PEDIATRIC/CRANIOFACIAL Planning Surgical Reconstruction in TreacherCollins Syndrome Using Virtual Simulation Dariush Nikkhah, M.Sc., M.R.C.S.(Eng.) Allan Ponniah, M.Sc., M.R.C.S.(Eng.) Cliff Ruff, M.Sc. David Dunaway, F.R.C.D.S., F.R.C.S. East Grinstead and London, United Kingdom

Background: Treacher-Collins syndrome is a rare autosomal dominant condition of varying phenotypic expression. The surgical correction in this syndrome is difficult, and the approach varies between craniofacial departments worldwide. The authors aimed to design standardized tools for planning orbitozygomatic and mandibular reconstruction in Treacher-Collins syndrome using geometric morphometrics. Methods: The Great Ormond Street Hospital database was retrospectively identified for patients with Treacher-Collins syndrome. Thirteen children (aged 2 to 15 years) who had suitable preoperative three-dimensional computed tomographic head scans were included. Six Treacher-Collins syndrome threedimensional computed tomographic head scans were quantitatively compared using a template of 96 anatomically defined landmarks to 26 age-matched normal dry skulls. Results: Thin-plate spline videos illustrated the characteristic deformities of retromicrognathia and maxillary and orbitozygomatic hypoplasia in the Treacher-Collins syndrome population. Geometric morphometrics was used in the virtual reconstruction of the orbitozygomatic and mandibular region in Treacher-Collins syndrome patients. Intrarater and interrater reliability of the landmarks was acceptable and within a standard deviation of less than 1 mm on 97 percent and 100 percent of 10 repeated scans, respectively. Conclusions: Virtual normalization of the Treacher-Collins syndrome skull effectively describes characteristic skeletal deformities and provides a useful guide to surgical reconstruction. Size-matched stereolithographic templates derived from thin-plate spline warps can provide effective intraoperative templates for zygomatic and mandibular reconstruction in the Treacher-Collins syndrome patient.  (Plast. Reconstr. Surg. 132: 790e, 2013.) CLINICAL QUESTION/LEVEL OF EVIDENCE: Diagnostic, V.

T

he face, with its complexity and unique quality in every individual, is difficult to quantify in an objective manner. Landmark-based geometric morphometrics can be used to define complex facial surface and bony anatomy, and principal components analysis provides a useful tool for describing and From East Grinstead Hospital; Great Ormond Street Hospital; and University College London. Received for publication January 6, 2013; accepted May 13, 2013. Presented in part at the Fourth Annual Meeting of the European Plastic Surgery Research Council, in Hamburg, Germany, August 23 through 26, 2012, and published as an abstract in Plastic and Reconstructive Surgery; and at the Craniofacial Society of Great Britain and Ireland, in York, United Kingdom, April 13 through 15, 2011. Copyright © 2013 by the American Society of Plastic Surgeons DOI: 10.1097/PRS.0b013e3182a48d33

790e

quantifying variations within and between populations. Geometric morphometrics can provide both subjective and objective outcome measures in craniofacial surgery. The premise for our study was to apply geometric morphometrics to the surgical planning Disclosure: None of the authors has a financial interest in any of the products or devices mentioned in this article. Supplemental digital content is available for this article. Direct URL citations appear in the text; simply type the URL address into any Web browser to access this content. Clickable links to the material are provided in the HTML text of this article on the Journal’s Web site (www. PRSJournal.com).

www.PRSJournal.com

Volume 132, Number 5 • Correction of Treacher-Collins Syndrome

Fig. 1. Phenotypic variation in Treacher-Collins syndrome skeletal tissues, from mild (left), to moderate (center), to severe (right).

of Treacher-Collins syndrome, a rare autosomal dominant craniofacial condition of varying phenotypic expression (Fig. 1).1 Congenital craniofacial anomalies may have significant functional and psychological impact on children, and appropriate presurgical planning in reconstructive surgery can have a profound impact on quality of life.2 Geometric morphometrics will also help us understand the craniofacial shape variations seen in Treacher-Collins syndrome and how best to correct deformities with surgical procedures available today. The presentation of patients with TreacherCollins syndrome varies from those with severe facial disfigurement and functional airway problems, to some who remain undiagnosed throughout their lifetime. The surgical correction in this syndrome remains challenging, and the approach and timing of reconstruction vary greatly. Data on Treacher-Collins syndrome are limited to case series and case reports, with a variety of staged surgical approaches described to correct the mandibular and orbitozygomatic deformities.3–7 This is most likely because of the rarity of the condition (i.e., one in 50,000 cases).8 We are the first group to use geometric morphometrics to analyze the facial skeleton in Treacher-Collins syndrome. The geometric morphometric algorithms used allow a visualization of what the individual would have appeared like had they not had the Treacher-Collins syndrome mutation. This technology is likely to prove useful in improving the effectiveness of surgery and reducing the number of stages required to treat patients with Treacher-Collins syndrome.

Aims 1. To produce a mathematical model based on geometric morphometrics and principal components analysis that describes the Treacher-Collins syndrome skeletal deformity and the variation seen within the study population. 2. To demonstrate the use of thin-plate spline warping to highlight the differences of the Treacher-Collins syndrome skull from normal and visually describe the variation seen within the Treacher-Collins syndrome population under study. 3. To demonstrate how geometric morphometrics may be useful in guiding the surgical correction of Treacher-Collins syndrome skeletal deformities.

PATIENTS AND METHODS The Great Ormond Street Hospital craniofacial database was used to identify patients with a diagnosis of Treacher-Collins syndrome between 1987 and 2011 who had undergone three-dimensional computed tomography head scans before surgery. Thirteen children between 2 and 15 years (mean, 8 years) who had suitable preoperative three-dimensional computed tomographic head scans were included. However, seven of these 13 children could only be assessed qualitatively because of the lack of homologous points (Table 1). The remaining six Treacher-Collins syndrome three-dimensional computed tomographic head scans were analyzed quantitatively using a template of 96 anatomically defined landmarks (Table 2) and compared to 26 normal

791e

Plastic and Reconstructive Surgery • November 2013 Table 1.  Patient Demographics Case

Age (yr)

Sex

Symmetry

Severity Based on Color Maps

Inclusion/Exclusion Criteria for Quantitative Analysis

1 2* 3* 4 5* 6 7 8 9 10* 11 12 13

2 3 3 3 4 8 8 10 10 12 12 13 15

M M F M F M M F M M M M F

Yes Yes Yes Yes Yes Yes No No No Yes Yes Yes Yes

Moderate Moderate Severe Moderate Severe; choanal atresia Moderate Moderate (asymmetric) Moderate (asymmetric) Severe Mild Mild Moderate Moderate

Included Included Excluded; no homologous points over zygoma Included Excluded; no homologous points over zygoma Included Excluded; no homologous points over zygoma Excluded; no homologous points over zygoma Excluded; no homologous points over zygoma Included Included Incomplete scan; posterior aspect of skull missing Incomplete scan; posterior aspect of skull missing

*Patients who needed a tracheostomy.

age-matched dry skulls. Three-dimensional computed tomographic head scans were taken of dry skulls donated from the National History Museum (London, United Kingdom) and the Johns Hopkins Bosma collection (Table 3). The scans were taken using a 16-slice Siemens Somatom Sensation (Siemens AG, Munich, Germany) spiral computed tomography scanner set to 0.75-mm collimation. These scans were then loaded onto Robin3D (London, United Kingdom) landmarking software for analysis.9 All children with Treacher-Collins syndrome who had preoperative three-dimensional computed tomographic head scans and who had been diagnosed with the TCOF1 gene were included. The whole spectrum of Treacher-Collins syndrome was included for qualitative and quantitative analysis. However, the subpopulation with complete absence of zygomas and temporomandibular joints would not be possible to analyze quantitatively using geometric morphometrics because of a lack of homologous points. Children with unconfirmed Treacher-Collins syndrome in clinic letters, and those conditions with phenotypic overlap such as Nagar and Miller syndromes, were excluded. Landmarking and Warping of the Craniofacial Skeleton The three-dimensional computed tomographic head scans were then landmarked with 96 homologous points each (Figs. 2 and 3). The landmarks were based on anthropometric definitions from the original works on craniofacial measurement by Farkas,10 with additional landmarks to describe the remaining dysmorphology. The 96 landmarks were defined using an iterative process to ensure that the number and distribution of the landmarks was sufficient to define variations in skull shape

792e

between individuals and the populations under investigation. This process has been described in previous publications.2,11 With this technique, a Treacher-Collins syndrome skull can be warped into a normal dry skull and thin-plate spline videos can be made to qualitatively visualize global changes in the craniofacial skeleton. (See Video, Supplemental Digital Content 1, which shows warping of a Treacher-Collins syndrome skull to a normal skull using 96 landmarks. This video illustrates the global deformities seen in the Treacher-Collins syndrome skeleton in comparison with an agematched control, http://links.lww.com/PRS/A877.) Modeling Variation in the Treacher-Collins Syndrome Population Principal components analysis can provide a simplified representation of the morphologic variations that occur in a given population. This is a useful tool that can be used to compare biological shapes. Principal components analysis identifies the most variable shape changes in a population and is performed by first making a point distribution model. The point distribution model is used to demonstrate shape within the patient population; the model is constructed from the given set of landmarks from each member of the population. From the point distribution model, total variance and mean shape can be calculated. The modes of variation, which represent the principal components, can be presented in thin-plate spline warps, which is a way of interpolating changes between homologous landmarks.12 This allows the statistically meaningful data to be represented in a video, which the clinician can assess qualitatively. The first mode of variation is usually the one that shows the greatest variability. This could be the varied skull growth in a wide age group of adults and children. In a thin-plate spline video, this

Volume 132, Number 5 • Correction of Treacher-Collins Syndrome Table 2.  List of 96 Craniofacial Landmarks with Descriptions Landmark

Position

Description

A B C D

Right mental foramen Left mental foramen Interdentale inferius* Mental protuberance*

E F

Gnathion Right incisive fossa

G H

Left incisive fossa Right gonion*

I

Left gonion*

Point in the center of the mental foramen Point in the center of the mental foramen Junction between the lower two incisors Pogonion (pg): most anterior point on the mandibular symphysis Most inferior point on the chin Just below incisor teeth; most concave ­depression incisive fossa As above Point on the angle of the mandible located by the bisection of the angle of the mandibular line and the posterior border of the ramus As above

J

Left coronoid process*

Most superior point on the coronoid process

K L M N O P

Left condylar process* Right condylar process* Right coronoid process* Right canine mental eminence Left canine mental eminence Right mandibular notch

Q R

Left mandibular notch Right oblique line of the mandible

S

Left oblique line of the mandible

T

V W X Y Z

Point directly beneath the left mental foramen Point directly beneath the right mental foramen Right mental tubercle Left mental tubercle Interdentale superius Superior part of the piriform aperture Left ZF suture line

Most lateral point on the condylar head Most lateral point on the condylar head Most superior point on the coronoid process Most concave portion of the canine eminence Most concave portion of the canine eminence Point of greatest concavity on the mandibular notch As above Measured using an imaginary line directly opposite the gonion Measured using an imaginary line directly opposite the gonion Directly below point B

A1

Right ZF suture line

B1

U

Point of View Frontal Frontal Frontal Frontal Frontal Frontal Frontal 90 degrees, 70 degrees rotated 90 degrees, 70 degrees rotated 90 degrees, 60 degrees rotated As above As above As above Frontal Frontal 90 degrees 90 degrees 90 degrees 90 degrees Frontal

Directly below point A

Frontal Frontal Frontal Frontal Frontal Frontal

Right mastoid process*

Most superior aspect of the triangular eminence Most superior aspect of the triangular eminence Junction between the upper two incisors Apex of the nasal aperture Point at which the frontozygomatic suture intersects the inner orbital rim Point at which the frontozygomatic suture intersects the inner orbital rim Most inferior point of the mastoid process

C1

Left mastoid process*

As above

D1 E1

Glabella External occipital protuberance

F1

Right supraorbital notch

G1

Left supraorbital notch

H1 I1 J1 K1

Right ZT suture line Left ZT suture line Right infraorbital foramen Right articular tubercle

L1

Right articular tubercle

M1

Right articular tubercle

N1 O1 P1

Right superior aspect of the ZT suture line Right inferior aspect of the ZT suture line Right zygomatic arch

Point between the supraorbital notches Most prominent part of the occipital bone at the posteroinferior part of the skull (inion) Point of greatest concavity on the supraorbital notch Point of greatest concavity on the supraorbital notch Midpoint of the ZT suture Midpoint of the ZT suture Point in the center of the infraorbital foramen Point posterior to the inferior point of the tubercle of the zygoma Most inferior point of the articular tubercle of the zygoma Point anterior to the inferior point of the articular tubercle of the zygoma Most superior point of the ZT suture Most inferior point of the ZT suture Midpoint between M1 and O1

Frontal 90 degrees, 60 degrees rotated 90 degrees, 60 degrees rotated Frontal Posterior Frontal Frontal 90 degrees 90 degrees Frontal 90 degrees 90 degrees 90 degrees 90 degrees 90 degrees 90 degrees (Continued)

793e

Plastic and Reconstructive Surgery • November 2013 Table 2.  (Continued) Landmark

Position

Q1 R1

Right ZM suture line Right jugal point

S1 T1

Left infraorbital foramen Right orbital rim

U1

Right orbitolateral corner

V1 W1 X1

Left orbital rim Anterior nasal spine* Left orbitolateral corner

Y1 Z1 A2

Left superior ZT suture line Left inferior ZT suture line Left articular tubercle

B2 C2

Left articular tubercle Left articular tubercle

D2 E2

Left zygomatic arch Left jugal point

F2 G2 H2 I2

Left ZM suture Right midpoint of the orbital wall Right supraorbital rim Midpoint of the right gonion (H) and point directly below U Midpoint of the left gonion (I) and point directly below T Superior part of insertion onto the squamous part of the right temporal process zygoma Inferior part of insertion of the right temporal process of the zygoma

J2 K2 L2 M2 N2

Midpoint between L2 and K2, insertion of the right temporal process Superior part of insertion of the left temporal process of the zygoma

O2

Inferior part of insertion of the left temporal process of the zygoma

P2

Midpoint of insertion of the left temporal process Left midpoint of the orbital wall Left supraorbital rim Right orbitolateral wall Left orbitolateral wall Midpoint between Q and K, left side over the condylar process Midpoint between J and Q, left side of the coronoid process Medial end of the left condylar process Apex of the left condylar process Midpoint between L and P on the right condylar process Midpoint between P and M on the right coronoid process Medial aspect of the right condylar process Superior aspect of the right condylar process Midpoint T1 and Q1 on the right maxilla

Q2 R2 S2 T2 U2 V2 W2 X2 Y2 Z2 A3 B3 C3

Description Most inferior point of the ZM suture line Point on the zygomatic bone where the superior border of the zygomatic arch changes from horizontal to vertical Point in the center of the infraorbital foramen Point directly above the right infraorbital foramen Point on the inner orbital rim where the lower margin of the orbit meets the lateral margin of the orbit Point directly above the left infraorbital foramen Apex of the anterior nasal spine Point on the inner orbital rim where the lower margin of the orbit meets the lateral margin of the orbit Most superior point of the ZT suture Most inferior point of the ZT suture Point posterior to the inferior point of the tubercle of the zygoma Most inferior point of the tubercle of the zygoma Point anterior to the inferior point of the articular tubercle of the zygoma Midpoint between A2 and Z1 Point on the zygomatic bone where the superior border of the zygomatic arch changes from horizontal to vertical Most inferior point of the ZM suture Midpoint G3 and R1 Midpoint of the supraorbital notch and right ZF Body of the mandible

Point of View Frontal 90 degrees Frontal Frontal Frontal Frontal Frontal Frontal 90 degrees 90 degrees 90 degrees 90 degrees 90 degrees 90 degrees 90 degrees Frontal 90 degrees Frontal 90 degrees

Body of the mandible

90 degrees

Point where the most superior aspect of the ZT process meets the squamous part of the temporal bone Point where the most inferior aspect of the ZT process meets the squamous part of the temporal bone Insertion of the ZT process

90 degrees

Point where the most superior aspect of the ZT process meets the squamous part of the temporal bone Point where the most inferior aspect of the ZT process meets the squamous part of the temporal bone Midpoint between N2 and O2

90 degrees

Midpoint between E2 and H3, side-on view Midpoint between G1 and left ZF Midpoint A1 and U1 Midpoint A and X1 Part of the condylar process

90 degrees Frontal Frontal Frontal 90 degrees

Part of the coronoid process

90 degrees

Most medial point of the left condylar process Most superior point on the left condylar process Part of the condylar process

90 degrees 90 degrees 90 degrees

Part of the coronoid process

90 degrees

Most medial point of the condylar process Most superior point of the condylar process

90 degrees 90 degrees

Maxilla

Frontal

90 degrees 90 degrees

90 degrees 90 degrees

(Continued)

794e

Volume 132, Number 5 • Correction of Treacher-Collins Syndrome Table 2.  (Continued) Landmark

Position

Description

D3 E3

Midpoint V1 and F2 on the left maxilla Right supraciliary arch above F1

F3 G3

Left supraciliary arch above G1 Right ZF suture from side-on*

H3 I3

Left ZF suture from side-on* Left asterion

J3 K3 L3 M3

O3

Right asterion Right pterion Left pterion Point opposite V2 describing the anterior part of the left coronoid process Point opposite Z2 describing the anterior part of the right coronoid process Right corner of the piriform aperture

P3 Q3

Left corner of the piriform aperture Right canine fossa

R3

Left canine fossa

N3

Point of View

Maxilla Arched elevations on the frontal bone directly above the supraorbital notch As above Frontomalare temporale; most lateral point of the frontozygomatic suture As above Where the lambdoid, parietomastoid, and ­occipitomastoid sutures meet As above Posterior end of the sphenoparietal suture As above Anterior aspect of the coronoid process

Frontal Frontal

As above

90 degrees

Where the lateral and inferior aspects of the piriform aperture meet As above Deepest concavity in the maxilla above the canine tooth As above

Frontal

Frontal 90 degrees 90 degrees Posterior Posterior 90 degrees 90 degrees 90 degrees

Frontal Frontal Frontal

ZF, zygomaticofrontal; ZT, zygomaticotemporal; ZM, zygomaticomaxillary. *Predefined anthropometric landmark.

would be represented by great changes in skull size. The next mode of variation may represent another factor such as mandibular length. Each video for each mode of variation helps the surgeon understand the most important shape changes in the craniofacial skeleton in that specific population. The latter can be useful in determining what reconstructive procedures are suitable to correct underlying deformities. Table 3.  Estimated Ages of Bosma Collection and NHM Scans Based on Dental Eruption Patterns* CT Label

Bosma Age

CT Label

NHM Age (yr)

Bosma9 Bosma10 Bosma11 Bosma12 Bosma13 Bosma14 Bosma15 Bosma16 Bosma17 Bosma18 Bosma19 Bosma20 Bosma21 Bosma22 Bosma23 Bosma24 Bosma25 Bosma27 Bosma28 Bosma29 Bosma30

4 yr ± 12 mo 3 yr ± 12 mo 4–5 yr 2 yr ± 8 mo 3 yr ± 12 mo 4 yr ± 12 mo 4–5 yr 6 yr ± 24 mo 4 yr ± 12 mo 6–7 yr 4 yr ± 16 mo 4–5 yr 3 yr ± 12 mo 5–6 yr 9 yr ± 24 mo 15 yr ± 36 mo 4–5 yr 7 yr ± 24 mo 12–15 yr 10–12 yr 10 yr ± 30 mo

NHM1 NHM2 NHM3 NHM4 NHM5 NHM6 NHM7 NHM8 NHM9 NHM10 NHM11 NHM12 NHM13 NHM14 NHM15 NHM16 NHM17 NHM18 NHM19 NHM20

10–15 16–19 2.5–5 6–9 2.5–5 2.5–5 2.5–5 2.5–5 6–9 10–15 10–15 10–15 2.5–5 10–15 6–9 10–15 6–9 16–19 2.5–5 10–15

CT, computed tomography; NHM, National History Museum. *The Bosma Collection is named after the dry skull collection of Dr. James Bosma.

Planning Correction of Skeletal Treacher-Collins Syndrome Deformities It is not possible to surgically correct all the deformities seen in Treacher-Collins syndrome. For example, correcting the skull base abnormalities or the position of the glenoid fossa would risk unacceptable morbidity. To have a useful surgical planning tool, it is necessary to create a model in which only correctable parts of the craniofacial skeleton are altered. The process of fixing immovable landmarks in three-dimensional virtual models has been described previously.2 In this study,

Fig. 2. Treacher-Collins syndrome patient with 96 landmark points on the facial skeleton (anterior view).

795e

Plastic and Reconstructive Surgery • November 2013

Fig. 3. Normal dry skull with homologous 96 landmarks on the facial skeleton (anterior view).

Video 2. Supplemental Digital Content 2 shows warping of a Treacher-Collins syndrome mandible to a normal age-matched mandible, http://links.lww.com/PRS/A878.

we have built on this concept to model what is surgically possible to correct in the Treacher-Collins syndrome skull. To correct the severely retromicrognathic mandible, the mandible can be isolated from the craniofacial skeleton. The template of 96 landmarks can then be placed on the isolated Treacher-Collins syndrome mandible, and some of the landmarks will be in free space and serve as anchor points to prevent the mandible from becoming distorted during the warping process. Using the “Morph by Lan” function in Robin 3D, the Treacher-Collins syndrome mandible is warped to the 96 landmarks placed on an age-matched dry skull. The virtually corrected Treacher-Collins syndrome mandible can then be edited onto

the original Treacher-Collins syndrome skull that remains unchanged. Thin-plate spline videos also allow us to visually assess the severity of the deformity in different subsections of the mandible and provide a virtual endpoint for distraction procedures. (See Video, Supplemental Digital Content 2, which shows warping of a Treacher-Collins syndrome mandible to a normal age-matched mandible. This video illustrates that there is no mandibular angle, and there is a short ramus and anterior height in the Treacher-Collins syndrome mandible compared with an age-matched control, http://links.lww.com/PRS/A878.) The same principles and techniques can be applied to correct the orbitozygomatic region in Treacher-Collins syndrome. As in the correction

Video 1. Supplemental Digital Content 1 shows warping of a Treacher-Collins syndrome skull to a normal skull using 96 landmarks, http://links.lww.com/PRS/A877.

Video 3. Supplemental Digital Content 3 demonstrates warping of a Treacher-Collins syndrome zygoma with a vestigial temporal process to an age-matched normal zygoma, http://links. lww.com/PRS/A879.

796e

Volume 132, Number 5 • Correction of Treacher-Collins Syndrome of the mandibular deformity, the dysplastic orbitozygomatic region is isolated from the original Treacher-Collins syndrome skull and warped into an age-matched normal orbitozygomatic region. (See Video, Supplemental Digital Content 3, which demonstrates warping of a Treacher-Collins syndrome zygoma with a vestigial temporal process to an age-matched normal zygoma, http:// links.lww.com/PRS/A879.) To virtually reconstruct the Treacher-Collins syndrome group with complete absence of the orbitozygomatic region, a modified process has to be used. First, an age-matched zygoma is isolated from a normal skull using two-dimensional and three-dimensional editing of computed tomographic scan axial slices. Two landmarks are placed over the inferior and superior aspects of the zygomatic body, two are placed over the superior and inferior ends of the temporal process of the zygoma, and finally two are placed over the supraorbital rim. The six landmarks on the edited-out normal zygoma are then warped into the six homologous landmarks placed on the Treacher-Collins syndrome skull. In essence, we are copying and pasting the zygoma onto the original Treacher-Collins syndrome skull but using landmark-based geometric morphometrics to accurately adjust the shape and size of the zygoma through interpolated transformation. The same technique can be used to make size-matched stereolithographic templates in all phenotypic expressions of Treacher-Collins syndrome.

RESULTS Modeling Variation in the Treacher-Collins Syndrome Population Using all of the landmarks from the six Treacher-Collins syndrome patients whose ages ranged from 2 to 12 years, a point distribution model was created. Principal components analysis was used to determine ranked modes of variation. The first three modes of variation accounted for 85.9 percent of the Treacher-Collins syndrome population and were used to illustrate shape variation. To demonstrate this variation graphically, the point distribution model was used to generate a new set of landmarks (±2 SD) along each mode of variation. Thin-plate spline videos were made of each mode of variation to illustrate changes within the Treacher-Collins syndrome population. The first mode of variation in the TreacherCollins syndrome group represents allometric skull growth and demonstrates that the size of the skull is the greatest variable in this Treacher-Collins

Video 4. Supplemental Digital Content 4 demonstrates the first mode of variation of the Treacher-Collins syndrome population along ±2 SD, http://links.lww.com/PRS/A880.

syndrome population of different ages. (See Video, Supplemental Digital Content 4, which demonstrates the first mode of variation of the Treacher-Collins syndrome population along ±2 SD. This video shows the dramatic change in skull size in our Treacher-Collins syndrome population, http://links.lww.com/PRS/A880.) The second mode of variation shows wide variation in the temporal process of the zygoma. The second mode also shows a wide variation in the vertical ramus height of the mandible and also the chin point. This mode of variation shows the Treacher-Collins syndrome population varying from mild mandibular hypoplasia to the more severe deformities of the mandible, with severe hypoplasia or absence of the temporomandibular joint along with a vestigial mandibular ramus. (See Video, Supplemental

Video 5. Supplemental Digital Content 5 demonstrates the second mode of variation in the Treacher-Collins syndrome population along ±2 SD, http://links.lww.com/PRS/A881.

797e

Plastic and Reconstructive Surgery • November 2013 in the Treacher-Collins syndrome population from mild to severe retrusion, http://links.lww. com/PRS/A882.) The first mode of variation in the normal and Treacher-Collins syndrome populations describes allometric skull growth. The higher order modes of variation in the Treacher-Collins syndrome population represent the multiple aspects of dysmorphology in the disease, which is not seen in the normal population. The higher modes in the normal population plateau, illustrating the lack of variability, which is not seen in the TreacherCollins syndrome group (Fig. 4). Video 6. Supplemental Digital Content 6 demonstrates the third mode of variation Treacher-Collins syndrome population along ±2 SD, http://links.lww.com/PRS/A882.

Digital Content 5, which demonstrates the second mode of variation in the Treacher-Collins syndrome population along ±2 SD. This video shows wide variation in the temporal process of the zygoma from complete absence to dysplasia. The second mode also shows significant variation in the vertical ramus height of the mandible and position of the chin point, http://links.lww. com/PRS/A881.) The third mode of variation in the Treacher-Collins syndrome population illustrated a wide spectrum of maxillary hypoplasia, from mild to severe retrusion. (See Video, Supplemental Digital Content 6, which demonstrates the third mode of variation Treacher-Collins syndrome population along ±2 SD. This video illustrates the wide spectrum of maxillary hypoplasia

Landmark Reliability Intraoperator and interoperator landmark reliability were established using univariate statistical analysis using a software package called Lan Analysis. The SD value signifies the mean standard deviation, which is used to determine the repeatability and reproducibility of the landmark. In the intraobserver reliability test, 93 of 96 landmarks were within the threshold of SD less than 1 mm, three of 96 produced an SD greater than 1 mm, and 71 of 96 produced an SD less than 0.5 mm. Ninety-seven percent of the landmarks had good or acceptable repeatability. In the interobserver reliability test, 96 of 96 landmarks were within the threshold of SD less than 1 mm, zero of 96 produced an SD greater than 1 mm, and 89 of 96 were within an SD less than 0.5 mm. One hundred percent of landmarks had good or acceptable repeatability. Having mathematically described the key skeletal features of Treacher-Collins syndrome, the

Fig. 4. Graph comparing modes of variation between Treacher-Collins syndrome (TCS) and normal populations. Mode of variation is shown on the x axis and total variance is shown on the y axis.

798e

Volume 132, Number 5 • Correction of Treacher-Collins Syndrome

Fig. 5. An 8-year-old child with severe Treacher-Collins syndrome and complete absence of the orbitozygomatic region and subtle asymmetry.

degree of variability seen within a Treacher-Collins syndrome population, and the way in which this population differs from a normal population, we have demonstrated two ways in which this is useful to the surgeon. Orbitozygomatic Virtual Reconstruction Case Example The simplified six-landmark technique as described above was used to virtually reconstruct the orbitozygomatic region in a severely affected individual (Figs. 5 and 6). Size-matched stereolithographic zygoma templates were generated

Fig. 7. Zygoma stereolithographic template printed with a rapid prototype device. This template can be used as a guide for calvarial bone graft harvest.

using this technique and printed with a rapid prototyping machine (Fig. 7). These templates can then be used to guide the harvest of calvarial bone graft for reconstruction of the orbitozygomatic region. The templates can also be fashioned to guide accurate placement of the bone graft. Growth in the orbitozygomatic region is nearly complete by the age of 7 years, and most surgeons delay definitive reconstruction to at least this age.13 Therefore, It may be prudent to base these templates on adult sized zygomas at this age without the risk of growth differences.14 This overcorrection could factor in possible bone resorption resulting from the lack of functional load in this region. Mandibular Virtual Reconstruction Case Examples The 96-landmark technique as described earlier under Patients and Methods was used to provide a virtual endpoint for the surgical correction of a case of mandibular deformity in Treacher-Collins syndrome (See Video, Supplemental Digital Content 2, http://links.lww.com/PRS/A878). Geometric morphometrics can also be used to virtually reconstruct the temporomandibular joint where the temporomandibular joint is absent. Through interpolated transformation, the age-matched normal temporomandibular joints are anchored accurately onto the original dysplastic Treacher-Collins syndrome mandible (Figs. 8 and 9).

Fig. 6. An 8-year-old child with Treacher-Collins syndrome after bilateral virtual zygoma reconstruction based on age-matched zygomas.

DISCUSSION In this study, point-based geometric morphometrics analyzed by principal components analysis

799e

Plastic and Reconstructive Surgery • November 2013

Fig. 8. A 4-year-old with severe expression of Treacher-Collins syndrome. The condyle, ascending ramus, and glenoid fossa are not developed bilaterally. There is evidence of vestigial temporomandibular joints.

appears to have been effective in mathematically describing the skeletal deformity seen in TreacherCollins syndrome and offers a method of quantifying and characterizing the variation in deformity seen within this small population studied. Previous studies have described Treacher-Collins syndrome skull morphology either with cephalography, qualitative analysis using three-dimensional computed tomographic head scans, or volume analysis.1,15–20 They all use traditional morphometric techniques, which measure lengths, angles, and

Fig. 9. A 4-year-old Treacher-Collins syndrome patient with virtual correction of vestigial temporomandibular joints. Hybrid temporomandibular joints based on an age-matched normal dry skull are slotted onto the original Treacher-Collins syndrome skull using interpolated transformation.

800e

volumes and relate them to specific defects. However, none of these tools can provide a global picture of the Treacher-Collins syndrome deformity. The advantage of geometric morphometrics based on the Cartesian coordinates of specific homologous points on the skull is that it does provide a very accurate description of the global deformity. This novel system of quantitative analysis takes the description of Treacher-Collins syndrome skull morphology a step further. However, the problem with the technique is the complexity of the results it produces. These results can be readily harnessed by computer software but, presented in raw form, are so complex that they are rendered meaningless to the human observer. To be useful to the clinician, they must be interpreted. In achieving this task, we have used principal components analysis, a form of multivariate analysis that separates out the major modes of variation within a population into various ranked components. When expressed mathematically, these modes of variation still have little meaning to the nonmathematician. The results showing change in shape and variation have therefore been represented in three-dimensional graphic form using thin-plate spline warping. This technique shows the changes that occur to the shape of the skull when one set of landmarked points is warped to another. It renders the differences between two shapes obvious to the observer. In this way, the complex changes described by the underlying analysis can be interpreted by the clinician. The thin-plate spline videos shown in this article demonstrate the differences between Treacher-Collins syndrome and normal skulls, and the modes by which Treacher-Collins syndrome skulls differ from one another. This study includes a relatively small number of subjects, which limits its ability to look with any confidence at multiple modes of variation within our Treacher-Collins syndrome group. It is interesting, however, to see the results of looking at the variations present with principal components analysis. Treacher-Collins syndrome is caused by an abnormality in a single gene and is known to have variable penetrance. The effects of the gene on facial development show well-described features that to the observer appear to produce extremely variable facial forms. Principal components analysis allows us to separate mathematically distinct modes of variation. When these variables are viewed with thin-plate spline warps, it can be seen that what to the human eye is a confusing variety of variation can be broken down to a few modes of variation confined to the first few eigenvectors. Separating

Volume 132, Number 5 • Correction of Treacher-Collins Syndrome out modes of variation in this way allows various aspects of the deformity to be studied in isolation from the global anomalies seen. The particular technique used to acquire and analyze data in this study had several limitations. For example, it could not quantify mathematically the severe cohort of Treacher-Collins syndrome with complete aplasia of the orbitozygomatic region. It can estimate those with vestigial remnants of the zygoma; however, if these remnants are not present, these scans have to be excluded from the principal components analysis. Nevertheless, the spectrum of severity was still quite pronounced in our small group of patients, from mild skeletal dysplasia to more pronounced dysmorphology resulting in functional problems. Other authors have also suffered from groups of patients not truly representative of the TreacherCollins syndrome population and low sample sizes because of the rarity of the condition.1,17,21 We have shown that geometric morphometrics still has a role in the severely affected child with absent zygomatic arches or complete absence of the orbitozygomatic region. If there is unilateral absence of the orbitozygomatic region secondary to trauma or malignancy, surgeons have used the contralateral normal facial skeleton to design a size-matched zygoma or absent temporomandibular joint process.22 However, in Treacher-Collins syndrome, using this virtual mirror technique is not possible because of bilateral aplasia of skeletal structures. If there is bilateral absence, morphing of age-matched zygomas using geometric morphometrics provides a solution (Figs. 5 and 6). Our cohort of age-matched controls was also limited in sample size and probably not truly representative of an entirely normal population. The Natural History Museum scans were from Victorian England, and we are not entirely certain how some of these children died and whether they were malnourished. Some of the scans from the Johns Hopkins Bosma (named after Dr. James Bosma) archive were also collected from the developed world; therefore, these could have suffered from health issues. To acquire a larger population of age-matched controls, one would need to retrieve trauma computed tomographic head scans. However, these will be of insufficient quality because of inadequate computed tomographic slices. This would pose a challenge to accurate landmarking. It is also not ethically feasible to perform computed tomographic scanning on normal patients because of the high radiation dose. Newer generations of computed tomographic scanners

with lower radiation doses may be the answer to this dilemma. Each mode of variation illustrates to us the greatest shape change in the Treacher-Collins syndrome skulls. The first mode of variation relates mostly to size and is a reflection of the different ages of patients in this series. The second mode of variation removes allometric growth and is the most elegant representation of the wide phenotypic expression in Treacher-Collins syndrome. The second mode demonstrated that those with more severe orbitozygomatic defects have the more severe mandibular deformities. The third mode of variation also demonstrated the retruded maxilla in Treacher-Collins syndrome and how this can be disguised by the retromicrognathic mandible. On direct comparison with the modes of variation in the normal population, it is clear that TreacherCollins syndrome is characterized mainly by mandibular, orbitozygomatic, and maxillary hypoplasia but with significant variation. Although other tools have described these changes, here we represent them in a dynamic fashion, which is particularly important in understanding the unique deformities in Treacher-Collins syndrome skeletal anatomy. Geometric morphometrics may have a significant role to play in guiding surgical reconstruction in Treacher-Collins syndrome, and in this study we have suggested ways in which the application of morphometrics in preoperative and intraoperative phases may aid craniofacial reconstruction. Stereolithographic templates are already used in clinical practice today, and many craniofacial procedures use templates to guide surgery in cancer excision, trauma, and also pediatric craniofacial reconstruction.23–25 Stereolithographic templates can be used with existing technology to guide and improve the precision of existing techniques.26 Geometric morphometrics may serve as a tool with which to enhance the correction of mandibular deformities that warrant bone grafting. Warping a Treacher-Collins syndrome mandible to an age-matched control can provide a stereolithographic template for surgical guidance. It may also serve as an adjunct to distraction osteogenesis by highlighting which subparts of the Treacher-Collins syndrome mandible need the most correction. However, this global correction of the dysplastic Treacher-Collins syndrome mandible is not achievable with present-day distraction procedures. The warped mandible does provide us with the “ideal” virtual endpoint of surgery; unfortunately, with current techniques, it is not feasible to achieve the perfect mandible as is

801e

Plastic and Reconstructive Surgery • November 2013 shown in the figures in this study. These videos are also instructive educational tools for younger surgical residents (e.g., one can see how the maxilla needs counterrotation, which can be corrected by the Tessier integral procedure).27 The videos of the isolated mandibles also illustrate that there is no mandibular angle and a short ramus and anterior height. Supplemental Digital Content 1 shows correction of a Treacher-Collins syndrome skull to an age-matched normal skull, but this fails to completely correct the class II malocclusion. The correction of the severe global deformities in Treacher-Collins syndrome using this technique is still far from perfect, and in some morph videos we cannot correct the anterior open bite because of a limited number of landmarks over the teeth in the growing child. The correction of the anterior open bite can only be addressed by isolating the Treacher-Collins syndrome mandible and morphing it to an age-matched normal mandible. The corrected mandible can then be slotted onto the original Treacher-Collins syndrome skull, correcting for the malocclusion (Fig. 10). Current techniques for shaping the zygoma harvested from the calvaria are reliant on the artistic abilities of the surgeon, which provides scope for error. Stereolithographic templates designed using geometric morphometrics confer an advantage to the surgeon on planning the shape, size,

and placement of the bone autograft with the surrounding bony anchor points. Posnick et al., in a series of eight Treacher-Collins syndrome patients, described the use of flat metallic templates for zygomatic reconstruction in Treacher-Collins syndrome.5,26 These were constructed from dry skulls but were not age or size matched. Unlike the flat metallic templates described in the literature, our method better follows the normal contours of a zygoma, with a three-dimensional virtual representation of the final outcome. This is useful particularly for contoured cranial bone grafts before inset into the facial skeleton. These also serve as a guide for determining the zygomatic arch and orbital wall length in the Treacher-Collins syndrome orbitozygomatic region. The technique also accounts for the subtle asymmetries that are seen in the orbitozygomatic region in TreacherCollins syndrome.19 With the advent of scaffolds fabricated with three-dimensional printing, these three-dimensional virtual models designed using geometric morphometrics have great applications for the ability to calculate the exact shape for customized implants for zygomatic or mandibular bone replacement in Treacher-Collins syndrome.28 The chief disadvantage of bone graft in this area, however, is its tendency to resorb. This is thought to be because the bone graft is not subjected to any functional load, as it would be in

Fig. 10. An example of a 2-year-old Treacher-Collins syndrome patient with severe retromicrognathia. The Treacher-Collins syndrome mandible has been isolated and corrected after warping to an age-matched control. The newly warped mandible has then been slotted onto the original Treacher-Collins syndrome skull, correcting for the anterior open bite.

802e

Volume 132, Number 5 • Correction of Treacher-Collins Syndrome most mandibular or maxillary osteotomies. One could therefore argue that the templates we design should take into account the degree of resorption in the calvarial bone grafts. However, this is only possible to quantify in long-term studies on Treacher-Collins syndrome patients, which are scarce in the craniofacial literature. Fan et al.14 assessed outcomes of calvarial bone graft reconstruction for the zygoma in Treacher-Collins syndrome and noted that the 0- to 5-year-old age group experienced the greatest bony resorption, with 100 percent of bony grafts undergoing either partial or complete resorption at 1 year. Their 6- to 12-year-old age group had 25 percent of patients undergoing either partial or complete resorption. The group aged 13 years and older experienced resorption of grafts in only 14 percent. Having corrected the bony abnormality in Treacher-Collins syndrome, the soft tissue is augmented as necessary. This study concentrated on the planning of the bone reconstruction. After skeletal reconstruction, there are still remaining soft-tissue deformities in the temporal and infraorbital regions of these children. Some authors have attempted to address the outward appearance of Treacher-Collins syndrome with soft-tissue augmentation. Procedures such as fat transfer29 and free muscular flap transfer have been described30 to correct the soft-tissue deformities. Formal rhinoplasty to correct the beaked nose is also delayed until all maxillary and mandibular operations are completed, to achieve the best cosmesis. Some children with Treacher-Collins syndrome also suffer from cleft palate, and this can be corrected by 12 months. This can complicate midface growth and should be considered when planning surgery. The aberrant craniofacial muscle structure in Treacher-Collins syndrome is also worthy of special mention and may explain the relatively unique bony deformities seen in the condition. Previous literature has described that the normal growth and development of the craniofacial skeleton can be affected by muscular function.31,32 In TreacherCollins syndrome, the muscles of mastication are absent or dysplastic, with abnormal insertions and origins.33 The result could explain the zygomatic and mandibular dysplasia that is relatively unique to Treacher-Collins syndrome. The abnormal pull of the muscles of mastication that insert onto the mandible could explain the temporomandibular joint disturbances and the underlying bony defects.34 In children with Treacher-Collins syndrome, there are few magnetic resonance imaging data describing the aberrant or dysplastic presence of the muscles of mastication. Further

magnetic resonance imaging work and biomechanical testing of aberrant muscle structures in animals may help prove the latter hypotheses. Unfortunately, our size-matched templates do not take into account the overlying dysplastic muscles in Treacher-Collins syndrome or the soft-tissue envelope. One can mask the soft-tissue defects in Treacher-Collins syndrome by augmenting the overlying soft tissues with autologous fat transfer.14 The technique has the advantage that it is repeatable and can be undertaken before skeletal maturity, as it does not interfere with bony growth. Fat transfer can be very useful in the growing child, who is psychologically affected by their appearance, and can be used for this purpose in all categories of zygomatic anomaly.20 The disadvantage of fat transfer used in isolation is that it is really effective only for minor defects. Fat transfer has a role in more severe defects used in combination with hard-tissue augmentation. Further work should also include the softtissue analysis of Treacher-Collins syndrome patients using geometric morphometrics. Warping Treacher-Collins syndrome faces to normal faces can help quantify the changes and the subtle variations in soft-tissue deformities in TreacherCollins syndrome. Indeed, previous authors have stated that the soft-tissue envelope reflects the underlying severity of the skeletal deformity.17 Here, standardized tools can be made to guide the correction of the deformities, quantifying the amount of soft-tissue augmentation needed for a desired virtual endpoint. The technology will be helpful in defining how much bone and softtissue envelope is needed in different subunits of the face. Our research group also aims to develop automatic landmarking tools to hasten the timely process of manual landmarking of normal scans, a process that can take up to 30 minutes for each scan. Automatic landmarking for normal scans is currently feasible; however, in complex cases such as Treacher-Collins syndrome, which have marked phenotypic variation, automatic landmarking may not be reliable. Further work would also involve using medical imaging to determine how the softtissue envelope in Treacher-Collins syndrome is altered when absent structures such as the zygoma are introduced with an autograft. This may be helpful in guiding the soft-tissue reconstruction for the orbital defects in Treacher-Collins syndrome. Currently, it is possible to alter the Hounsfield number in conventional three-dimensional computed tomography to bring up the soft-tissue envelope. This study has also not addressed the postoperative outcomes of Treacher-Collins syndrome reconstruction at our institute. All of the scans on these

803e

Plastic and Reconstructive Surgery • November 2013 children were preoperative three-dimensional computed tomographic scans. Further studies would include comparison of our reconstructive efforts with age-matched controls and identify the successes and failures of our technique compared with current methods of planning orbitozygomatic reconstruction in Treacher-Collins syndrome. As noted by Paul Tessier,27 Treacher-Collins syndrome is one of the most complicated craniofacial syndromes to correct. Craniofacial surgeons have experienced resorption of grafts, donor-site morbidity, malposition after osteotomies, and inappropriate vector choice of mandibular distraction.

CONCLUSIONS Geometric morphometrics and principal components analysis allow characterization and quantification of the craniofacial deformity in Treacher-Collins syndrome. Geometric morphometrics can be used for planning and guiding the surgical correction of complex craniofacial conditions such as Treacher-Collins syndrome. These techniques are likely to play an increasingly important role with the advent of three-dimensional printing techniques for bone scaffolds. We acknowledge that it is not possible to achieve the perfect global correction of the Treacher-Collins syndrome skull using geometric morphometrics. However, it can provide powerful tools that can bring us closer to understanding the specific reconstructive challenges that remain in this difficult condition and thereby provide the best possible result for the individual patient. Dariush Nikkhah, M.Sc., M.R.C.S.(Eng.) Queen Victoria Hospital Holtye Road East Grinstead RH19 3DZ, United Kingdom

REFERENCES 1. Chong DK, Murray DJ, Britto JA, Tompson B, Forrest CR, Phillips JH. A cephalometric analysis of maxillary and mandibular parameters in Treacher Collins syndrome. Plast Reconstr Surg. 2008;121:77e–84e. 2. Ponniah A, Witherow H, Evans R, Richards R, Ruff C, Dunaway D. Planning reconstruction for facial asymmetry. Int J Simulation 2006;7:32–39. 3. Freihofer HP. Variations in the correction of Treacher Collins syndrome. Plast Reconstr Surg. 1997;99:647–657. 4. van der Meulen JC, Hauben DJ, Vaandrager JM, BirgenhagerFrenkel DH. The use of a temporal osteoperiosteal flap for the reconstruction of malar hypoplasia in Treacher Collins syndrome. Plast Reconstr Surg. 1984;74:687–693. 5. Posnick JC, Ruiz RL. Treacher Collins syndrome: Current evaluation, treatment, and future directions. Cleft Palate Craniofac J. 2000;37:434. 6. Taylor JA. Bilateral orbitozygomatic reconstruction with tissue-engineered bone. J Craniofac Surg. 2010;21:1612–1614.

804e

7. Acosta HL, Stelnicki EJ, Boyd JB, Barnavon Y, Uecker C. Vertical mesenchymal distraction and bilateral free fibula transfer for severe Treacher Collins syndrome. Plast Reconstr Surg. 2004;113:1209–1217; discussion 1218. 8. Trainor PA, Dixon J, Dixon MJ. Treacher Collins syndrome: Etiology, pathogenesis and prevention. Eur J Hum Genet. 2009;17:275–283. 9. Robin R. Medical physics and bioengineering (Web site). Available at: http://www.robins3d.co.uk/. Accessed October 3, 2012. 10. Farkas LG. Anthropometry of the Head and Face. 2nd ed. New York: Raven Press; 1994. 11. Pluijmers BI, Ponniah AJ, Ruff C, Dunaway D. Using principal component analysis to describe the Apert skull deformity and simulate its correction. J Plast Reconstr Aesthet Surg. 2012;65:1750–1752. 12. Bookstein FL. Morphometrics in Evolutionary Biology: The Geometry of Size and Shape Change, with Examples from Fishes. Philadelphia, Pa: Academy of Natural Sciences of Philadelphia; 1985. 13. Waitzman AA, Posnick JC, Armstrong DC, Pron GE. Craniofacial skeletal measurements based on computed tomography: Part II. Normal values and growth trends. Cleft Palate Craniofac J. 1992;29:118–128. 14. Fan KL, Federico C, Kawamoto HK, Bradley JP. Optimizing the timing and technique of Treacher Collins orbital malar reconstruction. J Craniofac Surg. 2012;23(Suppl 1):2033–2037. 15. Travieso R, Terner J, Chang C, et al. Mandibular volumetric comparison of treacher collins syndrome and hemifacial microsomia. Plast Reconstr Surg. 2012;129:749e–751e. 16. Terner JS, Travieso R, Chang C, Bartlett SP, Steinbacher DM. An analysis of mandibular volume in treacher collins syndrome. Plast Reconstr Surg. 2012;129:751e–753e. 17. Marsh JL, Celin SE, Vannier MW, Gado M. The skeletal anatomy of mandibulofacial dysostosis (Treacher Collins syndrome). Plast Reconstr Surg. 1986;78:460–470. 18. Travieso R, Chang CC, Terner JS, et al. A range of condylar hypoplasia exists in Treacher Collins syndrome. J Oral Maxillofac Surg. 2013;71:393–397. 19. Wong KR, Pfaff MJ, Chang CC, Travieso R, Steinbacher DM. A range of malar and masseteric hypoplasia exists in Treacher Collins syndrome. J Plast Reconstr Aesthet Surg. 2013;66:43–46. 20. Nikkhah DP, Ponniah A, Ruff C, Dunaway D. A classification system to guide orbitozygomatic reconstruction in Treacher-Collins syndrome. J Plast Reconstr Aesthet Surg. 2013;66:1003–1005. 21. Posnick JC, al-Qattan MM, Moffat SM, Armstrong D. Cranio-orbito-zygomatic measurements from standard CT scans in unoperated Treacher Collins syndrome patients: Comparison with normal controls. Cleft Palate Craniofac J. 1995;32:20–24. 22. El-Khayat B, Eley KA, Shah KA, Watt-Smith SR. Ewings sarcoma of the zygoma reconstructed with a gold prosthesis: A rare tumor and unique reconstruction. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2010;109:e5–e10. 23. Gateno J, Xia JJ, Teichgraeber JF, et al. Clinical feasibility of computer-aided surgical simulation (CASS) in the treatment of complex cranio-maxillofacial deformities. J Oral Maxillofac Surg. 2007;65:728–734. 24. Mehra P, Miner J, D’Innocenzo R, Nadershah M. Use of 3-d stereolithographic models in oral and maxillofacial surgery. J Maxillofac Oral Surg. 2011;10:6–13. 25. Shen Y, Sun J, Li J, Ji T, Li MM. A revised approach for mandibular reconstruction with the vascularized iliac crest flap by virtual surgical planning. Plast Reconstr Surg. 2012;129:565e–566e. 26. Posnick JC, Goldstein JA, Waitzman AA. Surgical correc tion of the Treacher Collins malar deficiency: Quantitative

Volume 132, Number 5 • Correction of Treacher-Collins Syndrome CT scan analysis of long-term results. Plast Reconstr Surg. 1993;92:12–22. 27. Tulasne JF, Tessier PL. Results of the Tessier integral procedure for correction of Treacher Collins syndrome. Cleft Palate J. 1986;23(Suppl 1):40–49. 28. Seitz H, Rieder W, Irsen S, Leukers B, Tille C. Three dimensional printing of porous ceramic scaffolds for bone tissue engineering. J Biomed Mater Res B Appl Biomater. 2005;74:782–788. 29. Zhang Z, Niu F, Tang X, Yu B, Liu J, Gui L. Staged reconstruction for adult complete Treacher Collins syndrome. J Craniofac Surg. 2009;20:1433–1438. 30. Saadeh P, Reavey PL, Siebert JW. A soft-tissue approach to midfacial hypoplasia associated with Treacher Collins syndrome. Ann Plast Surg. 2006;56:522–525.

31. Ingervall B, Thilander B. Relation between facial morphology and activity of the masticatory muscles. J Oral Rehabil. 1974;1:131–147. 32. Kiliaridis S, Engström C, Thilander B. The relationship between masticatory function and craniofacial morphology: I. A cephalometric longitudinal analysis in the growing rat fed a soft diet. Eur J Orthod. 1985;7:273–283. 33. Poswillo D. The pathogenesis of the Treacher Collins syndrome (mandibulofacial dysostosis). Br J Oral Surg. 1975;13:1–26. 34. Stelnicki EJ, Lin WY, Lee C, Grayson BH, McCarthy JG. Long-term outcome study of bilateral mandibular distraction: A comparison of Treacher Collins and Nager syndromes to other types of micrognathia. Plast Reconstr Surg. 2002;109:1819–1825; discussion 1826.

805e