1. Introduction
Patients with Class III malocclusion with mandibular prognathism have a high prevalence of facial asymmetry [
1,
2,
3,
4]. The prevalence of skeletal Class III face asymmetry has been reported to range from 47.9% to as high as 85% in Asian cohorts [
3,
4]. The lower face, particularly the chin, deviates more frequently than the upper face [
2,
3]. The occurrence of facial skeleton asymmetry may result from variations in the morphology of bony structures and the positional deviation of the maxillomandibular complex, or only the mandible [
5,
6].
For patients with severe skeletal jaw discrepancy and facial asymmetry, a combination of orthodontic treatment and orthognathic surgery (OgS) is considered the only viable treatment option for improving facial appearance and restoring normal occlusal function [
5]. The goal of 2-jaw OgS is to correct the overall dentofacial asymmetry regarding dental occlusion, align maxillary midline and chin to the facial midline, level the oral commissure and occlusal plane cant, and correct vertical and sagittal discrepancy [
6]. Adjunctive surgeries or treatments, such as bone shaving, bone grafting, and soft tissue grafting, can improve contour asymmetry and enhance harmony of the face form [
7,
8]. With the development of 3D images, cone-beam computed tomography (CBCT), and computer-assisted surgical simulation, 2-jaw OgS can be precisely planned, assisting surgeons to achieve similar treatment outcomes to correct facial asymmetry [
9,
10,
11].
In many cases, however, residual asymmetry persists postoperatively [
12]. Lin et al. reported that mandibular asymmetry persisted in 45 patients with facial asymmetry undergoing OgS, even though the center point of the mandible was well aligned with the facial midline [
13]. According to the previous study from our team, Lin et al. analyzed 24 patients with skeletal Class III malocclusion with double-jaw OgS, who were classified according to the outcome of subjective visual perception scores [
14]. The findings demonstrated that midline parameter deviation, shape of the mandibular border, and contour of menton morphology influenced the visual perception of postoperative asymmetry.
Chen et al. categorized mandibular asymmetry into three groups according to the amount and direction of ramus asymmetry relative to menton deviation in 3D-CBCT analysis [
15,
16]. They concluded that ramus asymmetry was less predictable. However, no specific factors were found to contribute to residual facial asymmetry.
The study aims to explore the preoperative anatomical variables affecting the outcome of surgical correction in facial asymmetry in patients with skeletal Class III malocclusion. The null hypothesis was that the anatomical factors are not different between the surgical outcome of good facial symmetry and residual asymmetry.
2. Materials and Methods
2.1. Patients
This retrospective study evaluated 37 consecutive adult Taiwanese patients with skeletal Class III facial asymmetry, who received bimaxillary OgS from 2014 to 2016 at the Chang Gung Memorial Hospital, Taipei, Taiwan.
The inclusion criteria were as follows: (1) having skeletal Class III dentofacial deformities with preoperative menton deviation >4 mm from the midsagittal plane (MSP); (2) undergoing bimaxillary OgS (Le Fort I osteotomy and BSSO); (3) having complete preoperative and posttreatment CBCT data; and (4) receiving treatment from the same orthodontist and the same group of surgeons.
The exclusion criteria of the subjects were: (1) craniofacial anomalies, such as cleft lip and palate, hemifacial microsomia; (2) craniofacial trauma; and (3) untreated temporomandibular joint disorder.
The sample size was calculated using G*Power (version 3.1.9.4; Universität Kiel, Kiel, Germany) [
17]. On the basis of the difference in Go to FHP, significance of 95%, and power of 80%, the minimum sample size required for the two-sample
t-test was 36.
This study followed the guidelines of the Declaration of Helsinki. The research protocol was approved by the Institutional Review Board and medical ethics committee of the Chang Gung Memorial Hospital.
2.2. Data Collection
The patients’ 3D craniofacial images were taken in the natural head position with maximum intercuspation, by using an i-CAT scanner (Image Sciences International, Hatfield, PA), with the following settings: 120 kVp; 36.9 mA; field of view, 22 × 16 cm; scanning time, 40 s; and voxel size, 0.4 × 0.4 × 0.4 mm. All data were obtained 1 month before surgery (T0) and on completion of orthodontic treatment (T1).
The DICOM format from the CBCT scanner was imported into the Simplant O&O software (Materialise Dental NV, Leuven, Belgium) to reconstruct the 3D skull models, including the cranium with maxilla and mandible. Maxillary and mandibular dental casts were digitized by a 3D scanner (3Shape, Copenhagen, Denmark) and integrated into the CBCT model by surface matching.
2.3. Presurgical 3D Surgical Simulation
The guidelines and procedures for 3D surgical simulation were reported in a previous study [
11]. Both clinical facial and 3D cephalometric measurements were incorporated into the development of the 3D surgical plan.
The surgical stent for the 2-jaw OgS was fabricated according to the surgical occlusion setup in the maxillary and mandibular dental casts. An intermediate stent was fabricated with a 3D-printed technique to guide the surgery.
2.4. Surgical Technique
All patients underwent double-jaw OgS. The maxillo-mandibular complex (MMC) was completely mobilized with LeFort I osteotomy and BSSO. The sequence of surgery was based on the designated intermediate stent. The final 2-jaw movement was guided by the final occlusion stent with temporary intermaxillary fixation during surgery. The surgeons finalized the MMC pitch rotation according to the 3D surgical simulation and final assessment of the facial symmetry and profile. Next, the MMC was fixed to the maxillary base and mandibular proximal segments with rigid fixation. Genioplasty surgeries were conducted to enhance the chin contour as a last step if required.
The BSSO technique conducted at our center was modified from the Hunsuck technique, with a more anteriorly extended anterior osteotomy cut [
18,
19]. This technique facilitates later intraoral placement of the plates and screws, and allows for a greater amount of mandibular repositioning with good long-term stability.
For Asian women, simultaneous mandibular angle contouring is a common adjunct procedure to improve the “square face” [
20]. The medial cortex of the mandibular angle on the proximal segment can be harvested for a high-quality bone graft, if necessary [
21].
2.5. 3D Dentoskeletal Measurements
The 3D skull models at T0 and T1 were superimposed on the anterior cranial base, and frontal and periorbital surface. The deviation values were calculated, and a value of 0.5 mm was considered acceptable to ensure that the corresponding reference areas had the highest possible accuracy.
The 3D reference planes were constructed using the landmarks identified on the preoperative objects (
Figure 1).
Table 1 lists the definition of the 3D landmarks, and depicts the linear and angular measurements of each landmark relative to the reference planes (
Figure 2,
Figure 3,
Figure 4,
Figure 5,
Figure 6 and
Figure 7). The deviated side was defined as the side of the face, including the skeletal menton, and the opposite side was the contralateral half (
Figure 2). All landmarks were identified by the same investigator.
2.6. Classification of Patients after Treatment Completion (T1)
The distances from each landmark to the three reference planes were measured and used to calculate the asymmetry index: dx, dy, and dz represent the distance from the landmark to the MSP, FHP, and CP, respectively.
For bilateral asymmetry index, asymmetry index = √((d_d x − [d_o x)]^2 + (d_d y − [d_o y)]^2 + (d_d z − [d_o z)]^2),
where ddx, ddy and ddz represent the distance from the landmark on the deviated side to the MSP and dox, doy and doz represent the distance from the landmark on the opposite side to the MSP.
According to the asymmetry index after treatment completion, patients were categorized into the relative symmetry outcome group (group S) and the relative asymmetry outcome group (group A). One median was excluded, and each group had 18 patients.
Figure 8 illustrates the study flowchart.
2.7. Statistical Analyses
The Mann–Whitney U test was used to compare the between-group differences in preoperative anatomical variables. p < 0.05 was set as statistically significant. Spearman correlation analysis was performed to examine the relationship of all dentoskeletal variables in T0 with the facial symmetry outcome—asymmetry index (T1)—for all patients.
For error study, 10 randomly selected cases were re-landmarked and measured within 3 weeks. The intraexaminer reliability test was conducted using intraclass correlation (ICC) [
22] and Dahlberg’s formula [
23].
4. Discussion
This study used 3D linear and angular measurements to clarify the preoperative anatomical variables affecting the outcome of surgical correction in Class III facial asymmetry.
Patients with skeletal Class III malocclusion and facial asymmetry, with a menton deviation greater than 4 mm, were included in this study. According to Kwon et al., the perception of asymmetry of a chin deviation of 2 mm to the right and 4 mm to the left is not detected clinically [
24]. Alongside that, most individuals, whether dental professionals or laymen, can notice a chin deviation of more than 4 mm [
25]. Thus, it is acceptable to define 4 mm of menton deviation relative to MSP as facial asymmetry. The mean preoperative menton deviation in the two groups was 6.52 ± 5.36 mm in group S and 8.15 ± 4.19 mm in Group A (
Table 3). The patients included in our study showed a severe deviation of menton, indicating a greater severity of facial asymmetry.
After treatment completion, facial asymmetry significantly improved in all patients, but did not completely resolve. Lin et al. reported that preoperative 3D dentoskeletal variables were not related to subjective visual perceptions of postoperative asymmetry [
14]. In the present study, we used a more objective method to define postoperative asymmetry. However, the menton deviation decreased significantly after treatment, and the difference in midline deviation was not appropriate for defining postoperative symmetry and asymmetry. Currently, there is no objective 3D index to represent the degree of facial asymmetry. Therefore, we used the asymmetry index at treatment completion to classify patients into two groups and evaluate the preoperative variables.
We observed that only bilateral variables, and not midline variables, significantly affected the surgical outcomes. This may be because midline deviation is already corrected by surgeons with the aid of 3D surgical simulation [
10,
11]. For bilateral discrepancy, four variables were observed to significantly affect the surgical outcome: (1) maxillary anterior occlusal canting, (2) maxillary posterior occlusal canting, (3) ΔGo to the FHP, and (4) Δmandibular axis to the FHP. These variables indicate the “roll asymmetry” of the maxilla and mandible.
For maxillary occlusal canting, the greater the preoperative discrepancy, the higher the possibility of residual asymmetry. However, maxillary occlusal canting was corrected well in both groups. Thiesen et al. found that the odds of presenting with mandibular asymmetry were significantly higher in the presence of maxillary asymmetry [
26]. The mandible is often associated with craniofacial asymmetries, with maxillary asymmetries often secondary to asymmetrical mandibular growth [
27]. Severe maxillary canting might indicate more severe deviation of the mandible and more severe asymmetry of the whole face. The occlusal plane canting was usually the result of compensating growth to mandibular asymmetry. The reasons underlying the correlations are not certain and deserve future investigation.
Correction of bilateral gonion contour discrepancy is more difficult and relies more on the surgeon’s decision-making during surgery. Trimming or contouring the gonion angle or adaptation of the proximal and distal segments all affect the surgical outcome. Presurgical ΔGo to MSP did not affect the surgical outcome, but ΔGo to the FHP and Go of the deviated side to CP showed a significant between-group difference. This might indicate that the preoperative roll (ΔGo to the FHP) and yaw (Go of the deviated side to CP) mandibular asymmetry affects the outcome. In Asia, most patients request mandibular angle reduction during surgery due to the preference for a slim and oval face [
20]. Thus, the 3D gonion point would be altered after surgical contouring of the mandibular angle.
In mandibular variables on the deviated side, a significant difference was observed in the distances between the bilateral gonion, glenoid fossa, condyle, and medial and lateral condylar poles and the coronal plane between groups A and S. The distance of the deviated side variables to the coronal plane was smaller in group A than in group S. On the opposite side, the distances between the bilateral medial and lateral condylar poles to the MSP were significantly different between the two groups. The distance in group A is greater than that in group S. The different results between the deviated and opposite side variables indicate that the preoperative condyle position in the sagittal and transverse directions might relate to a higher asymmetry index after surgery. One reason is that the condyle and glenoid fossa might be anatomical limitations, because we could not move or change their original position during surgery. This should be considered during 3D planning. In clinical situations, when the distance of the CoL–MSP is greater in the opposite side than in the deviated side, the ramus axis is affected, causing facial asymmetry, even when the midline landmarks are perfectly aligned. In such cases, the patient should be informed preoperatively of a greater chance of residual asymmetry. The mandible border contouring or fat transplantation are choices for further asymmetry correction [
8].
Some studies have discussed different surgical techniques to improve the outcome of ramus symmetry. The greatest advantage of intraoral vertical ramus osteotomy (IVRO) over BSSO is a lower incidence of injury to the inferior alveolar nerve [
28,
29]. Other advantages of IVRO include technical simplicity, shorter operation time, and less adverse effects related to the temporomandibular joint [
30]. However, BSSO provides better bony interface with proximal and distal segments, and easier use of rigid fixation, thus minimizing the need for intermaxillary fixation [
31]. Therefore, there is larger interference between the proximal and distal segments in IVRO during mandibular setback, resulting in a more significant increase of the total ramus angle, compared with that in BSSO [
32]. However, Lee et al. had a different opinion: they reported a series of three cases with severe rotational asymmetry, using unilateral IVRO on the shorter side combined with sagittal split ramus osteotomies on the contralateral side for a greater setback [
33]. They proposed that IVRO causes less rotational displacement of the proximal segment on the deviated side. The condylar segments that were displaced or rotated during surgery might return to their original physiologic position. The outcome and stability of the two different combined osteotomies on the mandible, in correction of face asymmetry, require further investigation.
Many factors are related to residual facial asymmetry, such as anatomical factors, surgical factors, surgical relapse, soft tissue response and adaptation, and asymmetric dynamic facial movement. However, in this study, we only discussed anatomical factors. This study also had the following limitations: first, we only discussed skeletal and dental variables, while the information of real facial soft tissue appearance was lacking. Second, the 3D reference planes were designed by skeletal landmarks; the setting of the 3D reference plane may be different from real patients. In clinical situations, the whole face, including eyes and ears, should be considered when deciding the reference plane. Finally, considerable interindividual variation may exist in the etiology of facial asymmetry, and thus, variables accounting for each type of asymmetry classification. Further investigations might be needed to develop different surgical consideration and modification for outcome improvement in correction of facial asymmetry. Additionally, a larger sample size and soft tissue evaluation could provide more informative outcome assessment in the future study.