Dimensional analyses of frontal posed smile attractiveness in Japanese female patients

The goal of orthodontic treatment is to obtain balanced facial appearance, stable occlusion, healthy oral tissues, and efficient mastication.¹  Assessment of facial esthetics is important in orthodontic diagnosis and in evaluating results.^2,3 

In the past, the facial esthetics of orthodontic patients has been analyzed mostly by comparing angles and distances on lateral cephalograms or profile facial photographs with those of subjects having normal occlusion and ideal or acceptable facial esthetics.⁴  Recent studies, however, have reported that the frontal smile has a significantly greater influence on human judgment of esthetics than does the profile.⁵  Subsequently, an increasing number of studies have focused on the frontal smile.⁶  These studies have examined a variety of components of frontal posed smile attractiveness in both vertical and transverse dimensions. However, clinically valuable items for objectively analyzing treatment goals have not yet been established.⁷ 

Therefore, the present study was conducted with the aim of identifying dimensional diagnostic items for objective analysis that correlate with human judgments of postorthodontic frontal smile attractiveness in Japanese female patients.

This study was approved by the ethics committee of the Nippon Dental University (Approval T2010-30).

One hundred female patients (mean age, 19.68 ± 4.46 years) who had undergone conventional orthodontic treatment, including the control of second molars, and were under 30 years of age at the completion of active treatment in the Nippon Dental University Hospital were selected for the present study. Seventy-five of the patients had been treated with two- or four-premolar extractions. Patients having had cleft lip, cleft palate, or orthognathic surgery were excluded.

Within one month after active treatments, patients were seated with a natural head position, fixed with bilateral ear rods, and photographed from a standardized distance of 150 cm (Figure 1). Frontal photos of the patients smiling appropriately were selected and scanned using a flatbed scanner (GT-X970; Epson, Tokyo, Japan). The scanned images were then converted into 2000-dpi monochrome TIFF images depicting only the lower third of the face.

The sample was estimated to have an effect size of 0.96, which was determined according to a previous study⁸  using the G*power 3.1 statistical program (Heinrich Heine Universitat Dusserdolf Experimentelle Psycologie, Dusseldorf, Germany). Analysis revealed that the required minimum sample size for each group was 18 to detect this effect size with 0.8 power and a significance level of P < .05.

A total of 20 students (10 males, 10 females; mean age, 20.65 ± 2.41 years) enrolled at the Nippon Dental University who had not yet completed a course in smile esthetics in orthodontics were selected because of their similarity to laypersons.^9,10  An equal number of male and female evaluators were used.^10–12 

According to the method of McNamara et al.,¹³  selected images were printed on evaluation sheets at a fixed width of 100 mm along with a 100-mm visual analogue scale (VAS) (Figure 2), which was used by the students to evaluate the attractiveness of the images. Student evaluations were measured in millimeters from the left end of the 100-mm bar using a digital caliper (NTD12-15CX; Mitutoyo, Kawasaki, Japan).

The images were ranked according to the results of the students' evaluations to determine the 75th and 25th percentiles. Images with evaluations higher than the 75th percentile (ie, more attractive) were categorized as group A, while images in the 25th percentile or lower were placed in group B (ie, less attractive).

Subsequently, 10 postorthodontic dental casts were randomly selected. The mesiodistal crown diameters of both maxillary central incisors were measured in millimeters using the digital caliper. Then, on corresponding photographic images, the same measurements were made using image analysis software (Scion Image Beta 4.03; Scion Corp, Torrance, Calif). The mean enlargement ratio of these images was calculated as 6.9144 pixels/mm by comparing the means of the mesiodistal crown widths of the 20 central incisors obtained from the dental casts and the corresponding images.

Thirteen reference points were identified in the selected images in both groups A and B (Figure 3). Then 12 objective analysis items (seven vertical and five transverse dimensions)^7,8,13–18  were calculated (Table 1, Figures 4 and 5). The means and standard deviations of the 12 items were also calculated for both groups and compared between the groups using unpaired t-tests by a statistical analysis program (SPSS version 21.0 for Microsoft Windows, Redmond, Wash) with the significance level set at P < .05.

In order to examine the intraexaminer reliability of the reference point identifications, one examiner repeated the same identifications for 10 randomly selected cases at a 2-week interval, and the x-y coordinates of the two sets of reference points were analyzed using Dahlberg's method.¹⁹  Next, to check the interexaminer reliability of the reference point identifications, two different evaluators made reference point identifications for each photograph. The results of repeated analyses were all within 1.00 mm, indicating appropriate intra- and interexaminer reliability.

Among all 100 postorthodontic treatment images, the mean VAS evaluation at the 75th percentile was 57.04 ± 18.95 mm, while the mean evaluation at the 25th percentile was 47.06 ± 18.82 mm. Ranking of the results of the VAS evaluations of smile attractiveness for the 100 images from highest to lowest revealed an S-curve with a strong curvature at both ends (Figure 6).

All 25 images in groups A and B are shown in Figures 7 and 8, respectively. The mean ± SD of the VAS evaluations are shown below each individual image.

For the Height-1, 2, 3, 4, and (Table 1, Figure 5) measurements, the mean values were significantly smaller in group A than in group B (P < .05; Table 2).

The finding of an S-curve (Figure 6) when the results of the VAS evaluations of smile attractiveness were displayed suggests that these particular images tend to obtain higher or lower VAS evaluations than the middle-ranked images, which supports the comparison of group A and group B. Furthermore, the graph showed a more acute curvature in the unattractive end than at the attractive end. As all the images were taken after the active orthodontic treatment phase, few unattractive images may have been included.

In the present study, significant differences between groups A and B were found for only 5 of 12 diagnostic items for objective analysis and all 5 of the items were in the vertical dimension (Table 2). The mean and SD of the five vertical dimensional items calculated for group A might be clinically useful as diagnostic norms for treatment objectives in the frontal posed smile analysis using standardized photographs.

McNamara et al.¹³  reported that dimensional measurements have inferior accuracy compared with ratio analysis because of the risk of error when calculating the actual distance from the width of the maxillary right central incisor. However, ratio analysis has not been shown to be significantly correlated with smile attractiveness.^13,20  Therefore, we focused on dimensional items and tried to improve reliability by using standardized photographs and calculating the enlargement ratio from the subjects' maxillary central incisors.

Transverse dimensional items^{7,8,13,15–18}  and vertical dimensional items^7,8,13–18  have been proposed for analyzing the attractiveness of frontal posed smiles, and these items have been reported to affect the smile attractiveness of orthodontic patients. However, in the present study, significant differences between the groups were found only in the items related to the vertical dimension, not the transverse dimension. The reason for this tendency might be the nature of the sensitivity of evaluating the vertical dimension because the human field of view is wider horizontally than vertically.²¹  Therefore, the study results may have been influenced by the nature of human vision.

One controversial issue in orthodontics over the past decade has been the resultant widening of the dental arch after nonextraction orthodontic treatment.²²  Consequently, the importance of analyzing the buccal corridor has been emphasized. In such studies, digitally manipulated images have been used to avoid the influence of tooth shape and color, smile line, and lip shape which often has been reported to affect smile attractiveness.^{9–12,23–26}  In the literature, digitally altered images have often been created from a single image of the same patient to obtain a variety of sizes of the buccal corridor.^{9–12,23–26}  However, studies on the buccal corridor with nonmanipulated photographs^13,27  have failed to report a correlation between the buccal corridor and smile attractiveness. Janson⁶  recently reviewed previous articles on this matter and concluded that images of actual patients are more acceptable than digitally altered smiles. Therefore, we utilized nonmanipulated photographs of Japanese orthodontic patients to analyze smiles and found no significant difference between groups A and B for Width-4 (right buccal corridor^7,8,13,16 ), Width-5 (left buccal corridor^7,8,13,16 ), Width-6 (right posterior corridor¹³ ), and Width-7 (left posterior corridor¹³ ) evaluated by the dental students. This finding suggests that the buccal corridor is not correlated with human judgments of smile attractiveness in Japanese orthodontic patients when nonmanipulated images are used, and supports the results of previous studies.^13,26–28 

In the present study, a significantly smaller Height-1, also known as interlabial gap^7,8,13,15,17  or smile height,¹⁸  was observed in group A, which had attractive frontal smiles compared with group B. Schabel et al.⁷  evaluated the Height-1 of smile photographs in the United States, and found an insignificant difference between the most attractive and least attractive frontal posed smiles. In their study,⁷  similar results were also found for Height-4, also frequently referred to as maximum incisor exposure,^7,8  and for incisor exposure,¹³  upper lip to incisal edge,¹⁴  and maxillary incisal display.¹⁷  These differing results suggest that it is important to consider the ethnic and cultural background of both patients and evaluators when analyzing the smiles of orthodontic patients. Ioi et al.²⁹  pointed out that there is insufficient evidence to support what is considered an attractive smile in the Japanese population. Therefore, the present study is the first to report about Height-1 and Height-4 among the Japanese; further studies comparing these findings with other populations with different ethnic and cultural backgrounds are necessary.

In the present study, the mean of Height-3, also frequently referred to as upper lip drape,^7,8  upper lip smile line,¹⁴  gingival display,¹³  and maxillary gingival display,¹⁸  was 0.53 ± 0.90 mm in the nonmanipulated images of young female patients in group A. This finding supports the results of previous studies that indicated 0 to 2.0 mm of gum display in a smile is preferred.^{14,18,29–33}  Our results also support the findings of Isiksal et al.¹⁸  and Ioi et al.²⁹  who reported the correlation between gingival display and smile esthetics, but contradict the findings of McNamara et al.¹³  who reported no correlation between gingival display and smile esthetics.

Recently, it has been proposed that it is necessary to consider components other than smile arc and gingival display—that influence perceptions of smile attractiveness—to help clinicians develop more satisfying treatment plans for their patients.³⁰  For example, it has been suggested that vertical lip thickness is one of the most important factors affecting facial esthetics.^13,34  Height-2 (intervermilion distance) was defined by Krishnan et al.¹⁶  as including vertical lip thickness. In the present study, Height-2 was significantly smaller in group A than in group B. This finding suggests that Height-2 is a valid item for evaluating smile attractiveness in orthodontic patients when considering vertical lip thickness.

The results in the present study for Height-5, also often referred to as lower lip to incisor^7,8,13  or maxillary incisal edge to lower lip,¹⁷  support Schabel et al.,⁷  who reported that the least attractive smiles had a significantly greater distance from the lower lip to the incisor, a finding that also contradicts the findings of McNamara et al.,¹³  who reported that a greater distance between the lower lip and the incisor resulted in a more attractive smile for orthodontists, although there is known to be strong agreement between orthodontists and laypersons in their subjective evaluations of smile esthetics.

It has been suggested that ideal posttreatment occlusion and posttreatment smile esthetics are not related.³⁵  Therefore, the findings of the present study are limited by the fact that the occlusion of the sampled patients was not considered. This should be done in future research.

• The following five dimensional diagnostic items were found to be correlated with human judgments of postorthodontic frontal smile attractiveness in Japanese female patients: interlabial gap, intervermillion distance, maxillary gingival display, maximum incisor exposure, and lower lip to incisor. All five items were in the vertical dimension only.