Is Cone-Beam Computed Tomography (CBCT) Effective in Arch Discrepancy Measurement? A Comparative Study Using a Digital Program
Materials and Methods
50 patients were selected: 27 women and 23 men from the Orthodontics Department of the Faculty of Medicine and Dentistry at the University of Valencia, Spain. The mean age of the patients was 30.22 years. CBCTs were undertaken on all patients due to the fact that they were to undergo orthognathic surgery. In addition, plaster cast study models were made for all of them.
Inclusion criteria were as follows:
1. Permanent dentition from the first molar to the first contralateral molar.
2. Absence of anomalies in number, shape and size.
3. Good quality study models.
4. Absence of large-scale occlusal restorations or prostheses.
The Digital Model employed was developed by a work group of the University of Valencia. Its reliability and reproducibility had previously been tested. In this method, the plaster cast models were scanned using a conventional scanner. Once the images were obtained, they were then stored in the computer and analysed using a computer program, as can be observed on the left-hand side of Figure1.
The CBCT scanner employed was the Dental Picasso Master 3D® (EWOO technology, Republic of Korea. 2005), belonging to the Faculty of Medicine and Dentistry at the University of Valencia. All patients were scanned in maximum intercuspation, without inserting a wax bite, so avoiding appliances during the subsequent fragmentation.
The dimensions for the full head scan were 200×150 mm (12bits) over 15 seconds. Slice thickness was 0.1mm and scanning covered 360º. The field of view (FOV) used generated 496 images, with a voxel size of 0.4mm. In addition, a 50 kV tube voltage range and a 6 mA intensity range were used. The main radiation dose was 150 mSv.
The InVivoDental (Anatomage, San Jose, California) program was used for analysing the images from the CBCT. Once obtained from the CBCT, they were securely sent in DICOM format through the web page of the InVivoDental company, where they were segmented manually by a member of that company’s staff, so as to obtain images of the three-dimensional models as can be observed on the right-hand side of Figure1.
Once the sample was obtained, a single previously-trainedoperator proceeded to undertake dental measurementson each of the models described: Digital and Three-Dimensional. Having obtained these, arch discrepancymeasurement was then calculated.
In order to obtain arch discrepancy we measured toothsize in each arch between first molars. Then, we measuredarch length. And finally, we subtracted the sum of toothsize with the arch length. In both models, the softwareand the computed calculate arch discrepancy automatically,once we measured tooth sizes and arch discrepancy.
All the measurements obtained were introduced into aspreadsheet and analysed using the SPSS v. 15.0 statisticsprogram for Windows.
The intraobserver and interobserver error were calculatedfor tooth sizes, as they are the basis for calculating archdiscrepancy measurement. In order to calculate theintraobserver error, 15 of the 50 patients were randomlyselected and a single observer measured the toothsizes three times at intervals of at least one week.Reproducibility of Three-Dimensional Models was 1.08%for tooth sizes and 1.1% for Digital Models. The twomethods were, therefore, perfectly comparable.
The interobserver error was also calculated. To do so, asecond well-trained observer undertook the measurementson three occasions at intervals of at least one week.Reproducibility for tooth sizes of the Three DimensionalModels was 1.2% and 1.4 for Digital Models.
The correlation between the variables of both methodswas determined by means of Pearson’s correlationcoefficient, the slope and ordinate at origin and theirrespective confidence intervals of 95%. Table 1 shows thedata for upper and lower arch discrepancy measurement,and both jointly. Pearson’s correlation coefficients werevery high for all of them (0.924, 0.974 and 0.972 forupper and lower arch discrepancy measurement, andjointly). Figure 2 shows the line of fit for all these values.
Table 1. Ordinate at origin and slope with their respective confidence levels (95%) and r-Pearson correlation coefficients for the superior, lower and joint arch discrepancy measurement.
Figure 2. Dispersion diagram (%): of the Digital Method (ordinate axis) and of the Three-Dimensional Method or CBCT (abscissa
axis), of upper osteo-dental discrepancy (in orange), and lower (in blue).
The difference between both Methods was calculated as the differences between the mean values calculated using each Method (Three-Dimensional -Digital). The mean differences of the upper and lower arch discrepancy measurement with their standard deviations and their respective confidence intervals at 95% are shown in Table 2 and a graphic representation of these differences in Figure 3.
Figure 3. Bar chart of different means (mm) of osteo-dental discrepancy between both Methods.
The arch discrepancy measurements being positive and negative, we decided to calculate the mean of the absolute differences, in order to avoid error compensation. The mean absolute differences of upper and lower arch discrepancy measurement with their standard deviations and respective maximum and minimum intervals are shown in Table 3 and a graphic representation of these differences in Figure 4.
However, the advantages of CBCT should be assessed in terms of additional cost compared to the conventional radiograph. Segmentation of the models further increases the cost (the expenses of InVivo segmentation are around 70 dollars per patient). It would be better if we could measure directly form the CBCT avoiding segmentation, but nowadays this is not possible. We can not obtain good images for measuring tooth sizes.
Moreover, the use of CBCT exploration exposes the patient to ionizing radiation. In those patients with implants, prostheses, amalgams… image quality is not as accurate.Lastly, CBCTs are not justified for all orthodontic patients.So we could conclude that arch discrepancy could be calculated in 3D models in those patients who had a CBCT because is necessary for his diagnosis and treatment plan. In the remaining patients, we can continue using other type of study models.
The conclusions of the study are as follows:
CBCT allows us to determine upper and lower arch discrepancy measurement accurately and reproducibly in comparison to measurements obtained from Digital Models, which, in turn, are obtained from the digitalization of traditional plaster cast models. The differences existing between both Methods are clinically acceptable.
5.Tarazona B, Llamas JM, Cibrian R, Gandia JL, Paredes V. A comparison between dental measurements taken from CBCT models and those taken from a Digital Method. Eur J Orthod. 2013, 35 (1): 1-6.