Three-dimensional assessment of facial asymmetry among pre-pubertal class III subjects: a controlled study

European Journal of Orthodontics 36 (2014) 431 435 doi:10.1093/ejo/cjt069 Advance Access publication 23 September 2013 The Author 2013. Published by Oxford University Press on behalf of the European Orthodontic Society. All rights reserved. For permissions, please email: journals.permissions@oup.com Three-dimensional assessment of facial asymmetry among pre-pubertal class III subjects: a controlled study Maja Ovsenik*, Giuseppe Perinetti**, Alexei Zhurov***, Stephen Richmond*** and Jasmina Primozic* *Department of Orthodontics and Jaw Orthopaedics, Medical faculty, University of Ljubljana, Ljubljana, Slovenia, **Department of Medical, Surgical and Health Sciences, School of Dentistry, University of Trieste, Trieste, Italy, ***Dental Health and Biological Sciences, Dental School, Cardiff University, Cardiff, UK Correspondence to: Jasmina Primozic, Department of Orthodontics, Medical Faculty, Vrazov trg 2, 1000 Ljubljana, Slovenia. E-mail: jasminaprimozic@gmail.com Summary Background/objective: Facial asymmetry is very common in adult class III patients; however, the degree of facial asymmetry in growing class III subjects has been poorly investigated. Therefore, the aim was to assess the degree of facial asymmetry of growing class III subjects and to compare it with a sample of growing subjects without malocclusion, recorded using a three-dimensional laser scanning method and classified according to the dentition phase. Subjects/methods: A group of 156 Caucasian subjects, 52 with class III malocclusion (28 females and 24 males) and 104 without malocclusion (control, 51 females and 53 males), with an overall mean age 6.7 ± 1.4 years (range 4.3 10.3 years), were included. The subjects were further subdivided according to the presence of either primary or early/intermediate mixed dentitions. Facial asymmetry was assessed on three-dimensional surface facial images obtained using a laser scanning device in terms of mirrored face distances and percentages of asymmetry. Multivariate analyses were used to assess the differences among class III and control groups. Results: Complete facial symmetry was not seen in any subject. No significant differences of facial asymmetry were observed between the class III and control groups, neither in the primary nor in the early/ intermediate mixed dentition phases. Limitations: No conclusions about longitudinal changes in the degree of facial asymmetry among class III subjects could be drawn from the present cross-sectional study. Conclusions: Based on three-dimensional facial asymmetry analysis, class III subjects do not show clinically relevant facial asymmetry, at least during the pre-pubertal growth period. Introduction Facial asymmetry can be recognized as differences in the size or relationship of the two sides of the face. This may be the result of discrepancies either in the form of individual bones or a malposition of one or more bones in the craniofacial complex (Sutton, 1968). It has been reported that the mandibular region may have relevant degree of asymmetry (Vig and Hewitt, 1975; Shah and Joshi, 1978; Primozic et al., 2012, 2013b) due to factors related to both dental arches and skeletal dysgnathia, which may be causes of craniofacial asymmetries (Lundström, 1961; Primozic et al., 2009; Primozic et al., 2013b,c). In the United States, 25 34 per cent of facial asymmetry is associated with skeletal dysgnathia (Proffit et al., 1990; Severt and Proffit, 1997), whereas in Europe, facial asymmetry was observed among 18 23 per cent dysgnathic patients (Willems et al., 2001; Good et al., 2006). class III malocclusion is defined as a sagittal discrepancy; therefore, its morphological characteristics were mainly investigated in the sagittal plane (Chen et al., 2007; Choi et al., 2010). On the contrary, very little data have been reported about any associated asymmetry in the frontal plane (Sutton, 1968; Haraguchi et al., 2002; Staudt and Kiliaridis, 2010). Facial asymmetry was observed in 40 per cent (Severt and Proffit, 1997) to 80 per cent (Haraguchi et al., 2002) adult class III patients as assessed by extra-oral photographs or X-ray analysis. Of interest, in the very same sample of adult class III patients, a noteworthy difference in the prevalence of skeletal (80 per cent) and soft tissue (56 per cent) asymmetry on the frontal plane has been reported (Haraguchi et al., 2002). However, the reported studies were performed on adult populations, with only one reporting a prevalence of 58 per cent facial asymmetry associated with class III malocclusion in growing, i.e. pre-pubertal, subjects (Haraguchi et al., 2008). Further, prevalence rather than the degree of asymmetry was assessed using two-dimensional methods; therefore, the degree of facial asymmetry using an objective and accurate method among pre-pubertal class III subjects remains to be determined.

432 M. OVSENIK ET AL. It has been reported that the characteristics of class III malocclusion have the tendency to worsen along with growth and development, especially during the pubertal growth spurt (Kuc-Michalska and Baccetti, 2010; Almeida et al., 2011; Primozic et al., 2013a), resulting in severe facial deformity not only in the sagittal but also in the frontal plane and therefore increasing treatment difficulty with age (Campbell, 1983; Guyer et al., 1986; Burns et al., 2010). Although the optimal treatment timing of class III malocclusion is still controversial, it has been accepted that orthopaedic approach for such malocclusion is effective when performed early at a pre-pubertal stage of development (Masucci et al., 2011). Considering that growth in the transverse plane ends up before those in the vertical and sagittal planes, an early diagnosis of any facial asymmetry in the frontal plane associated with class III malocclusion in growing subjects would be recommended. Therefore, the aim of the present study was to compare the degree of facial asymmetry of growing class III subjects with a sample of growing subjects without malocclusion, recorded using a three-dimensional laser scanning method and classified according to the dentition phase. Subjects and methods Ethical approval for this study was gained from the Slovenian Ethical Committee of the Medical University in Ljubljana, Slovenia, and informed consent was obtained from the parents of all subjects. A group of 156 Caucasian subjects, 52 with class III malocclusion (28 females and 24 males) and 104 without malocclusion (control, 51 females and 53 males), with an overall mean age 6.7 ± 1.4 years (range 4.3 10.3 years), were included. The subjects were further subdivided according to the presence of either primary or early/ intermediate mixed dentitions. Within each of the dentition phase, the number of subjects was 26 and 52 for the class III and control groups, respectively. The mean age for the class III group were 5.5 ± 0.5 years and 7.9 ± 0.6 years, whereas the mean age for the control groups were 5.4 ± 0.4 years and 8.0 ± 0.9 years in the primary and early/ intermediate mixed dentition phases, respectively. The class III subjects were randomly selected from a pool of patients referred to the Department of Orthodontics, University Clinical Centre. Only the subjects with a reverse overjet, at least a bilateral half-cusp class III molar (either primary or first permanent) relationship and a class III profile, were included. Further, Wits appraisal among all class III subjects had to be below a threshold value of 1 mm. Subjects with craniofacial syndromes or with unilateral posterior crossbite due to a lateral mandibular shift were excluded. The control subjects were randomly selected from a local school and only subjects without malocclusion and a normal profile associated with a good general health without any respiratory, mastication, or deglutition problems were included. All the clinical assessments were performed by an experienced operator (MO). The appraisal of the dentition phase was carried out according to the following definitions: 1. primary dentition, when all the primary teeth were erupted, 2. early mixed dentition, when the first permanent molars and incisors were erupting, 3. intermediate mixed dentition, when the permanent incisors and first molars had fully erupted, with presence of the primary canine, and first and second molars (Franchi et al., 2008). Assessment of facial asymmetry Surface facial images were obtained using two Konica/ Minolta Vivid 910 eye safe laser scanners (Konica Minolta, Inc., Japan) angled to capture left and right sides of the face with significant overlap in the anterior part of the face to facilitate registration and merging of the two images to produce one facial image (Kau et al., 2004). Natural head posture was adopted for this study as this has been shown to be clinically reproducible (Chiu and Clark, 1991). The technique for positioning the patient and image capture has been validated and described elsewhere and their use in growing subjects has been previously validated (Kau et al., 2004). The three-dimensional data were imported to a reverse modelling software package, Rapidform 2006 (INUS Technology Inc, Seoul, Korea). Each scan (left and right images) of the face was processed in order to remove unwanted data, registered and merged to produce a complete facial image. Left and right scans were merged only if there was at least 70 per cent matching between them in the overlap area with ±0.5 mm tolerance (Toma et al., 2009; Djordjevic et al., 2011; Primozic et al., 2012; Primozic et al., 2013c). The facial image was divided into three functional parts: 1. the upper part (forehead), defined as the part of the face above the endocanthion plane, 2. the middle part (maxillary), from the endocanthion plane to the plane through the outer commissures of the lips, and 3. the lower part (mandibular), below this plane (Figure 1). To check for left/right symmetry, the face was mirrored across the Y Z plane and the mirrored images were superimposed using the automatic best-fit procedure of the mirrored facial image surfaces (Kau and Richmond, 2010). Asymmetry was assessed quantitatively as the average distances (in millimetres) between the mirrored images and as the percentage of mirrored images not coinciding within 0.5 mm (percentage of asymmetry). The greater the average distance was between the mirrored images and the greater the percentage of asymmetry the greater the asymmetry of the face was. The parameters were calculated for the whole face and for each part of the face separately. The assessment was performed by an experienced operator.

FACIAL ASYMMETRY IN class III MALOCCLUSION 433 methods. For each part of the face two-way multivariate analysis of covariance (MANCOVA) were performed to assess differences in the average distance and asymmetry. Dentition phase and group were entered as independent fixed factors, whereas age and gender were entered as covariates. A P-value less than 0.05 was accepted for rejection of the null hypothesis. Figure 1 Assessment of facial asymmetry on colour deviation maps. Percentages of asymmetry were assessed on colour deviation maps of the mirrored facial images for the upper, middle, and lower parts of the face: dark gray colour indicates deviations within 0.5 mm, which we considered symmetric (histogram). The smaller the facial surface area is in dark gray colour the greater the percentage of facial asymmetry is detected. Sample size calculation and method error analysis A sample size of at least 26 subjects per group was necessary to detect an effect size coefficient (Cohen, 1992) of 0.8 for either of the asymmetry parameters (average distance or percentage of asymmetry) in any comparison between the groups, with an alpha set at 0.05 and a power of 0.80 (Perinetti et al., 2011). The effect size coefficient is the ratio of the difference between the recordings of the two groups, divided by the within-subject standard deviation. An effect size of at least 0.8 is regarded to as a large effect (Cohen, 1992). The method error for each asymmetry parameter was calculated using the intraclass correlation coefficients (ICC) on a random sample of 10 replicate measurements. With the aim of quantifying the full method error of the recordings for both of these asymmetry parameters, the method of moments (MME) variance estimator was used (Springate, 2012). Therefore, the mean error and 95% confidence intervals (CIs) between the repeated recordings were calculated using the MME variance estimator, and were expressed as percentages (Perinetti et al., 2012). The MME variance estimator has the advantages of not being affected by any unknown bias, i.e. systematic errors, between pairs of measurements (Springate, 2012). Statistical analysis The Statistical Package for Social Sciences (SPSS Inc., Chicago, Illinois, USA) was used to perform the data analysis. Each data set was tested for the normality of the data by means of the Shapiro Wilk test and by Q-Q normality plots; equality of variance was also tested by means of the Levene test and Q-Q normality plots of the residuals. Through this analysis, a root-square transformation of the data sets was required to meet the assumption for using parametric Results The ICC for the asymmetry parameters ranged from 0.85 to 0.92 for the average distance and percentage of asymmetry, respectively. Method errors as mean (95% CI) ranged from 5.6 per cent (2.6 9.2) to 7.5 per cent (3.5 12.4) for the percentage of asymmetry and average distance, respectively. No face was perfectly symmetric (Table 1). In the whole group (n = 156), the average distance between the mirrored images for the whole face ranged 0.22 0.85 mm and the percentage of asymmetry ranged 7.8 66.9. At the two-way MANCOVAs for the upper, middle, and lower parts of the face, no significant interaction was seen. The only exception was the whole face, in which the groups yielded a significant interaction at the multivariate level (F 2;149 = 4.01; P = 0.020). However, at the corresponding subsequent univariate analyses, no significant differences were seen. Moreover, no significant difference was seen among the different parts of the face (upper, middle, and Table 1 The facial asymmetry parameters in the class III and control groups for the whole face and each part of the face separately, according to the dentition phase. Parameter Portion Group Dentition phase Average distance (mm) Asymmetry (%) Primary Early/ intermediate mixed Whole class III 0.48 ± 0.20 0.42 ± 0.09 Control 0.40 ± 0.10 0.42 ± 0.08 Upper class III 0.42 ± 0.15 0.40 ± 0.09 Control 0.40 ± 0.10 0.41 ± 0.11 Medium class III 0.45 ± 0.22 0.42 ± 0.15 Control 0.41 ± 0.15 0.40 ± 0.14 Lower class III 0.39 ± 0.19 0.39 ± 0.14 Control 0.39 ± 0.14 0.45 ± 0.16 Whole class III 31.69 ± 14.38 30.34 ± 9.31 Control 28.06 ± 9.45 29.85 ± 7.87 Upper class III 29.16 ± 14.79 28.36 ± 10.40 Control 28.19 ± 10.81 29.00 ± 11.22 Medium class III 30.73 ± 19.37 31.30 ± 15.93 Control 28.48 ± 14.22 28.74 ± 13.00 Lower class III 26.94 ± 17.59 27.41 ± 14.32 Control 26.89 ± 13.56 33.91 ± 16.15 Data are presented as mean ± standard deviation; n = 52 and 104 for the class III and control group, respectively, equally divided into the primary and early/intermediate mixed dentition phase group. No significant differences between the groups were seen in any of the comparisons.

434 M. OVSENIK ET AL. lower) when this classification criterion was entered as a repeated factor in a further MANCOVA. Discussion In the present cross-sectional controlled study, facial asymmetry of class III subjects was assessed in the primary and early/intermediate mixed dentition phases and compared with a control group of subjects without malocclusion using a non-invasive three-dimensional laser scanning method. The results show no significant differences of facial asymmetry between the class III and control groups in the prepubertal growth period. Previous evidence shows that facial asymmetry is common among class III patients, as 40 per cent of Caucasian (Severt and Proffit, 1997) and 66 per cent of Japanese (Haraguchi et al., 2008) class III patients exhibit some degree of soft tissue asymmetry. However, previous investigations regarding facial asymmetry associated with class III malocclusion were limited by the use of twodimensional recordings on standard extra-oral photographs (Severt and Proffit, 1997; Haraguchi et al., 2008) or postero-anterior head films (Haraguchi et al., 2002). Moreover, only in these studies (Severt and Proffit, 1997; Haraguchi et al., 2002, 2008), facial asymmetry was not reported as absolute scores, but only as a positive when above a certain threshold, thus missing information about the degree of asymmetry in the examined sample. The present study made use of non-invasive three-dimensional recordings of the face with a full quantitative analysis and also included a control group of subjects with normal occlusion instead of class I malocclusion as previously reported (Severt and Proffit, 1997; Haraguchi et al., 2008). In this regard, to avoid biases in the recording of the lower face asymmetry due to lateral mandibular shift, class III subjects with unilateral posterior crossbite were not included in the present study. Further, the degree of facial asymmetry, rather than its prevalence, was examined in class III and control prepubertal subjects. According to the present results, no face (including those of the control group) was perfectly symmetric, neither in the primary dentition nor in the early/intermediate mixed dentition phase, as minor non-pathologic facial asymmetry is relatively common (Shah and Joshi, 1978). Further, previous studies have shown that gender and age do not have any effect on facial asymmetry (Ferrario, 2001; Djordjevic et al., 2011; Primozic et al., 2012, 2013b,c), which is in accordance with the results of the present study. Generally, no significant differences were observed in the degree of facial asymmetry among class III and control subjects. On the contrary, facial asymmetry parameters of the whole face yielded a significant interaction at the multivariate level of the MANCOVA. However, this statistically significant interaction does not appear to have clinical relevance. Furthermore, no significant differences were seen in the facial parameters within each group (class III and controls) according to different dentition phases for the upper, middle, and lower parts of the face. However, as previously reported, soft tissues may mask the underlying skeletal asymmetry (Peck et al., 1991), and a higher degree of skeletal asymmetry might had been found if a radiographic examination had been performed (Haraguchi et al., 2002). The subjects included in the present study were in the pre-pubertal growth period, when as suggested by previous reports the prevalence of facial asymmetry is lower compared with pubertal and post-pubertal growth phases (Haraguchi et al., 2008). However, Haraguchi et al. (2008) classified pre-pubertal subjects as being up to 11 years old, which may have caused a bias since this age is compatible with pubertal growth spurt (Baccetti et al., 2006). Instead, in the present study, the subjects included were all at ages and dentition phases compatible with only a pre-pubertal growth phase (Baccetti et al., 2006; Franchi et al., 2008). Indeed, facial asymmetry has the tendency to increase during growth and development, leading to greater facial disharmony in the post-pubertal growth period, regardless of malocclusion type (Haraguchi et al., 2008; Primozic et al., 2013b). Clinical implication As facial asymmetry in class III subjects does not have tendency to appear at a clinically relevant level in prepubertal subjects, while this becomes evident at later growth stages (Haraguchi et al., 2008), a proper interceptive treatment for this skeletal malocclusion may be indicated during early developmental stages, i.e. pre-pubertal. The treatment of class III malocclusion with an orthopaedic approach during childhood has been shown to be effective and with more long-term effects regarding the sagittal growth correction (Masucci et al., 2011). Although not shown in the present study, the results reported herein are consistent with prevention of the establishment of facial asymmetry if the class III growth tendency is treated by current orthopaedic modalities. In this regard, a comprehensive diagnosis of class III malocclusion should include also an analysis of the morphological features in the frontal plane. Conclusions Based on three-dimensional facial soft tissue recordings in pre-pubertal subjects, class III subjects do not exhibit a greater degree of facial asymmetry compared with subjects without malocclusion, irrespective of the different parts of the face. Further, the primary or early/intermediate mixed dentition phases have no influence on the degree of facial asymmetry, regardless of the presence of a class III malocclusion.

FACIAL ASYMMETRY IN class III MALOCCLUSION References Almeida M R, Almeida R R, Oltramari-Navarro P V, Conti A C, Navarro Rde L, Camacho J G 2011 Early treatment of class III malocclusion: 10-year clinical follow-up. Journal of Applied Oral Science 19: 431 439 Baccetti T, Franchi L, De Toffol L, Ghiozzi B, Cozza P 2006 The diagnostic performance of chronologic age in the assessment of skeletal maturity. Progress in Orthodontics 7: 176 188 Burns N R, Musich D R, Martin C, Razmus T, Gunel E, Ngan P 2010 Class III camouflage treatment: what are the limits? American Journal of Orthodontics and Dentofacial Orthopedics 137: 9.e1 9.e13 Campbell P M 1983 The dilemma of class III treatment. Early or late? Angle Orthodontist 53: 175 191 Chen F, Wu L, Terada K, Saito I 2007 Longitudinal intermaxillary relationships in class III malocclusions with low and high mandibular plane angles. Angle Orthodontist 77: 397 403 Chiu C S, Clark R K 1991 Reproducibility of natural head position. Journal of Dentistry 19: 130 131 Choi H J, Kim J Y, Yoo S E, Kwon J H, Park K 2010 Cephalometric characteristics of Korean children with class III malocclusion in the deciduous dentition. Angle Orthodontist 80: 86 90 Cohen J 1992 A power primer. Psychological Bulletin 112: 155 159 Djordjevic J, Toma A M, Zhurov A I, Richmond S 2011 Three-dimensional quantification of facial symmetry in adolescents using laser surface scanning. European Journal of Orthodontics 36: 125 132 Ferrario V 2001 The effect of sex and age on facial asymmetry in healthy subjects: a cross-sectional study from adolescence to mid-adulthood. Journal of Oral Maxillofacial Surgery 59: 382 388 Franchi L, Baccetti T, De Toffol L, Polimeni A, Cozza P 2008 Phases of the dentition for the assessment of skeletal maturity: a diagnostic performance study. American Journal of Orthodontics and Dentofacial Orthopedics 133: 395 400 Good S, Edler R, Wertheim D, Greenhill D 2006 A computerized photographic assessment of the relationship between skeletal discrepancy and mandibular outline asymmetry. European Journal of Orthodontics 28: 97 102 Guyer E C, Ellis E E, McNamara J A, Behrents R G 1986 Components of class III malocclusion in juveniles and adolescents. Angle Orthodontist 56: 7 30 Haraguchi S, Iguchi Y, Takada K 2008 Asymmetry of the face in orthodontic patients. Angle Orthodontist 78: 421 426 Haraguchi S, Takada K, Yasuda Y 2002 Facial asymmetry in subjects with skeletal class III deformity. Angle Orthodontist 72: 28 35 Kau C, Richmond S 2010 Three-dimensional imaging for orthodontics and maxillofacial surgery. Blackwell Publishing Ltd, Oxford Kau C H, Zhurov A, Scheer R, Bouwman S, Richmond S 2004 The feasibility of measuring three-dimensional facial morphology in children. Orthodontics and Craniofacial Research 7: 198 204 Kuc-Michalska M, Baccetti T 2010 Duration of the pubertal peak in skeletal class I and class III subjects. Angle Orthodontist 80: 54 57 Lundström A 1961 Some asymmetries of the dental arches, jaws, and skull, and their etiological significance. American Journal of Orthodontics and Dentofacial Orthopedics 47: 81 106 Masucci C, Franchi L, Defraia E, Mucedero M, Cozza P, Baccetti T 2011 Stability of rapid maxillary expansion and facemask therapy: a 435 long-term controlled study. American Journal of Orthodontics and Dentofacial Orthopedics 140: 493 500 Peck S, Peck L, Kataja M 1991 Skeletal asymmetry in esthetically pleasing faces. Angle Orthodontist 61: 43 48 Perinetti G, Baccetti T, Di Leonardo B, Di Lenarda R, Contardo L 2011 Dentition phase and chronological age in relation to gingival crevicular fluid alkaline phosphatase activity in growing subjects. Progress in Orthodontics 12: 100 106 Perinetti G, Marsi L, Di Lenarda R, Contardo L 2012 Is postural platform suited to study correlations between the masticatory system and body posture? A study of repeatability and a meta-analysis of reported variations. Progress in Orthodontics 13: 273 280 Primozic J, Farcnik F, Perinetti G, Richmond S, Ovsenik M 2013a The association of tongue posture with the dentoalveolar maxillary and mandibular morphology in class III malocclusion: a controlled study. European Journal of Orthodontics 35: 388 393 Primozic J, Ovsenik M, Richmond S, Kau C H, Zhurov A 2009 Early crossbite correction: a three-dimensional evaluation. European Journal of Orthodontics 31: 352 356 Primozic J, Perinetti G, Richmond S, Ovsenik M 2013b Three-dimensional evaluation of facial asymmetry in association with unilateral functional crossbite in the primary, early, and late mixed dentition phases. Angle Orthodontist 83: 253 258 Primozic J, Perinetti G, Zhurov A, Richmond S, Ovsenik M 2012 Assessment of facial asymmetry in growing subjects with a threedimensional laser scanning system. Orthodontics and Craniofacial Research 15: 237 244 Primozic J, Richmond S, Kau C H, Zhurov A, Ovsenik M 2013c Threedimensional evaluation of early crossbite correction: a longitudinal study. European Journal of Orthodontics 35: 7 13 Proffit W R, Phillips C, Dann C T 1990 Who seeks surgical-orthodontic treatment? International Journal of Adult Orthodontics and Orthognathic Surgery 5: 153 160 Severt T R, Proffit W R 1997 The prevalence of facial asymmetry in the dentofacial deformities population at the University of North Carolina. International Journal of Adult Orthodontics and Orthognathic Surgery 12: 171 176 Shah S M, Joshi M R 1978 An assessment of asymmetry in the normal craniofacial complex. Angle Orthodontist 48: 141 148 Springate S D 2012 The effect of sample size and bias on the reliability of estimates of error: a comparative study of Dahlberg s formula. European Journal of Orthodontics 34: 158 163 Staudt C B, Kiliaridis S 2010 Association between mandibular asymmetry and occlusal asymmetry in young adult males with class III malocclusion. Acta Odontologica Scandinavica 68: 131 140 Sutton P R 1968 Lateral facial asymmetry-methods of assessment. Angle Orthodontist 38: 82 92 Toma A M, Zhurov A, Playle R, Ong E, Richmond S 2009 Reproducibility of facial soft tissue landmarks on 3D laser-scanned facial images. Orthodontics and Craniofacial Research 12: 33 42 Vig P S, Hewitt A B 1975 Asymmetry of the human facial skeleton. Angle Orthodontist 45: 125 129 Willems G, De Bruyne I, Verdonck A, Fieuws S, Carels C 2001 Prevalence of dentofacial characteristics in a Belgian orthodontic population. Clinical Oral Investigations 5: 220 226