RESEARCH

Clinical–embryological and radiological correlations of

oculo-auriculo-vertebral spectrum using 3D-CT

DT Santos1, O Miyazaki2 and MGP Cavalcanti*,1

1Department of Radiology, College of Dentistry, University of Sao Paulo, Sao Paulo, Brazil; 2Department of Radiology, Children’sHospital of Tokyo, Tokyo, Japan

Objectives: The purpose of this paper is to present a variety of imaging findings of oculo-auriculo-vertebral spectrum (Goldenhar syndrome) using three-dimensional reconstructed images fromcomputed tomography (3D-CT), associating clinical and embryological patterns of the syndrome.Methods: The study population consisted of 10 patients with oculo-auriculo-vertebral spectrumwith clinically identified hemifacial microsomia. The patients were examined using spiral CT, andabnormal imaging features were grouped under facial, ear and temporal bone, vertebral, and skullbase anomalies. The original CT data were transferred to a networked computer workstation with acomputer graphics system to generate 3D-CT volume rendered images of the skull and vertebra.Two observers analysed the bone and muscular setting protocols to assess the relationship betweenbone and muscular structures.Results: Asymmetric underdevelopment was a characteristic pattern of this syndrome resultingfrom hypoplasia of the mandibular ramus and condyle, the zygomatic, sphenoid and auricularconduct bones, and the temporal and masseter muscles. The syndrome was associated with localatrophy seen on 3D-CT images using specific bone and muscles protocols in all cases.Conclusions: Understanding the aetiology, embryology and wide imaging spectrum of thissyndrome is essential to make a correct diagnosis, for treatment planning, and for evaluation whenassociated with a 3D-CT computer graphics system.Dentomaxillofacial Radiology (2003) 32, 8–14. doi: 10.1259/dmfr/36409607

Keywords: tomography, X-ray computed; image processing, computer assisted; syndrome, head;Goldenhar syndrome

Introduction

Oculo-auriculo-vertebral spectrum (OAVS) is a complexsyndrome characterized by an association of maxilloman-dibular hypoplasia, hypoplasia and deformity of the ear,ocular dermoid with vertebral anomalies, and the mostsevere form of hemifacial microsomia.1 – 14 This non-random association of anomalies represents abnormalmorphogenesis of the first and second branchial arches,with vertebral and/or ocular anomalies.1 – 14 The clinicalmanifestations of this syndrome are epibulbar dermoid,vertebral anomalies and hemifacial microsomia, and it isconsidered to represent a spectrum of a similar error inmorphogenesis.1,3,7,10 – 13 This syndrome has also beenknown as hemifacial microsomia, Goldenhar–Gorlin’s

syndrome, first arch syndrome, first and second branchialarch syndrome, lateral facial dysplasia, unilateral man-dibulofacial dysostosis, unilateral intrauterine facialnecrosis, auriculo-branchiogenic dysplasia and facio-auriculo-vertebral malformation complex.2,5,7,9,15 Whilethere is no agreement on the minimal diagnostic criteriaof OAVS, the facial phenotype is characteristic whenenough manifestations are present.5

Owing to the difficulties of visualization of the entireskull base in routine radiographic examinations, fewliterature reports mention OAVS skull base changes.16

However, three-dimensional reconstructed images fromcomputed tomography (3D-CT) are able to demonstratethe entire skull base configuration.17,18 3D-CT imagesusing volume rendering techniques associated with acomputer graphics system allow the pre-operative evalu-ation and post-surgical management of patients withcraniofacial anomalies.19 Previous significant articles on

*Correspondence to: Dr Marcelo Cavalcanti, University of Sao Paulo, Faculty of

Odontology, Department of Radiology, Av. Prof. Lineu Prestes, 2227, FOUSP,

Sao Paulo, SP, 05508-900, Brazil; E-mail: mgpcaval@usp.br

Received 10 October 2002; accepted 20 November 2002

Dentomaxillofacial Radiology (2003) 32, 8–14q 2003 The British Institute of Radiology

http://dmfr.birjournals.org

OAVS have focused on the survey of clinical and geneticviewpoints,2,3,5,6,9 – 12 with no correlation of the 3D-CTimaging findings.

The aim of this paper is to present a variety of imagingfindings of OAVS using 3D-CT, linking together theclinical and embryological patterns of this syndrome.In addition, the aim was to emphasize the importance ofthis new technology in aiding diagnosis, treatmentplanning and follow-up of this anomaly.

Materials and methods

Ten cases of hemifacial microsomia were identified andthe diagnosis was established by clinical criteria. Patientswere referred for imaging on a spiral CT imaging system(Toshiba X/Press; America Medical Systems Inc., Tustin,CA) beginning superior to the vertex of the skull andextending inferiorly to below the mandible. High resolu-tion, continuous, 3 mm thick axial slices were producedwith the following parameters: 1.5 mm s21 table feed,1.5 mm reconstructed slice interval, 100 kVp, 120 mA,field of view 20.3 cm and 512 £ 512 matrix. CT data weretransferred to a networked computer workstation (DELLPrecision 420 Windows NT 4.0) to generate 3D images ofthe skull and vertebra using 2.3 version Vitreaw software(Vital Images Inc., Plymouth, MN). To assess therelationship between those structures and to comparemuscular atrophy and bone hypoplasia of both sides, twoobservers independently analysed the 3D-CT images usingbone and muscular setting protocols.

The computer graphics system allowed 3D-CT analysisfrom different views using rotation, translation andsegmentation of the rendered image. The 3D volumerendering package offers a wide choice of settings thatapply different colour tables and transparency functions tothe CT data, based on CT density of tissues. After analysisof the images using 2D-CT and applying the bone protocolof the software to the reconstructed images, bony aspectsof the syndrome were demonstrated (Figure 1). Abnormalimaging features were identified such as: facial, ear andtemporal bone, vertebral, and skull base anomalies. Theseabnormal imaging features were correlated with the localdisturbances of the first and second branchial arch, whichoriginated from specific bone and muscle structures(Table 1). Initially, the bone structures of both sides ofthe mandible, and the temporal, sphenoid and zygomaticbones were analysed and compared (Figure 2).

Using the cross-hair tool, the largest images of themasseter muscle in the axial, coronal and sagitalreconstructed views were selected. Volumetric measure-ments were made by manual contour delineation of thesethree images (Figure 3) using the Vitreaw software tools,which automatically defined the range of CT density in theregion of interest. Two observers, on two occasionsseparated by several weeks, traced the contour using acomputer mouse, making their own decision as to theboundaries. The software then automatically displayed themuscles as 3D reconstructed images with the correspond-ing measurements of area and volume. Using the muscular

setting, which colorizes the muscles owing to the softtissue, it was possible to localize and visualize all of itsborders in relation to the anatomical structures. We couldalso quantify and qualify the atrophy of these muscles,comparing both sides (Figure 4).

Results

We found only one patient with all the classical clinicalcharacteristics of Goldenhar syndrome (ear abnormality,mandibular hypoplasia, vertebral and skull baseanomalies). The other nine patients were diagnosedusing particular protocols such as bone and muscularsettings to facilitate visualization.

From analysis of the 2D-CT and 3D-CT images,asymmetry of the face was seen to be a particularfeature of this syndrome resulting from hypoplasia ofthe mandibular ramus and condyle (Figure 2b). Theseverity appeared to be related to the degree of localdisturbance. In 65% of cases, facial asymmetry wasevident and in 20% of cases it was considered severe(Figure 1a). Our investigation suggested that hypoplasiaof the greater wing of the sphenoid bone wascharacterized in patients with the syndrome and itmay be responsible for the vertical underdevelopment ofthe face (Figure 1b). Furthermore, the maxillary,temporal and zygomatic bones on the involved sidewere reduced in size and were flattened. Hypoplasia ofthe internal auditory canal and ocular disturbances werealso demonstrated on the images (Figure 2b).

Regarding the skin and muscle layers, 3D-CT coulddepict the patient’s abnormal morphology of phenotype.Muscles of mastication and facial expression appearedhypotrophied on the facial side affected by thesyndrome. We found significant differences betweencontralateral sides of the masseter muscles in allpatients. The side affected by the syndrome showed areduction from 56% to 66% in area (mean 61%) andfrom 67% to 78% in volume (mean 72%) of themasseter muscles (Table 2; Figure 4).

Table 3 shows a summary of the clinical findings for our10 patients with OAVS. Of the 10 patients, 20% presentedwith coloboma unilateral (left) and 30% with a unilateral orbilateral dermoid. Atresia of the pinna was found in 50% ofpatients (unilateral or bilateral), and 20% of patientsdemonstrated a small unilateral pinna (left). With regardto the face, 20% of patients showed a unilateral cleft (left),20% with bilateral ear tags and 10% with bilateral skin tags.

Discussion

The first branchial arch forms the “Meckel’s cartilage”,which further develops into the mandible and muscles ofmastication. The major portion of the malleus and incus arealso derived from Meckel’s cartilage. The second branchialarch forms the “Reichert’s cartilage”, which furtherdevelops into the hyoid and styloid bone and the musclesof facial expression. A major portion of the stapes is also

Goldenhar syndromeDT Santos et al 9

Dentomaxillofacial Radiology

Table 1 Correlation between local destruction (first and second branchial arches) and involved structures

Arch Muscles Skeletal structure

First branchial arch Muscles of mastication MalleusMylohyoid and anterior belly of digastric Incus

MandibleTensor tympani MaxillaTensor veli palatini Zygoma

Squamous part of temporal boneSecond branchial arch Muscles of facial expression Stapes

Stapedius Styloid processStyloid Hyoid bonePosterior belly of digastric

Figure 1 (a) The 3D-CT bone protocol reconstructed image shows marked deviation of the mandibular symphysis to the left caused by ipsilateralhemifacial microsomia. (b) Superoinferior view in 3D-CT imaging using bone protocol reveals hypoplasia of the greater wing of the left sphenoid andzygomatic bone compared with the unaffected side

Figure 2 (a) Lateral view in 3D-CT bone protocol of the non-affected side of the syndrome. (b) Lateral view in 3D-CT bone protocol on the affected sideof the syndrome depicting hypoplasia of the auditory canal, zygomatic arch and mandibular ramus

Goldenhar syndromeDT Santos et al10

Dentomaxillofacial Radiology

derived from Reichert’s cartilage. There is great variationin both the extent and the degree of facial deformity,2,3,20,21

and isolated microtia or auricular or periauricular abnorm-ality may represent the mildest manifestation.2 Themorphology of the craniofacial bones has mainly beenanalysed by X-ray cephalometry, as well as panoramic andconventional tomography.22 However, with these tech-niques morphological analysis of bones has been difficultin individuals with complex deformities such as facialasymmetry and distortions.23

3D-CT images have been much used as a diagnostictool for craniofacial anomalies, identifying paediatriccraniofacial malformations involving bone, and 3D-CTimaging should be used for morphological mapping to aidsurgical treatment planning.8,18,24,25 A quantitative studywas performed to evaluate linear and volumetric measure-ments of soft tissue lesions using 3D-CT and demonstrated

that these measurements were precise and accurate in vitroand in vivo.26 Following this experimental study werendered the 3D-CT images of the masseter musclesfrom multiplanar reconstructed images to obtain thevolume measurements of each muscle on both sides.Using this methodology we compared the volume of thenormal muscle with the atrophic one, and then determinedthe difference for OAVS patients. In our opinion, the newmethodology enables the radiologist to visualize andmanipulate volumetric data quickly and allows the use ofdifferent protocols to demonstrate muscular atrophy andhypoplasia of bones.

To demonstrate the correlation between local disturb-ance (first and second branchial arches) and involvedstructures (Table 1), axial scans and multiplanar recon-struction images were used. Then the computer graphicsystem tools such as segmentation and transparency

Figure 3 Original axial slice (a), and multiplanar reconstructed images in coronal view (b) and in sagittal view (c) depicting the contour of the massetermuscle (white line)

Goldenhar syndromeDT Santos et al 11

Dentomaxillofacial Radiology

functions were applied to improve visualization of the boneand soft tissue anatomical structures. Following imageanalysis, we were able to describe disturbances of themuscles of mastication (mylohyoid and anterior belly ofdigastric, tensor tympani, tensor veli palatini) anddisturbances of the bone structures (malleus, incus,mandible, maxilla, zygoma and squamous part of thetemporal bone). These are all anatomical defects of the firstbranchial arch. Regarding the second branchial arch, wefound disturbances of the muscles of facial expression,stapedius, styloid and posterior belly of digastric. Also,bone disturbances of the stapes, styloid process and hyoidbone were identified (Table 3). This methodology forvisualizing the entire skull base uses a 3D volumerendering technique used by several authors19,26 – 28 fordiagnosis of maxillofacial lesions.

Another important advantage of computer-assistedmedical imaging is for assessing muscle volume andbone shape and size in vivo. Marsh and Vannier18 usedthis technology in patients with unilateral hemifacialmicrosomia, demonstrating the relationship between the

Figure 4 (a) Lateral view in 3D-CT muscular protocol of the non-affected side of the syndrome, with measurement of area ðA ¼ 19:0 cm2Þ and volumeðV ¼ 4:7 mlÞ of the masseter muscle. (b) Lateral view in 3D-CT muscular protocol in the affected side, with measurement of area ðA ¼ 8:3 cm2Þ andvolume ðV ¼ 1:3 mlÞ of the masseter muscle. The muscle layer, shown as a grey region, shows atrophy of the temporal and masseter muscles on theleft side

Table 2 Measurements of area and volume of the masseter muscles

Patient Area (cm2) Volume (ml)

1 (a) 19.0 (a) 4.7(b) 8.3 (b) 1.3

2 (a) 17.5 (a) 4.5(b) 6.2 (b) 1.0

3 (a) 23.8 (a) 7.9(b) 9.0 (b) 2.1

4 (a) 21.0 (a) 6.5(b) 8.8 (b) 1.6

5 (a) 19.5 (a) 5.2(b) 7.2 (b) 1.5

6 (a) 22.5 (a) 7.9(b) 10.0 (b) 2.0

7 (a) 21.6 (a) 7.5(b) 8.0 (b) 1.9

8 (a) 19.8 (a) 6.8(b) 6.9 (b) 1.9

9 (a) 24.3 (a) 8.5(b) 10.0 (b) 2.8

10 (a) 20.5 (a) 8.0(b) 7.0 (b) 2.5

(a) Normal side; (b) abnormal side

Table 3 Clinical summary of the 10 patients with oculo-auriculo-vertebral spectrum

Patient (age, sex) HFM Pinna Eye Face Neck Other

1 (0 d, F) left atresia coloboma — — —2 (23 d, M) left small — cleft — —3 (19 d, F) left/right — — — — —4 (14 y, F) left small — cleft — rib5 (19 y, F) right atresia dermoid ear tag — —6 (6 y, M) right anterior position — — scoliosis rib7 (2 m, F) left/right atresia coloboma ear tag — ASD8 (2 m, F) left/right atresia dermoid skin tag K–F rib9 (16 y, F) — atresia dermoid — K–F HC10 (9 y, M) left/right — — — K–F rib

HFM, hemifacial microsomia; K–F, Klippel–Feil syndrome; rib, rib anomalies; HC, hydrocepharus; ASD, axial skeletal dysplasia

Goldenhar syndromeDT Santos et al12

Dentomaxillofacial Radiology

volume of a muscle of mastication. They stated that theshape and size of its bone origin and insertion in thesepatients varied widely. According to those authors,assumptions of muscle mass and, in turn, function cannotbe made regarding an individual patient on the basis of bonedysmorphology using skull radiographs. In our currentpaper we used a similar methodology to obtain measure-ments of the muscles and to compare dysmorphology ofsides. Table 1 shows measurements of area and volume ofmasseter muscles of all patients and as a result we foundsignificant differences between the affected and non-affected sides in the syndrome.

Computer software allowed the use of protocols thatmade it possible to link automatically the volumerendering to the scan acquisition and to create appropriate3D reconstructed images with high imaging quality. The3D-CT volume rendering images using the computergraphics system further improved through use of themuscular setting protocol, which permits muscular tissuedifferentiation based upon CT density of tissue (musclesfibres), and then confirmed the clinical diagnosis (in 9(90%) of our cases).

According to the literature, other hypoplastic disturb-ances of the glenoid fossa, zygoma, macrostomia, cleft lipand palate and dental malformations can be identified inthis syndrome.16 Narrow external auditory canals werefound in mild cases and atresic canals were observed inmore severe cases in our study. Inner ear abnormalitieswere not considered to be a feature of OAVS.21 However,abnormal shaped and malpositioned internal auditorycanals may be found and may be considered to be a

manifestation of abnormal cranial base development.12,16

In agreement with those authors, we found atresia of thepinna in 50% of our patients (unilateral or bilateral), and20% of patients demonstrated small unilateral pinna (left).With regard to the face, 20% of patients showed unilateralcleft (left), 20% bilateral ear tags and 10% bilateral skintags. In our clinical record we correlated the patterns of ourpatients with OAVS with those described above(Table 2).1,2,3,7,9 We also observed 50% of our patientswith ocular disturbances (unilateral coloboma and dermoidunilateral or bilateral), as also reported by severalauthors.5,10,11,15,16

In conclusion, 3D-CT volume rendering images usingspecific protocols such as bone and muscular settingsallowed the correlation between clinical–embryologicaland radiological findings of Goldenhar syndrome. The3D-CT computer graphics system contributed to therefinement of imaging diagnostic methods and can beused in anatomical and morphological mapping to helpin diagnosis, treatment planning and post-surgicalmanagement of patients. In addition, it may be used asan important differential factor to increase treatmentoptions in patients who have significant craniofacialanomalies.

Acknowledgments

Grants from the National Council of Research and Development(CNPq-520425/01-5), Brasılia, Brazil are gratefullyacknowledged.

References

1. Stark RB, Saunders DE. The first branchial syndrome. Plast ReconstrSurg 1962; 29: 229–239.

2. Gorlin RJ, Kenneth LJ, Jacobsen U, Goldschmidt E. Oculoauriculo-vertebral dysplasia. J Pediatr 1963; 63: 991–999.

3. Grabb WC. The first and second branchial arch syndrome. PlastReconstr Surg 1965; 36: 485–508.

4. Russell LJ, Weaver DD, Bull MJ. The axial mesodermal dysplasiaspectrum. Pediatrics 1981; 67: 176–182.

5. Hodes ME, Gleiser S, Paul de Rosa G, Yune HY, Girod DA, WeaverDD, et al. Trisomy 7 mosaicism and manifestations of Goldenharsyndrome with unilateral hypoplasia. J Craniofac Genet Dev Biol1981; 1: 49–55.

6. Wilson GN, Mason B Jr. Trisomy 9 mosaicism: another etiology forthe manifestations of Goldenhar syndrome. J Craniofac Genet DevBiol 1983; 3: 313–316.

7. Converse JM, Coccaro PJ, Becker M, Smith DW. On hemifacialmicrosomia. Plast Reconstr Surg 1984; 51: 268–273.

8. Heffez DMD, Doku CH. The Goldenhar syndrome: diagnosis andearly surgical management. Oral Surg Oral Med Oral Pathol 1984;58: 2–9.

9. Rollnick BR, Kaye CI, Nagatoshi K, Hauck W, Martin AO.Oculoauriculovertebral dysplasia and variants: phenotypiccharacteristics of 294 patients. Am J Med Genet 1987; 26:361–375.

10. Van Bever Y, Van den Ende JJ, Richieri-Costa A. Oculo-auriculo-vertebral complex and uncommon associated anomalies: report on 8unrelated Brazilian patients. Am J Med Genet 1992; 44: 683–690.

11. Gibson JNA, Sillence DO, Taylor TKF. Abnormalities of the spine inGoldenhar’s syndrome. J Pediatr Orthop 1996; 16: 344–349.

12. Goret-Nicaise M, Baertz G, Saussoy P, Dhem A. Oculo-auriculo-vertebral spectrum: cranial and vertebral malformations due to focaldisturbed chondrogenesis. J Craniofac Genet Dev Biol 1997; 17:35–42.

13. Stanojevic M, Stipoljev F, Kopreina B, Kurjak A. Oculo-auriculo-vertebral (Goldenhar) spectrum associated with pericentric inversion9: coincidental finding or etiologic factor? J Craniofac Genet DevBiol 2000; 20: 150–154.

14. Monahan R, Seder K, Patel P, Alder M, Grud S, O’Gara M.Hemifacial microsomia. Etiology, diagnosis and treatment. J AmDent Assoc 2001; 132: 1402–1408.

15. Smahel Z. Craniofacial changes in hemifacial microsomia.J Craniofac Genet Dev Biol 1986; 6: 151–170.

16. Herring SW, Rowlatt UF, Pruzansky S. Anatomical abnormalities inmandibulofacial dysostosis. Am J Med Genet 1979; 3: 225–229.

17. Binaghi S, Gudinchet F, Rilliet B. Three-dimensional spiral CT ofcraniofacial malformations in children. Pediatr Radiol 2000; 30:856–860.

18. Marsh JL, Vannier MW. Three-dimensional surface imaging from CTscans for the study of craniofacial dysmorphology. J Craniofac GenetDev Biol 1989; 9: 61–75.

19. Vannier MW, Marsh JL. Three-dimensional imaging, surgicalplanning, and image-guided therapy. Radiol Clin North Am 1996;34: 545–563.

20. Cohen MM, Rollnick BR, Kaye CI. Oculoauriculovertebral spectrum:an updated critique. Cleft Palate J 1989; 26: 276–286.

21. Phelps PD, Lloyd GA, Poswillo DE. The ear deformities incraniofacial microsomia and oculo-auriculo-vertebral dysplasia.J Laryngol Otol 1983; 97: 995–1005.

Goldenhar syndromeDT Santos et al 13

Dentomaxillofacial Radiology

22. Kaban LB, Moses MH, Mulliken JB. Surgical correction ofhemifacial microsomia in the growing child. Plast Reconstr Surg1988; 82: 9–19.

23. Kearns GJ, Padwa BL, Mulliken JB, Kaban LB. Progression of facialasymmetry in hemifacial microsomia. Plast Reconstr Surg 2000; 105:492–498.

24. Ono I, Ohura T, Narumi E, et al. Three-dimensional analysis ofcraniofacial bones using three-dimensional computer tomography.J Craniomaxillofac Surg 1992; 20: 49–60.

25. Fishman EK, Magid D, Ney DR, Chaney EL, Pizer SM, RosenmanJG, et al. Three-dimensional imaging: state of the art. Radiology1991; 181: 321–337.

26. Cavalcanti MG, Vannier MW. Measurement of the volume of oraltumors by three-dimensional spiral computed tomography. Dento-maxillofac Radiol 2000; 29: 35–40.

27. Cavalcanti MG, Antunes JL. 3D-CT imaging processing forqualitative and quantitative analysis of maxillofacial cysts andtumors. Pesqui Odontol Bras 2002; 16: 189–194.

28. Xia J, Samman N, Yeung RW, Wang D, Shen SG, Ip HH, et al.Computer-assisted three-dimensional surgical planning and simu-lation. 3D soft tissue planning and prediction. Int J Oral MaxillofacSurg 2000; 29: 250–258.

Goldenhar syndromeDT Santos et al14

Dentomaxillofacial Radiology