A CBCT Evaluation of Esthetic Preference Regarding the Perceived Facial Attractiveness of Young Korean Female Adults with a Normal Skeletal Pattern

The aim of this study was to determine the factors that affect esthetic preference regarding the perceived facial attractiveness of young Korean female adults with a normal skeletal pattern using cone-beam computed tomography (CBCT). After reorienting the CBCT images of 40 young Korean female adults, three-dimensional (3D) measurements were analyzed, and five 3D facial images were digitally constructed. A computer-based questionnaire was used to determine perceived facial attractiveness by scoring from 1 (very unattractive) to 10 (very attractive). Esthetic perception and the most influential facial view and component were examined and compared by orthodontists, general dentists, and laypeople to evaluate facial attractiveness. Compared to the unattractive group, the attractive group had significantly greater values for Pog-NB, L1SI-FH, U lip thickness, L lip-U lip, and Nasolabial angle, along with significantly lower values for U1SI-FH and Cant of U lip; the attractive group also had a more retruded U1MP (p < 0.01), L1MP (p < 0.01), U lip (p < 0.05), and L lip (p < 0.01) relative to N, as well as a more retruded U lip (p < 0.05) and L lip (p <0.001) relative to esthetic lines. Orthodontists and general dentists were more likely to consider the profile view. When evaluating facial attractiveness, orthodontists and laypeople more often focused on the lips and chin. When evaluating facial attractiveness, all evaluators showed a high esthetic preference for retroclination and retrusion of the incisors and lips, but there were some differences in how they perceived facial attractiveness. Therefore, the ultimate perception of facial attractiveness needs to be based on the esthetic perceptions of laypeople.

Transverse: −4 mm < Pog(x) < 4 mm 2.2. Standardized Reorientation of CBCT Images, 3D Coordinate System, and 3D Measurement ON3D (3DONS, INC., Seoul, Korea) software was used for this study. First, reorientation of the head position of each CBCT scan was performed to minimize any measurement errors from non-standard head postures [24]. The 3D coordinates (x, y, z) of all landmarks represent their 3D position relative to N (0,0,0) ( Figure 1). 3D landmarks were automatically digitized with manual modification, and 3D measurements were made to determine the facial, skeletal, and dental characteristics of the 40 subjects using ON3D software (Figures 2 and 3). Definitions of the 3D landmarks and values are described in Table 1A

Evaluation of Perceived Facial Attractiveness Using a Questionnaire
Images in soft tissue mode were captured from the frontal, oblique (left and rig and profile (left and right) views of 40 subjects using ON3D software. Thus, 3D soft tis images consisting of a total of five photographs of each sample were constructed. H eyes, and ears were excluded from the image to eliminate distractions, thus allowing ev uators to focus on the middle and lower face ( Figure 4A).
The evaluators consisted of 42 orthodontists, 42 general dentists, and 42 laypeop Data were collected via a computer-based questionnaire distributed personally to the p ticipants by cell phone. The inclusion criteria for all evaluators included being at least years of age and willing to participate in this study. The evaluators were asked to exam

Evaluation of Perceived Facial Attractiveness Using a Questionnaire
Images in soft tissue mode were captured from the frontal, oblique (left and right), and profile (left and right) views of 40 subjects using ON3D software. Thus, 3D soft tissue images consisting of a total of five photographs of each sample were constructed. Hair, eyes, and ears were excluded from the image to eliminate distractions, thus allowing evaluators to focus on the middle and lower face ( Figure 4A).
The evaluators consisted of 42 orthodontists, 42 general dentists, and 42 laypeople. Data were collected via a computer-based questionnaire distributed personally to the participants by cell phone. The inclusion criteria for all evaluators included being at least 20 years of age and willing to participate in this study. The evaluators were asked to examine the facial 3D images for a sufficient time and then rank them from 1 (very unattractive) to 10 (very attractive) using a numeric rating scale (NRS) ( Figure 4A). After that, they were asked the two questions shown on the last page of the questionnaire ( Figure 4B). According to the ranking of facial attractiveness, some of the 40 subjects were divided into either the attractive group (AG, top 20%, n = 8) or the unattractive group (UAG, bottom 20%, n = 8). Three-dimensional measurements were analyzed, and esthetic perception was compared between the three groups of evaluators. Additionally, it was determined which facial view was the most influential and which components were the most important in evaluating facial attractiveness.

Statistical Analysis
A power analysis using G*Power (version 3.1.9.7: Franz Faul, Christan-Albrechts-Universitat, Kiel, Germany) was performed to determine the sample size required for this study. It was determined that a total number of 126 subjects would provide a power of 0.87 and a two-tailed alpha value of 0.05.
To assess the reliability of the reorientation and digitizing process, CBCT scans of 10 subjects were re-digitized by the same operator after 3 weeks. Intraexaminer reliability was assessed by the intraclass correlation coefficient, which showed excellent reliability (ICC = 0.981~0.983). Statistical analysis was performed using SPSS software (version 27.0 for Windows; SPSS Corp., Chicago, IL, USA) for statistical analysis. An independent sample t-test was performed to determine the difference in 3D values according to perceived facial attractiveness between the AG and the UAG. Spearman's rank correlation analysis was performed to determine the correlations between the ranks of the three evaluator groups. Cross-analysis was performed to determine the association between the evaluator groups for the image view and facial components when judging facial attractiveness. After the Shapiro-Wilk normality test was performed, if normality was not satisfied, a nonparametric test was performed. If the analysis of variance was significant, Scheffe's post hoc test was performed. The significance level was 0.05. According to the ranking of facial attractiveness, some of the 40 subjects were divided into either the attractive group (AG, top 20%, n = 8) or the unattractive group (UAG, bottom 20%, n = 8). Three-dimensional measurements were analyzed, and esthetic perception was compared between the three groups of evaluators. Additionally, it was determined which facial view was the most influential and which components were the most important in evaluating facial attractiveness.

Statistical Analysis
A power analysis using G*Power (version 3.1.9.7: Franz Faul, Christan-Albrechts-Universitat, Kiel, Germany) was performed to determine the sample size required for this study. It was determined that a total number of 126 subjects would provide a power of 0.87 and a two-tailed alpha value of 0.05.
To assess the reliability of the reorientation and digitizing process, CBCT scans of 10 subjects were re-digitized by the same operator after 3 weeks. Intraexaminer reliability was assessed by the intraclass correlation coefficient, which showed excellent reliability (ICC = 0.981~0.983). Statistical analysis was performed using SPSS software (version 27.0 for Windows; SPSS Corp., Chicago, IL, USA) for statistical analysis. An independent sample t-test was performed to determine the difference in 3D values according to perceived facial attractiveness between the AG and the UAG. Spearman's rank correlation analysis was performed to determine the correlations between the ranks of the three evaluator groups. Cross-analysis was performed to determine the association between the evaluator groups for the image view and facial components when judging facial attractiveness. After the Shapiro-Wilk normality test was performed, if normality was not satisfied, a nonparametric test was performed. If the analysis of variance was significant, Scheffe's post hoc test was performed. The significance level was 0.05.

Results
Compared to the UAG, the AG showed greater values for Pog-NB (p < 0.05), L1SI-FH (p < 0.01), U lip thickness (p < 0.05), L lip-U lip (p < 0.05), and Nasolabial angle (p < 0.001), and lower values for U1SI-FH (p < 0.001) and Cant of U lip (p < 0.001); the AG also exhibited a more retruded U1MP (p < 0.01), L1MP (p < 0.01), U lip (p < 0.05), L lip (p < 0.01), and ML S (p < 0.05) relative to N, as well as a more retruded U lip (p < 0.05) and L lip (p < 0.001) relative to esthetic lines (S-line and E-line). The 3D cephalometric results obtained from each of the groups of evaluators were almost the same as those obtained from the evaluators as a whole. The AG showed more retroclination and retrusion of incisors and lips than the UAG (Table 2A,B).
The most influential facial view (p < 0.05) and facial component (p < 0.001) with respect to perceived facial esthetics were significantly different between the three groups. Orthodontists (59.5%, p < 0.001) and general dentists (57.2%, p < 0.01) considered the profile view to be the most important, while laypeople (40.5%, p > 0.05) considered the frontal view the most important. Orthodontists (73.8%, p < 0.001) focused on lips, while general dentists (38.1%, p > 0.05) and laypeople (45.2%, p < 0.05) focused on the chin. The three groups of evaluators showed some differences in how they perceived facial attractiveness (Table 3).

Discussion
In this study, we evaluated facial attractiveness using 3D images obtained from CBCT and then derived 3D measurements and coordinates of faces that were perceived to be beautiful at the time of the study. Furthermore, we were able to acquire information about the normative mean values of the 3D coordinates and measurements of Korean female adults using data for the specific ethnic group [21,22].
Pog-NB and Pog(y)-B(y) values were greater in the AG than in the UAG in our study. This means that a more prominent Pog relative to the NB line and B(y) was considered to be more attractive; this is consistent with the results of a previous study [27]. On the other hand, Kambara et al. [27] concluded that a chin prominence (Pog-NB) greater than 4 mm could result in an unattractive face. From an esthetic standpoint, Pog-NB was related to lower lip position [6,7] and vertical skeletal pattern [28], which should be considered in order to identify the ideal lip position during treatment planning. Even so, in our study, the AG showed a more posterior position of Pog(y) relative to N than the UAG, which is consistent with the results of previous studies [1][2][3]21]. These results in young female adults may differ from those observed in young male adults [1][2][3], so further studies are necessary.
The positions of the tips of the cheeks and nose were not significantly different between the AG and the UAG in our study. However, some previous studies [12,26] suggested that maxillary skeletal expansion was correlated with forward movement of the cheek point and that cheek volume should be considered when evaluating facial attractiveness. Our study also showed that the lips were more posteriorly positioned in the AG than in the UAG, which means that somewhat retruded lips were perceived as more attractive. These results are consistent with those of previous studies [2,4] on the preference of facial esthetics in the Asian population. These studies reported that the public tends to prefer lips that are more retruded than average, although they have been distinguished by convex facial profiles.
In this study, significant differences were also found in U lip thickness and L lip(y)-U lip(y). Compared to the UAG, the AG showed greater upper lip thickness, which is consistent with some previous studies [11,29] that reported fuller lips were preferable. In addition, the L lip(y)-U lip(y) values were significantly greater in the AG. This indicates that it was considered more esthetic when the lower lip was positioned posteriorly relative to the upper lip. In addition, the preference for the lower lip position is closely related to the chin position [3,7]. Jang et al. [21] and Bayome et al. [23] reported that women chosen as Miss Korea had less chin and cheek volume than the general population. In our study, the attractive group tended to have a retrusive chin, but there were no significant differences in the values for the cheek or zygoma areas. The size and position of the cheeks and nose vary greatly from one individual to the next and should not be considered as a single factor that affects facial esthetics.
In this study, the most influential facial views and components used to judge facial esthetics were significantly different between the three evaluator groups, which is consistent with a previous study [30]. It is quite possible that esthetic perceptions are different among the evaluators depending on their training, educational background, and knowledge [2,8,15,17,20,30].
Yin et al. [2] suggested that orthodontists' concepts of profile esthetics are influenced by their specialist training and preference for using cephalometric measurements to assess facial profiles. However, orthodontists should not forget that the ultimate perception of facial attractiveness should be based on the perceptions of laypeople and patients, not orthodontists [11,14]. Recognizing the differences in their esthetic perceptions may help orthodontists understand their patients' expectations for orthodontic treatment and move them to consult with patients about the effects or limitations of treatment in terms of various facial areas during the diagnosis and treatment planning processes.
Smile esthetics, skin texture and color, lip color and form, etc., may affect the esthetic perception of facial attractiveness [31]. However, they were not included in this study because we could not take CBCT images that included them; this is a limitation of this study. Therefore, 3D photographs including these factors should be combined with CBCT images in future studies [32].

Conclusions
The null hypothesis was rejected; some differences were identified between the three groups of evaluators.

•
When evaluating facial attractiveness, all three groups of evaluators showed a high esthetic preference for retroclination and retrusion of incisors and lips. Still, they showed differences in how they perceived facial attractiveness. The results are valid only in Asian populations, within the limitations of this study. • The most influential facial views used to judge facial esthetics were the profile view for orthodontists (59.5%) and general dentists (57.2%) and the frontal view for laypeople (40.5%).

•
The most influential facial components used to judge facial esthetics were the lips for orthodontists (73.8%) and the chin for general dentists (38.1%) and laypeople (45.2%).

•
The ultimate perception of facial attractiveness needs to be based on the esthetic perceptions of laypeople. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The authors declare that the materials are available.

Conflicts of Interest:
The authors declare no conflict of interest.