Correlation of Different Cephalometric Measurements to Define Facial Type By Luiz R Paranhos, DMD, MSc, PhD; Eduardo N Benedicto, DMD; Marcos F Nunes, Silvana A Kairalla; Danilo Furquim Siqueira, DM D, MSc, PhD; Fernando Cesar Torres, DMD, MDS, PhD

Abstract: The aim of this study was to analyze a possible correlation of different cephalometric measurements to define facial type and to evaluate the best measurement for diagnosis. The sample consisted of 95 lateral cephalograms of caucasian patients with normal occlusion, aged between 15 years and two months and 2.1 years and four months, ofwhich 54 were male and 41 female. The faciaL types were divided into dolichofaciaL, mesofocial and brachyfacial, according to the standards stated by different authors, and a relationship among them was investigated using the Kappa and Total agreement methods. The highest agreement found was between FMA and SN GoGn; the Lowest was between SN Gn and VERT index. According to literature review and the interpretation of the results, SN GoGn appears to be the best measurement to define faciaL type.

Key WOrth: Face; FaciaL bones; Cephalometry.

ntroduction and literature Review Defining the facial cype of an individual is an essential key LO obtaining an accurate orthodontic diagnosis, and the lateral

cephalometric x-ray is presented in literature as a relevant resource for this purpose. In orthodontics, professionals have several analyses tO assist in diagnosis, but sometimes more than one analysis becomes necessary to confirm an uncertain diagnosis, which can lead co misinterpretations.

The studies of correlation among dilferenr cephalometric analyses that define facial cypes help orthodontic practitioners observe the many variations of these analyses, allowing them to decide the best measure tO more accurately define the diagnosis and treatment for their patients.

In 1948, Downs, 1 in pursuit of determinant patterns between the teeth and alveolar process with the facial skeleton by means of a lateral radiograph, revealed the existence of a facial partern that represents an average of individuals with ideal occlusions.

Riedel2 informed in studies about the importance of the angle between the cranium base (SN) at the base of the mandi~le (GoGo) to determine the aspects of current and future growth. For Tweed,3 the direction of facial growth is considered normal if the FMA (Frankfurt Mandibular Plane Angle) presents values between 20° and 30. 0 Later, Steine.-4 developed a cephalometric analysis adjusting the "Y"- axis of Downs,1 thus defining the results of anterior and lower growth vector of the mandible by using the angle SN.Gn.

An important method for qualifYing vertical facial pattern is Sassouni Archial analysis. lr dates from 1955, and the planes such author used were anterior cran ial base plane, occlusal plane and mand ibular base plane, which in a proportional face, would be located posteriorly, at a "0" point, defined by the author.

The correlation between different measurements had already been studied,6 and it was found 66.67% of agreement berween the methods suggested by Rickercs and Siriwat-Jarabak, considered statistically significant, but only moderately so.

In order to optimize working time, avoid errors in diagnosis and reduce the cost of the orthodon tic practitioner's work, this study aimed to investigate a possible correlation between different cephalometric measurements to define facial cypes, in addition ro researching the best variable to define facial cypes in a sample of Caucasian individuals with normal occlusion.

Materials and Methods

Sample The sample consisted of 95 lateral cephalomeu·ic

x-rays of Caucasian subjectS with normal occlusion, aged from 15 years and 2 months to 2 1 years and 4 months, of which 54 (56.84%) were male and 4 1 (43.16%) female. All individuals had at lease four of the six keys of Andrews/ when analyzed in dental casts, and presented only permanem teeth, with the exception of third molars. The first key of Andrews7 was considered essential for the sample (Angle Molar Class I).

Ethics approval was obtained after analysis from the Committee for Ethics in Research of che Methodist University of Sao Paulo (Brazil) . The protocol number is 245760-09 and dated March 30, 2009.

Methods A lateral radiograph of each individual was taken using

a Rotograph Plus""' (Rotograph Plus, Villa Sistemi Medicali S.p.A., Buccinasco, Italy) x-ray machine, operated by a single technician. The processing of radiographs was analyzed by a time/temperature method in an automatic processor, model130 Runzamatic (Eletro Medica Brasileira) maintaining the quality of chemicals used.

The digitalization of the lateral radiographs was obtained with the use of an AMD""' (Advanced Micro Device, Sunnyvale, California, USA) computer with a 4C Hewlett Packard"' Scanjet 6100/CT (Hewlett­Packard Development Company, Palo Alco, CA, USA) scanner with 150 dpi resolution. The images were

transferred ro the CeOC" computer program (Cefalometria Compuradorizada, COT Informatica LTDA, Sao Paulo, Brazil) which generated the requested cephalometric analysis. All linear and angular measures were automatically defined afrer the landmarks were made by a single investigator. Values were transferred to Microsoft Excel2000"' (Microsofr Corporation, USA) for statistical application.

To determine the facial type of each subject, the classification suggested by Ricketts was used (VERT index), which separates the types into mesofacial, dolichofacial and brachyfacial. The measurements used are presented on Table 1.

The sample was also separated in meso, dolicho and brachyfacial, following Tweed3, Steiner" and RiedeJ2 criteria. According co RiedeP, the standard is 32°, and rhe deviation is± 5°, what means that angles of27" to

37° determine mesofacial individuals. More than 3r indicates dolychofacial and less than 2r, brachyfacial. To

Figure 1 -A-) Facial axis angle (Rick­etts). B-) Mandibular plane angle (Ricketts) e C-) Lower face height (Ricketts).

Figure 2 -A-) Facial angle or facial depth (Ricketts), B-) SN.GoGn (Riedel) and C-) Mandibular arch (Ricketts).

Figure 3- A-) FMA (Tweed) and B-) SN.Gn (Steiner).

Fig'ure 4 - (A) SN.GoGn in brachyfacial individual; (B) SN.GoGn in mesofacial individual; (C) SN.GoGn in dolychofacial individual.

32 IJO • VOL n • NO I • SPRING 2012

Tweed,3 individuals with angles between 20o and 30o are mesofacial, with more than 30° are dolycho, and with less than 20° are brachyfacial. Steiner4 afirmed that SN.Gn angle between 64° e 70° determines a mesofacial type, while more than 70° and less than 64° indicates dolycho and brachyfacial, respectively.

Ricketts8 is an exception, because the VERT index considers the average of 5 measurements obtained from the difference between the value found and the individual standard, so the amount must be divided by the clinical standard (which varies according to angle). A positive sign is given when the value indicates a trend of brachyfacial growth, or negative if indicates a dolichofacial trend. The representation of all lines and planes used are shown in Figures 1, 2 and 3. Figure 1 shows the Facial axis angle (A), the Mandibular plane angle (B) and The Lower face height (C), all of them are Rickett's planes. In Figure 2 are presenced the Facial angle or facial depth (A) and mandibular arch (C), both needed for Ricketts analysis and SN.GoGn form Riedel analysis (B). Figure 3, shows FMA angle from Tweed analysis (A) and SN.Gn from Steiner 's (B).

Statistical Analysis Systematic error was verified with a paired t-test

by adopting a 5% level (P < .05) of significance. Error analysis was determined by retracing and re-measuring 19 teleradiographs randomly. A minimum interval of two weeks was taken between the first and second measurements. The casual error was calculated using the suggested formula (Error= ..JI,d2/2n) by Dahlberg.9 No

random errors were found, as the error analysis showed no significant differences while the systematic bias was rested (P < .05) and correlations were greater than 0.95.

To check the correlation between the 4 methods used to establish facial type, the Kappa10 and the Total agreement methods were used. The interpretation of Kappa was achieved according to recommendation of Landis and Koch. 11

Results The results were separated in tables numbered from 2

to 7 confronting measure to measure, showing each facial type found according with its own measurement. At the bottom, Total agreement and Kappa values were given.

Discussion Facial type is an important aspect to consider during

orthodontic treatment planning. The presented study relates to different results to define facial types according to Co ben, 12 who affirmed rhat a single variable is not enough to understand the differences between facial types. Thus, the role of each variable and its integration in face morphology must be known, not judging any characteristics as "normal" or "abnormal" without an appreciation of the value presented in the sum of the facial complex.

The main objective of this study is to correlate the facial growth panern (VERT) of Ricketts et al. 8 and other measurements used to determine facial type, such as: SN.GoGn2, FMN , SN.Gn4• The classification of facial types followed the suggestions made by Ricketts et al.8

Table 1: Classification of facial types according to the authors.

RICKETTS

RIEDEL

TWEED

STEINER

· Frankfurt Mandibular Angle

Facial axis angle

Mandibular plane angle

Lower face height

Mandibular arch

Facial angle or facial depth

SN.GoGn

FMA•

SN.Gn

The angle formed by the facial axis (Pt-Gn) and Basion-Nasion p lan (Bo-Na).

(Standard: 90° ± 3°);

The angle formed by the mandibular p lane (GaMe) and Frankfurt p lane (PoOr)

(Standard: 26° + 4°}; Angle formed by the plans Xi-ENA and Xi-Pm

(Standard: 47° ± 4°); Angle between the axis of the body of the mandible (Xi-Pm)

and condylar axis (De-Xi) (Standard: 26° + 4°);

Formed by the facial plane (N-Pog) and the Frankfurt p lane (PoOr) (Standard: 8JO ± 3°);

The angle formed by the SN p lane and the Mandibular p lane (GoGn) (Standard: 32° ± 5).

It is the angle formed by Frankfurt plane (PoOr) and the mandibular plane (GoMe)

(Standard: 25° ± 5}.

The angle formed by the SN p lan with the axis "Y" of Downs (S-Gn) (Standard: 67° ± 3).

Table 2. Correlation between FMA and SN.GoGn with Agreement and Kappa values.

SN.GoGn FMA' x SN.GoGn

Mesofacial Brachyfacial Dolichofacial Total

5 4 67 Mesofacial

(4 ,2%) (70,5%)

0 24 Brachyfacial

FMA' (6,3%) (25,3%)

4 Dolichofacial

(1 , 1%) (0,0%) (4,2%)

65 23 7 95 Total

(68,4%) (24,2%) (7,4%) (100,0%)

Agreement= 83,2% kappa= 0,63

· Frankfurt Mandibular Angle

Table 3. Correlation between FMA and SN.Gn with Agreement and Kappa values.

SN.Gn FMA' xSN.Gn

Mesofacial Brachyfacial Dolichofacial Total

8 13 67 Mesofacial

(1 3,7%) (70,5%)

0 24 Brachyfacial

(14,7%) (25,3%) FMA'

1 4 Dolichofacial

(1,1%) (0,0%) (4,2%)

61 18 16 95 Total

(64,2%) (18,9%) (1 6,8%) (100,0%)

Agreement= 62,1% kappa= 0,24

·Frankfurt Mandibular Angle

Table 4. Correlation between FMA and VERT with Agreement and Kappa values

FMA' x VERT VERT

Mesofacial Brachyfacial Dolichofacial Total

28 7 67 Mesofacial

(7,4%) (70,5%)

24 Brachyfacial

(25,3%) FMA'

4 Dolichofacial

(0,0%) (0,0%) (4,2%)

34 50 11 95 Total

(35,8%) (52,6%) (11 ,6%) (100,0%)

Agreement= 61 ,1% kappa= 0,36

· Frankfurt Mandibular Angle

that separate the types into dolichofacial, mesofacial and brachyfacial. Other analysis as Sassouni Archial, Downs and Jarabak, should be object of future studies, once they are relevant in classifying facial panern.

The subjects selected were in permanent denrition because it is after the conclusion of the deciduous den tition that the patterns no longer change, according ro Broadbent. 13

In this study, the Kappa method was used due ro irs severity when compared co other statistical methods for excluding rhe possibili ty of zero (or accidenral). This measurement results in values ranging between -1 and + 1, establishing a perfect correlation with + 1 and when the agreemenr is lower than expected in an accident. Under ideal conditions, the correlation berween analysis is around 60%, as stared by authors. 14

Considerations on the planes used in this study

Steiner,4 in 1953, affirmed that the choice of rhe Frankfurt plane for cephalometric appreciation by amhropologists is understandable due to their traditional established methods. Arnhropologisrs used that plane, originally, because the landmarks Porion and Orbital are visible in rhe dry cranium and therefore easily available co them. The points Sella and Nasion would not be clearly visible to them without opening the cranium or using x-rays. In orthodon tics, we are not dealing with dry cranium, and Porion and Orbital are nor as clearly visible, easy ro locate and precise in radiographs as landmarks Sella and Nasion.

The literature reported the variability of location in landmarks such as Porion 15 and Basion 16 resulting in great difficulty and diversity of associations of planes related ro them, which allows us to declare the absence of a constant relationship between the SN line with other planes of rhe face, such as the Palatal and Frankfurt planes.

In 1968, Wey17 investigated the variabili ty of reference lines of the cranial base and affirmed that the SN line promoted lower instability; then Bjork,16

Table 6. Correlation between SN.GoGn and VERT with Agreement and Kappa values.

SN.GoGn x VERT VERT

Mesofacial Brachyfacial Dolichofacial

Mesofacial 29 5

(5,3%)

0 Brachyfacial

(2,1%) SN.GoGn

Dolichofacial 1

(1,1%) (0,0%)

34 50 11 Total

(35,8%) (52,6%) (11,6%)

Agreement= 61,1% kappa= 0,37

Table 7. Correlation between SN.Gn and VERT with Agreement and Kappa values

VERT SN.Gn x VERT

Mesofacial Brachyfacial Dolichofacial

Mesofacial

Brachyfacial

SN.Gn

Dolichofacial

34 50 11 Total

(35,8%) (52,6%) (1 1,6%)

Agreement = 50,5% kappa= 0,23

Total

65

(68,4%)

23

(24,2%)

7

(7,4%)

95

(100,0%)

Total

61

(64,2%)

18

(18,9%)

16

(16,8%)

95

(100,0%)

Negreiros and Siqueira18 used this line in their studies in disagreement with Ricketts et al. , 19 Lundstrom and Cooke,2° which . provided higher reliability to the Frankfui-r plane.

The Frankfurt plane was reported in literature by Mattila and Haataja,21 as a plane that presented many disadvantages in its use by the difficult location of landmarks Porion and Orbital, reaffirming the studies ofLundstrom and Lundstrom.22

In 199 1, Yiazis23 affirmed that lines and planes used as references in cephalometric analysis create concerns about the precision of the results by their presented variability, because the Sella point may vary vertically and anterior or posterior, while the Frankfurt plane may vary by the difficulty in locating Orbital and Porion.

Correlation between FMA, SN.GoGn and SN.Gn

These three measurements allow the definition of facial type by elucidating the mandible's rotation. The FMA shows the vertical growth of the individual's face.3

SN.GoGn, as well as FMA, shows the relation of the mandibular base to the cranium base, their higher values indicate a trend to vertical growth and the lower values indicates horizontal growth. 24 The SN. Gn angle is a measurement of mainly vertical growth very sensible to anterior and posterior changes, which collaborates in defining the vertical of anterior inferior growth of the mandible, detecting its degree of opening.4 The interpretation of SN.Gn and SN.GoGn together allows a "real detection of the facial growth vector."24

The highest agreement found, with 83.2% and Kappa = 0.63, was between FMA and SN.GoGn. This higher agreement, considered substantial, may refer to the reference angles with fixed points in both mandible and cranium (SN and PoOr), providing no significant differences in their values, although there is an inconstancy while marking their points.20

•2

L25 It is important to observe that measurements FMA and SN.GoGn seem co be not affected by the sagital variations of mandible, as well as Sassouni Archial analysis, once they use a plane and not only a point at the chin.

SN.Gn and SN.GoGn reached 70.5% agreement and Kappa = 0.41; these two angles were classified as moderate. The agreement is justified by the use of the exactly same fixed cranium base (SN). The more obtuse these two angles are, vertical growth predominates, while values lower than the standard indicate a predominance of the horizonral compound.

Correlation between SN.GoGn, SN.Gn, FMA in relation to VERT

The growth pattern of Ricketts (VERT index) is computed by an arithmetic average of 5 factors recommended by Ricketts et al. 8 Their factors involve the angles: facial axis angle, mandibular plane angle, lower face height, mandibular arch and facial angle (facial depth). The agreement ofSN.GoGn with VERT was 61.1 o/o and Kappa= 0.37, classified as fair. The SN.GoGn indicates the facial growth vector showing the behavior of the mandibular base related to the cranium base, as the measurements suggested by Ricketts to evaluate the entire facial pattern of growth.

FMA and VERT presented 61 .1 o/o and Kappa = 0.36; these results are extremely close to the obtained value of SN.GoGn to VERT. The classification was again named fair. Such value reminds the great correlation between FMA and SN.GoGn, because when both were compared to VERT, they showed almost identical values. VERT seeks the determination of the growth pattern by an arithmetical average, while FMA is exclusively a measure to evidence the direction of the facial growth of the individual.

The lowest agreement of all occurred between SN.Gn and VERT; its agreement was 50.5% and Kappa = 0.23. Considered as fair, the discrepant value compared to the others may be justified by the differences of objective­SN.Gn defines the anterior inferior vector of growth, while VERT evaluates the facial pattern as a whole.

This study used a sample composed of individuals with no important skeletal alterations. It is important to consider that if such alterations are present the angles that determine facial patterns can be influenced, indicating a facial pattern different from the real. For example, a skeletal Class II can present higher values of SN.Gn. Figure 4 (A, Band C) show brachy, meso and dolychofacial individuals, respectively, with SN.GoGn planes.

Establishing facial type is essential for treatment, due to its influence on morphological characteristics of the face capable of increasing or decreasing the success in results of ce~tain orthodontic procedu.res.

? We suggest the use of the SN.GoGn angle to

detelimine facial type in orde.r to make the definition of orthodontic diagnosis increasingly effective, for being

an angle that has few points in comparison with the VERT index, allowing a relative decrease in errors during location, and the Sella and Nasion points can be located more easily and precisely if compared co Porion and Orbital. Another disadvantage of the VERT index is that the Facial axis angles and the Lower face height from Ricketts analysis are sensitive to sagittal changes. Using only SN.GoGn presumes an optimization of working time, simplification in diagnoses and reduced costs by choosing an easy measurement which avoids requesting data with discordant results.

Conclusion Considering the results, in a sample of individuals

with natural normal occlusion, it was possible to verify that the highest agreement was between FMA and SN.GoGn; the lowest correlation found was between SN. Gn and VERT index:.

The results and the literature provided important data that allow us to choose the best variables - in this case SN.GoGn- to establish the facial type, making orthodontic diagnoses more effective.

References 1. Downs WB. Variations in facial relationship: their significance

in treatment and prognosis. Am j Orthod DemofiJcial Orthop 1948;34:812-40.

2. Riedel RA. The relation of maxillary structures to cranium in malocclusion and in normal oclusion. Angle Orthod 1952;22:142-145.

3. Tweed CH. Evolutionary trends in orthodontics, past, present and future. Am] Orthod Dentofocial Orthop 1953;39:81-108.

4. Steiner CC. Cephalometries in clinical practice. Angle Orthod 1959;29:8-29.

5. Sassouni V. A roentgenographic cephalometric analysis os cephalo­facio-dencal relationships. Am] Orthod 1955;41:735-64.

6. Yamagmo OT, Urbano AL, Vasconcelos MHF. The correlation between the methods applied to establish the facial rype preconized by RickettS and by Siriwat; Jarabak. Odonro (Sao Bernardo do Campo) 2004;23:100-113.

7. Andrews LF. The six keys to normal occlusion. Am J Orthod Dentofocial Orthop 1972;62:296-279.

8. Rickens RM, Bench RW, Gugino CF, Hilgers JJ, Schulhof RJ. Tecnica bioprogresiva de Ricketts. Buenos Aires: Panamericana; 1983.

9. Dahlberg G. Statistical methods for medical and biological Students. New York. Interscience publications, 1940.Fleiss JL. Statistical methods for rates and proportions. New York: John Wiley & Sons; 1973.

10. Landis JR, Koch GG. The measurement of observer agreemenr for categorical clara. Biometrics; 1977.

11. Coben SE. The inregration offacial skeleral variantS. Am j Orthod Dentofocial Orthop 1955;41 :407-434.

12. Broadbent BH. The face of the normal child. Angle Orthod 1937;7: 183-208.

13. Biitow K-W, Muller WG, Mi.ielenaere JGG. Profilocephalometric analysis: a combination of the cephalophorometric and the architecrural-srrurural craniofacial analysis. !nt J Adult Orthotkn OrthognathSurg 1989; 4:87-104.

14. Downs WB. Analysis of the dencofacial profile. Angle Orthod 1956;26(4):190-211.

i I

15. Bjork A. C ranial base development: a follow-up x-ray srudy of the individual variation in growth occurring berween the ages of 12 and 20 years and its relation to brain case and face development.

16.

17.

18.

19.

20.

21.

22.

23.

24.

Am] OrthodDmtofacial Orthop 1955;41:198-225. Wey SH. The variability of coentnographic cephalometric lines of reference. Angle Orthod l%8;38:74-78. Negreiros PE, Siqueira VCV. The effect of the alteration the narura l head position (NHP) on the cephalometric measurements. Rev Dental Press Orrodon Ortoped Facial 2004;9(3):59-76. Ricketrs R.ivl, Schulhof BA, Bagha L. Orientation-sella-nasion or Frankfurr horizon tal. Am j Orthod Den to facial Orthop 1976;69(6):648-654. Lundstrom A, Cooke M S. Proportional analysis of the facial profile in natural head position in caucasian and chinese children. Br J Orthod 1991;18(1) :43-49. Mattila K, Haaraja J. On the accurancy of determining cerrain reference points cephalometric radiography. Odonrol T idskr 1968;76:249-259. Lundstrom A, Lundstrom F. The Frankfort horizontal as basis for cephalometric analysis. Am j Orthod Dentofacial Orthop 1995;107:537-540. Viazis AD. A cephalometric analysis based on natural head position. j Clin Orthod 199 1;25: 172-181. Bishara SE, Augspurger EF. The Role of Mandibular Plane Incl ination in Orthodomic Diagnosis. Angle Orthod 1975;45(4) :273- 281. Scanavini C, Vigoriro JW, Cephalomerric-racliographic study of the possible correlations existem between Vigoriw, RickettS e Siriwat & Jarabak analysis on facial parrerns determination. Ortodontia 2001;34(3):27-41.

Fernando Cesar Torres, DMD, MD~ PhD holds Masters and Doctorate Degrees in Orthodontics, Bauru DentaL School - USP is a Professm; Graduate and Postgraduate Program in Orthodontics at Methodist University of Sao Paulo, Brazil. He is the corresponding author for this article and can be reached at fernandozzy@hotmail. com.

Luiz R Paranhos, DMD, MSc, PhD, is Assistant Professor, Postgraduate Program in Oral BioLogy, University Sagrado Corafda- USC, Bauru, Sao PauLo, Brazil.

Edumdo N Benedicta, DMD, is a Master's student, Bucodental BioLogJ~ University of Campinas, Sao Paulo, Brazil.

Marcos F Nunes is a Master's student, Bucodental Biology, University of Campinas, Sao Paulo, Brazil.

Silvana A KairaLLa is a Master's student, Orthodontics, Methodist University ofSdo Paulo, Brazil.

Danilo Furquim Siqueira, DMD, MSc, PhD is Assistant Professor, Postgraduate Program in Orthodontics, University Sagrado CorafaO - USC, Bauru, Sao Paulo, Brazil.

How to cite this article:

Paranhos LR, Benedicta EN, Nunes MF, Kairalla SA, Siqueira DF, Torres FC. Correlation of different cephalometric measurements to define facial type. lnt J Orthod Milwaukee. 2012 Spring;23(1 ):31-7. PubMed PMID: 22533026.