Effect of saliva isolation and intraoral light levels on performance of intraoral scanners


The use of digital models in orthodontics is becoming increasingly widespread. This study aimed to evaluate the accuracy and performance of digital intraoral scanning under 4 different intraoral environmental conditions.


Four digital models were acquired with TRIOS intraoral scanner (3Shape, Copenhagen, Denmark) for 50 subjects. A total of 200 digital models were divided into 4 groups as follows: daylight and saliva (group 1), daylight with saliva isolation (group 2), reflector light and saliva (group 3), and relatively dark oral environment and saliva (group 4). The 4 digital models were superimposed, and the edges of the models were trimmed to create common boundaries (Geomagic Control X; 3D Systems, Rock Hill, SC). Group 2 models were used as a reference and superimposed separately with the models of the other 3 groups. Deviations between corresponding models were compared as means of negative deviation, means of positive deviation, in total area, out total area, positively positioned areas, and negatively positioned areas. In addition, all groups were compared in terms of scanning time, the total number of images, and the mesiodistal width of teeth.


Overlapping of group 1 with the reference model (group 2), a surface deviation of 13.1% (out total area) was observed. This analysis revealed that a 13% deviation was caused by the presence of saliva alone. This rate was 12.6% in group 3 and 15.5% in group 4, respectively. The values for means of negative deviation were −55 μ in group 1,−63 μ in group 3, and −68 μ in group 4. Means of positive deviation values were distributed among groups as follows: 68 μ in group 1, 69 μ in group 3, and 78 μ in group 4. The total number of images was observed, at least in group 4.


The intraoral scanner performance was affected by different environmental conditions, and that caused variations on the surface of digital models. However, the performance of the intraoral scanner was independent of the scanning time and mesiodistal width of the teeth.


  • The use of digital models in orthodontics has become increasingly widespread.

  • Light levels and saliva affect the performance of intraoral scanners.

  • Scanning time is not affected by saliva or light level.

  • Some environmental conditions cause variations on the surface of digital models.

Traditional plaster models are routinely used by orthodontists for diagnosis and treatment planning. These models are convenient for 3-dimensional (3D) dentoalveolar diagnostic analysis. However, classical plaster models have disadvantages such as breakability, storage problem, cost, and laboratory procedure. In the 21st century, named digital age, technological devices have developed very rapidly, and the usage of plaster models has been replaced by 3D intraoral scanners (IOS). , A digital impression is obtained slower than the current alginate impression system because of longer scanning time, but patient compliance and acceptance are higher in the digital impression, , and it enables easy digital storage of data. In addition, 3D images obtained with these scanners might be used for digital model setup, customized bracket, and specialized appliance construction such as expanders, aligners, retainers. These advantages have made IOS widespread in the last 2 decades.

Thanks to specialized software, various analyses such as Bolton and Hayce-Nance could be made on digital study models. In recent studies, the researchers tested the accuracy and reproducibility of these 3D images by using the software. The investigations were conducted either by comparing the performance of 2 different devices or by comparing the traditional plaster model to the IOS. ,

Many factors affect the performance of intraoral scanners such as the presence of saliva or water, scanning path, increased curvature or complex geometry of an object, hidden surfaces, and patient head movements. For instance, the presence of thick saliva or water film on the tooth surface may affect the accuracy of the measurements, and it can cause negative results especially on the margin of the digital impression ( Fig 1 ). Before assessment of the accuracy and reproducibility of IOS, it is necessary to understand the working principles of these devices and image acquisition mechanisms that will be described below.

Fig 1
Schematic description of the effect of saliva on refraction. P1 , predicted point; P2 , real point.

This study was based on the null hypothesis that saliva isolation and intraoral light level do not affect the performance of IOS devices in terms of accuracy, speed, and image quality. It was thought that this kind of study would allow more efficient usage of IOS devices.

IOS devices consist of a handheld camera, computer, and software. These devices, used in many fields of dentistry, such as prosthodontics, restorative, and orthodontics, have acquired digital 3D models with a variety of complex mechanisms. Some of the devices currently available on the market and their methods of image acquisition include the following : True Definition Scanner (3M ESPE, Maplewood, Minn) with wavefront sampling; TRIOS 3 (3Shape, Copenhagen, Denmark) with confocal laser technology; iTero Element (Align Technologies, San Jose, Calif) with confocal microscopy; CS 3500 (Carestream Dental, Atlanta, Ga) with triangulation; Lythos (Ormco, Orange, Calif) with triangulation; and Apollo DI (Dentsply Sirona, York, Pa) with confocal microscopy, etc.

Although the devices have different mechanisms of image acquisition, all devices collect data by using a reflection light that goes from their cameras to scanned objects. During images acquisition, triangular data called points of interest (POIs) are saved as images or videos, and the information is often combined by software in stereolithography (STL) format. The first 2 coordinates of the POI (x and y) are defined on the objects, and the third coordinate (z) is determined by the triangulation principle (a calculation that involves dividing the area by triangles to find an unknown length, height, or mapping coordinate). With the help of the IOS camera movement, images of the same POI that are taken from different angles are overlapped after the similarity calculation. Cartesian coordinates of overlapping POI are defined as a triangle. The STL file consists of a combination of thousands of triangles (or POIs) whose x-, y-, and z-coordinates are defined as described ( Fig 2 ).

Fig 2
STL file generation mechanism.

Confocal microscopy is the basis of confocal laser technology and 1 of the image acquisition mechanisms of optical image stabilizer. The basic mechanism of the confocal microscope was defined by Marvin Minsky in 1957 ( Fig 3 ). Advances in laser and computer technology have increased interest in this method over time. Modern confocal laser scanning microscopes have been developed with the integration of some other technologies such as beam scanning mechanisms and wavelength selection devices into this mechanism.

Fig 3
The basic mechanism of the confocal microscope.

Obtaining a digital model with the light reflection might be affected by surface properties, brightness, and wetness of the scanned object. For example, the reflection of the beam on a tooth surface illuminated by the reflector light differs from other beam reflection on a darker tooth surface. In addition, the light reflected from the wet tooth surface is refracted by the effect of water on the surface. Do all of these affect the performance and accuracy of IOS? Our study was designed to find the answer to this question.

Material and methods

Experimental protocols of this study were approved by the Afyonkarahisar University Clinical Research Ethics Committee. The research was conducted on 50 patients who applied to the orthodontic department of our faculty. Informed consent forms were obtained from all the patients included in the study. Only the maxillae were scanned, and maxillary digital models were analyzed. The maxilla was preferred because saliva isolation is easier and does not have movable structures such as tongue that might adversely affect scanning. The exclusion criteria of the study included the following: missing teeth in the maxilla, the presence of partially erupted teeth, or decayed teeth. The maxilla of each patient were scanned 4 times in 4 different oral environmental conditions. The obtained 200 digital models were divided into 4 different groups and analyzed. Definitions of the groups are as follows: scanning in daylight without isolation of saliva (group 1), daylight scanning after saliva isolation (group 2), scanning under reflector light without saliva isolation (group 3), and scanning in a relatively dark oral environment without isolation of saliva (group 4). During the fourth group of scans, curtains of clinics were closed, and the clinician placed 1 of his hands in front of the mouth to prevent the passage of daylight. In group 2, a cheek retractor and cotton rolls were used for the saliva isolation during the entire scanning process, and all the teeth and soft tissues, including the roof of palate, were dried with air spray just before the scanning.

TRIOS (3Shape) device with confocal laser technology was used for intraoral scanning. Scanning duration was measured for each scan by using a stopwatch. In addition, the total number of images the device used to create each digital model was recorded. Mesiodistal width of the 6 maxillary teeth (canine to canine) was measured with Ortho Analyzer software (3Shape) on 200 digital models. All 4 scans were compared for the scanning time, the mesiodistal width of the teeth, and the number of images used for the generation of each digital model. All oral scanning procedures and measurements on digital models were performed by a single researcher (F.S.). IOS companies guide on the scan path that clinicians must follow to use the devices efficiently. Therefore, the researcher followed the same scan path recommended by 3Shape in all scans ( Fig 4 ), and color calibration adjustment of the device was renewed in each patient.

Fig 4
Recommended scan path by the company.

The models obtained from each patient under 4 different environmental conditions were individually overlapped. Group 2 digital models (daylight scanning after saliva isolation) were used as a reference because they had been acquired under the appropriate environmental condition recommended by the company. All models were imported into Geomagic Control X (3D Systems, Rock Hill, SC) in STL format. First, 4 models overlapped for each patient individually. Subsequently, negligible areas, such as above the mucogingival junction and beyond the field of interest, were extracted simultaneously from all 4 models to ensure accurate superimposition ( Fig 5 ). The digital models in groups 1, 3, and 4 were superimposed separately using the best-fit algorithm with the individual reference model in group 2 ( Fig 6 ).

Fig 5
Removing negligible areas from all 4 models simultaneously.

Fig 6
Overlapping of group 1 and reference model.

Three-dimensional surface deviations were shown with color-coded maps. The color codes had the following meanings: green = perfectly aligned areas; red = positively positioned areas (PPA) relative to the reference model; blue = negatively positioned areas (NPA). To convert these qualitative data into a quantitative form, we calculated the surfaces deviations between the reference models (group 2) and test models (groups 1, 3, and 4) as means of negative deviation, means of positive deviation, in total area, out total area (OTA), PPA, and NPA. The data were compared between the 3 test groups.

Statistical analysis

SPSS statistical software (version 22; IBM, Armonk, NY) was used to calculate the mean values and standard deviations of each parameter. One-way analysis of variance and post-hoc Tukey tests were performed to compare data among the groups.


Scanning duration for all groups was very close to each other ( Table I ). These minimal differences in the scanning duration between the groups were not statistically significant. It was found that the number of images was at least in group 4 ( Table I ). This finding meant that the program could achieve the same digital model while scanning in the dark by combining fewer images. There was only a statistically significant difference between group 4 and other groups.

Table I
Scanning durations and the total number of images during the acquisition of the virtual models (multiple comparisons: Tukey HSD)
Dependent variable (I) Group (J) Group Mean difference (I−J) SE P 95% confidence interval
Lower bound Upper bound
Duration of scanning 1.00 2.00 0:00:02 0:00:03 0.951 −0:00:07 0:00:11
3.00 −0:00:03 0:00:03 0.745 −0:00:13 0:00:06
4.00 −0:00:00 0:00:03 0.996 −0:00:10 0:00:09
2.00 1.00 −0:00:02 0:00:03 0.951 −0:00:11 0:00:07
3.00 −0:00:05 0:00:03 0.416 −0:00:15 0:00:03
4.00 −0:00:02 0:00:03 0.877 −0:00:12 0:00:06
3.00 1.00 0:00:03 0:00:03 0.745 −0:00:06 0:00:13
2.00 0:00:05 0:00:03 0.416 −0:00:03 0:00:15
4.00 0:00:03 0:00:03 0.858 −0:00:06 0:00:12
4.00 1.00 0:00:00 0:00:03 0.996 −0:00:09 0:00:10
2.00 0:00:02 0:00:03 0.877 −0:00:06 0:00:12
3.00 −0:00:03 0:00:03 0.858 −0:00:12 0:00:06
Total image number 1.00 2.00 −38.96000 26.63431 0.462 −107.9750 30.0550
3.00 5.70000 26.63431 0.997 −63.3150 74.7150
4.00 77.58000 26.63431 0.021 8.5650 146.5950
2.00 1.00 38.96000 26.63431 0.462 −30.0550 107.9750
3.00 44.66000 26.63431 0.339 −24.3550 113.6750
4.00 116.54000 26.63431 0.000 47.5250 185.5550
3.00 1.00 −5.70000 26.63431 0.997 −74.7150 63.3150
2.00 −44.66000 26.63431 0.339 −113.6750 24.3550
4.00 71.88000 26.63431 0.038 2.8650 140.8950
4.00 1.00 −77.58000 26.63431 0.021 −146.5950 −8.5650
2.00 −116.54000 26.63431 0.000 −185.5550 −47.5250
3.00 −71.88000 26.63431 0.038 −140.8950 −2.8650
Only gold members can continue reading. Log In or Register to continue

Jan 9, 2021 | Posted by in Orthodontics | Comments Off on Effect of saliva isolation and intraoral light levels on performance of intraoral scanners
Premium Wordpress Themes by UFO Themes