ORIGINAL ARTICLE

Assessing soft-tissue characteristics of facialasymmetry with photographs

Mi-song Lee,a Dong Hwa Chung,b Jin-woo Lee,c and Kyung-suk Chad

Cheonan, Korea

Introduction: Precise diagnosis and treatment of facial asymmetry are important in orthodontics. The aims ofthis study were to determine the soft-tissue characteristics of patients perceived to have severe asymmetryrequiring treatment and the soft-tissue factors affecting the subjective assessment of facial asymmetry.Methods: In the first part of this study, 5 observers examined 1000 photographs of patients receiving ortho-dontic treatment and selected 100 for further assessment. These photographs showed 50 patients who wereconsidered to have little or moderate asymmetry and 50 who were considered to have severe asymmetry. Apilot study was performed to select the reference photographs representing the most symmetric (score of 0)and the most asymmetric (score of 100). A panel of 9 orthodontists then rated the facial asymmetry of the 100patients on a 100-mm visual analog scale. The scale was divided into 3 equal regions. Region 1 included pa-tients with the least facial asymmetry; according to the orthodontists, these patients did not require treatment.Region 2 included patients with moderate facial asymmetry who did not require treatment. Region 3 includedpatients with the most facial asymmetry who did require treatment. Results: One-way analysis of varianceshowed that lip canting, chin deviation, body inclination difference, and gonial angle difference had significantdifferences between the groups. Chin deviation and gonial angle difference were significant factors affectingthe assessment of facial asymmetry, according to stepwise linear regression analysis. Conclusions: Theseresults will help in the diagnosis and treatment planning for patients with asymmetry. (Am J OrthodDentofacial Orthop 2010;138:23-31)

About 80% of adults who seek orthodontic treat-ment want to improve facial aspects that are notassociated with structural or functional prob-

lems.1 Patients who do not perceive their facial asym-metry in the early stage of orthodontic treatmentmight begin to recognize it as treatment progresses.2,3

De Smit and Dermaut4 and Foster5 reported that faceswith ideal proportions and symmetry are not beautiful.On the other hand, they reported that abnormal propor-tions and asymmetric faces have unfavorable aspects.

The correction of facial asymmetry is becoming animportant goal of orthodontic treatment and orthog-nathic surgery. It has been reported that all patientshave some craniofacial asymmetry, including those per-ceived as normal.6 Ferrario et al7 found variable degrees

From the Graduate School of Dentistry, Dankook University, Cheonan, Korea.aPostgraduate student.bAssociate professor, Department of Orthodontics, School of Dentistry.cProfessor and chairman, Department of Orthodontics, School of Dentistry.dProfessor, Department of Orthodontics, School of Dentistry.

The authors report no commercial, proprietary, or financial interest in the

products or companies described in this article.

Reprint requests to: Dong Hwa Chung, Dankook University, School of

Dentistry, Department of Orthodontics, San 7-1, Shin-Bu Dong, Cheonan,

ChungNam, South Korea; e-mail, abeh@dankook.ac.kr.

Submitted, March 2008; revised and accepted, August 2008.

0889-5406/$36.00

Copyright � 2010 by the American Association of Orthodontists.

doi:10.1016/j.ajodo.2008.08.029

of soft-tissue facial asymmetry in healthy white dentalstudents with normal dentition. Therefore, it is widelyrecognized that facial asymmetry is often present inthe normal craniofacial complex. In addition, it wasreported that an occlusal cant of 0� to 3� has beenobserved in normal, healthy patients.8,9

There is no invisible demarcation line, or evena range, marking the distinction between normal andabnormal asymmetry in terms of soft-tissue characteris-tics. A demarcation line determining the necessity oforthognathic surgery for facial asymmetry is alsounknown. The borderline between normal and abnormalasymmetry might be determined by subjective evalua-tion such as the patient’s, or the orthodontist’s, percep-tion of facial asymmetry.

Although cephalometric measurements address skel-etal symmetry,10 subjective evaluations, such as the per-ception of facial asymmetry, might be based on soft-tissue features, including the outline of the face.11-13

Robinson et al14 reported that a beautiful face should beharmonious with comparable size and position ofthe skeletal structures and soft tissues. They stated thata favorable face can be shown by the soft tissues.

A diagnosis of facial asymmetry is normally madebased on the measurements from a posteroanteriorcephalogram or 3-dimensional (3D) skeletal computedtomography (CT) images by a clinician. However, the

23

24 Lee et al American Journal of Orthodontics and Dentofacial Orthopedics

July 2010

patient’s decision for treatment or satisfaction with theresult is subjective and can be assessed by a perceptiveassessment.15 There is a gap between the patient and theclinician in terms of the assessment of facial asymmetry.Therefore, the first assessable tool for patients withfacial asymmetry must be the soft tissues.

There has only been 1 study on the correlationbetween the perception of facial asymmetry and thecephalometric measurements, and no reports of therelationship between the perception of facial asymmetryand the soft-tissue measurements.16

The aims of this study were to survey the facial fron-tal photographs by subjectively evaluating facial asym-metry with a visual analog scale (VAS) and to determinethe differences in soft-tissue measurements between thegroups. The next steps were to determine the soft-tissuecharacteristics of patients perceived as having severeasymmetry and the borderline of abnormal asymmetry.Finally, the factors affecting the severity of the subjec-tive assessment of facial asymmetry were determined.This study should help clinicians make a differentialdiagnosis of asymmetric patients and develop anappropriate treatment plan.

MATERIAL AND METHODS

Five observers screened 1000 standardized facialfrontal photographs and selected 50 with little or no fa-cial asymmetry and another 50 with moderate or severefacial asymmetry. The inclusion criteria were asfollows: over 18 years of age, no congenital abnormal-ities in the maxillofacial region, no prior surgery for aninjury involving the maxilla or the mandible, and stan-dardized facial photographs taken before treatmentwith sufficient quality for evaluation.

The final sample consisted of photographs of 61women and 39 men (ages, 18.0-30.1 years; mean, 25.1years). The photographs had been taken with a digitalcamera (350D, Canon, Seoul, Koea) with a distance of1.5 m between the patient and the focus. The patientswere seated upright on a special chair, with the Frank-fort plane parallel to the floor. The patients lookeddirectly at the camera lens. Both ears were exposed,and the right and left distances between the exocanthusand the hairline were the same. The patients were inocclusal rest postion.

The color photographs were printed individually onsize A4 paper. The printed photographs were used tomeasure several soft-tissue measurements.

Five orthodontists in the Department of Orthodontics,Dankook University, Cheonan, Korea, selected the 5 mostsymmetric facial photographs and the 5 most asymmetricfacial photographs. The photographs selected most

frequently were called the reference photographs torepresent the most symmetric and the most asymmetricfaces; they were assigned scores of 0 and 100.

A panel of 9 orthodontists from the Department of Or-thodontics, Dankook University, were asked to judge the100 frontal photographs. The photographs were pre-sented for 1 minute each, in random order. The orthodon-tists rated the facial asymmetry on a 100-mm VAS,marked ‘‘most symmetric’’ (0) on the left and ‘‘mostasymmetric’’ (100) on the right. A 100-mm ruler was di-vided into 3 equal regions, and the meaning of the regionswas explained to the panel. Region 1 included patientswith little or no facial asymmetry and not requiring treat-ment. Region 2 included patients with moderate facialasymmetry but still not requiring treatment. Region 3 in-cluded patients with the most facial asymmetry who re-quired treatment. All assessments were subjective.

During the session, the 2 reference slides represent-ing 0 and 100 on the VAS were projected continuously.The panel was told of these set scores. The assessmentswere repeated twice with 2 weeks between sessions.

When each judge had finished, the 100-mm VASvalues were measured to the nearest 0.1 mm by usinga caliper and recorded on a data sheet. The raw datawere entered in an Excel spreadsheet (Microsoft,Redmond, Wash).

The 9 assessment scores of each patient were aver-aged. The average scores from first and second sessionsof each patient were compared by using the Student ttest. The difference between assessments was not signif-icant. Therefore, the second-session scores were used.

The patients were classified into 3 groups accordingto the average assessment scores in the previously fixedregions. Group I patients had scores of 0 to 33.3, groupII patients had scores of 33.4 to 66.7, and group IIIpatients had scores of 66.8 to 100. The printed frontalphotographs were measured to analyze the soft-tissuecharacteristics of each group.

The soft-tissue landmarks used in this study aredescribed in Table I and Figure 1. Most points used inthis study were proposed by Farkas.17

The midsagittal reference line was defined as theline from glabella to subnasale. The horizontal refer-ence line was the line perpendicular to the midsagittalline passing through the midpoint of both pupils. Tenlines, except for the midsagittal and horizontal referencelines, were established. These lines were used for theangular measurements. These were the bipupillaryline, otobasion inferius line, lip line, gonion line, prona-sale line, chin line, ramus line (right and left), and man-dibular body line (right and left) (Table II).

Nine angular and 2 linear measurements were made.The angular measurements were eye canting, otobasion

Table I. Definitions of landmarks used in this study

Landmark Definition

Pp (pupil) The apparently black circular opening in the center of the iris of the eye

G (glabella) The most forward projecting point of the forehead in the midline of the supraorbital ridges

Na (soft-tissue nasion) The middle point of the soft-tissue frontonasal suture

O (otobasion inferius) The inferior insertion of the ear

Sn (subnasale) The point at which the columella merges with the upper lip in the midsagittal plane

Pr (pronasale) The middle point of the outline of the nose tip

Ch (cheilion) The most lateral extent of the outline of the lips

Me (soft-tissue menton) The most inferior point of the soft-tissue outline on the chin

Go (soft-tissue gonion) The most everted point of the soft-tissue outline of the angle of the mandible

Pre (preaureculare) The most lateral point of the soft-tissue facial outline in front of tragus

Zero point The intersection of midsagittal line and horizontal line

American Journal of Orthodontics and Dentofacial Orthopedics Lee et al 25Volume 138, Number 1

canting, lip canting, gonion canting, nose deviation,chin deviation, ramal inclination difference, body incli-nation difference, and gonial angle difference. Table IIIgivess the definitions of these measurements. Eye cant-ing, otobasion canting, lip canting, and gonion cantingwere measured by using the horizontal reference line.Nose deviation, chin deviation, ramal inclination, andbody inclination were measured by using the midsagit-tal reference line. The ramal inclination difference,body inclination difference, and gonial angle differencewere measured to calculate the difference between theright and left angles. All angular measurements wereused as absolute quantities.

Statistical analysis

The linear measurements were calculated by usingthe asymmetry index (Table III). Each distance for rightand left soft-tissue gonion (Go) was measured from themidsagittal and horizontal reference planes, respec-tively. The asymmetry indexes were calculated withthe following formula:

asymmetry index ð%Þ5jR� Lj=M3100;

where R is the value of the right distance, L is the valueof the left, and M is the average of the right and leftvalues.

The distance for Go from the midsagittal plane wasdefined as the horizontal Go value. The distance fromthe horizontal plane was defined as the vertical Govalue. Therefore, there were 2 types of asymmetry in-dexes: for horizontal Go and vertical Go. The asymme-try index for horizontal Go is the ratio of the right andleft horizontal Go lengths; the horizontal Go lengthmeans the length from the midsagittal reference lineto Go (right and left). The asymmetry index for verticalGo is the ratio of the right and left vertical Go length; thevertical Go length means the length from the horizontalreference line to Go (right and left).

The SPSS software program (version 10.0, SPSS,Chicago, Ill) was used. One-way analysis of variance(ANOVA) was used to analyze the differences amongthe 3 groups. Bonferroni tests were performed forpost-hoc analysis of the differences. The sex differenceswere examined with the Student t test.

Stepwise linear regression analysis was performedto determine the independent variables that wereassociated most closely with the assessment of facialasymmetry (dependent variables). The linear regres-sion equation was:

Y 5 a 1 b1X1 1 b2X2 1 . 1 bkXk

where, a, b1, and b2 . bk are constants, and X1, X2 .,and Xk are independent variables that are combined lin-early to explain the variation in the dependent variable(Y).

Reliability was assessed by using duplicate mea-sures on a subset of randomly selected photographsscored 2 weeks apart. Systemic error, assessed by com-paring the mean differences between replicates to thestandard error, was not statistically significant. The ran-dom technical error was calculated by using Dahlberg’sformula.18 The method error was\2.0� for all measure-ments except for ramus inclination difference, whichwas 2.5�.

RESULTS

The patients were classified into 3 groups by usingthe assessment scores of facial asymmetry. Group I in-cluded patients with little or no facial asymmetry, notrequiring treatment. Group II included patients withmoderate facial asymmetry but not requiring treatment.Group III included patients with severe asymmetry, re-quiring treatment. One-way ANOVA showed that lipcanting, chin deviation, body inclination difference,and gonial angle difference were significantly differentbetween the groups. In group I, the mean scores of the

Fig 1. A drawing used for the subjective evaluation andthe landmarks on the facial soft tissues: 1, Pp (pupil); 2, O(otobasion inferius); 3, Ch (cheilion); 4, Go (soft-tissuegonion); 5, Me (soft-tissue menton); 6, G (glabella); 7,Sn (subnasale); 8, Pr (pronasale); 9, Na (soft-tissuenasion); 10, Pre (preaureculare).

Table II. Definitions of reference lines constructed inthis study

Reference line Definition

Midsagittal line G-Sn

Horizontal line Perpendicular to midsagittal

line passing through

the midsagittal line

Bipupillary line Right Pp-left Pp

Otobasion inferius line Right O-left O

Lip line Right Ch-left Ch

Gonion line Right Go-left Go

Pronasale line Na-Pr

Chin line Zero point-Me

Ramus line (right and left) Pre-Go (right and left)

Mandibular body line (right and left) Go (right and left)-Me

26 Lee et al American Journal of Orthodontics and Dentofacial Orthopedics

July 2010

factors for lip canting, chin deviation, body inclinationdifference, and gonial angle difference were 1.61�,1.11�, 2.80�, and 2.31�, respectively. In group II, themean scores were 2.29�, 2.00�, 3.75�, and 3.05�, respec-tively. In group III, the mean scores were 3.09�, 3.64�,7.11�, and 6.14�, respectively (Table IV).

The sex differences were not significant statisticallyaccording to the Student t test. Chin deviation tendencyof each group was evaluated in Table V.

Pearson correlation coefficients were examined todetermine the factors affecting the assessment of facialasymmetry (Table VI). The affecting factors were foundto be lip canting, chin deviation, ramus inclination dif-ference, body inclination difference, and gonial angledifference. The Pearson correlation coefficients for lipcanting, chin deviation, ramus inclination difference,body inclination difference, and goninal angle were0.33�, 0.59�, 0.21�, 0.48�, and 0.44�, respectively.

Stepwise linear regression analysis was used tounderstand the affecting factors better (Table VII).

The analysis showed that chin deviation and gonialangle difference significantly affected the assessmentof facial asymmetry. Therefore, 2 regression equationswere made. The following regression equation wasused to include both factors:

Y 5 6:8X1 1 1:7X2 1 27:9

where Y means the assessment of facial asymmetry(scores), X1 means chin deviation (degrees), and X2

means gonial angle difference (degrees). The coeffi-cients of these factors were 6.8� and 1.7� for chin devi-ation and gonial angle difference, respectively, and theconstant was 27.9.

If we include the most powerful independent factor,we can make a simple equation about chin deviation andthe assement of facial asymmetry (Fig 2).

DISCUSSION

Facial asymmetry was investigated subjectivelyin relation to the soft-tissue features by using facialfrontal photographic measurements.12,13 Naoya16

reported a subjective evaluation of facial asymmetryusing frontal photographs. In that study, 10 orthodon-tists classified 100 facial frontal photographs; therewas a correlation between the subjective evaluation offacial asymmetry and the cephalometric indexes. How-ever, the author did not report the soft-tissue factorsresponsible for the assessment. We examined the rela-tionship between the assessment of facial asymmetryand the soft-tissue characteristics using frontal facialphotographs. Nine orthodontists were asked to scorethe frontal photographs using a VAS, and the factorsthat affect facial asymmetry assessment were estab-lished. The frontal photographs were classified into 3groups according to the assessment scores.

Group I contained patients considered almostsymmetric who needed no treatment. Group II included

Table III. Definitions of angular and linear measurements used in this study

Measurement Definition

Eye canting (�) :Horizontal reference line-bipupillary line

Otobasion canting (�) :Horizontal reference line-otobasion inferius

Lip canting (�) :Horizontal reference line-lip line

Gonion canting (�) :Horizontal reference line-gonion line

Nose deviation (�) :Midsagittal reference line-pronasale line

Chin deviation (�) :Midsagittal reference line-chin line

Ramus inclination difference (�) The difference of right and left :midsagittal reference line-ramus line

Body inclination difference (�) The difference of right and left :midsagittal reference line-mandibular body line

Gonial angle difference (�) The difference of right and left :Pre-Go-Me

Asymmetry index for horizontal Go (%) The ratio of right and left horizontal Go’ length

Asymmetry index for vertical Go (%) The ratio of right and left vertical Go length

Table IV. Comparisons of measurements among the groups

Group I Group II Group III

P valueMean SD Mean SD Mean SD

Eye canting 1.48 0.93 1.47 1.15 1.55 1.05 NS

Otobasion canting 1.36 0.90 1.61 2.03 1.36 1.19 NS

Lip canting 1.61 1.12 2.29 1.47 3.09 2.31 0.00*

Gonial canting 2.02 1.20 1.96 1.31 2.34 2.08 NS

Nose deviation 1.42 1.23 1.37 1.17 1.73 1.18 NS

Chin deviation 1.11 0.96 2.00 1.38 3.64 2.19 0.00*

Ramus inclination difference 2.55 2.33 3.73 2.93 4.18 3.66 NS

Body inclination difference 2.80 1.90 3.75 2.53 7.11 4.86 0.00*

Gonial angle difference 2.31 2.05 3.05 2.62 6.14 5.95 0.00*

Asymmetry index of H-Go 6.34 4.42 9.20 7.12 8.50 7.79 NS

Asymmetry index of V-Go 5.78 3.53 5.30 3.58 6.68 6.17 NS

According to 1-way ANOVA, H-Go and V-Go indicate horizontal and vertical Go.

*P\.01; NS, not significant.

American Journal of Orthodontics and Dentofacial Orthopedics Lee et al 27Volume 138, Number 1

patients with moderate asymmetry who did not requiretreatment. Group III comprised patients with severeasymmetry who required treatment.

ANOVA (Table IV) showed that chin deviation,body inclination difference, gonial angle difference,and lip canting were significantly distinct variables be-tween the groups. The mean values in group III were3.1�, 3.6�, 7.1�, and 6.1� for lip canting, chin deviation,body inclination difference, and gonial angle difference,respectively. These values from experts suggest criteriafor assessing asymmetric patients.

Skeletal analysis with x-rays has been used to diag-nose facial asymmetry. Generally, orthodontists evalu-ate facial asymmetry by analyzing the facial skeletonquantitatively using frontal cephalometry. They estab-lish a midsagittal reference plane and compare thedifference between the right and left values. Rickettset al19 and Svanholt and Solow20 evaluated facial asym-metry by measuring the angles in frontal cephalometry.Solow21 and Nakasima and Ichinose22 measured height

and dimensions. Vig and Hewitt23 measured areas andcompared both sides. These authors basically used theskeletal measurements of asymmetric faces.

However, Michaels and Tourne24 and Yogosawa25

reported that skeletal deformities can be hidden bysoft tissues such as muscles and skin. Peck et al26 sug-gested that many people with skeletal asymmetry havea symmetric face, and that there are differences betweenskeletal and soft-tissue asymmetries. In addition, Hara-guchi et al27 reported differences between the degrees ofactual skeletal asymmetry and soft-tissue asymmetryperceived in Class III patients. They emphasized thenecessity of soft-tissue analysis. It has been suggestedthat people considered to be symmetric and harmoniousin a clinical examination have some facial asymmetry,as shown in radiographic examinations.7,8,10,23,27,28

Ferrario et al7,8 and Haraguchi et al27 suggested thatsoft-tissue asymmetry appeared to be less severe thanskeletal symmetry. Lee29 evaluated facial asymmetryusing facial photographs and frontal cephalometric

Fig 2. Scattergram of facial asymmetry scores accord-ing to chin deviation. This simple functional graphincludes only the first independent variable as chin devi-ation was drawn according to stepwise linear regressionanalysis.

28 Lee et al American Journal of Orthodontics and Dentofacial Orthopedics

July 2010

radiographs, and reported less difference in the horizon-tal and vertical lengths on the soft tissues than on thehard tissues.

Therefore, we suggest that there is a difference be-tween the asymmetry perceived in the soft tissues andthe actual skeletal asymmetry in patients with facialasymmetry. Examining the subjective assessments forfacial asymmetry is an important clinical procedurefor orthodontic diagnosis.

The soft-tissue features can be quantified by mea-suring frontal facial photographs or 3-dimensional(3D) CT images. Patients with a chief complaint of fa-cial asymmetry visit the clinic after seeing their imagesin the mirror or photographs of themselves. Further-more, photographs are easy to take, feasible, cheap,and useful compared with 3D CT images. Therefore,a frontal photograph would be an effective tool for diag-nosing facial asymmetry and making a decision fortreatment.

Few studies have examined the reference plane forphotographs in facial asymmetry analysis. Most studieson soft-tissue facial asymmetry suggested referenceplanes related to the eyes (horizontal) and nose (verti-cal). Ras et al30,31 used a horizontal plane passingthrough the exocanthi and pupils, and a vertical planeperpendicular to and bisecting a line connecting theexocanthi. O’Grady and Antonyshyn32 defined a verticalmidsagittal plane passing through nasion, pronasale,and labrale superius, and an axial plane connectingthe exocanthi and the bridge of the nose.

Determining a midsagittal reference plane is essen-tial for diagnosing asymmetry, because all differencesbetween the right and left sides are measured and com-pared from the midsagittal plane. Paek et al33 suggestedthe midsagittal reference plane as a line connecting gla-bella and anterior nasal spine because the differencesbetween the right and left measurements with this linewere the smallest in normal subjects. Therefore, inthis study, the midsagittal reference plane was definedas a line connecting soft-tissue nasion and subnasale.Soft-tissue nasion was chosen because glabella is diffi-cult to identify in frontal photographs, and subnasale isthe counter-landmark of the anterior nasal spine. Thehorizontal reference plane was then determined perpen-dicular to the midsagittal plane, bisecting the bipupil-lary line. The reason that the vertical reference planewas set in advance was that the effects of eye cantingcould not be evaluated if the horizontal plane was deter-mined first by using the eyes. The landmarks defined inthis study described the facial soft tissues with almostthe same weight for all features (eyes, ears, nose, mouth,gonion). People can mask facial asymmetry by theirposture.

A VAS was used to subjectively assess facial asym-metry. VAS values provide rapid, convenient, valid,reproducible, and representative ratings of dental andfacial appearance.34 For quantification of the subjectiveassessments, Faure et al35 suggested 3 methods forassessing facial esthetics: plane VAS, ratio scale, andranking scale. The plane VAS has no reference or noranking; it just scores the subjective evaluation. The ra-tio scale uses the reference photograph with a score of60 with the typical procedure for the VAS. The rankingscale is used to arrange the photographs from the least tothe most esthetic.

Phillips et al36 used the ranking scale. However, itwas less preferable because it allowed for too manytied scores, which impaired the assumption of continu-ous data. In addition, the plane VAS tends to lack objec-tivity and reproducibility. Therefore, among the 3methods, the ratio scale with a typical VAS procedurewas used, giving the reference photographs scores of0 and 100 scores.

Johnston et al37 standardized the raw scores toz-scores to remove interexaminer variations in scalewhile removing any differential effects between panelsand judges. However, Schlosser et al38 reported no sig-nificant difference between raw scores and z-scores.Therefore, the raw scores were used as the data in thisstudy.

Farkas39 suggested that the maximum normal asym-metry would furnish a threshold value for identifyingasymmetric subjects. Obviously, this threshold is

Table V. Differences in photometric measurementsamong the groups

Group I vsgroup II

Group I vsgroup III

Group II vsgroup III

Eye canting

Otobasion canting

Lip canting *

Gonial anting

Nose deviation

Chin deviation * † †

Ramus inclination difference

Body inclination difference † †

Gonial angle difference * †

Asymmetry index of H-Go

Asymmetry index of V-Go

According to 1-way ANOVA, H-Go and V-Go indicate horizontal and

vertical Go.

Bonferroni post-hoc analysis: *P \0.05; †P \0.001.

Table VI. Correlation coefficients between the assess-ment of facial asymmety and photographic measure-ments

Measurement r P value

Eye canting �0.09 NS

Otobasion canting �0.07 NS

Lip canting 0.33 0.00†

Gonial canting 0.05 NS

Nose deviation 0.05 NS

Chin deviation 0.59 0.00†

Ramus inclination difference 0.21 0.02*

Body inclination difference 0.48 0.00†

Gonial angle difference 0.44 0.00†

Asymmetry index of H-Go 0.14 NS

Asymmetry index of V-Go 0.02 NS

NS, not significant. *P \0.05; †P \0.001.

American Journal of Orthodontics and Dentofacial Orthopedics Lee et al 29Volume 138, Number 1

important, particularly in borderline patients whose fa-cial asymmetry is not clinically discernible.40 Ferrarioet al28 suggested a borderline for normal facial asymme-try. They examined 314 normal patients, identified land-marks on their facial soft tissues directly, measured thedistance from central point to the landmark to obtain thelength and percentage differences between the right andleft sides, and defined the asymmetric borderline.However, they used a 3D method—3D computerizedelectromagnetic digitizer—that is difficult to use asa standard. Hence, their borderline is difficult to applygenerally. In this study, we attempted to detect the bor-derline between normal and abnormal asymmetry as inprevious studies but in a feasible way. The mean valuesin the 3 groups were suggested instead of the border-lines. The values we obtained are significant and shouldbe helpful in diagnosing asymmetry because they distin-guish specific groups. Groups I and II were consideredto have normal asymmetry, and group III had abnormalasymmetry because of the need for treatment. The needfor treatment reflects the perception of facial asymmetryand could be a clinical guideline for treating facialasymmetry.

Head tilting might slightly correct a deviated chin.Kim and Hwang41 reported that asymmetric patientstended to tilt their heads to compensate for a deviatedmenton. The more asymmetric the patients, the morethey tilted their heads. Those authors suggested that,in a patient’s clinical examination, some tilting of theeyes can create doubt as to whether the patient hasasymmetry. Therefore, it was hypothesized that eyecanting might be an affecting factor of facial asymmetryand is clinically meaningful. However, eye canting wasnot a significant factor affecting the assessment of facial

asymmetry (Tables VI and VII). In addition, nosedeviation and otobasion canting were not significantfactors.

The Bonferroni post-hoc analysis (Table V) showedthat only chin deviation was different between thegroups. We suggest that chin deviation is the most influ-encing factor in assessing facial asymmetry. The differ-ence in body inclination and gonial angle showed somedissimilarity between groups I and III, and betweengroups II and III. There were differences lip canting be-tween groups I and III. Chin deviation, body inclinationdifference, gonial angle difference, and lip canting areall associated with asymmetry of the mandible in thelower third of a face. Asymmetry of the lower third ofthe face, the mandibular area, tends to be perceivedmore than asymmetry in the upper or middle third ofthe face. Lip canting was reported to be associatedwith mandibular asymmetry as well as maxillary heightdifferences and occlusal plane canting.42 At the restingposition, lip canting (deviated labial comissure oralar base) on 1 side often indicates vertical skeletalasymmetry.43

The Pearson correlation coefficient and stepwiselinear regression analysis were used to examine the fac-tors affecting the assessment of facial asymmetry(Tables VI and VII). The factors significantly relatedto facial asymmetry in descending order according tothe Pearson correlation coefficients were chindeviation, body inclination difference, gonial angledifference, lip canting, and ramus inclination difference.

Regression analysis showed that chin deviation andgonial angle difference are factors significantly affect-ing facial asymmetry (Table VII). The body inclinationdifference was excluded after stepwise linear regressionanalysis because it was closely related to chin deviation.

Table VII. Factors affecting the assessment of facial asymmetry

Step Constant (a) Variable 1 b Variable 2 R2 P value

1 31.06 Chin deviation 8.00 0.35 0.00

2 27.86 Chin deviation 6.77 Gonial angle difference 1.65 0.41 0.00

Stepwise linear regression analysis.

30 Lee et al American Journal of Orthodontics and Dentofacial Orthopedics

July 2010

Asymmetric indexes of Go and linear measurementswere not significantly affecting factors. Only the angu-lar measurements affected the assessment of facialasymmetry. With all these significant affecting factors,mandibular factors were overwhelming in surveyingfacial asymmetry. This finding has been suggested inseveral studies.8,26,39

Lee et al44 and Ahn and Hwang2 reported that themandibular chin point deviation has the most effecton facial asymmetry assessment. Severt and Proffit45

found that most differences were in the lower third ofthe lateral facial surface. In the normal population,the lowest mean frequency of asymmetry was observedin the orbital region (\2%), followed by the nose (7%)and mouth (approximately 12%).39 In addition, Severtand Proffit45 reported the frequencies of facial asymme-try in the facial portions: 5%, 36%, and 74% in the up-per, middle, and lower thirds of the face, respectively.Therefore, they concluded that most asymmetryoccurred in the mandible. Vig and Herwitt,23 Shahand Joshi,10 Grayson et al,46 and Peck et al26 suggestedthat craniofacial asymmetry becomes more severe fromthe deep portion to the superficial portion by the refer-ence plane as the sagittal plane, such as from thecranium to the mandible. Many studies reported that,overall, gonion and tragion were the most asymmetriclandmarks, and endocanthion was the most symmetriclandmark. These results are consistent with previousstudies.

In our study, the observers were experts. Therefore,the results should be helpful in diagnosis and treatmentplanning. However, nonexperts’ opinions and patients’perceptions are important because the decision abouttreatment for asymmetry is up to each patient.

Therefore, nonexperts’ assessments of facial asym-metry are worthwhile. Bonnie43 suggested that thedifferences in the detection of occlusal canting ina clinical examination depend on the degree of cantingand not necessarily on the level of expertise of the ob-servers. Therefore, the experts’ assessments of facialasymmetry might be similar to those of nonexperts.However, since their study was restricted to lip canting,further studies on the difference in the assessment offacial asymmetry on the overall facial aspects betweenexperts and nonexperts are needed.

CONCLUSIONS

We examined the soft-tissue characteristics ofasymmetric patients assessed by orthodontists to deter-mine the soft-tissue factors affecting the recognition offacial asymmetry. In the view of experts, chin deviation,body inclination difference, gonial angle difference, andlip canting had significant differences in the 3 groupsaccording to the assessment of facial asymmetry. Ingroup III, containing patients with severe asymmetry re-quiring treatment, the mean values were 3.1�, 3.6�, 7.1�,and 6.1� for lip canting, chin deviation, body inclinationdifference, and gonial angle difference, respectively.Chin deviation and gonial angle difference affectedthe assessment of facial asymmetry. Overall, we believethat these results will help clinicians make differentialdiagnoses and treatment plans for patients with facialasymmetry.

REFERENCES

1. Baldwin DC. Appearance and aesthetics in oral health. Commu-

nity Dent Oral Epidemiol 1980;9:244-56.

2. Ahn JS, Hwang HS. Relationship between perception of facial

asymmetry and posteroanterior cephalometric measurements.

Korean J Orthod 2001;31:489-98.

3. Epker BN, Fish LC. Dentofacial deformities: integrated orthodon-

tic and surgical correction. 2nd ed., Volume 3. St Louis: C.V.

Mosby; 1986.

4. De Smit A, Dermaut L. Soft tissue profile preference. Am J Or-

thod 1984;86:67-73.

5. Foster EJ. Profile preferences among diversified groups. Angle

Orthod 1973;43:34-40.

6. Erickson KL, Bell WH, Goldsmith DH. Analysis model surgery.

In: Bell WH, editor. Modern practice in orthognathic and recon-

structive surgery. Philadelphia: Saunders; 1992. p. 156.

7. Ferrario VF, Sforza C, Poggio CE, Tartaglia G. Distance from

symmetry: a three-dimensional evaluation of facial asymmetry.

J Oral Maxillofac Surg 1994;52:1126-32.

8. Ferrario VF, Sforza C, Miani A, Tartaglia G. Craniofacial mor-

phometry by photographic evaluation. Am J Orthod Dentofacial

Orthoped 1993;103:327-37.

9. El Mangoury NH, Shaheen SI, Mostafa YA. Landmark identifica-

tion in computerized posteroanterior cephalometrics. Am J

Orthod Dentofacial Orthop 1987;91:57-61.

10. Shah SM, Joshi MR. An assessment of asymmetry in the normal

craniofacial complex. Angle Orthod 1978;48:141-8.

11. Kobayakawa M. The experimental evaluation of facial asymmetry

of mandibular prognathic patients. Jpn J Oral Maxillofac Surg

1990;36:112-26.

12. Dahan J. A simple digital procedure to assess facial asymmetry.

Am J Orthod Dentofacial Orthop 2002;122:110-6.

American Journal of Orthodontics and Dentofacial Orthopedics Lee et al 31Volume 138, Number 1

13. Edler R, Wertheim D, Greenhill D. Comparison of radiographic

and photographic measurement of mandibular asymmetry. Am J

Orthod Dentofacial Orthop 2003;123:167-74.

14. Robinson SW, Speidel TM, Isaacson RJ, Worms FW. Soft tissue

profile change produced by reduction of mandibular prognathism.

Angle Orthod 1972;42:227-35.

15. Proffit WR, Phillips C, Dann CIV. Who seeks surgical-orthodontic

treatment? Int J Adult Orthod Orthognath Surg 1990;5:153-60.

16. Masuoka N, Momoib Y, Arijic Y, NawadH, Muramatsua A, Gotoe S,

Ariji E. Can cephalometric indices and subjective evaluation be

consistent for facial asymmetry? Angle Orthod 2005;75:651-5.

17. Farkas LG. Examination. In: Farkas LG, editor. Anthropometry of

the head and face. 2nd ed. New York: Raven Press; 1994. p. 3-56.

18. Dahlberg G. Statistical methods for medical and biological

students. London: George Allen & Unwin; 1940.

19. Ricketts RM, Bench RW, Hilgers JJ, Schulhof R. An overview of

computerized cephalometrics. Am J Orthod 1972;61:1-28.

20. Svanholt P, Solow B. Assessment of midline discrepancies on the

postero-anterior cephalometric radiograph. Trans Eur Orthod Soc

1977;25:261-8.

21. Solow B. The pattern of craniofacial associations. Acta Odont

Scand 1966;24:24.

22. Nakasima A, Ichinose M. Size of the cranium in parents and their

children with cleft lip. Cleft Palate J 1984;21:193-203.

23. Vig PS, Hewitt AB. Asymmetry of the human facial skeleton.

Angle Orthod 1975;45:125-9.

24. Michaels LYF, Tourne LPM. Nasion true vertical: a proposed

method for testing the clinical validity of cephalometric measure-

ments applied to a new cephalometric reference line. Int J Adult

Orthod Orthognath Surg 1990;5:43-52.

25. Yogosawa F. Predicting soft tissue profile changes concurrent with

orthodontic treatment. Angle Orthod 1990;60:199-206.

26. Peck S, Peck L, Kataja M. Skeletal asymmetry in esthetically

pleasing faces. Angle Orthod 1991;61:43-8.

27. Haraguchi S, Takada K, Yasuda Y. Facial asymmetry in subjects

with skeletal Class III deformity. Angle Orthod 2002;72:28-35.

28. Ferrario VF, Sforza C, Ciusa V, Dellavia C, Tartalgia G. The effect

of sex and age on facial asymmetry in healthy subjects: a cross-

sectional study from adolescence to mid-adulthood. J Oral

Maxillofac Surg 2001;59:382-8.

29. Lee SM. The evaluation of mandibular asymmetry using the

radiography and facial photography. J Korean Acad Dent Radiol

2001;31:199-204.

30. Ras F, Habets LL, van Ginkel FC, Prahl-Anderson B. Method for

quantifying facial asymmetry in three dimensions using stereo-

photogrammetry. Angle Orthod 1995;65:233-9.

31. Ras F, Habets LL, van Ginkel FC, Prahl-Anderson B. Three di-

mensional evaluation of facial asymmetry in cleft lip and palate.

Cleft Palate Craniofac J 1994;31:116-21.

32. O’Grady KF, Antonyshyn OM. Facial asymmetry: three-

dimensional analysis using laser surface scanning. Plast Reconstr

Sur 1999;104:928-37.

33. Paek SH, Ahn BK, Kim SH, Hong BS, Ho JH, Kang SM. A frontal

cephalometric study on the reference lines to assess the cranio-

maxillofacial asymmetry. Korean J Orthod 1993;23:1-15.

34. Howells DJ, Shaw WC. The validity and reliability of ratings of

dental and facial attractiveness for epidemiological use. Am J

Orthod 1985;88:402-8.

35. Faure JC, Rieffe C, Maltha JC. The influence of different facial

components on facial aesthetics. Eur J Orthod 2002;24:1-7.

36. Phillips C, Tulloch C, Dann C. Rating of facial attractiveness.

Community Dent Oral Epidemiol 1992;20:214-20.

37. Johnston CD, Burden DJ, Stevenson MR. The influence of dental

to facial midline discrepancies on dental attractiveness ratings.

Eur J Orthod 1999;21:517-22.

38. Schlosser JB, Preston CB, Lampasso J. The effects of computer-

aided anteroposterior maxillary incisor movement on ratings of facial

attractiveness. Am J Orthod Dentofacial Orthop 2005;127:17-24.

39. Farkas LG. Valuable views in photogrammetry. In: Farkas LG,

editor. Anthropometry of the head and face. 2nd ed. New

York: Raven Press; 1994. p. 103-11.

40. Lu KH. Harmonic analysis of the human face. Biometrics 1965;

21:491-505.

41. Kim H, Hwang HS. Frontal cephalogram study on the natural

head position of facial asymmetry patients. Korean J Orthod

2000;30:535-42.

42. Hwang HS. A comparative study of PA cephalometry and three-

dimensional morphometry according to the degree of facial asym-

metry. Korean J Orthod 2005;35:163-73.

43. Padwa BL, Kaiser MO, Kaban LB. Occlusal cant in the frontal

plane as a reflection of facial asymmetry. J Oral Maxillofac

Surg 1997;55:811-6.

44. Lee GH, Cho HG, Hwang HS, Kim JC. The relationship between

the perceiveness of facial asymmetry and the measurements of

frontal cephalometry. Korean J Phys Anthropol 1998;11:41-8.

45. Severt TR, Proffit WR. The prevalence of facial asymmetry in the

dentofacial deformities population at the University of North

Carolina. Int J Adult Orthod Orthognath Surg 1997;12:171-6.

46. Grayson B, Cutting C, Bookstein FL, Kim H, McCarthy JG.

The three-dimensional cephalogram: theory, technique, and

clinical application. Am J Orthod Dentofacial Orthop 1988;

94:327-37.