3-D volume rendering maxillofacial analysis of angularmeasurements by multislice CTPatricia M. L. Lopes, DDS, PhD,a Carla R. Moreira, DDS, MS,b Andréia Perrella, DDS, MS,c

José L. Antunes, PhD,d and Marcelo G. P. Cavalcanti, DDS, PhD,e João Pessoa, Brazil, Bauru,Brazil, and São Paulo, BrazilCENTRO UNIVERSITÁRIO DE JOÃO PESSOA AND UNIVERSITY OF SÃO PAULO

Objective. This study was designed to determine the precision and accuracy of angular measurements using three-dimensional computed tomography (3D-CT) volume rendering by computer systems.Study design. The study population consisted of 28 dried skulls that were scanned with a 64-row multislice CT, and3D-CT images were generated. Angular measurements, (n � 6) based upon conventional craniometric anatomicallandmarks (n � 9), were identified independently in 3D-CT images by 2 radiologists, twice each, and were thenperformed by 3D-CT imaging. Subsequently, physical measurements were made by a third examiner using a BeyondCrysta-C9168 series 900 device.Results. The results demonstrated no statistically significant difference between interexaminer and intraexamineranalysis. The mean difference between the physical and 3-D–based angular measurements was �1.18% and �0.89%,respectively, for both examiners, demonstrating high accuracy.Conclusion. Maxillofacial analysis of angular measurements using 3D-CT volume rendering by 64-row multislice CT isestablished and can be used for orthodontic and dentofacial orthopedic applications. (Oral Surg Oral Med Oral

Pathol Oral Radiol Endod 2008;105:224-30)

With the improvement of three-dimensional (3-D) im-aging technology, more accurate analysis of craniofa-cial complex can be achieved. Several computer 3-Dmethods have been developed to assist diagnosis, treat-ment planning, and predict outcomes.1-7

Despite evolution in the orthodontic field, evalua-tion of facial growth and dental relationship throughcephalometrics is still a difficult frontier to cross. Inthis regard, 3-D analysis is essential for preciselyassessing craniofacial morphology. For orthodonticpurposes, software tools to facilitate accurate land-mark identification for quantitative measurementsand the 3-D segmentation of regions of interest are

This research was supported by Foundation of Research of São PauloState (grants 2005/02157-8 to M. G. P. C.) and CAPES (PhD schol-arships to A. P., C. R. M., and P. M. L. L.).aProfessor, Department of Social Dentistry, College of Dentistry,Centro Universitário de João Pessoa, João Pessoa, Brazil.bStudent, Department of Stomatology, College of Dentistry, Univer-sity of São Paulo, Bauru, Brazil.cStudent, Department of Stomatology, College of Dentistry, Univer-sity of São Paulo, São Paulo, Brazil.dProfessor, Department of Social and Preventive Dentistry, Collegeof Dentistry, University of São Paulo, São Paulo, Brazil.eProfessor, Department of Stomatology, College of Dentistry, Uni-versity of São Paulo, São Paulo, Brazil.Received for publication Jun 12, 2007; returned for revision Aug 24,2007; accepted for publication Aug 28, 2007.1079-2104/$ - see front matter© 2008 Mosby, Inc. All rights reserved.

doi:10.1016/j.tripleo.2007.08.036

224

progressing.3-5,8 Three-dimensional analysis is es-sential for precisely assessing craniofacial morphol-ogy. Several 3-D techniques have been developed tocompensate for the drawbacks of two-dimensionalmeasurements.3

The accuracy of selected craniometric linear mea-surements in a 3-D surface-shaded rendering tech-nique from a single-slice computed tomographic(CT) system has been reported.9 Recently, a three-dimensional computed tomographic (3D-CT) vol-ume-rendering protocol regarding several linearcraniometric measurements from multislice com-puted tomography (MSCT) was evaluated for accu-racy and precision of the system.10 There has, how-ever, been no report regarding the validation of theangular measurements using specific computer sys-tem 3D-CT volume-rendering tools in associationwith the newer generations of multislice CT(MSCT).

The purpose of this research is to demonstrate theprecision and accuracy of angular measurements byusing 3-D volume-rendering imaging using 64-rowmultislice CT for orthodontic applications.

MATERIAL AND METHODSThe study population consisted of 28 intact dry

skulls (5 females and 23 males), from 20 years old to56 years old, provided by the Department of Anat-

omy, Federal University of Sao Paulo, of a medical

OOOOEVolume 105, Number 2 Lopes et al. 225

school, following approval by the Institutional Re-view Committee/Human Specimens Committee(School of Dentistry, University of Sao Paulo). Nohistory of bone disease was identified in the medicalrecords.

Before scanning, the mandibles were fixed to theskulls in central occlusion by using tape. The axial slicedirection was assured by placing the skulls in the scan-ner so that the Frankfurt plane was perpendicular to thetable and parallel to the sagittal plane. For imaging,each skull was placed into a plastic bucket with waterto attenuate the radiation in a manner resembling that ofsoft tissue. The scanning position was kept close to thein vivo situation.

The specimens were scanned in 64-row multisliceCT (Aquilion, Toshiba-America Medical Systems Inc.,Tustin, CA), beginning superior to the vertex and extend-ing inferiorly to below the mandible. High-resolutioncontiguous images were obtained with 0.5-mm thickaxial slices with 0.3 mm of reconstruction interval per0.4 seconds of time, 512 � 512 matrix, and 120 kVpand 300 mA, and a low frequency filter cutoff was usedin the reconstruction algorithm (bone filter). The fieldof view was 24.0 cm. The original CT data were storedon a CD-ROM to allow full retrospective review of anydata and image processing and transferred to an inde-pendent workstation with Vitrea version 3.5 software(Vital Images Inc., Plymouth, MN) to generate 3-Dvolumetric images for visualization, manipulation, andanalysis.

Multiplanar (coronal and sagittal) and 3-D recon-structed images using bone protocol were obtained simul-taneously based on the original axial slices. In axial and inthe correspondent multiplanar reconstructed images, itwas possible to estimate certain landmark locations,and these were automatically marked in the 3-D im-ages.

Angular measurements of 3-D coordinates were ob-tained using several unique and conventional cranio-metric anatomic landmarks previously used by Caval-canti and Vannier9 and by Cavalcanti et al.10 (Table I).The computer graphics system allowed 3D-CT mea-surements from different views by using rotation andtranslation of the image rendered, and by using trans-parency tools (Fig. 1). The craniometric points (n � 9;Table I) were located and the correspondent angularmeasurements (n � 6; Fig. 1 and Table II) were deter-mined electronically by 2 examiners twice each, inde-pendently, by using computer software tools. For eachskull, all points were located during 1 reading session,and then the correspondent angular measurements wereobtained by computer tools. The time interval betweenthe repeated measurements for intraexaminer reliability

determination was 7 days for each examiner. The 28

skulls were scanned by the same MSCT scanner on thesame day. All measurements were made in an averagetime of 20 minutes per skull.

In total, 672 image measurements were made. Thevisualization, required adjustments, and analysis werepossible by using preestablished criteria for bone tissueprotocols. The parameters used were (1) the right sidewas used to locate lateral points, (2) frontal and lateralpositions were standardized using program functions,and (3) the same head position was utilized to deter-mine the measurements in the 2 types of protocolsexamined.10

To obtain the physical data, the mandibles werepositioned at the same occlusal relationship as theywere during image acquisition, and the set of landmarkmeasurements was performed once by a third examiner.A Beyond Crysta-C9168 series 900 device (MitutoyoSul Americana Ltda, Suzano, SP, Brazil) was used for3-D coordinates. This device was located at the Labo-ratory for Reference in Measures for Coordinates (Mi-tutoyo, Suzano, SP, Brazil). This equipment function isbased on a Cartesian 3-D system with 3 reading scales,which allows the mapping of any point within the limitsof the equipment. The device had an active point that is0.3-mm thick, to be equal to the thickness interval of CTreconstruction. The data was imported to a personal com-puter with Windows XP (Microsoft Corporation, Red-mond, WA) operating system (Fig. 2, A, B). Geopak-Win Version 2.4 R.8 Edition 10 (Mitutoyo, Neuss,Germany) was used for making measurements. Thissoftware collects all information from the coordinatesof each landmark and performs the correspondent cal-culations to generate the results (angular measure-ments). The device was calibration certified by Inmetro(National Institute of Metrology, Rio de Janeiro, RJ,

Table I. Craniometric anatomical points1. A Alveolare, deepest point of the maxillary alveolar bone

concavity2. G Glabella, most anterior point of the skull in the sagittal

plane3. Me Menton, most inferior and anterior point of the mental

region4. N Nasion, medial point between the nasofrontal suture

and the nasal origin in midsagittal plane5. Co Condylion, most posterior and superior point of the

mandibular condyle6. ENA Anterior nasal spine, highest point on nasal spine7. Go Gonion, most posterior and inferior point of the curve

between the body and ascending ramus of themandible

8. Pg Pogonion, most anterior point of the mental symphysis9. B The deepest point of the mandibular alveolar concavity

Brazil). The direct measurements of the dry skulls were

n-alve

OOOOE226 Lopes et al. February 2008

considered as the gold standard for physical measure-ments.

The validity of 3D-CT measurements was assessedin terms of their differences, both in degree and per-centages, to corresponding physical measurements. Theprecision of 3D-CT was assessed in terms of the in-traexaminer and interexaminer measurement, both interms of angular degree and percentage. Both validityand precision analysis followed the criteria of Blandand Altman.11 This method is based on simple calcu-lation of the difference between 2 compared measure-ments, following the formulas below:

● for precision:

intraexaminer difference (in degrees) � measure-ment 1 � measurement 2

intraexaminer difference (in percentage) � 100 �(measurement 1 � measurement 2)/measurement 1

interexaminer difference � examiner 1 � examiner 2

interexaminer difference (in percentage) � 100 �

Fig. 1. Three-dimensional computed tomographic measuremeB, Lateral view of glabella-anterior nasal spine-pogonionspine-menton (N.ANS.Me; 166.9°). D, Frontal view of nasiogonion-menton (Co.Go.Me; 126.7°). F, Frontal view of nasio

Table II. Angular measurements1. N.A.Pg2. G.ENA.Pg3. N.ENA.Me4. N.Me.Go5. Co.Go.Me6. N.A.B

(examiner 1 � examiner 2)/examiner 1

● for validation:

validation in degrees � examiner � gold standard

validation (in percentage) � 100 � (examiner �gold standard)/gold standard

Systematic errors of the intraexaminer and interex-aminer analysis were assessed by the Wilcoxon test andthe Mann-Whitney U test, respectively. The compari-son between MSCT and physical measurements wasmade using a paired t test. Statistical analyses weremade using the SPSS version 8.0 software (SPSS Inc.,Chicago, IL). A P value �.05 was accepted to indicatestatistical significance.

RESULTSMean degree values for each measurement obtained

by physical and 3D-CT methods are shown in Table III.The highest mean differences in degrees and percentageranked (Table IV), are �0.03° and �0.07% (intraex-aminer error), and 0.55° and 0.29% (interexaminererror). For each measurement, the highest mean per-centage difference was 1.31% (nasion-alveolare-pogo-nion) for the interexaminer and 0.46% (nasion-menton-gonion) for the intraexaminer assessment (Table IV).

The largest difference between the physical and 3-D–based angular measurements (Table V) was �3.16%(nasion-menton-gonion), and the lowest value was�0.10% (nasion-alveolare-point B). No statisticallysignificant difference was observed for both the intraex-

Lateral view of nasion-alveolare-pogonion (N.A.Pg; 174.5°).NS.Pg; 173.7°). C, Lateral view of nasion-anterior nasalton-gonion (N.Me.Go; 69.9°). E, Lateral view of condylion-olare-point B (N.A.B; 171.7°).

nts. A,(G.A

n-men

aminer (examiner 1, P � .111; examiner 2, P � .712)

physic

OOOOEVolume 105, Number 2 Lopes et al. 227

and the interexaminer (P � .567) assessment. A statis-tically significant difference was observed for the mea-surement nasion-menton-gonion, referring to the com-parison between CT and physical measurements forboth examiners (examiner 1, P � .027; examiner 2, P� .029). Differences related to the other measurementswere not statistically significant (P � .05). The mini-mum P values observed in this study (Table VI) rankedfrom .027 (nasion-menton-gonion) to .789 (glabella-anterior nasal spine-pogonion).

DISCUSSIONSeveral authors have studied CT imaging for analy-

ses using craniometric points, establishing the precisionand accuracy of this method by using different CT

Fig. 2. A, The Beyond Crysta-C9168 device used to perform

Table III. Mean degrees of each measurement in goldstandard and in 3D-CT for both examiners

MeasurementPhysical

measurements,°

3D-CT

Examiner 1,° Examiner 2,°

N.A.Pg 167.1 167.2 165.1G.ENA.Pg 167.9 167.5 167.6Co.Go.Me 123.2 121.9 122.9N.Me.Go 72.9 70.6 70.6N.A.B 165.8 165.6 163.5N.ENA.Me 160.6 159.2 158.9

3D-CT, three-dimensional computed tomography.

technology with 3-D reconstructed images.9,10,12-14

Furthermore, it was established that CT generations,data acquisition, and parameters such as slice thicknessand interval of the reconstruction can determine a high3-D imaging resolution,15,16 which corroborate to qual-itative patterns of our 3-D images, because the sourcewas 64-row multislice CT using 0.3-mm intervals ofreconstruction. Our results highlight favorable indica-tions of precision and accuracy for the 3-D volume-rendering angular measurements from 64-row MSCT.

Regarding linear measurements, our previous studyin MSCT10 found 0.83% and 1.78% of error for boneand soft tissue measurements, respectively, when com-pared with the physical measurements. These resultsdemonstrated high accuracy of both 3D-CT protocols.

Following Swennen et al.,2 a reconstruction intervalof 0.75 mm is acceptable for 3-D cephalometry. Mul-tislice computed tomography 3-D cephalometry wasconsidered highly accurate and reliable. Intraexaminermeasurement errors were low: 0.88, 0.76, and 0.84 mmfor horizontal, vertical, and transverse orthogonal mea-surements, respectively. Interexaminer measurementerror was also low: 0.78, 0.86, and 1.26 mm for hori-zontal, vertical, and transverse orthogonal measure-ments, respectively. Squared correlation coefficientsshowed high intraexaminer and interexaminer reliabil-ity.2

Kusnoto et al.4 assessed the accuracy of angularmeasurements in 3-D projections and found an averageabsolute error of 3.5°. They observed that errors in

al measurements in a dried skull. B, Nasion point landmark.

linear measurements were lower than those of the an-

1.4

OOOOE228 Lopes et al. February 2008

gular measurements. In our research, using newest CTmultislice technology, there was high precision andaccuracy of angular measurements—we found �1.18%of difference for accuracy and 0.29% of difference forintraexaminer and interexaminer analyses.

Cavalcanti et al.,10 Ward and Jamison,12 and Jamisonand Ward13 considered small dimensional measure-ments of up to 6 cm to be less reliable than largerdimensional measures. Our results demonstrated agree-

Table IV. Intraexaminers and interexaminers mean dePrecision intraexaminer 1 P

Degree % De

Mean SD Mean SD Mean

N.A.Pg 0.50 0.68 0.30 0.41 �0.56G.ENA.Pg �0.04 0.95 �0.03 0.56 �0.18Co.Go.Me �0.47 0.71 �0.39 0.58 �0.33N.Me.Go �0.30 0.56 �0.43 0.80 0.31N.A.B 0.62 1.95 0.37 1.17 0.20N.ENA.Me �0.39 0.65 �0.24 0.41 0.40Total �0.01 1.10 �0.07 0.76 �0.03

Table V. Mean degree and percentage differences of cdent physical measurements

Validation examiner 1

Degree %

Mean SD Mean

N.A.Pg 0.17 1.18 0.11G.ENA.Pg �0.42 1.17 �0.26Co.Go.Me �1.25 1.91 �1.00N.Me.Go �2.34 1.58 �3.16N.A.B �0.16 1.68 �0.10N.ENA.Me �1.49 1.45 �0.92Total �0.91 1.73 �0.89

3D-CT, three-dimensional computed tomography.

Table VI. Minimum P values for comparison between3D-CT and physical measurementsMeasurements P values

N.A.Pg .254G.ENA.Pg .789Co.Go.Me .343N.Me.Go .027*N.A.B .173N.ENA.Me .336

3D-CT, three-dimensional computed tomography.*Statistically significant.†Two examiners performing 2 examinations each. P � .05 forsignificance.

ment with these earlier reports, because the smallest

angular measurements of the study (Table III: nasion-menton-gonion, mean 70.6°) showed the greatest dif-ferences in the validation analysis, which represents thelowest accuracy value (Table V).

Cephalometric analysis with 3D-CT was used forfacial asymmetry by Maeda et al.17 and Katsumataet al.18 These authors considered 3D-CT as an appli-cable method that can substitute conventional tech-niques. Park et al.3 proposed a new cephalometricanalysis with 3D-CT and showed the method usedfor a patient with facial asymmetry. The CT imageswere made with a single slice CT. The thickness ofthe axial images was 3 mm, the table speed was 6mm per second, and the 3-D image was reconstructedwith a 2-mm slice thickness. For this protocol, nostatistically significant differences were found. Alllandmarks were reproducible, and there was no sig-nificant intraexaminer error between the 2 sessions(P � .01). There are some limitations when using3D-CT as a diagnostic tool, as they found relativelylarge errors in the vertical position (z-coordinate)compared with the anteroposterior (y-coordinate)and transverse (x-coordinate) positions. However,these errors can be decreased if thin slices are used

nd percentual differences for each measurementintraexaminer 2 Precision interexaminer

% Degree %

Mean SD Mean SD Mean SD

7 �0.35 0.85 2.17 2.10 1.31 1.286 �0.12 0.70 �0.15 1.23 �0.10 0.730 �0.27 1.15 �0.97 1.65 �0.82 1.399 0.46 1.79 �0.04 1.29 �0.07 1.869 0.12 0.98 2.06 2.55 1.26 1.578 0.24 1.01 0.21 1.65 0.14 1.053 0.01 1.15 0.55 2.13 0.29 1.55

ison of all measurements in 3D�CT to the correspon-

Validation examiner 2

Degree %

Mean SD Mean SD

0 �2.00 2.26 �1.21 1.361 �0.27 1.22 �0.16 0.724 �0.27 2.63 �0.19 2.146 �2.30 1.99 �3.09 2.682 �2.21 2.76 �1.36 1.719 �1.70 1.83 �1.07 1.165 �1.46 2.31 �1.18 1.98

gree arecision

gree

SD

1.31.11.41.21.51.5

ompar

SD

0.70.71.52.01.00.81.6

during the reconstruction.3 In comparison, our re-

OOOOEVolume 105, Number 2 Lopes et al. 229

search used a CT multislice with 0.5-mm axial slicethicknesses and 0.3 mm for interval of reconstruc-tion, which allowed higher 3-D resolution. Conse-quently, we found high precision and accuracy ofangular measurements, because the 3-D images wereobtained from volume-rendering technique software.

Newer 64-row MSCTs yield several technologicalimprovements over earlier systems, permitting greaterfidelity in anatomical structure. A volumetric image setwith isotropic properties can be obtained in multipleacquisitions, with a 0.5-mm slice width. Multislicecomputed tomography allows extended anatomic cov-erage with thin slices, in a very short time. The nearfuture will include 256-row MSCT19 for craniofacialapplications. This technology will allow dramaticshortening of the acquisition time.

The substitution of CT for cephalometric skull radi-ography can become practical when high-quality 3-Dreconstructions of the skull can be obtained with suit-able geometric accuracy, typically imprinted on hard-copy in a life-size film format. However, the radiationdose and the cost effectiveness are still the particularpoints that limit this specific application, turning thistechnique mainly toward surgical purposes. All modernCT scanners possess some form of 3-D reconstructionimaging capability, and manufacturers now offer awide range of products.20,21 The computer graphicsmethods that serve as the basis for 3-D reconstructionimaging have evolved rapidly, and dramatic improve-ments in image quality, efficiency, versatility, ease ofuse, and reduction in the cost of systems have beenrealized.20,21 These intrinsically quantitative 3-D imag-ing methods have enabled the development of objectivemeasurement protocols for the study of skull form andvolume change. It is important to point out that mea-surements using 3-D technology can be widely appliedin many areas, especially involving surgical procedure,forensics, and radiotherapy.

Regarding our research, 3D-CT images were used todevelop measurements that were then employed in pro-cedures to aid in the displacement or reorientation ofthe facial skeletal components, which were freed fromtheir normal attachments through orthodontic treatmentplanning. The acceptance of these systems for orth-odontic treatment and the dentofacial orthopedic field iswidespread. These imaging techniques are playing animportant role in the orthodontic field, representingcutting-edge technology and improving diagnostic andclinical treatment.

CONCLUSIONSAngular measurements obtained from 64-row MSCT

3-D reconstructions are precise and accurate. The 3-D

volume-rendering technique can distinguish craniofa-

cial anatomy accurately in bone tissue protocols, add-ing important sources for maxillofacial analyses andapplications.

We thank University Center of João Pessoa, Paraíba,Brazil, for software upgrade support.

REFERENCES

1. Lagravere MO, Hansen L, Harzer W, Majord PW. Plane orien-tation for standardization in 3-dimensional cephalometric analy-sis with computerized tomography imaging. Am J Orthod Dento-facial Orthop 2006;129:601-4.

2. Swennen GRJ, Schutyserb F. Three-dimensional cephalometry:spiral multi-slice vs cone-beam computed tomography. Am JOrthod Dentofacial Orthop 2006;130:410-6.

3. Park SH, Yu HS, Kim KD, Lee KJ, Baike HS A proposal for a newanalysis of craniofacial morphology by 3-dimensional computedtomography. Am J Orthod Dentofacial Orthop 2006;623:634.

4. Kusnoto B, Evans CA, BeGole EA, De Rijk W. Assessment of3-dimensional computer-generated cephalometric measurements.Am J Orthod Dentofacial Orthop 1999;116:390-9.

5. Togashi K, Kitaura H, Yonetsu K, Yoshida N, Nakamura T.Three-dimensional cephalometry by using helical computer to-mography: measurement error caused by head inclination. AngleOrthod 2002;72:513-20.

6. Macchi A, Carrafielo G, Cacciafesta V, Norcini A. Three-dimen-sional digital modeling and setup. Am J Orthod DentofacialOrthop 2006;129:605-10.

7. Halazonetis DJ. From 2-dimensional cephalograms to 3-dimen-sional computed tomography scans. Am J Orthod DentofacialOrthop 2005;127:627-37.

8. Kitaura H, Yonetsu K, Kitamori H, Kobayashi K, Nakamura T.Standardization of 3-D CT measurements for length and anglesby matrix transformation in the 3-D coordinate system. CleftPalate Craniofac J 2000;37:349-56.

9. Cavalcanti MGP, Vannier MW. Quantitative analysis of spiralcomputed tomography for craniofacial clinical applications.Dentomaxillofac Radiol 1998;27:344-50.

10. Cavalcanti MGP, Rocha SS, Vannier MW. Craniofacial mea-surements based on 3D-CT volume rendering. Implications forclinical applications. Dentomaxillofac Radiol 2004;33:170-6.

11. Bland JM, Altman DG. Statistical methods for assessing agree-ment between two methods of clinical measurement. Lancet1986;8:307-10.

12. Ward RE, Jamison PL. Measurement precision and reliabilityin craniofacial anthropometry: implications and suggestionsfor clinical applications. J Craniofac Gen Dev Biol 1991;11:156-64.

13. Jamison PL, Ward RE. Brief communication: measurement size,precision, and reability in craniofacial anthropometry: bigger isbetter. Am J Phys Anthropol 1993;90:495-500.

14. Krasgov J, Bosch C, Gyldensted C, Sindet-Pedersen S. Compar-ison of the reliability of craniofacial anatomic landmarks basedon cephalometric radiographs and three-dimensional CT scans.Cleft Palate Craniofac J 1997;34:111-6.

15. Jung H, Kim HJ, Kim DO, Hong SI, Jeong HK, Kim KD, et al.Quantitative analysis of three-dimensional rendered imaging ofthe human skull acquired from multi-detector row computedtomography. J Digit Imaging 2002;15:232-9.

16. Kim DO, Kim HJ, Jung H, Jeong HK, Hong SI, Kim KD.Quantitative evaluation of acquisition parameters in three-dimensional imaging with multidetector computed tomogra-phy using human skull phantom. J Digit Imaging 2002;

15:254-7.

OOOOE230 Lopes et al. February 2008

17. Maeda M, Katsumata A, Ariji Y, Muramatsu A, Yoshida K, GotoS, et al. 3D-CT evaluation of facial asymmetry in patients withmaxillofacial deformities. Oral Surg Oral Med Oral Pathol OralRadiol Endod 2006;102:382-90.

18. Katsumata A, Fujishita M, Maeda M, Ariji Y, Ariji E, LanglaisRP. 3D-CT evaluation of facial asymmetry. Oral Surg Oral MedOral Pathol Oral Radiol Endod 2005;99:212-20.

19. Mather R. The next revolution: 256-Slice CT. [cited 2007Feb 27]. Available from: http://madeforlife.toshiba.com/ct256/TMS082_CT_WP_256.pdf. Toshiba America Medical Systems,Inc. Tustin, CA.

20. Vannier MW. Craniofacial imaging informatics and technology

development. Orthod Craniofac Res 2003;6 Suppl 1:73-81.

21. Vannier MW. Craniofacial computed tomography scanning:technology, applications and future trends. Orthod Craniofac Res2003;6 Suppl 1:23-30.

Reprint requests:

Marcelo G. P. Cavalcanti, DDS, PhDDepartment of RadiologyUniversity of São Paulo, College of DentistryAv. Prof. Lineu Prestes, 2227CEP 05508-900São Paulo, SP, Brazil

mgpcaval@usp.br