Keywords
CBCT - All-on-Four - precision - degree of deviation
Introduction
The cone-beam computed tomography (CBCT) machine is a well-established adjunctive
diagnostic, virtual simulation, and treatment planning tool with numerous clinical
applications in multiple disciplines. This device can produce three-dimensional images
with a significant reduction in radiation exposure, a shorter scan time, and lower
costs compared with conventional CT.[1] The information obtained from CBCT scans allows the measurement of bone density,
height, and the buccolingual width of the alveolar bone at any specified jaw location[2]; allows a comprehensive understanding of accurate jaw dimensions and anatomical
structures, decreasing potential risks and dramatically enhancing predictability of
treatment results[3]; offers multiple advantages compared with traditional two-dimensional radiography,
including a lack of superimposition, 1:1 measurement, absence of geometric distortions,
and three-dimensional (3D) display.[4] The efficiency of CBCT machines is on the rise, and there is also a rapid growth
in current software packages designed for processing, managing, and analyzing 3D images.
Moreover, CBCT is quickly spreading and has become the imaging modality for planning
implant placement.[5] Virtual treatment planning includes using software (primary or third party) available
with CBCT images, which allows virtual implant planning that can be transferred to
the surgical site directly by using image-guided navigation or indirectly via the
construction of a surgical guide.[6]
CBCT images are appraised to allow highly accurate and reliable linear measurements.
Furthermore, the accuracy of reformatted CBCT images is affected by several factors.
These include the characteristics of the machine itself (e.g., nominal resolution
and image quality), radiation exposure (kV, mA, and the number of basis images, the
software for reconstruction and dimensional measurements, patient motion artifacts,
and the limitations of the clinician's interpretation. To determine the best application
of CBCT in dentistry, it is necessary to analyze the accuracy of the data acquired
related to distance measurements.[7] CBCT is considered to yield accurate volumetric data, and its multiplanar reformatted
images are generally regarded as reliable for linear measurements. However, it is
important to acknowledge that the accuracy of these measurements may be influenced
by a variety of factors, including metallic artifacts, patient motion, and device-specific
exposure parameters.[1]
[8] However, there is no universal standard considering the exposure geometry and parameters
for CBCT scanners.[3] As previously mentioned, the presence of metallic objects will likely show artifacts
and will thus jeopardize the image quality and hamper the visualization of the implant–bone
surface.[9]
[10] Therefore, it is still questionable whether the measurement performed near dental
implants is accurate.[11] Parameters of the CBCT devices may be of utmost importance to achieve better image
quality and enhance evaluation accuracy.[11]
[12] The accuracy of measurements must be defined as CBCT imaging is commonly used to
determine linear dimensions in different clinical dental applications.[1] The anatomic radiographic fidelity of bone structures and the accuracy of linear
measurements are pivotal for primary preoperative implant planning and even more so
when applied in guided implant surgery imaging.[13] Guided surgery systems are not perfectly precise, which can cause deviations in
both horizontal and vertical directions from the planned implant site to the virtual
position before the surgery.[14]
A plethora of CBCT machines are available in the dental market. These different machines
use different acquisition times, resulting in different effective radiation doses
for the patient. They differ in three significant properties: field of view, voxel
size, and focal point. The success and survival of implants crucially depend on thorough
diagnosis and treatment planning, which can be achieved by using a CBCT machine.[2] The increasing need for dental implants to replace missing teeth requires a sensitive
technique to obtain highly accurate alveolar and implant site measurements to assist
in treatment planning and avoid damage to the adjacent vital structures during surgery.[6] The rapid advancement of CBCT scans and computed-aided implant planning programs
is believed to enhance CBCT technology and improve implant placement accuracy.[15]
Knowledge of the measurement accuracy of CBCT images using software programs is essential
for validating their use for examining implant sites and for understanding and associating
all possible sources of error in the multistep and complex implant surgery process.
Such complex treatment planning sequences may result in potential errors, so it is
crucial to address the issue of possible deviations between the preoperative plan
and the postoperative implant location.[16]
Most in vivo clinical studies rarely quantify measurement accuracy, as this would require an intervention
to control the radiographic measurements themselves.[17] Thus, this study aims to investigate the reliability of different available CBCT
machines for digital implant planning by retrospectively evaluating the precision
of position and dimension of previously placed implants recorded from those CBCT machines.
Significant deviation from reality will affect implant planning as distances measured
from the CBCT images are crucial, such as proximity to adjacent implants, teeth, and
vital structures. Two null hypotheses were formulated for this study. The first hypothesis
was that there would be no significant difference in the dimensions of the All-on-Four
implant zone between different CBCT machines, and the second hypothesis was that there
would be no significant difference in the total degree of deviation of the All-on-Four
implant position between different CBCT machines.
Materials and Methods
Sample Preparation and CBCT Imaging
An artificial mandibular jaw (Straumann, Switzerland) was selected to have four implants
(4.5 × 10 mm Superline II, Dentium, South Korea) placed in an All-on-Four concept.
The implants were placed freehand following the manufacturer's instructions. Three
different CBCT machines of variable voxel size were chosen for this experiment (Rainbow
CT, Dentium, South Korea; Veraview X800, Morita, Japan; Planmeca Viso G3, Planmeca
OY, Finland). CBCT machines' properties are listed in [Table 1]. A putty mix (Easy Putty, Variotime, Germany) was used to stabilize the jaw on the
CBCT mount for imaging. The jaw was radiated 10 times per CBCT, following the manufacturer's
instructions. A total of 30 Digital Imaging and Communications in Medicine (DICOM)
files were exported, n = 10.
Table 1
CBCT machines' properties
|
Properties
|
Rainbow CT
|
Vera view ×800
|
Planmeca Viso G3
|
|
FOV
|
Ø 5 × 5, 16 × 10, 16 × 18 mm (stitching)
|
Ø 170 × 120 mm
Ø 170 × 50 mm
Ø 140 × 100 mm
Ø 140 × 50 mm
Ø 100 × 100 mm
Ø 100 × 50 mm
Ø 80 × 80 mm
Ø 60 × 60 mm
Ø 40 × 40 mm
|
Ø 40 × 50 mm
|
|
Focal spot (mm)
|
0.5
|
0.5
|
0.5
|
|
Voxel size
|
300 µm (high 200 µm)
|
80, 125, 160, 200, 250 µm
|
150–300 µm
|
|
Tube voltage
|
60–100 kVp
|
60–90 kV
|
90 kV
|
Abbreviations: CBCT, cone-beam computed tomography; CT, computed tomography; FOV,
field of view.
Implant Segmentation
All-on-Four implants from each DICOM file were segmented and exported as an STL file
(3D image) using Blue Sky Plan software (version 4.12.13/Blue Sky Bio, United States).
The segmentation was manually performed, and the density was set at 3,000 HU. Since
the jaw was radiated every time in the same position per CBCT machine, this ensures
perfect alignment on all DICOM files of each CBCT machine, hence, accurate measurement.
The segmentation was then exported as an STL file for analysis.
Dimensional Discrepancy
The analysis was performed using Autodesk Meshmixer software (version 3.5.474/California,
US). Analysis was applied by measuring the All-on-Four implant zone using the X, Y, and Z axes in the software to investigate the discrepancy in each dimension. The X-axis represents the mesiodistal dimension, the Y-axis represents the apico-occlusal dimension, and the Z-axis represents the buccolingual dimension. The recorded readings included 90 readings,
10 for each CBCT machine in each dimension.
Total Degree of Deviation Analysis
The total degree of deviation of All-on-Four implants was measured to evaluate the
precision of the All-on-Four implant position of each CBCT imaging. Measurement was
done using Autodesk Meshmixer software (version 3.5.474). The deviation was measured
between the implants' 3D images of each CBCT machine. Forty-five deviation readings
were recorded for each CBCT machine, for a total of 135 readings. A diagram summarizing
the methodology is shown in [Fig. 1].
Fig. 1 A diagram of the methodology.
Statistical Analysis
The data distribution was tested for normality, and the variances were tested for
equality with Shapiro–Wilk's and Levene's tests, respectively (p-value < 0.05). The data were only normally distributed in the Y-axis values (p-value > 0.05), and variances of the values were not equal either in the implant zone
analysis or in the degree of deviation analysis (p-value < 0.05); hence, Brown–Forsythe one-way analysis of variance and Tamhane's post
hoc tests were chosen to compare the differences between the groups (p-value <0.05). The analysis of statistics was performed by utilizing IBM SPSS 22 (SPSS
Inc., Chicago, Illinois, United States).
Results
Despite maintaining a constant jaw position and machine setting for 10 exposures,
there was a noticeable variation in the amount of distortion present in the Planmeca
CBCT and Dentium images. Morita exhibited the highest degree of distortion, followed
by Dentium and Planmeca, resulting in a longer segmentation process. Morita and Dentium
images boasted high resolution and clear implant surface details, resulting in simplified
segmentation. Conversely, Planmeca images exhibited significantly lower resolution
as illustrated in [Figs. 2]
[3] to [4].
Fig. 2 Cone-beam computed tomography images of the specimen taken by Planmeca. (A) Three-dimensional image; (B) horizontal section; and (C) implant cross-section.
Fig. 3 Cone-beam computed tomography images of the specimen taken by Morita. (A) Three-dimensional image; (B) horizontal section; and (C) implant cross-section.
Fig. 4 Cone-beam computed tomography images of the specimen taken by Dentium. (A) Three-dimensional image; (B) horizontal section; and (C) implant cross-section.
Implant Zone
The respective X, Y, and Z mean dimensions of the All-on-Four implant zone were: Dentium (34.95, 14.71, and
9.97); Morita (34.88, 14.74, and 10.56); and Planmeca (34.73, 15.15, and 12.33) ([Table 2]). Planmeca readings exhibited the least consistency in all axes (standard deviation
[SD] 0.16–0.35). Significant differences between CBCT machines were found in all axes
(p-value < 0.05); however, the Z-axis had the most discrepancy ([Table 3]). Only the Planmeca–Dentium comparison was significant in the X-axis; Morita–Planmeca and Planmeca–Dentium comparisons were significant in the Y-axis; the Z-axis showed significant differences in all comparisons. [Figs. 5]
[6] to [7] illustrate the distribution of the values and means and spot any outliers. Only
two outliers were spotted, one on the Y-axis and one on the Z-axis.
Table 2
Descriptive statistics of All-on-Four implant zone analysis
|
Axis
|
CBCT machine
|
N
|
Mean
|
Standard deviation
|
|
X
|
Dentium
|
10
|
34.95
|
0.08
|
|
Morita
|
10
|
34.88
|
0.06
|
|
Planmeca
|
10
|
34.73
|
0.23
|
|
Total
|
30
|
34.85
|
0.16
|
|
Y
|
Dentium
|
10
|
14.71
|
0.07
|
|
Morita
|
10
|
14.74
|
0.01
|
|
Planmeca
|
10
|
15.15
|
0.17
|
|
Total
|
30
|
14.87
|
0.23
|
|
Z
|
Dentium
|
10
|
9.97
|
0.02
|
|
Morita
|
10
|
10.56
|
0.09
|
|
Planmeca
|
10
|
12.33
|
0.35
|
|
Total
|
30
|
10.95
|
1.04
|
Abbreviation: CBCT, cone-beam computed tomography.
Table 3
Means comparison of implants zone analysis
|
Axis
|
CBCT machine
|
Mean difference
|
Standard error
|
Significance
|
|
X (mesiodistal)
|
Dentium
|
Morita
|
0.07
|
0.03
|
0.11
|
|
Morita
|
Planmeca
|
0.15
|
0.07
|
0.20
|
|
Planmeca
|
Dentium
|
0.22
|
0.07
|
0.04[a]
|
|
Y (apico-occlusal)
|
Dentium
|
Morita
|
0.03
|
0.02
|
0.49
|
|
Morita
|
Planmeca
|
0.40
|
0.05
|
0.00[a]
|
|
Planmeca
|
Dentium
|
0.44
|
0.06
|
0.00[a]
|
|
Z (buccolingual)
|
Dentium
|
Morita
|
0.59
|
0.03
|
0.00[a]
|
|
Morita
|
Planmeca
|
1.77
|
0.11
|
0.00[a]
|
|
Planmeca
|
Dentium
|
2.36
|
0.11
|
0.00[a]
|
Abbreviation: CBCT, cone-beam computed tomography.
a The mean difference is significant at the 0.05 level.
Fig. 5 Box plot of the All-on-Four implant zone analysis in the X-axis of the tested cone-beam computed tomography machines.
Fig. 6 Box plot of the All-on-Four implant zone analysis in the Y-axis of the tested cone-beam computed tomography machines. One outlier was spotted
in Dentium.
Fig. 7 Box plot of the All-on-Four implant zone analysis in the Z-axis of the tested cone-beam computed tomography machines. One outlier was spotted
in Morita.
Degree of Deviation
As shown in [Table 4] and [Fig. 8], Dentium exhibited the lowest deviation in the implant position (0.61), while Morita
showed the worst (0.85); Planmeca had the most consistent readings (SD 0.13). A significant
difference in the total degree of deviation was spotted when only Morita was included
in the comparison ([Table 5]).
Table 4
Descriptive statistics of total degree of deviation analysis
|
CBCT machine
|
N
|
Mean
|
Standard deviation
|
|
Dentium
|
45
|
0.61
|
0.32
|
|
Morita
|
45
|
0.85
|
0.40
|
|
Planmeca
|
45
|
0.66
|
0.13
|
|
Total
|
135
|
0.71
|
0.32
|
Abbreviation: CBCT, cone-beam computed tomography.
Fig. 8 Box plot of the All-on-Four implant total degree of deviation of the tested cone-beam
computed tomography machines.
Table 5
Means comparison of total degree of deviation analysis
|
Machine
|
Mean difference
|
Standard error
|
Significance
|
|
Dentium–Morita
|
0.24
|
0.07
|
0.00a
|
|
Morita–Planmeca
|
0.19
|
0.06
|
0.01a
|
|
Planmeca–Dentium
|
0.04
|
0.05
|
0.73
|
a The mean difference is significant at the 0.05 level.
Discussion
The authors have found variations in the accuracy between the tested CBCT machines.
Differences were spotted in the All-on-Four implant zone dimensions and the total
degree of deviation between the different CBCT machines; thus, both null hypotheses
were rejected. These findings influence implant planning, which relies mainly on the
accuracy of CBCT readings, especially when the digital approach is favored. Errors
are still inevitable in digital implant planning; however, trivial errors should not
affect the outcome significantly, especially when they are counted during the plan.
Reflection 3D readings are of significant influence on the success of digital planning
and guided implant placement.
There are many articles evaluating CBCT accuracy; however, based on this, the authors
have decided to generate a novel method for assessing the CBCT of previously placed
implants retrospectively by evaluating the discrepancy in the three dimensions of
the overall implant zone and also the total degree of deviation. The analytical interpretation
of All-on-Four implant zone includes the following: readings recorded from the Planmeca
machine were remarkably deviated and less consistent in all dimensions (X, Y, and Z axes) compared with the other CBCT machines; Morita exhibited the most consistent
values through all axes; the means of Morita and Dentium were closer to each other
in general in all axes compared with that of Planmeca; the Z-axis showed the least intramachine discrepancy (narrowest SDs), followed by the Y and X axes, respectively; however, it showed the most intermachine discrepancy (widest
SDs) ([Table 2]; [Figs. 5]
[6]
[7]). It can be noticed that when the implant zone means comparison was significant,
Planmeca and Dentium were involved, except in the Z-axis, where all comparisons were significant ([Table 3]). The total degree of deviation analysis showed that Dentium exhibited the lowest
deviation in the All-on-Four implants position, with the most consistent readings.
In contrast, the Morita machine exhibited the highest deviation, with the least consistent
readings ([Table 4], [Fig. 8]). A significant difference in the total degree of deviation was spotted when only
Morita was included in the comparison ([Table 5]). Upon comparison of the two analyses, it is evident that Morita presents the least
dimensional discrepancy. However, it is worth noting that it also exhibited the least
favorable readings regarding implant position. Such variations hold the potential
to result in clinical complications. As such, it is recommended to exercise caution
and maintain a 2-mm safety margin near adjacent anatomic structures when using CBCT.
The following studies shed light on the accuracy of linear measurements obtained from
various CBCT machines and highlight the importance of maintaining safe distances while
placing implants. Azhari (2019)[7] evaluated the amount of linear measurement inaccuracy of four different CBCT machines
compared with clinical measurement. His study showed that three CBCT machines did
not exceed a clinically acceptable threshold of 1 mm when compared with each other.
Similarly, Kosalagood et al (2015)[3] investigated the accuracy of linear measurements from multiple CBCT devices and
found that the radiographs slightly underestimated the actual distance when compared
with physical measurements. Wanderley et al evaluated dimensional changes in dental
implants using 13 CBCT devices, each set to specific scanning protocols. They found
that the visualization of the implant's dimensional changes varied across the CBCT
devices and scanning protocols. Specifically, they observed an increase in diameter
ranging from 0.27 to 1.04 mm.[18] Finally, a recent study explored the differences among four types of 3D X-ray machines
used for implant planning and found that each machine had deviations from the exact
measurements, emphasizing the importance of maintaining safe distances while placing
implants.[19] The investigations made could potentially hold significant weight for clinicians
and researchers who rely on CBCT machines for implant planning and treatment purposes.
Several previous studies also showed that the linear measurements of CBCT images from
other CBCT machines underestimate the actual distance. Underestimation is considered
clinically safer than overestimation as it will preserve the vital structures when
placing the dental implants. It was indicated that the difference in voxel size has
no effect on the accuracy of linear measurements.[20] For example, Bohner et al[20] conducted a CBCT study of dry mandibles using voxel sizes of 0.2, 0.25, and 0.4 mm,
revealing no statistical difference between the image measurements. The study suggested
protocols using CBCT images with large voxel sizes (0.3 and 0.4 mm) were preferable
for evaluating linear measurements for implant treatment planning due to the lower
radiation dose. Therefore, the study's authors recommend that the protocol that uses
the lowest effective dose is preferable for linear measurements in multiple implant
planning. Additionally, in the work of Waltrick et al,[21] the accuracy of linear measurements and visibility of the mandibular canal on CBCT
images were evaluated using varying voxel sizes. The experiment involved scanning
12 dry human mandibles using voxel sizes of 0.2, 0.3, and 0.4. The results indicated
an average SD error between measurements on images and direct measurements of 0.23
± 0.20 mm. Additionally, the CBCT measurements underestimated direct measurements
in 390 cases. On the other hand, in 2022, the Kehrwald et al's[22] study reported no substantial variation in measurements obtained with different
voxel sizes. Their study was performed to investigate the effect of voxel size on
CBCT images used for dental implant planning. This was done by utilizing synthetic
human mandibles with different degrees of bone resorption. Digital calipers were used
for each mandible to measure the bone thickness and height. It may be recommended
that larger voxel images (0.40 mm) be used when necessary for bone thickness and height
measurements without negatively impacting the patient's clinical planning quality.
Accurately measuring CBCT 3D images can be challenging due to metal distortion. This
study has encountered this issue. It was found that while Planmeca images had high
distortion levels, CBCT images from the same machine exhibited noticeable variability
in distortion levels. On the other hand, Morita images displayed clearer implant threads,
but higher distortion levels resulted in a more time-consuming and difficult segmentation
process. Research by Gurjar et al in 2024[23] analyzed the precision of CBCT in implant-supported prostheses and investigated
metal artifacts in the presence or absence of implants or prostheses. The study assessed
accuracy and artifacts at three points on the buccal and lingual cortical plates on
the mandible's body near the crest and base, using physical and radiographic measurements.
The results showed that CBCT artifacts were most prominent in full-arch prostheses,
while single implants with a prosthesis produced the least artifacts. Although metal
in implants or implant-supported prostheses can negatively impact the accuracy of
peri-implant area assessment, the study suggests using lower voxel integration scales
in the presence of implants or implant prostheses for more precise measurements. These
findings are significant as they will aid in developing better strategies to minimize
metal distortion and enhance the precision of CBCT 3D image measurements.
It is important to recognize that the segmentation process is subject to limitations
due to the variability in CBCT machines, potentially leading to variable degrees of
metal distortion and introducing potential bias. Also, as there was no reference file,
the authors could not test how accurate the implants' positions were compared with
reality. The alignment of segmented implants with digital analogs extracted from a
digital impression of the implants could have provided additional reliability to this
study. As a result of metal distortion, the distortion of implant surface details
was unavoidable, compromising the accuracy of segmentation and potentially impacting
precise alignment. It is worth considering that these limitations may compromise the
accuracy of the recorded measurements.
Overall, these studies have contributed to our understanding of the accuracy and precision
of linear measurements obtained through CBCT imaging. They highlighted the importance
of selecting and using the appropriate CBCT device to obtain reliable and accurate
measurements. Such knowledge can be helpful in enhancing the quality of diagnosis,
treatment planning, and patient care.
Conclusion
After analyzing the data, it can be concluded that Planmeca had the least dimensional
accuracy among the tested CBCT machines. The most significant difference between the
machines was in the buccolingual dimension (Z-axis), while the smallest difference was in the mesiodistal dimension (X-axis). In terms of All-on-Four implant position accuracy, Dentium had the highest
accuracy, while Morita had the lowest. However, there was no statistically significant
difference between the accuracy of Dentium and Planmeca machines. This new information
can aid clinicians in achieving a more predictable outcome when precise full arch
implant placement is crucial.
