Morphometric evaluation of soft-tissue profile shape

ORIGINAL ARTICLE

Morphometric evaluation of soft-tissueprofile shapeDemetrios J. HalazonetisAthens, Greece

Introduction: Soft-tissue facial outline has been studied by conventional cephalometric methods, anddifferences between the 2 sexes have been identified, mainly related to size and timing of growth. However,shape per se was not sufficiently evaluated, especially regarding variability, age-related changes, and sexualdimorphism. The purpose of this study was to evaluate shape variability and sexual dimorphism of thesoft-tissue outline by using morphometric methods. Methods: Pretreatment lateral cephalograms from 170consecutive patients (82 male, 88 female) aged 7 to 17 years were used. Fifteen skeletal and 22 soft-tissuelandmarks were digitized and processed with Procrustes superimposition and principal component analysis.The principal components (PCs) of the soft-tissue shape were analyzed in relation to age and sex. Results:The first 8 PCs explained approximately 90% of the total shape variability. The first coefficient (PC1) related to lip,nose, and chin prominence and included 36% of total shape variability. It was significantly correlated to age, butwith a low coefficient of determination (r2 � 13%). The second coefficient (PC2) related to facial convexityand explained 18% of shape variability. The next 2 coefficients were mainly related to lower lip shape.Statistically significant sexual dimorphism was detected, but the overall shape differences between theaverage profiles of boys and girls were minor and barely detectable visually. Shape dimorphism was presentboth before and after the age of 12 years. Conclusions: Shape variability related mainly to relative lipprotrusion, convexity of the face, and lower lip shape. Shape differences between the sexes seemed to existeven before the pubertal growth spurt, but they were small. Age changes in shape appeared more significant.
(Am J Orthod Dentofacial Orthop 2007;131:481-9)
Facial esthetics is receiving ever increasing atten-tion in orthodontics. Diagnostic methodology isplacing more emphasis on soft tissues, and many

studies measured soft-tissue relationships to establishnormal values and treatment goals.1-9 The facial soft-tissue outline was studied by cephalometric and photo-graphic methods, and differences between the 2 sexeswere identified, mainly related to overall size and timingof growth.10-13 However, shape per se was not sufficientlyevaluated, regarding either age-related changes or sexualdimorphism. The main reason for this might be thatconventional measurements are not particularly suitablefor shape assessment.14 To overcome some problems ofconventional measurements, unconventional techniqueswere studied, but the results are inconsistent, and compar-isons between these published articles are difficult be-cause of their methodological differences.11,15-17

Another area that has been insufficiently investigatedis shape variability. Almost all reports in the literature are

Assistant professor, Orthodontic Department, University of Athens DentalSchool, Athens, Greece.Reprint requests to: D. Halazonetis, 6 Menandrou St, Kifissia 145 61, Greece;e-mail, [email protected], April 2005; revised and accepted, June 2005.0889-5406/$32.00Copyright © 2007 by the American Association of Orthodontists.
doi:10.1016/j.ajodo.2005.06.031
based on evaluation of average shape, with almost noinformation on how soft-tissue shape varies from personto person. Soft-tissue shape patterns, similar to skeletalshape patterns (eg, Class II, Class III, or hyperdivergent/hypodivergent patterns), have not been identified.

The purpose of this article was to evaluate the shapeof the soft-tissue outline during adolescence by usingmorphometric methods, to establish patterns of shapevariability and to determine whether sexual dimor-phism is present.

MATERIAL AND METHODS

The original sample consisted of initial (pretreat-ment) cephalometric radiographs of 200 consecutivelytreated white patients of Greek ethnic origin. Patientrecords were selected from a private orthodontic prac-tice irrespective of sex, age, and type of malocclusion.Only radiographs of good quality with a reference ruleron the cephalostat for exact measurement of the mag-nification factor were included. Patients with congenitalmalformations or syndromes were excluded. The radio-graphs of the sample were inspected for signs ofstrained lip posture, as evidenced by a flattened men-talis area, and 12 were removed, after confirmationfrom the patients’ photographs. The sample was further
restricted to the age range of 7 to 17 years, because of
481

American Journal of Orthodontics and Dentofacial OrthopedicsApril 2007

482 Halazonetis

the small number of patients outside these extremes.The final sample consisted of 170 radiographs (88female, 82 male).

The radiographs were scanned at 150 dpi anddigitized by using Viewbox 3 software (dHAL soft-ware, Kifissia, Greece). A comprehensive set of land-marks was digitized, but the following were used in thisinvestigation (Fig 1): 15 skeletal landmarks: basion(Ba), sella (S), sphenoethmoidale (Se), nasion (N),porion (Po), orbitale (O), anterior nasal spine (ANS),A-point (A), posterior nasal spine (PNS), articulare(Ar), gonion (Go), antegonial notch (Ag), menton (Me),pogonion (Pg), and B-point (B); and 22 soft-tissuelandmarks. Of the soft-tissue landmarks, 12 were iden-tified as lying at locally extreme positions on thesoft-tissue outline, relative to the Frankfort horizontal:N=, the most posterior point at the concavity at the rootof the nose; pronasale (Pr), the most anterior point onthe nose; subnasale (Sn), the most posterior-superiorpoint where the columella meets the upper lip; A=, themost posterior point between the nose and upper lip;labrale superior (Ls), the most anterior point of theupper lip; stomion superior (Ss), the most inferior pointof the upper lip; stomion inferior (Si), the most superiorpoint of the lower lip; labrale inferior (Li), the mostanterior point of the lower lip, labiomental (Lm), themost posterior point on the concavity betweenthe lower lip and Pg=; Pg=, the most anterior point of the

Fig 1. Landmarks (squares) and constructed landmarks(circles) used.

chin; gnathion= (Gn=), the most anterior-inferior point

of the chin; and Me=, the most inferior point of the chin,directly below the skeletal point Me. The other 10soft-tissue points (constructed landmarks) lay betweenthe extreme points at various distances along thesoft-tissue outline. The distances were specified aspercentages of the total distance (along the curvedoutline) between the defining points—eg, N=-Pr 15%was defined as lying between N= and Pr at a distance of15% of the total distance between these 2 points,starting from N=. With this notation, the 10 constructedlandmarks were N=-Pr-15%, N=-Pr-50%, N=-Pr-85%,Pr-Sn-50%, A=-Ls-50%, Ls-Ss-50%, Si- Li-50%, Li-Lm-30%, Li- Lm-80%, and Lm-Pg=-30%. Constructedpoints were used in addition to conventional points tobetter represent the shape of the soft-tissue outline. Thetotal number of these points and their positions werearbitrary and were determined by attempting to arrive ata compromise between the following goals: minimumnumber of points, equal distances between points, andgood shape representation.

The x and y coordinates of all points were scaledaccording to the magnification of each radiograph, tocorrespond to true life size.

The tracings were superimposed by using the Pro-crustes superimposition, which aims to achieve the bestfit by minimizing the sum of the squared distancesbetween all points. The average of the soft-tissue shapewas calculated, and the residuals (differences in the xand y coordinates) between each tracing and the aver-age were computed for all points. Principal componentanalysis (PCA) was used on the residuals to extract themost significant shape coefficients (principal compo-nents of shape, PCs).18

To estimate the error of the method, 15 radiographswere selected at random and redigitized. PCs werecalculated, systematic error was assessed by paired ttests, and random error was calculated according to themethod of Houston.19

RESULTS

Procrustes superimposition and PCA for calculationof the shape coefficients were carried out by Viewbox3. Statistical tests were performed by StatsDirect (Stats-Direct Ltd., Cheshire, United Kingdom) and PAST

Table I. Age descriptive statistics (in years)

Male (n � 82) Female (n � 88)

Mean 11.59 11.91SD 1.68 1.80Range 7.34-16.07 7.86-16.68Median 11.52 11.89

(PAleontological STatistics).20

American Journal of Orthodontics and Dentofacial OrthopedicsVolume 131, Number 4

Halazonetis 483

Descriptive statistics and method error

Descriptive statistics relating to the age of the maleand female groups are shown in Table I. The unpairedt test, testing for difference in age of the 2 groups, wasnot significant at the 5% level (t statistic, 1.19; 2-sidedP value, .235; 95% confidence interval for differencebetween means, –0.85 to 0.21).

Replicate measurements of the 15 radiographsshowed no systematic errors. The coefficient of reli-ability for the first 7 PCs ranged from 97% (for PC1) to87% (for PC7), showing an increasing trend for the lesssignificant PCs. This trend was expected, because theless significant PCs are associated with smaller changesin shape and smaller residuals of landmark positions.To reduce random error, it was decided to extendreplicate measurements to all radiographs. Thus, eachradiograph was digitized a second time, and the aver-ages of the duplicate digitizations were used for sub-sequent analysis.

Selected cephalometric measurements are shown inTable II. The Hotelling T2 test was performed todetermine whether the 2 groups could be discriminatedon these measurements, but the result was not statisti-cally significant (P � .7329). Furthermore, unpairedt tests did not show a significant difference between themale and female groups for these variables (Table II).

PCs of shape

The whole sample was processed as 1 group for thecalculation of the PCs of shape. PCA produced 40 PCsfrom the 22 pairs of residual coordinates. Of the 40PCs, the first 8 explained 89% of the total shapevariability (Table III). These were used for all subse-quent evaluations.

A graphic depiction of the first 8 PCs is shown in

Table II. Selected skeletal and dental cephalometric va

Male (n � 82)

Mean (SD) Median

SNA angle (°) 80.5 (3.54) 80.9SNB angle (°) 76.1 (3.16) 76.2ANB angle (°) 4.4 (2.53) 4.5GoGn to SN (°) 33.9 (5.70) 33.7Upper incisor to NA (°) 20.5 (7.97) 20.6Upper incisor to NA (mm) 3.51 (2.99) 3.5Lower incisor to NB (°) 25.6 (6.07) 24.5Lower incisor to NB (mm) 4.3 (2.13) 4.1Overjet (mm) 5.3 (2.67) 5.4Overbite (mm) 4.2 (2.03) 4.7

Hotelling T2 test with all variables was not significant (P � .61).NS, Not significant.

Figure 2. The first PC, accounting for 36% of total

shape variability, related to lip protrusion and relativenose and chin size. PC2 seemed to relate to facialconvexity. PC3 mainly concerned the vertical positionof the lower lip, and PC4 related to lower lip protrusion(more details in “Discussion”).

Size

The size of each soft-tissue outline was calculatedas the square root of the sum of the squared distances ofeach point from the center of gravity of the shape(centroid size21). Size was found to be significantlycorrelated to age, but the correlation coefficient waslow (r � 0.299, r2 � 0.089, P �.0001). The male grouphad larger sizes than the female group, as confirmed byan unpaired t test (t � 2.038, P � .0431). Sizedifference was approximately 1.8%.

To test whether the size difference was present fromyoung ages, the sample was divided into 2 subsamples:patients below 12 years of age (46 male, 50 female),and patients above 12 years (36 male, 38 female). In theyounger group, size was similar between boys and girls(t � 0.729, P � .467). In the older group, size wassignificantly larger in the male group (t � 2.324, P �

Female (n � 88)

P value, unpaired t-testMean (SD) Median

80.2 (3.17) 79.9 .4250 (NS)76.1 (3.42) 75.3 .8616 (NS)4.13 (2.23) 4.5 .3754 (NS)34.0 (5.28) 34.6 .7411 (NS)21.3 (7.43) 21.3 .5230 (NS)3.9 (2.67) 4.1 .3865 (NS)

24.0 (5.79) 24.0 .2014 (NS)4.0 (1.99) 3.6 .2893 (NS)5.5 (2.31) 4.9 .5942 (NS)3.8 (1.85) 3.9 .1155 (NS)

Table III. Percent variance and cumulative varianceexplained by first 8 PCs (total number of PCs: 40)

Component % variance Cumulative variance

PC1 36.0% 36.0%PC2 18.4% 54.4%PC3 10.6% 65.0%PC4 7.0% 72.0%PC5 5.7% 77.7%PC6 4.9% 82.6%PC7 3.6% 86.2%PC8 2.8% 89.0%

riables

.023), by approximately 3%.


484 Halazonetis

Correlation analysis (simple linear regression) be-tween size and each PC showed no statistically signif-icant findings, except for PC1 (r � 0.154, r2 � 0.024,P � .0454). Because of the low value of the coefficientof determination (2%), it was concluded that shape andsize were not related in any clinically significant way.

Comparison of shape between sexes

Comparison of the first 8 PCs between the male and

Fig 2. First 8 principal components of shape. Mwas constructed by setting component at 3 Scorresponding PC at 3 SD above average.

female groups was performed by the multivariate Ho-

telling T2 test. The difference was found significant atP �.0001. Discriminant analysis showed that 71% ofthe subjects could be classified correctly. Unpairedt tests between the PCs of the male and female groupsshowed that PC1 and PC6 were statistically different(Table IV). Discriminant analysis with these 2 coeffi-cients could classify 63% of the subjects correctly.Figure 3 shows the 95% confidence ellipses for themale and female groups, as separated by canonical

shape in each set is average shape. Left shapeow average; right shape resulted from setting

iddleD bel

variates analysis. From this graph, it can be seen that


Halazonetis 485

positive values of the 2 PCs signified male shape, andnegative values signified female shape. The 2 groupsshowed considerable overlap; their difference wereapproximately half of a standard deviation in each ofthe 2 PCs. Figure 4 shows the average tracing of themale and female groups for comparison. Figure 5shows the average shape warped toward the male andfemale shapes but exaggerated to highlight the differ-ences.

To test whether the difference in shape between thesexes was related to maturation changes, the sameanalysis was applied separately to subjects less than 12years of age and those above 12 years of age. Resultsvirtually identical to the total sample were obtained.For the age group of 7 to 12, the Hotelling T2 test (P �.0068) classified 73% correctly. For the age group of 12

Table IV. Unpaired t test results comparing PCs be-tween sexes

Male FemaleSD

(whole sample) t value P

PC1 0.009619 �0.008954 0.042046 2.942 .0037PC2 0.002951 �0.002664 0.030045 1.219 .2244 (NS)PC3 0.000212 �0.000171 0.022790 1.109 .9133 (NS)PC4 �0.002241 0.002109 0.018593 1.531 .1277 (NS)PC5 �0.000994 0.000882 0.016670 0.732 .4650 (NS)PC6 0.004254 �0.00412 0.015566 3.629 .0004PC7 0.001702 �0.001655 0.013277 1.656 .0996 (NS)PC8 0.001585 �0.001435 0.011737 1.685 .0938 (NS)

NS, Not significant.

Fig 3. 95% confidence ellipses for male (black) andfemale (red) groups, as calculated by using canonicalvariates analysis.

to 17, the Hotelling T2 test (P � .0099) classified 73%

correctly. For comparison with other investigations (see“Discussion”), 14 conventional soft-tissue measure-ments were also made (Table V). Sexual dimorphismwas found in 6 of these with t tests.

Age and shape

The relationship of age to profile shape was inves-tigated by using correlation analysis. Only PC1 wasfound to be statistically correlated to age, albeit with arather low coefficient of determination (r � �0.364;r2 � 0.132; P �.0001). Analyzing the sexes separatelyshowed similar correlations (males: r � �0.358, r2 �0.128, P � .0009; females: r � �0.350, r2 � 0.122,P � .0008) and parallel regression lines (Fig 6), but thefemale group had the regression line positioned at alower level, due to smaller PC1 of this group, asmentioned previously (Table IV).

DISCUSSION

The findings of this study need to be interpreted inlight of the characteristics of the sample. The samplerepresents orthodontic patients, and thus normal soft-tissue profile shapes might be underrepresented or

Fig 4. Average tracings of male (broken line) and female(solid line) groups. Tracings were aligned and adjustedfor size by Procrustes superimposition on soft-tissueoutline (otherwise, male tracing would be approximately2% larger).

skewed toward the most common orthodontic maloc-


486 Halazonetis

clusions. For this reason, the results should not be usedas normative data. However, because the subjects wereaccumulated consecutively, they were assumed to rep-resent a broad range of shape patterns and be suitablefor evaluating shape variability. A sample of ideal oresthetic subjects would be unsuitable for this research,because it would not incorporate the full extent of shapevariability in the population. A sample selected ran-domly would be ideal.

The similarity of the male and female subgroups, asfar as age and skeletal pattern are concerned, wasverified statistically (Table II), thus establishing grosssimilarity, but detailed discrepancies cannot be ex-cluded. The subjects’ ages ranged from prepuberty topostpuberty but did not include the full range of malegrowth, which can last a few more years. Chronological

Fig 5. Exaggerated female and male shapes. Maleshape resulted from average shape by increasing effectof PC1 and PC6 by 2 SD. Female shape resulted fromopposite change. Shapes are scaled to same centroidsize.

age was used, which might not correspond to matura-

tional or skeletal age. This factor might be a confound-ing variable in the results.

Morphometric methods were used because conven-tional cephalometric measurements present difficultiesin differentiating between shape and size.14 Procrustessuperimposition and analysis of shape by computingthe PCs of the residuals of the landmarks had advan-tages over other methods used in the past. Rose et al22

and Ferrario et al3 used Fourier analysis, but manycoefficients were needed to adequately represent lipoutline, making analysis difficult. Also, the lips had tobe in contact; this can impose strain and destroy thenatural outline, especially in subjects with increasedlower facial height and relatively incompetent lips. Thepolar coordinate method of Ferrario and Sforza7 alsosuffered from the same problem. The method used heredid not impose such restrictions. Another major advan-tage was that the percentage of shape variance ex-plained by each PC is calculated, allowing us to retainonly the most significant components for further anal-ysis.

The PCs of shape variability have some interestingand familiar patterns. PC1 was mainly associated withrelative protrusion and size of lips, nose, and chin (Fig2). Small values of PC1 correspond to a relatively largenose and chin with retrusive lips, producing a concavelower third of the face. This profile shape is consideredtypical of low-angle Class II Division 2 malocclusionswith reduced lower facial height. Large values of PC1describe the opposite condition—convex face withprotruding lips, deficient chin, and relatively smallnose. The nose appears small because the total size ofthe profile shape is held constant by the Procrustessuperimposition. This profile shape would probablycorrespond to a high-angle skeletal pattern with anormal-sized nose but increased lower facial height andpossible bimaxillary protrusion. The lips are not touch-ing, a finding frequently associated with increasedlower facial height.

PC2 seemed to be associated with facial convexity.As the component increases in value, the face becomesless convex, eventually becoming orthognathic. Thissoft-tissue morphology can reflect an underlying skel-etal pattern from Class II to slightly Class III.

In a previous investigation, with almost the samesample but looking at the skeletal pattern, the 2 mostsignificant components of shape were closely analo-gous to the first 2 components of this study.18 The firstcomponent represented the spectrum of low-angle tohigh-angle skeletal patterns, whereas the second com-ponent represented anteroposterior intermaxillary rela-
tionship. Thus, it seems that underlying skeletal pattern


Halazonetis 487

might be closely associated with facial soft-tissueshape.

PC3 related mainly to lower lip length (fromstomion to Me), and PC4 related to lower lip protru-sion. Both components accounted for approximately18% of total shape variability. The lower lip seemed tobe an area of high variability between subjects, to anextent that would perhaps not be expected. Such vari-ability might be due to postural factors and also becausethe lower lip is affected by not only the position of themandibular incisor, but also the maxillary incisor,especially in patients with large overbites and overjets.The lower lip is also known to be the most difficult to

Table V. Selected soft-tissue cephalometric variables

Male (

Mean (SD)

Total facial convexity (Gl-Pr-Pg=) (°) 142.2 (5.21)Facial convexity without nose (Gl-Sn-Pg=) (°) 163.4 (6.17)Holdaway angle (NB-LsPg=) (°) 15.0 (5.70)Merrifield’s Z-angle (FH-LsPg=) (°) 70.8 (7.18)�Upper lip to E-plane (mm) �0.5 (2.56)�Lower lip to E-plane (mm) 0.4 (2.68)Nose height (N=-Sn) (mm) 49.4 (4.34)�Nose length (N=-Pr) (mm) 42.8 (4.05)Nasiolabial angle (°) 114.7 (10.35Nose protrusion (ANS-Pr) (mm) 23.7 (2.32)�Upper lip thickness (maxillary incisor-Ls) (mm) 12.9 (1.91)�Upper lip length (Sn-Ls) (mm) 14.9 (1.81)Lower lip length (Li-Gn=) (mm) 32.6 (3.11)�Labiomental sulcus depth (Lm-Pg=Li) (mm) �4.9 (1.60)

Hotelling T2 test using all variables was significant at P �.0001; 71statistically significant variables marked by �, 72% could be discrimNS, Not significant.

Fig 6. Regression plot of age vs PC1. Black diamonds,males; red triangles, females.

predict in orthognathic surgery.23

Regarding sex dimorphism of shape, the results ofthis study showed subtle differences between the sexes;these are shown, exaggerated, in Figure 5. Girls appearto have more prominent noses and chins, less protrudedlips, and shallower labial sulci. However, the followingcaveats must be noted. The differences were very smalland barely perceptible from the average tracings of the2 groups (Fig 4). This might be why other investigatorsdid not detect shape dimorphism.1,6,19,16,24 Girls matureearlier than boys and might therefore be at differentstages of facial development in this sample, with noappreciable chronologic age difference between the 2groups. However, splitting the sample into younger andolder subgroups did not change the results, showingthat shape differences are present from early ages. Thesample did not include subjects older than 17 years.Boys, especially, continue to grow appreciably afterthis, so the results do not apply to adults.

The results seem to run contrary to popular belief,holding that boys have larger noses and more promi-nent chins than girls. The opposite was found here, butthis was only in relative sizes, after scaling to acommon overall size by the Procrustes superimposi-tion. Because boys’ facial growth continues after 17years of age, increases of mandibular and nasal protru-sion are expected, bringing the adult shapes moretoward each other and possibly to virtually identicalshapes.6

The literature contains conflicting reports aboutdifferences in shape between the sexes. Such discrep-ancies might arise from various methodologies. Sprad-

) Female (n � 88)

P value, unpaired t testedian Mean (SD) Median

142.1 141.1 (5.01) 141.3 .1656 (NS)163.2 164.3 (6.05) 163.9 .3414 (NS)15.2 14.3 (5.26) 14.3 .3822 (NS)70.5 71.5 (6.99) 71.0 .5472 (NS)

�0.1 �1.6 (2.59) �1.1 .0039�

0.6 �0.6 (2.79) �0.6 .0169�

49.4 50.6 (4.41) 50.9 .0788 (NS)42.3 44.2 (4.32) 44.3 .0215�

116.0 112.6 (9.21) 113.1 .1595 (NS)23.5 24.3 (2.35) 24.5 .1101 (NS)12.6 12.2 (1.85) 12.2 .0110�

14.8 13.6 (1.73) 13.6 �.0001�

32.4 32.10 (2.71) 32.1 .2818 (NS)�5.0 �4.4 (1.38) �4.5 .0273�

ubjects were discriminated correctly. By using Hotelling T2 test oncorrectly.

n � 82

M

)

% of sinated

ley et al5 used vertical lines from N and Sn to assess the


488 Halazonetis

relative anteroposterior position of the lips, labial sulci,and chin in a sample of young adults (�19 years),selected on the basis of esthetic profiles. Statistical testswere not used, but means and standard deviations aregiven, thus permitting post-hoc application of unpairedt tests. These show that the 2 groups were similar as faras lip and chin protrusions are concerned, but males haddeeper labial sulci, especially of the lower lip. Figure 2shows that PC6 was related to lower labial sulcus depth,so the results are similar to this study, even though thesample was selected according to other criteria.

Bishara et al1,10 studied a longitudinal sample fromthe ages of 5 to 45 years. The sample was relativelysmall, and only 4 angular and 2 linear measurementswere taken. The soft-tissue profiles changed with age,but similar changes were reported for both men andwomen with no statistically significant differences atany time point (5, 10, 15, 25, 45 years). However,looking at the detailed tables1 and applying unpairedt tests, one finds that the distance of the upper lip to theE-plane was significantly different at 15 years (P �.014) and marginally not so at 16 and 17 years. Thismight indicate that a slight difference could have beenmissed because of the small sample size. The same 6measurements were also used in this sample (Table V),and statistically significant differences were noted forthe position of the lips relative to the E-plane, with thegreatest difference that of the upper lip. Again, the lipsof the male group were found to be more retruded thanthe female group.

Prahl-Andersen et al13 also studied a longitudinalsample and measured a more comprehensive set ofvariables. Statistical comparisons were not given, butthe plots of the measurements show differences intiming and rates of growth. From the individual mea-surements, it is difficult to deduce the overall effect onthe soft-tissue outline form. The authors used regionalsuperimpositions and concluded that the upper lip ishigher in girls than in boys, in relation to the maxillaryincisor, whereas the lower lip is more protruded in boysthan in girls. These findings agree with this studyregarding both the position of the lower lip and thehigher vertical position of the upper lip, as can bededuced from the shape coefficient PC6 (Fig 2) and theoverall sex shape difference (Fig 5).

The study of Nanda et al12 is another longitudinalinvestigation with the main purpose of studying soft-tissue growth in harmonious facial patterns. Sexualdimorphism was reported by the authors in mostsoft-tissue growth changes, especially concerning tim-ing and rates of growth. Girls were found to essentiallystop growing at 15, but boys continued even after 18
years. Overall shape differences of the soft-tissue out-
line were not reported and are difficult to infer from theindividual measurements. The distances of the lips tothe E-plane were similar between the sexes, althoughno statistical tests were given to substantiate this.

Skinazi et al17 used an unconventional method tomeasure the areas of each component (nose, chin, andlips) of the soft-tissue outline in young adults. Althoughthe men had larger absolute sizes, the percentagecontribution of the nose area was larger in the women.Furthermore, the lips occupied a larger percentage ofthe space between the nose and the chin in the malesubjects. These results agree with those of this study,but 1 finding does not: the chin area was larger in themen. However, this could be explained by the agedifference in the samples of the studies, as explainedabove.

Other studies with samples of young adults consis-tently reported that men have larger dimensions thanwomen.2,6,8,9,15,24 As already stated, even if thesestudies were directly comparable with this study re-garding age range, it is difficult to infer shape differ-ences from linear measurements. Ratios and angles aremore appropriate shape descriptors, but they weresparsely used. Scheideman et al9 reported relativelylarger lower facial heights and lower lip lengths in menthan women, a result that is opposite to the findings ofthis study. However, those subjects were all older than20 years. Lundström et al,8 in a sample of 25-year-oldstudents, used linear measurements, but also ratios andindexes. In agreement with the present findings, soft-tissue B-point and the chin were found more prominentin women than in men. In contrast, Budai et al24 foundno differences in any proportion indexes examined.Starck and Epker6 calculated ratios of linear dimen-sions in a sample of young adults and found nosignificant shape differences, except for the verticalposition of Sn relative to upper facial height and thenasolabial angle. They pointed out that nasal size andshape were proportionately not statistically different,contrary to popular belief.

Ferrario et al3,7,11,15,16,25 studied the soft tissuesextensively. They used various unconventional meth-ods to examine size and shape of the profile and the3-dimensional structure of the face. Ferrario andSforza7 studied the soft-tissue profile using a novelpolar coordinate system. Their sample included orth-odontic patients from 8 to 14 years of age. Althoughthey reported sex differences in absolute size andgrowth rates, no direct comparison of shape betweenthe sexes was given. Ferrario et al15 performed 3-di-mensional analysis by registering the positions of faciallandmarks using an optical system. In a sample of
young adults with normal occlusions, absolute size of


Halazonetis 489

men was larger than that of women by approximately6%, but no difference in shape was detected. The same3-dimensional system was used in a large sample ofsubjects from age 6 to adulthood to study growth of thelips.11 The expected differences in timing of growthand absolute sizes were found once again, but no directshape assessment was made, and no shape differenceswere reported. Fourier analysis of the facial profile ofadults showed significant shape changes during agingin both sexes.3,25 However, no significant shape differ-ence between the sexes was detected with Fourierdescriptors.16

The main results of this investigation are the shapepatterns that were uncovered by the PCA. These pat-terns seem to reflect well-known underlying skeletalpatterns (eg, PC1 and PC2) but also include patternsthat might be specific to soft-tissue variability (eg, PC3and PC4). Such components of shape could be usedclinically for comprehensive soft-tissue shape descrip-tion and classification purposes. Further research isneeded to determine their relevance to treatment plan-ning.

CONCLUSIONS

Shape variability of the soft-tissue profile outlinemainly concerned protrusion of the nose and chinrelative to lip protrusion, convexity of the face, andlower lip shape.

Relative nose and chin protrusion increased withage, similarly for both sexes.

Shape dimorphism was found between the sexes,both before and after the age of 12 years. However, sexdifferences in shape were small.

REFERENCES

1. Bishara SE, Hession TJ, Peterson LC. Longitudinal soft tissueprofile changes: a study of three analyses. Am J Orthod 1985;88:209-23.

2. Fernández-Riveiro P, Suárez-Quintanilla D, Smyth-Chamosa E,Suárez-Cunqueiro M. Linear photogrammetric analysis of thesoft tissue facial profile. Am J Orthod Dentofacial Orthop2002;122:59-66.

3. Ferrario VF, Sforza C, Dellavia C, Schmitz JH. Harmonicanalysis of facial profiles: a 10-year evaluation. Int J AdultOrthod Orthognath Surg 2001;16:54-60.

4. Smith SL, Buschang PH. Midsagittal facial soft-tissue growth ofFrench Canadian adolescents. Am J Hum Biol 2002;14:457-67.

5. Spradley FL, Jacobs JD, Crowe DP. Assessment of the antero-posterior soft tissue contour of the lower facial third in the idealyoung adult. Am J Orthod 1981;79:316-25.

6. Starck WJ, Epker BN. Cephalometric analysis of profile nasalesthetics. Part I. Method and normative data. Int J Adult Orthod
Orthognath Surg 1996;11:91-103.
7. Ferrario VF, Sforza C. Size and shape of soft-tissue facial profile:effects of age, gender and skeletal class. Cleft Palate Craniofac J1997;34:498-504.

8. Lundström A, Forsberg CM, Peck S, McWilliam J. A propor-tional analysis of the soft tissue profile in young adults withnormal occlusion. Angle Orthod 1992;62:127-33.

9. Scheideman GB, Bell WH, Legan HL, Finn RA, Reisch JS.Cephalometric analysis of dentofacial normals. Am J Orthod1980;78:404-20.

10. Bishara SE, Jakobsen JR, Hession, TJ, Treder JE. Soft tissueprofile changes from 5 to 45 years of age. Am J OrthodDentofacial Orthop 1998;114:698-706.

11. Ferrario VF, Sforza C, Schmitz JH, Ciusa V, Colombo A.Normal growth and development of the lips: a 3-dimensionalstudy from 6 years to adulthood using a geometric model. J Anat2000;196:415-23.

12. Nanda RS, Hanspeter M, Kapila S, Goorhuis J. Growth changesin the soft tissue facial profile. Angle Orthod 1990;60:177-90.

13. Prahl-Andersen B, Ligthelm-Bakker AS, Wattel E, Nanda R.Adolescent growth changes in soft tissue profile. Am J OrthodDentofacial Orthop 1995;107:476-83.

14. Moyers RE, Bookstein FL. The inappropriateness of conven-tional cephalometrics. Am J Orthod 1979;75:599-617.

15. Ferrario VF, Sforza C, Schmitz JH, Miani A Jr, Serrao G. Athree-dimensional computerized mesh diagram analysis and itsapplication in soft tissue facial morphometry. Am J OrthodDentofacial Orthop 1998;114:404-13.

16. Ferrario VF, Sforza C, Schmitz JH, Miani A Jr, Taroni G. Fourieranalysis of human soft tissue facial shape: sex differences innormal adults. J Anat 1995;187:593-602.

17. Skinazi GLS, Lindauer SJ, Isaacson RJ. Chin, nose, and lips:normal ratios in young men and women. Am J Orthod Dentofa-cial Orthop 1994;106:518-23.

18. Halazonetis DJ. Morphometrics for cephalometric diagnosis.Am J Orthod Dentofacial Orthop 2004;125:571-81.

19. Houston WJB. The analysis of errors in orthodontic measure-ments. Am J Orthod 1983;83:382-9.

20. Hammer Ø, Harper DAT, Ryan PD. PAST: paleontologicalstatistics software package for education and data analysis.Palaeontologia Electronica 2001;4:9. Available at: http://palaeo-electronica.org/2001_1/past/issue1_01.htm. Accessed on June12, 2005.

21. Dryden IL, Mardia KV. Statistical shape analysis. Chichester,UK: John Wiley & Sons; 1998.

22. Rose AD, Woods MG, Clement JG, Thomas CDL. Lateral facialsoft-tissue prediction model: analysis using Fourier shape de-scriptors and traditional cephalometric methods. Am J PhysAnthropol 2003;121:172-80.

23. Konstiantos KA, O’Reilly MT, Close J. The validity of theprediction of soft tissue profile changes after LeFort I osteotomyusing the dentofacial planner (computer software). Am J OrthodDentofacial Orthop 1994;105:241-9.

24. Budai M, Farkas LG, Tompson B, Katic M, Forrest CR. Relationbetween anthropometric and cephalometric measurements andproportions of the face of healthy young white adult men andwomen. J Craniofac Surg 2003;14:154-61.

25. Ferrario VF, Sforza C, Miani A, Poggio CE, Schmitz J. Har-monic analysis and clustering of facial profiles. Int J Adult
Orthod Orthognath Surg 1992;7:171-9.
http://palaeo-electronica.org/2001_1/past/issue1_01.htm

http://palaeo-electronica.org/2001_1/past/issue1_01.htm

Documents

Morphometric evaluation of soft-tissue profile shape