This research aimed to compare the estimation error of the root axis using 3-dimensional (3D) tooth models at the midtreatment stage between the whole-surface scan (WSS) and lingual-surface scan (LSS) methods.
The sample consisted of 208 teeth (26 each of central incisors, canines, second premolars, and first molars in the maxillary and mandibular dentition) from 13 patients whose pre- and midtreatment intraoral scan and cone-beam computed tomography (CBCT) were available. The 3D tooth models were constructed by merging the intraoral-scan crowns and the CBCT-scan roots obtained at the pretreatment stage. To estimate the root axis at the midtreatment stage, we superimposed the individual 3D tooth models onto the midtreatment intraoral scan obtained by the WSS and LSS methods. The midtreatment CBCT scan was used as the gold standard to determine the real root axis. The estimated root axis in terms of mesiodistal angulation and buccolingual inclination was measured in the WSS and LSS methods, and statistical analysis was performed.
The estimation errors of the mesiodistal angulation and buccolingual inclination were <2.0° in both methods. The LSS method demonstrated a statistically larger but clinically insignificant estimation error than the WSS method in the mandibular canine (mesiodistal angulation, 1.95° vs 1.62°) and the total tested teeth (mesiodistal angulation, 1.40° vs 1.29°; buccolingual inclination, 1.51° vs 1.41°).
Because the estimation errors of the root axis angle using the 3D tooth model by the WSS and LSS methods were within the clinically acceptable range, the root axis can be estimated by both methods.
The 3-dimensional tooth model was constructed using pretreatment 3-dimensional data.
The model was used to monitor changes in the root axis via intraoral scanning.
Lingual-surface scanning and whole-surface scanning methods were used.
Mesiodistal angulation and buccolingual inclination estimation errors were <2.0°.
The root axis can be estimated with whole-surface scanning and lingual-surface scanning methods.
With the development of 3-dimensional (3D) imaging technology, the crown alignment and root parallelism can be examined accurately. Cone-beam computed tomography (CBCT) can represent hard tissues of the jaws and teeth with lower-dose radiation than conventional computed tomography. , However, the crown morphology and occlusal relationship are hard to delineate in a CBCT image. , A detailed 3D image of the crown can be acquired using an intraoral scanner with high resolution. By combining complementary characteristics of the 3D imaging technologies, construction of 3D tooth models with precise crown and root morphology has been attempted previously.
The 3D tooth model has been applied mainly for treatment simulation and customized appliance fabrication in orthodontic dentistry. Recently, its application has been extended to monitor root movement, during or after the orthodontic treatment ( Fig 1 ). This application can be achieved by superimposing the individual 3D tooth models onto the mid- or posttreatment intraoral scan using the unaltered crown morphology as a reference. Because the incorporated root tracks the crown superimposition, the orthodontically moved root position can be visualized at any time of the treatment by intraoral scanning. This method is clinically valuable because it eliminates the need for subsequent radiation exposure after the initial CBCT scan.
However, there were several difficulties arising from the presence of brackets while obtaining the midtreatment intraoral scan as follows. When direct intraoral scanning patients, device inaccessibility and optical reflection affected the acquisition of the intraoral scan. In contrast, when indirect intraoral scanning the plaster model, the tooth surface beneath the bracket could not be replicated completely. Hence, intraoral scanning without the bracket-bonded tooth surface can be taken into consideration. The estimation error of the 3D tooth model by only lingual-surface intraoral scan has never been compared with the estimation error by the whole tooth surface intraoral scan. Therefore, the purpose of this study was to compare the estimation error of the root axis angle using the 3D tooth model at the midtreatment stage between the whole-surface scan (WSS) and lingual-surface scan (LSS) methods.
Material and methods
The inclusion criteria were as follows: (1) patients who were treated with multibracket appliance, (2) patients whose intraoral scan and CBCT scan at the pre- and midtreatment stages were available (which were acquired for orthosurgical treatment), and (3) patients whose midtreatment intraoral scans were obtained in the WSS and LSS methods. The exclusion criteria were as follows: (1) patients treated with lingual brackets, (2) patients treated using orthodontic bands, (3) patients who underwent morphologic changes of the crown via interproximal reduction or enameloplasty during the treatment, and (4) patients who had restorative or prosthodontic treatment during the treatment.
As a result, 13 adult patients were retrospectively recruited in this study. From the recruited subjects, a total of 208 teeth (26 each of central incisors, canines, second premolars, and first molars in the maxillary and mandibular dentition, respectively) were tested.
A power analysis using G∗Power software (version 3.1.3; Franz Faul University, Kiel, Germany) determined that a sample size of 26 teeth per group would provide a power of 91% with an α of 0.05 to detect significantly different root axis of 2.5° (an effect size of 0.71), on the basis of the previous reports. , ,
This retrospective study was reviewed and approved by the Institutional Review Board of the Chonnam National University Dental Hospital in Gwangju, South Korea (approval no. CNUDH-2019-024).
The individual 3D tooth model was constructed by merging the pretreatment intraoral-scan crown and the CBCT-scan root ( Fig 2 ). The intraoral scan was performed using a TRIOS scanner (3Shape, Copenhagen, Denmark) and processed into stereolithography (STL) format using OrthoAnalyzer (3Shape) software. The CBCT scan (Alphard Vega; Asahi Roentgen, Kyoto, Japan; 80 kVp; 5 mA; 0.39-mm voxel size; scan time, 17 seconds; and field of view, 200 mm × 179 mm) was performed, and the digital imaging and communication in medicine files were exported to the Invivo5 software (Anatomage, San Jose, Calif) for 3D volume rendering. In the “medical design studio” module of the program, the individual tooth, including the root, was segmented using the “sculpt” function by a single operator (H.H.C) and exported as individual STL files for each tooth type. The intraoral-scan crown and the segmented CBCT-scan teeth were exported into the reverse engineering program (Rapidform; 3D Systems, Rock Hill, SC). Crown registration was then performed to superimpose the 2 3D images. Segmented CBCT-scan teeth of the maxillary and mandibular arches were registered with their corresponding intraoral-scan arches. Finally, the CBCT-scan crown was removed from the superimposed images, followed by the merging of the remaining intraoral-scan crown and the allocated CBCT-scan root. Thus, the final 3D tooth model consisting of the intraoral-scan crown and CBCT-scan root was constructed ( Fig 2 ).
The direct intraoral scan (TRIOS) was performed at the midtreatment stage by the WSS and LSS methods ( Fig 3 , A ). The WSS was started from the occlusal surface of the second molar to the contralateral second molar along the occlusal surface, followed by the lingual and buccal surfaces. The orthodontic brackets, which were bonded on the buccal surface, were included in the scanning scope. For the LSS, the same procedure was followed, from the occlusal to the lingual surface, but was completed without scanning the buccal surface. The maxillary and mandibular arch scans were exported as independent STL files (OrthoAnalyzer).
The individual 3D tooth models were superimposed 1 by 1 onto the midtreatment intraoral scans obtained by the WSS and LSS methods ( Fig 3 , B ). Regional superimposition was performed, which commands the 3D tooth models to superimpose onto the intraoral scan using a defined region as a reference. When superimposing the 3D tooth models onto the WSS image, the whole crown surface except the bracket area was employed as reference. In contrast, only the lingual and occlusal surfaces were employed when superimposing onto the LSS image. After that, the maxillary and mandibular intraoral-scan crowns with their superimposed 3D tooth models were assembled as a single entity in each arch ( Fig 3 , C ).
Midtreatment CBCT scan was used as a gold standard to determine the real root axis ( Fig 2 ), which was obtained on the same day as the midtreatment intraoral scans. The individual tooth, including the root, was segmented by the same operator (H.H.C) who processed the pretreatment CBCT. All the segmented teeth in each maxillary and mandibular arch were assembled as a single entity maintaining their original position. The midtreatment intraoral scan, attached with the 3D tooth models, was reoriented in accordance with the gold standard CBCT by best-fit crown registration ( Fig 3 , D ). The accuracy of the crown registration was validated by measuring shell/shell deviation, which measures the absolute values of the 3D Euclidean distances between the surface points on the 2 images. The mean shell/shell deviations were 0.20 mm in the maxillary arch and 0.21 mm in the mandibular arch ( Table I ). Because the registration errors were less than the clinically acceptable value of 0.22 mm, the crowns of the midtreatment intraoral scan and the gold standard CBCT were considered to be in the same position, and subsequent measurements were conducted.
|Shell/shell deviation (n = 13)||Mean||SD||Range|
The mesiodistal angulation and buccolingual inclination were measured in the estimated root axis by the WSS and LSS methods, and the root axis of the gold standard CBCT ( Fig 4 ). First, the occlusal plane was constructed by connecting the mesiobuccal cusp tips of the right and maxillary left first molars and the midpoint of the maxillary right central incisor edge. Second, the root axis plane of each tooth type was constructed by connecting 3 points: 1 point in the root apex and 2 points in the crown ( Table II ). The distal angle between the occlusal plane and the root axis plane was measured for the mesiodistal angulation, and the lingual angle between the occlusal plane and the root axis plane was measured for the buccolingual inclination ( Fig 4 ). In each subject, once the occlusal plane was constructed, the same occlusal plane was used for the measurement of the estimated and CBCT root axis angle.
|Tooth type||Root axis plane||Definition of the landmarks|
|Two points in the crown||One point in the root|
|Central incisor||Mesiodistal angulation||The midpoint of the incisor edge and the central pit of the lingual surface||The most apical point of the root|
|Buccolingual inclination||The most mesial and distal points of the incisor edge|
|Canine||Mesiodistal angulation||The cusp tip and cingulum||The most apical point of the root|
|Buccolingual inclination||The most mesial and distal points of the marginal ridge|
|Second premolar||Mesiodistal angulation||The buccal and lingual cusp tips||The most apical point of the root or the buccal root in case of 2 roots|
|Buccolingual inclination||The most mesial and distal points in the line of occlusion|
|First molar||Mesiodistal angulation||Maxilla: The mesiobuccal and mesiopalatal cusp tips
Mandible: The mesiobuccal and mesiolingual cusp tips
|Maxilla: The most apical point of the palatal root
Mandible: The most apical point of the mesial or mesiobuccal root
|Buccolingual inclination||The mesiobuccal and distobuccal cusp tips|
One investigator (S.W.L) performed all measurements of the 13 subjects. To test intra- and interobserver reliability of the measurements, the measurement including the construction of the occlusal and the root axis plane was repeated by 2 investigators (S.W.L and H.H.C) with an interval of 3-weeks after the initial measurement. The intraclass correlation coefficient was calculated to test the intra- and interobserver reliability.
The mean and standard deviation (SD) of the mesiodistal angulation and buccolingual inclination were calculated in each tooth type. Because of the nonnormality of data for several tooth types, nonparametric tests were used. To assess the agreement of the mesiodistal angulation and buccolingual inclination measured in the WSS and LSS methods and the gold standard CBCT, we used the Kruskal-Wallis test. To compare the estimation error between the WSS and LSS methods, Wilcoxon signed rank test was used. A P value of <0.05 was considered statistically significant. All the statistical analyses were performed using the SPSS software (version 20.0; IBM, Armonk, NY).
The intraclass correlation coefficient values of mesiodistal angulation and buccolingual inclination ranged from 0.883 to 0.988, indicating good reproducibility of measurements in both intra- and interobserver reliability ( Table III ).
|Tooth type (n = 52 per each tooth type)||Interobserver ICC||Intraobserver ICC|
|Mesiodistal angulation||Buccolingual inclination||Mesiodistal angulation||Buccolingual inclination|