Three-dimensional evaluation of mid-facial soft tissue changes after expansion using micro-implant-supported maxillary skeletal expander

Background The aim of this study was to assess the mid-facial soft tissue changes induced by a micro-implant-supported maxillary skeletal expander in late adolescents and young adults by cone-beam computerized tomography and the correlations between hard and soft tissue changes after expansion with maxillary skeletal expander. Subjects and methods Twenty patients with maxillary transverse deficiency treated with maxillary skeletal expander were selected. Cone-beam computerized tomography images taken before and after expansion were superimposed to measure the changes in soft and hard tissue landmarks. Results Anterior nasal spine, posterior nasal spine, and alveolar bone width were significantly increased after expansion with maxillary skeletal expander (p < 0.05). The average lateral movement of the cheek points was 1.13 ± 0.33 mm (left) and 1.41 ± 0.39 mm (right), while that of the alar curvature points was 1.07 ± 0.72 mm (left) and 1.06 ± 0.68 (right) (p < 0.05). The average forward displacement of the cheek points was 0.42 ± 0.66 mm (left) and 0.60 ± 0.58 mm (right), whereas that of the alar curvature points was 0.80 ± 0.67 mm (left) and 0.68 ± 0.56 mm (right) side (p < 0.05). The average downward movement of the subnasale was 0.40 ± 0.37 mm (p < 0.05). The changes in cheek points and alar curvature points on both sides significantly correlated with hard-tissue changes (p < 0.05). Conclusions Maxillary expansion using maxillary skeletal expander resulted in significant lateral and forward movement of soft tissues of the cheek and alar curvature points on both sides and correlated with the maxillary suture opening at the anterior and posterior nasal spines.

0.05). The average downward movement of the subnasale was 0.40 ± 0.37 mm (p < 0.05). The changes in cheek points and alar curvature points on both sides significantly correlated with hardtissue changes (p < 0.05).
Conclusions Maxillary expansion using maxillary skeletal expander resulted in significant lateral and forward movement of soft tissues of the cheek and alar curvature points on both sides and correlated with the maxillary suture opening at the anterior and posterior nasal spines.

Background
Maxillary transverse deficiency is one of the skeletal problems in the craniofacial region encountered during orthodontic treatment [1]. To correct this malocclusion, rapid maxillary expansion (RME) treatments have been used for more than a century [2][3][4]. Several studies have proved that RME increased arch width and perimeter to allow the correction of posterior crossbite and to provide space for alleviation of crowding of the dentition [5][6][7]. However, limited skeletal movement, dentoalveolar tipping, root resorption, detrimental periodontal consequences, and lack of long-term stability were 3 reported as undesirable side effects [8].
With the widespread use of temporary anchorage devices, the micro-implant-supported maxillary skeletal expander (MSE) was recently developed to open the midpalatal suture and achieve skeletal expansion [9][10][11][12][13][14]. Numerous positive results were reported about the skeletal effects of MSE in different studies [11,12,15]. However, limited data have been published regarding soft tissue changes and their correlation with hard tissue changes [16].
Three-dimensional soft tissue analysis using the 3dMD Face system after MSE showed significant changes in the paranasal region, the upper lip, and in both cheeks [16]. Cone-beam computerized tomography (CBCT) is a useful 3D imaging method for the accurate evaluation of hard and soft tissue changes [17,18]. Progress in software development has allowed improved manipulation and visualization of CBCT images, and thus, collection of reliable and precise information [19].
The aim of this study was to assess the mid-facial soft tissue changes induced by MSE use in late adolescents and young adults by CBCT, and to evaluate the correlations between hard and soft tissue changes after expansion with MSE.

Subjects And Methods
This retrospective study analysed CBCT images of patients who received orthodontic treatment for resolution of maxillary transverse deficiency from June 2016 to January 2018 at Samsung Medical A total of 20 patients with a mean age of 21.9 years (range, 15.0-27.1 years) were selected. CBCT images before the delivery of appliance (T0) and post-MSE (T1) were taken. The mean treatment duration was 48.5 days (range, 31-80 days), and the mean amount of expansion was 6.5 mm (range, 5.2-8.0 mm). The post-MSE CBCT images were acquired within 4 weeks (mean duration, 21 days; range, 1-28 days) of completion of expansion. The interval of CBCT images was 4.05 months (range, 3.5-9.1 months).
The inclusion criteria were the following: transverse maxillary deficiency; absence of any previous mid-facial trauma; absence of functional shift; lack of previous orthodontic treatment. The exclusion criteria were the following: systemic disease or syndromes; growing patients.
The MSE (BioMaterials Korea Inc., Seoul, Korea) device is a specific type of bone-borne expander using four micro-implants in the posterior area of the palate with bicortical engagement, including an 8-mm expansion jackscrew unit supported by four palatal micro-implants with diameter of 1.8 mm and length of 11 mm attached to the first molars with connecting arms and molar bands [9,12]. The rate of expansion was two turns per day (0.20 mm per turn) until diastema appeared, after which the rate was changed to one turn per day. The expansion was stopped when the required amount of expansion was achieved. After completion, the MSE was retained for at least 3 months for stabilization.
All CBCT images were taken by the same technician using a CBCT scanner (CS 9300, Carestream Using the anatomical structures of the anterior cranial base, the superimposition of the postexpansion CBCT onto the pre-expansion CBCT was performed, as proposed by Cevidanes et al. [20][21][22] (Fig. 1). The superimposition method uses the voxel grayscale and is fully automated by the 'Automatic Registration' tool of the software to avoid operator-related errors (Fig. 2), a procedure whose accuracy has been previously validated [22,23].
The coordinate system featured three axes (x, y and z) with the origin (0, 0, 0) registered at Nasion (N). The x-axis, the transverse axis, was parallel to the Orbital (Or) line passing through the left and right Or. The y-axis, the anteroposterior axis, was perpendicular to the Or line and parallel to the right Frankfort horizontal (R FH) line. The z-axis, the vertical axis, was perpendicular to both the Or line and the R FH line (Fig. 3). Assuming the subject was in an anatomical position, positive values were to the left, posterior, and superior to the N point of the subject (Fig. 4), and negative values were to the right, anterior, and inferior to the N point. The three-dimensional coordinates (x, y, z) of any landmark 5 represented its 3D position relative to N (0, 0, 0).
In the OnDemand3D software, landmark points were defined using the (x, y, z) Cartesian coordinate system, based on the 3 orientation planes. A series of 20 points were marked on each pre-and postexpansion CBCT. On each CBCT, three points with three pairs of soft tissue landmarks and five pairs of hard tissue landmarks were identified based on previous reports [18,24] (Table 1; Fig. 5 and Fig.   6). Their three-dimensional changes between T0 and T1 were defined as the differences in their three coordinates. To investigate the amount of skeletal expansion, the following parameters were evaluated: amount of expansion at the anterior nasal spine (ANS) and the posterior nasal spine (PNS) [25], and alveolar width [26] (Table 2; Fig. 7A, 7B).

Statistical analysis
The normality of the data was determined using the Shapiro-Wilk test. Comparison of the coordinates of the soft tissue landmarks and hard tissue changes before and after expansion was performed using paired t-tests, according to the normality of data distribution. Pearson's correlation analysis was used to assess the correlation between hard tissue and soft tissue changes. P-values < 0.05 were considered statistically significant. All statistical analyses were performed using SPSS software, version 23 (SPSS Inc, Chicago, IL, USA).
A single examiner performed all measurements. In order to determine the intraexaminer reliability, the same examiner reanalysed ten randomly selected patients within a 2-week interval. The resultant correlation coefficient (ICC) indicated high reliability (ICC > 0.90). Cronbach's Alpha was used to evaluate the reliability of the measurements, and was equal to 0.92, showing appropriate superimposition and measurement agreement.

Results
In this study, the cheek points on left side and right side moved laterally by 1.14 and 1.41 mm and forward by 0.43 mm and 0.61 mm, respectively; these changes were statistically significant (p < 0.001; Table 3). The alar curvature points on both sides moved laterally by 1.11 and 1.07 mm, and forward by 0.82 and 0.70 mm, respectively and the subnasale point moved downward by 0.34 mm (p < 0.001; Table 3).
According to the result of superimposition of the 3D skull models, all the investigated hard tissue landmarks showed significant lateral movement (p < 0.05; Table 4). Furthermore, the A point, the ectocanine, and the prosthion of each half of the maxilla shifted forward (p < 0.05; Table 4), whereas the A point moved downward by 0.30 mm on the left side and by 0.28 mm on the right side following MSE (p < 0.05; Table 4).
Regarding the midpalatal suture, the amount of splits at the PNS (4.01 mm) and ANS (4.94 mm) were statistically significant (p < 0.001; Table 5) and the amount of PNS split corresponded to 81.17% of that of the ANS. Additionally, at the furcation of the first upper molars, the alveolar bone width after treatment showed a statistically significant expansion of 3.61 mm (p < 0.001; Table 5). Table 6 shows that while the changes in the x-coordinate of the cheek and alar curvature points on the left side were positively affected by changes in ANS and PNS, those on the right side were negatively affected (p < 0.01). Moreover, the changes in the y-coordinate of the cheek points and alar curvature points on both sides were negatively affected by the separation at ANS and PNS (p < 0.05).
The comparison of the effects of the ANS and PNS separation with the movement of cheek and alar curvature points are shown in Table 7. Only the regression coefficients between movement in the xaxis of the cheek points on both sides and the separation at ANS were statistically significant (p < 0.05). For these soft tissue points on both sides, the separation measured at the ANS had a stronger effect than that at the PNS.
The unstandardized coefficients produced by regression analysis, shown in Table 8, reflect the expected lateral movement of both cheek points depending on the separation at the ANS, and have high statistical significance (p < 0.001). Thus, they can be used to predict the lateral displacement of the cheek points from the split at the ANS using the following regression equations: Cheek point L(x) change = -1.377 + 0.508 × ANS separation Cheek point R(x) change = 1.441 -0.578 × ANS separation

Discussion
In this study, we evaluated the changes in soft tissue landmarks after MSE treatment and their correlation with changes in hard tissues in late adolescents and young adults with transverse 7 maxillary discrepancy using CBCT. We used voxel-based registration, one of six methods to obtain a rigid registration [27], on the cranial base of pre-and post-expansion CBCT images to evaluate soft tissue displacement. A previous study indicated that CBCT overcomes the limitations of 2D radiographs without exposing the patient to high levels of radiation, and allows the simultaneous evaluation of hard and soft tissues [28]. In recent years, the evaluation of soft tissue displacement using CBCT has become more common [17,18,[29][30][31][32][33][34][35].
Regarding soft tissue changes, we observed a forward and lateral movement of the cheek points and alar curvature points on both sides; and a significant downward movement of the subnasale. These trends of soft tissue displacement after MSE were similar to those observed in previous studies [16,19]. Kim et al reported that alar soft tissue over the infraorbital foramen, and soft tissue zygion were expanded significantly after conventional RPE [17]. The anterior and lateral movement of these soft tissue landmarks could be the result of the corresponding movement of the maxilla, which could be explained by the location of the maxillary rotational fulcrum. The maxilla is located medially and anteriorly relative to this fulcrum. Due to the outward rotation of the zygomaticomaxillary complex around the proximal part of the zygomatic process, each half of the maxilla will be displaced forward and laterally [36]. Additionally, the asymmetrical lateral movement of the cheek points and the alar curvature points between the left and right side could be explained by the unequal movement of the two maxillary bones. Previous studies by Cantarella et al and Torun et al showed that the displacement of the anterior part of the maxillary bones could affect mid-facial soft tissue during maxillary expansion [33] and might result in aesthetic alterations in this region, which could become asymmetrical [25]. In this study, a downward movement of the subnasale by 0.40 mm was observed. This could be explained by the median and inferior location of the maxilla relative to the fulcrum. Due to the outward rotation of the zygomaticomaxillary complex around the frontozygomatic suture area, the maxillary halves moved downward and outward [37]. Concerning hard tissue changes, there were significant lateral movements of all evaluated hard tissue landmarks, a significant forward displacement of the A point, the prosthion, the ectocanine and a significant downward shift of the A point following MSE. These types of movement were also observed in previous studies [36,37]. 8 The data obtained in this study showed that the midpalatal suture was successfully split in all patients after MSE. The mean separation at the PNS was about 81.17% of that at the ANS. This ratio demonstrates that MSE created an almost parallel split of the midpalatal suture, in agreement with previous studies [25,38]. The amount of separation at the midpalatal suture produced by MSE in this study (4.94 mm at the ANS and 4.01 mm at the PNS) was nearly equivalent to that in the study by . The reason is probably found in the mechanism of action of MSE. In fact, this appliance is located in the posterior portion of the hard palate, with 4 bicortical miniscrews positioned medial to the zygomatic buttress bones, allowing the opening force to be distributed along the entire suture length [25]. As a result, while the tooth-borne maxillary expanders (Hyrax-type expander, RME) created a V-shaped expansion pattern of the maxillary palatal suture, MSE could produce an almost parallel split at the midpalatal suture [26].
The amount of expansion in the appliance (6.5 mm; range, 5. The regression coefficients derived from these correlations were used to investigate to what extent the separation at the ANS and PNS affects the changes in soft tissue landmarks. The ANS split had greater effect than the PNS split on the cheek points and alar curvature points in the x-and y-axis.
The reason could lie in the fact that the ANS is anatomically closer to the cheek points and alar curvature points than the PNS. Additionally, the only statistically significant regression coefficients were seen between the lateral movement of the cheek points on both sides and the separation at the ANS.
The prediction of soft tissue changes is crucial for treatment planning, and the results of this study imply that MSE would be beneficial in late adolescents and young patients with mid-facial deficiency, including paranasal depression without maxillary surgery. However, this study has limitations, for example relatively small subject size and short follow-up period. Further research with a larger sample size, using CBCT, is required to evaluate the long-term stability of these changes.

Consent for publication
Not applicable.

Competing interests
There is no conflict of interest to be declared concerning this study.

Availability of data and materials
The datasets used and analysed in the current study are available from the corresponding author on reasonable requests.    x, x-axis, the transverse axis; y, y-axis, the anteroposterior axis; z, z-axis, the vertical axis; ANS, Anterior nasal spine; PNS, P * Statistically significant at p < 0.05. x, x-axis, the transverse axis; y, y-axis, the anteroposterior axis; ANS, Anterior nasal spine; PNS, Posterior nasal spine * Statistically significant at p < 0.05. x, x-axis, the transverse axis; ANS, Anterior nasal spine * Statistically significant at p < 0.05. Figure 1 OnDemand 3D voxel-based superimposition on the cranial base. The yellow box is used to determine the area of the cranial base to be used as a reference for the superimposition.
24 Figure 4 Positive values are to the left, posterior, and superior to the N point of the subject