Statement of problem
Measuring both the trueness and precision of an intraoral scanner (IOS) will provide a thorough understanding of its accuracy.
The purpose of this in vitro study was to measure the complete-arch trueness and precision of 3 commercially available intraoral scanners equipped with the latest software version and compare them by using a laboratory scanner as reference.
Material and methods
Nineteen maxillary and 19 mandibular completely dentate stone casts previously acquired from 19 patients by using a polyvinyl siloxane (PVS) dual mix impression and stock trays were scanned with 3 intraoral scanners (TRIOS 3; 3Shape A/S, i500; Medit, and Emerald; Planmeca) using their latest software versions. The same casts were also scanned with a laboratory scanner (E3; 3Shape A/S) that served as the reference scanner. Files were exported in standard tessellation language (STL) format and inserted into a metrology 3D mesh comparison software program (CloudCompare).
In terms of trueness, a significant difference was found among the scanners (F (2.37)=239.7, P <.001). Planmeca Emerald had significantly lower trueness values than either the Medit i500 ( P <.001) or the 3Shape A/S TRIOS 3 ( P <.001). No significant difference in trueness was found between the Medit i500 and the 3Shape A/S TRIOS 3 scanner ( P =.365). In terms of precision, a significant difference was found among the scanners (F (2.89)=301.2, P <.001). The 3Shape A/S TRIOS 3 scanner was significantly more precise than the other scanners ( P <.001 for both the Medit i500 and Planmeca Emerald). The Planmeca scanner was significantly more precise than the Medit i500 scanner ( P <.001). Concerning the ability of the scanners to reproduce the files of the reference scanner without overestimation or underestimation, the Medit i500 produced files that significantly underestimated the reference scanner’s files (t (37)=−12.4, P <.001). The other scanners did not significantly either underestimate or overestimate the files of the standard (t (37)=−1.91, P =.062 for the TRIOS 3 and t (37)=1.64, P =.101 for the Planmeca)
With regard to completely dentate arch trueness, the Planmeca Emerald IOS had statistically lower trueness. With regard to complete dentate arch precision, the 3Shape A/S TRIOS 3 IOS was the statistically more precise scanner. With regard to reference scanner file estimation, the Medit i500 IOS produced files that significantly underestimated the reference scanner files. All 3 tested scanners exhibited a completely dentate arch average accuracy below 100 μm in vitro.
Newest generation intraoral scanners exhibit a completely dentate arch accuracy of under 100 μm in vitro. Some IOSs tend to underestimate the arch size.
The use of digital methods, such as computer-aided design and computer-aided manufacturing (CAD-CAM), has increased rapidly in dentistry in recent years. The first step in this digital workflow is the acquisition of a digital scan by means of an intraoral scanner, a method that has been reported to provide excellent accuracy for short-span prostheses, both tooth- and implant-supported, compared with conventional impression methods. , Controversy still exists, however, regarding the accuracy of IOSs for scanning complete arches. Evidence has supported the superiority of conventional impressions for complete arches, but data from newer studies testing the latest IOS hardware and software versions tend to support the implementation of digital scan for complete arches. An IOS should achieve clinically acceptable levels of accuracy, often specified at 100 μm, although a definitive consensus and a scientific correlation between global deviation and actual marginal prosthesis misfit is lacking.
Trueness and precision are terms used for direct and indirect dental digitization. According to the ISO international standard number 5725, trueness is the ability of a measurement or measuring device to match the actual value of the quantity being measured, whereas precision is the ability of a measurement or measuring device to consistently repeat a particular measurement. Trueness and precision are both measures of accuracy.
New scanners are being introduced to the dental market every year. The TRIOS 3 color Pod, now in its fourth generation, was launched by 3Shape A/S in 2016, the Emerald (Planmeca) in 2017, and the i500 (Medit) in 2018. However, studies that compared different intraoral scanners in terms of dentate complete-arch accuracy are sparse and have reported conflicting results because of methodological, statistical, and technical issues. , , , , , The authors are unaware of studies comparing these 3 scanners for accuracy in dentate complete arches.
The purpose of this in vitro study was to measure the complete-arch trueness and precision of 3 recently introduced intraoral scanners, the TRIOS 3 color Pod (3Shape A/S), the Emerald (Planmeca), and the i500 (Medit) equipped with their latest software versions and to compare them with a laboratory scanner as reference. The null hypotheses were that no statistically significant difference would be found in the complete-arch trueness of the tested scanners and that no statistically significant difference would be found in the complete-arch precision of the tested scanners.
Material and methods
Thirty-eight Type IV stone casts (Hera Moldastone; Kulzer GmbH) recently acquired from completely dentate adult patients were used in the study. The stone casts (19 maxillary, 19 mandibular) were scanned with the desktop laser scanner (E3; 3Shape A/S), as a reference against which the meshes were compared, and the 3 intraoral scanners. The E3 is a scanner used in dental laboratories and commonly used for the digitization of stone casts to design and manufacture CAD-CAM dental prostheses. Its accuracy, as reported by the manufacturer, is 7 μm. The resultant triangular meshes of the stone casts (STL files) were used as the control. For the IOS digital scans, the recommended scan strategy per manufacturer was used to ensure optimal accuracy.
For the TRIOS 3 scanner, the maxillary scanning initiated from the left posterior area and proceeded occlusally with a zig-zag movement in the anterior teeth toward the right posterior area. It then turned buccally toward the contralateral side, and the scan was completed on the palatal side with a left to right direction of scan. For the mandible, scanning initiated from the posterior left quadrant and proceeded occlusally with a zig-zag movement in the anterior teeth toward the contralateral side. It then turned lingually toward the left quadrant and was completed on the buccal side with a left-to-right movement.
For the i500 scanner, the maxillary scan started on the posterior left occlusal area and proceeded toward the contralateral side with a zig-zag movement in the anterior teeth area. It then turned palatal and ended on the buccal side of the right side. For the mandibular casts, the scan was initiated on the posterior left occlusal area and proceeded toward the contralateral side with a zig-zag movement in the anterior teeth. It then turned lingually toward the left quadrant and terminated on the buccal side of the right posterior quadrant.
For the Emerald scanner, the maxillary and mandibular scans followed identical paths. They initiated from the left posterior occlusal surfaces in the maxilla and the left posterior occlusal surfaces in the mandible and proceeded toward the contralateral side with a 45-degree movement against the incisal area of the anterior teeth. It then turned buccally toward the left side and terminated on the palatal and lingual side of the right maxillary and right mandibular posterior teeth. Digital scanning was performed at room temperature by 1 experienced operator (G.M.) proficient with the Medit i500 and 3Shape A/S TRIOS 3 scanners and by a different experienced operator (A.T.) proficient with the Planmeca Emerald scanner. This minimized the operator experience bias reported to influence scan accuracy.
The scanners were calibrated before each scan session according to the manufacturers’ instructions. The scanning mode was set to model scan for all 3 scanners. For the i500 IOS, the scanning parameters used were a blue light mode with a filtering level 2 and a focal length of 17 mm.
All digital scans were automatically postprocessed by using the proprietary software before being exported and saved as STL files. For the i500, the Fill Major Holes option was elected before postprocessing. Software versions for the IOS used are shown in Table 1 . All the files were coded and sent to the second author (D.A.) for analysis. As a result, all the analyses were blinded to the brand of each scanner.
|Intraoral Scanner||Software Version|
|Medit i500||Medit Link version 2.0.3 build 376 Revision 27 520|
|3Shape A/S TRIOS 3||Dental Desktop 22.214.171.124 (insane mode)|
|Planmeca Emerald||Romexis 126.96.36.199|
Four sets of triangular meshes were available for comparison, totaling 152 STL files, n=38 for each IOS and n=38 for the control desktop scanner. For every arch, 4 meshes (3S, IM, PE, and GS) were imported for computational manipulation in a dedicated mesh and point cloud handling software program (CloudCompare, version 2.11 alpha; Anoia). The triangular mesh derived from the desktop laser scanning was used as a reference, and no other manipulation was performed. The 3S, IM, and PE originated meshes were then initially roughly registered together by using a minimum (3 to 5) number of points and then were again finely registered with each other by using the iterative closest point (ICP) algorithm, calculated on a sample of 50 000 pairs of points. This resulted in 3 meshes for each arch overlapping one another. The meshes were then simultaneously cropped, thus leaving only the teeth up to approximately the middle of the clinical crown of the second molar bilaterally and 3 to 5 mm of the gingiva. The result was 3 triangular meshes representing the same arch, with clinically relevant and almost identical remaining anatomy. Finally, each of these meshes was again separately, roughly, and finely registered to the GS (control). This resulted in 3 different meshes for each arch, which were finely registered to the control. The absolute distance of every face of each test mesh to a point on the surface of the reference mesh was computed, indicating the difference between this mesh and the control. The median value of the differences and the interquartile range (IQR) for each pair were noted.
Additionally, the signed (that is, positive and negative) distances of each mesh to the reference file were calculated, and the mean and standard deviation of the measurements were noted. These measurements were only used to estimate the ability of each scanner to correctly replicate (without overestimation or underestimation) the file produced by the control and were not used to calculate the accuracy of the intraoral devices.
The first stone cast (patient 1, maxillary) was scanned with each of the IOS scanners 10 times to estimate the precision of the intraoral scanners. The 10 meshes were simultaneously cropped and were finely registered with each other, following the same procedure described previously. Each of the meshes acquired was sequentially used as a reference, resulting in a total of 90 pairs of meshes for each intraoral scanning device. The average standard deviation of the differences of the meshes was used as a measure of repeatability for the scanners. All the handling and analysis of the STL files was performed by the same operator (D.A.), who was blinded to the brand of the scanner.
To estimate the repeatability of the reference scanner, the first stone cast (patient 1, maxillary) was scanned with the desktop scanner 10 times. The 10 meshes were simultaneously cropped and then finely registered with each other. Pairwise comparisons were conducted between the meshes. This resulted in 90 pairs of meshes whereby each of the meshes acquired was sequentially used as a reference. The average standard deviation of the differences of the meshes was used as a measure of repeatability for the desktop scanner. The handling of the files from the reference scanner was not blinded.
To estimate the precision of the registration software, 1 laser-scanned mesh was used. The mesh was imported into the software and cloned 4 times; the clones were then roughly and finely registered with each other and with the original mesh by using the same procedure described previously. This resulted in 20 pairs of meshes, with each of the meshes sequentially used as a reference. The standard deviation of the differences was used as a measure of its precision.
Because of the relatively large sample size and by virtue of the central limit theorem, parametric methods were used to draw inferences. For the estimation of trueness, descriptive statistics were calculated, and inferences were drawn by using repeated-measures 1-way ANOVA with a fixed factor “brand of intraoral scanner” (TRIOS 3, i500, and Emerald) and a dependent variable “the difference between the intraoral scanners and the control.” Post hoc analysis was conducted by pairwise t tests. To estimate precision, descriptive statistics were calculated and inferences were drawn by using repeated-measures 1-way ANOVA with a fixed factor “brand of scanner” (TRIOS 3, i500, Emerald, and E3) and a dependent variable “the pairwise differences between the meshes for each of the scanners.” Post hoc analysis was conducted by pairwise t tests. In relation to the ability of the scanners to correctly reproduce (without overestimation or underestimation) the files produced by the reference desktop scanner, 1 sample t test was used to draw inferences about the mean distance of the differences from zero for each of the scanners (α=.05 with familywise Bonferroni correction where appropriate). A spreadsheet (Excel 2016; Microsoft Corp) with the XRealStats add-in was used for the statistical analysis. All values were reported in μm. The statistical analysis concerning the accuracy of the intraoral scanner was performed blinded by 1 of the authors (D.A.), and the brands of the scanners were revealed after the results had been computed.