 for both quantity measured. This means that there is no
significant different between two software for the airway volume and minimum
area. While for ICC test, the intrarater value is more than 0.90 indicating
excellent agreement. According to 15, the ICC value of 0.50 to 0.74 was good
and 0.75 and above is considered as excellent 15. So, the correlation values
obtained from this study indicate that they are reproducible. The results
obtained are supported by other studies 2, 3, 4 and 7. Reference 16 had use
Romexis software to measure the airway volume to find the correlation between
3D airway and 2D. They found that the correlation value between area in 2D and
volume in 3D are very high correlation 16. While for Invivo software, 10
had used this software to measure pharyngeal airway volumes in healthy children
with retrognathic mandible and those with normal craniofacial growth.

The measurement
from these two software differs slightly due to the fact that this 2 software
programs did not use the same methods for calculation of the airway volume and
the minimum area. In Invivo5, the segmentation of the airway base on the point
the user click on the airway space and the upper and lower level are follows
the shape of the airway. However, in Romexis, the segmentation was done base on
the region growing in a cube, thus the upper and lower level does not follow
the shape of the airway. This gives a slightly vary measurement for both
software. The Invivo5 software allows more control where the user can “sculpt
out” the desired airway volume from the rest of the 3D structures. User also
can adjust the brightness and opacity values, clean out the unwanted voxels
before calculating the final airway volume. The software also lets the user to
change the threshold values to obtain a solid airway volume. This also might be
the reason to why the measurement of volume using Invivo5 software is more
variable than Romexis software.

For automatic
segmentation, volume measurements should be done with proper technique and
diligence. This is because the measurement changes depend on the image
threshold chosen. This is proved by 2. The proper technique also important as
different position will significantly increase or decrease the measurement
11. A study had proved that the CBCT-based 3D analysis gives a better picture
of the anatomical characteristics of the upper airways and therefore can lead
to an improvement of the diagnosis 4. The automatic segmentation of the
airway imaged using CBCT is feasible and this method can be used to evaluate
airway cross-section and volume comparable to measurements extracted using
manual segmentation 5. Reference
3 had suggested
that the three-dimensional CBCT digital measurements of the airway volume and
the most constricted area of the airway are reliable and accurate. The use of
CBCT imaging for the assessment of the airway can provide clinically useful
information in orthodontics and for assessing the airway after surgery. This is
proved by 17 where they conclude that the use of point-based analysis (from
3D CBCT) measures are better explained the changes in clinical symptoms
compared to conventional measures.

 The Bland & Altman plot are created to compare the two
measurements that each provides some errors in their measure. The plot also
allows the identification of any systematic difference between the measurements
or possible outliers. The dotted horizontal lines represent the 95% confidence
limits (limits of agreement). Thus, if the differences between methods were
distributed normally, 95% of the differences from the bias in the sample are
expected to be between upper and lower limit of agreement. As the confidence limits
are not exceeded, it can be concluded that the repeatability of the method is
acceptable and the two methods
are considered to be in agreement and may be used interchangeably.