Computerassisted zygoma reconstruction with vascularized iliac …€¦ · a surgical guide was...

Computer-assisted zygoma reconstruction withvascularized iliac crest bone graft

Ali Modabber1*Marcus Gerressen1

Nassim Ayoub1

Dirk Elvers1

Jan-Philipp Stromps2

Dieter Riediger1

Frank Hölzle1

Alireza Ghassemi1

1Department of Oral, Maxillofacialand Plastic Facial Surgery, RWTHAachen University Hospital, Germany2Department of Plastic, Hand andBurn Surgery, RWTH AachenUniversity Hospital, Germany

*Correspondence to: A. Modabber,Department of Oral, Maxillofacial andPlastic Facial Surgery, RWTH AachenUniversity Hospital, Pauwelsstrasse30, 52074 Aachen, Germany. E-mail:[email protected]

Abstract

Background The reconstruction of zygoma is a challenge with regard to aes-thetic and reconstructive demands.

Methods Pre-operative CT data were imported into specific surgical planningsoftware. The mirror-imaging technique was used. A surgical guide transferredthe virtual surgery plan to the operation site, whereby it fitted uniquely to theiliac donor site. A postoperative CT scan was obtained for comparing the actualpostoperative graft position and shape with the pre-operative virtual simulation.

Results A mean difference of 0.71 mm (SD±1.42) for the shape analysis and3.53 mm (SD±3.14) for the graft position was determined. The calculation ofthe closest point distance showed a surface deviation of< 2 mm for the shapeanalysis with 83.6% of values and for the graft position with 35.7% of values.

Conclusion Virtual surgical planning is a suitable method for zygoma recon-structionwith vascularized iliac crest bone graft, with good accuracy for restoringthe three-dimensional anatomy. Copyright © 2013 John Wiley & Sons, Ltd.

Keywords computer-assisted surgery; virtual planning; vascularized iliac crestbone graft; surgical guide; zygomatic reconstruction

Introduction

Malar and zygomatic arch defects often occur after cancer surgery, resection ofbenign maxillary tumours and trauma. The localization, extension and dimen-sions of the defect provide important information for the surgical treatment.Brown and Shaw (1) described a new classification for maxilla and mid-facedefects. However, this classification does not include the zygomatic arch, andthis region represents a complex anatomical structure. There are manytherapeutic options to restore complex craniofacial defects. Microvascularizedbone flaps present an excellent therapeutic alternative for zygoma reconstruc-tion with autologous transplants. Microsurgically revascularized iliac crestbone grafts have the benefit of a rich cancellous blood supply, a large amountof bone and a compact cortex (2), providing an ideal site for the reconstructionof malar and zygomatic arch defects. The aesthetic outcome and satisfyingfacial appearance of the reconstruction depends on the position and shape ofthe graft.

Three-dimensional (3D) modelling, assisted by computed tomography (CT),can be an ideal method for obtaining precise information for reconstructivesurgery. The transformation of digital CT data to 3D software for the simula-tion of the operative field and the donor region provides a detailed and preciseanalysis. It serves as a diagnostic tool to plan the size, shape and exact

ORIGINAL ARTICLE

Accepted: 10 October 2013

Copyright © 2013 John Wiley & Sons, Ltd.

THE INTERNATIONAL JOURNAL OF MEDICAL ROBOTICS AND COMPUTER ASSISTED SURGERYInt J Med Robotics Comput Assist Surg 2013; 9: 497–502.Published online 6 November 2013 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/rcs.1557

placement of the bone graft (3), which may be of impor-tant clinical benefit. Virtual pre-operative planningprovides accuracy in detail without loss of information,and a number of alternative surgical approaches can bevisualized (4). Using surgical planning software to simu-late various surgical scenarios can be convenient andcost-effective (5).

For an accurate translation of the virtual pre-operativeplanning to real-time surgery, operation templatesfabricated by a selective laser-sintering technique providea precise modelling of the detailed anatomical structuresof the defect. Various approaches demonstrate the benefitof prefabricated 3D image templates, linking the virtualoperation plan to the actual surgical procedure (6–9).

Computer-assisted surgery can help facilitate the shap-ing procedure of the bone graft and to increase precisionin order to achieve an optimal aesthetic outcome (10),as well as a shortening of transplant ischaemia time (11).

The aim of this study was to determine the accuracy ofcomputer-assisted zygoma reconstructionwith vascularizediliac crest bone graft and, consequently, whether it shouldbe routinely used for all patients undergoing bony recon-struction of malar and zygomatic arch defects.

Materials and methods

After institutional approval and written, informed consent,the translation from a virtual plan to the operating site witha surgical guide was performed in a patient with a gunshotdefect of the left malar and zygomatic arch.

Pre-operative CT scans of the facial skeleton and iliaccrest were performed with a 128 row multislice CT scan-ner (Somatom Definition Flash, Siemens, Erlangen, Ger-many). Reconstructions were carried out in a bone andsoft tissue window, kernel 30/60 for head and neck and70 for the pelvis. Acquisition of scans for head and neckis done in 0.5 mm slice thickness, for the pelvis in 1 mmslice thickness.

The CT data of the facial skeleton in digital imaging andcommunications in medicine (DICOM) file format wereimported into ProPlan CMF Planning Software (MaterialiseN.V., Leuven, Belgium). The process of segmentationfollowed, in which artifacts were removed and all bonystructures of interest were isolated. A high-quality 3D visu-alization of the mid-face was calculated. The mirror imageof the healthy side served as reference for the virtual recon-struction of the affected malar and zygomatic arch.

Figure 1. Comparison of the actual postoperative graft position (violet) with the virtually planned situation (green): (A) bottomview; (B) frontal view

498 A. Modabber et al.

Copyright © 2013 John Wiley & Sons, Ltd. Int J Med Robotics Comput Assist Surg 2013; 9: 497–502.DOI: 10.1002/rcs

Additionally, angiographic CT scans of the iliac donor site,which also allowed investigation of the arteries, were readin with the software. Depending on the vascularization,the position of the graft selected afterwards wasdetermined at the right iliac crest. After virtual reconstruc-tion of the defect area, as mentioned above, the best-fittingpart of the iliac crest was selected. The positions of thevessels nourishing the iliac graft were also taken intoaccount. The donor site was virtually osteotomized andreplaced into the zygoma defect. After fine adjustment,the data were imported into 3-matic software (MaterialiseN.V.) as an STL file. Thus, based on the final plan, a surgicalguide fitting uniquely to the iliac crest, and indicating thedesired osteotomy lines, graft size and angulation, wasdesigned. With the aid of the rapid prototyping selectivelaser-sintering method, the guide was produced out ofpolyamide powder and solidified using a carboxide laser.The surgical guide linked the computer-assisted surgicalplan to real-time surgery.

A postoperative CT scan was obtained after 6 months tocompare 3D computer models of the final reconstructionwith the pre-operative virtual plan. Using 3-matic software,the pre- and postoperative objects were first aligned usingpoint registration. Thereafter, automatic global surface reg-istration was performed; this registration is based on an

iterative closest point (ICP) algorithm. The remaining skull,after resection, was used to register the postoperative to thepre-operative situation, as it is important to use those ob-jects that remain unchanged through the surgery.

The actual postoperative graft position was comparedwith the virtual simulation (Figure 1). This was donebased on a part comparison algorithm that measures thedistance of every triangle corner of the postoperative sur-face against the planned surface. This resulted in a set ofmeasurements that were analysed in a histogram. Thecolour map overlay histograms represented a close prox-imity of objects coloured green and red to show the in-crease of distance differences from the pre-operative plan.

To compare the actual postoperative graft shape withthe virtual simulation, again the global surface registra-tion was performed, using only the actual postoperativeand simulated graft (Figure 2A, B). The same processwas used in 3-matic as for the graft position comparison.

Results

The presented surgical method of zygoma reconstructionallowed the implementation of predetermination of the iliac

Figure 2. Comparison of the actual postoperative graft shape (violet) with the pre-operative planned shape of the graft (green): (A)lateral view; (B) medial view

Computer-assisted zygoma reconstruction 499


graft with regard to its shape, size and the site of osteotomyduring surgery. Its temporary fixation on the donor site sim-plified the surgical procedure (Figure 3). Guided surgicalsawing of the iliac crest reduced the amount of removedbone to the determined level (Figure 4) and it fitted intothe zygoma defect without major adjustments (Figure 5).No complications were encountered during the surgery orhealing phase. Themicrosurgically revascularized iliac crestbone graft showed excellent perfusion. Comparison ofthe pre- and postoperative 3D computer models using a partcomparison algorithm showed amean difference of 0.71mm(SD±1.42) for the shape analysis and 3.53 mm (SD±3.14)for the graft position. The part comparison-based closestpoint distance of the absolute values indicated that 83.6%of surface deviation was< 2 mm for the shape analysis.The surface deviation showed a 35.7% smaller differencethan 2 mm for the position analysis. The absolute count ofmeasured points in each calculation using the part compari-son algorithm was 7113 points (ED, element distribution).The colour map overlay histograms (Figure 6A, B) showed

the results of the surface deviation analysis for shape andposition of the iliac crest bone graft.

DiscussionThere are many therapeutic options for restoring the facialskeleton. The most commonly used alloplastic implants forzygomatic reconstruction can cause postoperative compli-cations, such as foreign body reaction, swelling, infectionand replacement (12). Using autologous bone flaps such

Figure 3. The surgical guide is temporarily fixed on the exter-nal side of the right iliac, using osteosynthesis screws. Arrowpoints to the deep circumflex iliac artery at the medial side ofthe iliac crest

Figure 4. The iliac crest bone graft, exactly sawed and osteotomizedwith the help of the surgical guide

Figure 5. Positioning and fixation of the iliac crest bone graft forreconstruction of the left malar and zygomatic arch. Arrowpoints to the anastomosis of the deep circumflex iliac vessel tothe left temporal vessels

Figure 6. The surface deviation analysis for shape (A) and posi-tion (B) of the iliac crest bone graft. The calculation showed asurface deviation of<2 mm (red line) for the shape analysis with83.6% of values and for the graft position with 35.7% of values.ED, element distribution



as fibula, iliac crest or scapula is an excellent way ofreconstructing complex craniofacial defects.

Brown and Shaw (1) described the use of microvascu-lar iliac crest bone grafts for the reconstruction of defectsafter maxillectomy, with and without involving the orbit(class I–IV). However, this classification has a weaknesswith respect to zygomatic defects which involve thezygomatic arch. Microvascular iliac crest bone graft alsoprovides a perfect anatomy and bony structure to restorethe curvature of the zygomatic arch (Figure 7A, B). Thevascularized iliac crest graft has a short pedicle, whichmeans that careful planning is required (13).

The goal of zygoma reconstruction is to achieve themaximum aesthetic outcome. Therefore, pre-operativevirtual planning can help to evaluate the defect size andthe relation to neighbouring structures in order to choosethe best possible reconstruction plan. The analysis of theentire facial skeleton in 3D provides more accuracy, makesan overall evaluation possible and supplies valuableinformation which greatly facilitates further treatment.The virtual surgery plan requires a precise 3D model,conforming to the standards of the defect, as the basis forthe design of the transplant in shape, position and angula-tion. Different surgical tools are required from the softwareto perfect the virtual plan. Themirroring tool allows the useof the healthy side as a template for computational super-imposition on the affected side, for the restoration of facialsymmetry (14). However, the use of the mirroring tool caninfluence the accuracy of the pre-operative virtual planbecause of the natural asymmetry of humans skulls (15).

The present study, which uses the computer-assistedmethod ProPlan CMF (Materialise N.V.), previouslydescribed for mandibular and maxillary reconstructionwith microvascular bone flaps (10,11,16), is the first for ma-lar and zygomatic arch reconstruction. It delivers a surgicalguide for intra-operative use, based on an accurate virtualoperation plan, which is made with the aid of simulationbefore surgery. The goalwas to evaluate the accuracy of thismethod for malar and zygomatic arch reconstructionwith amicrovascular iliac crest bone graft.

The registration of the pre- and postoperative data for acomparison of the virtual and actual postoperative situations

can give an idea of whether the surgery has been performedin accordancewith the pre-operative virtual plan. In order tocompare the 3D computer models from the final reconstruc-tion with the pre-operative virtual plan, a postoperative CTscan was obtained after 6 months. The remodelling andresorption of the graft take placemainly in the first 6monthsafter transplantation (17) However, the measurements rep-resent the result after the remodelling phase has been com-pleted, which is very important for the long-term aestheticoutcome. A mean difference of 0.71 mm (SD±1.42) with83.6% surface deviation< 2 mm for the shape analysiswas calculated. It seems that the actual shape comes veryclose to the virtual plan. The surgical guide ensures accuratesawing of the iliac graft during the surgical procedure, witha determined transplant shape, size and number and site ofosteotomies. The immediate insertion of the graft into thezygoma defect followed the explantation, as no time forshaping was necessary.

The measurements for the graft position showed a meandifference of 3.53 mm (SD±3.14) with a 35.7% smallerdifference than 2 mm in the surface deviation calculation.It appears that the exact placement of the bone graft intothe zygoma defect is very difficult without predefined bonymargins. Roser et al. showed a mean maximum distance of2.00mm(SD±1.12) of postoperativemandibular resectionmargins to that of the virtual plan (18). This reflects the factthat the anatomy and 3D structure of the zygoma seems tobemore complex. However, the achieved aesthetic outcomewas very satisfactory (Figure 8A, B). It appears that thenatural asymmetry of faces in humans obscures some inac-curacy in surgical reconstructions to a certain degree (15).

The use of computer-assisted techniques in combinationwith free flaps provides good functional and aesthetic resultswith predictable outcome (19). A surgical guide transfersthe computer-based surgical plan to real-time surgery, whichis the required procedure to achieve the best possible result.The clinical benefits of computer-assisted surgery are likelyto outweigh the expenditure for technology (20). Surgicalguides shorten transplantation time, increase precision andcontrol and minimize the shaping process of the transplant(11). The increased accuracy ensures outcomes of constanthigh quality as regards both shape and aesthetics. Growing

Figure 7. Cone beam computed tomography (CBCT) of the pre-operative zygomatic bone defect (A) and postoperative result (B) ofthe anatomical reconstruction, with contouring accuracy of the zygomatic arch

Computer-assisted zygoma reconstruction 501


complexity in extensive bony defects may raise this method’sindication, as the surgery might be well simplified for suchdefect location, which is difficult to achieve through manualplacement of the graft, even for experienced surgeons (18).

The presentedmethod shows that using custom-made sur-gery guides using vascularized iliac crest bone graft can helpto restore complex areas, such as the malar and zygomaticarch, with good accuracy. We are aware that a randomizedprospective trial with larger sample sizes will be required toevaluate further benefits of computer-assisted zygoma recon-structions with vascularized iliac crest bone grafts.

Acknowledgements

The authors thank Annelies Genbrugge and Joris Bellinckx(MaterialiseN.V., Leuven, Belgium) for their valuable support.

Conflict of Interest

The authors have stated explicitly that there are no conflictsof interest in connection with this article.

References

1. Brown JS, Shaw RJ. Reconstruction of the maxilla andmidface: in-troducing a new classification. Lancet Oncol 2010; 11: 1001–1008.

2. Riediger D. Restoration of masticatory function by microsurgicallyrevascularized iliac crest bone grafts using enosseous implants.Plast Reconstr Surg 1988; 81: 861–876.

3. Marentette LJ, Maisel RH. Three-dimensional CT reconstruction inmidfacial surgery. Otolaryngol Head Neck Surg 1988; 98: 48–52.

4. Schramm A, Gellrich NC, Schmelzeisen R. Navigational Surgeryof the Facial Skeleton. Springer-Verlag: Berlin, Heidelberg, NewYork, 1997; 1–51.

5. Hallermann W, Olsen S, Bardyn T, et al. A new method forcomputer-aided planning for extensive mandibular reconstruc-tion. Plast Reconstr Surg 2006; 117(7): 2431–2437.

6. Yang X, Hu J, Zhu S, et al. Computer-assisted surgical planningand simulation for condylar reconstruction in patients withosteochondroma. Br J Oral Maxillifac Surg 2011; 49: 203–208.

7. Rose EH, Norris MS, Rosen JM. Application of high-tech three-dimensional imaging and computer-generated models in complexfacial reconstructions with vascularized bone grafts. Plast ReconstrSurg 1993; 91: 252–264.

8. Liu XJ, Gui L, Mao C, et al. Applying computer techniques inmaxillofacial reconstruction using a fibula flap: a messengerand an evaluation method. J Craniofac Surg 2009; 20: 372–377.

9. Feng F, Wang H, Guan X, et al. Mirror imaging and preshaped ti-tanium plates in the treatment of unilateral malar and zygomaticarch fractures. Oral Surg Oral Med Oral Pathol Oral Radiol Endod2011; 112: 188–194.

10. Modabber A, Gerressen M, Stiller MB, et al. Computer-assistedmandibular reconstruction with vascularised iliac crest bonegraft. Aesthet Plast Surg 2012; 36: 653–659.

11. Modabber A, Legros C, Rana M, et al. Evaluation of computer-assisted jaw reconstruction with free vascularized fibular flapcompared to conventional surgery: a clinical pilot study. Int JMed Robot 2012; 8: 215–220.

12. Binder WJ, Azizzadeh B. Malar and submalar augmentation.Facial Plast Surg Clin North Am 2008; 16: 11–32.

13. Brown JS, Jones DC, Summerwill A, et al. Vascularized iliac crestwith internal oblique muscle for immediate reconstruction aftermaxillectomy. Br J Oral Maxillofac Surg 2002; 40: 183–190.

14. Scolozzi P. Maxillofacial reconstruction using polyetheretherketonepatient-specific implants by ’mirroring’ computational planning.Aesthet Plast Surg 2012; 36: 660–665.

15. Metzger MC, Hohlweg-Majert B, Schön R, et al. Verification ofclinical precision after computer-aided reconstruction incraniomaxillofacial surgery. Oral Surg Oral Med Oral Pathol OralRadiol Endod 2007; 104: e1–e10.

16. Leiggener C, Messo E, Thor A, et al. A selective laser sinteringguide for transferring a virtual plan to real time surgery in com-posite mandibular reconstruction with free fibula osseous flaps.Int J Oral Maxillofac Surg 2009; 38: 187–192.

17. Verhoeven JW, Ruijter J, Cune MS, et al. Onlay grafts in combina-tion with endosseous implants in severe mandibular atrophy:one year results of a prospective, quantitative radiological study.Clin Oral Implants Res 2000; 11: 583–594.

18. Roser SM, Ramachandra S, Blair H, et al. Accuracy of virtualsurgical planning in free fibula mandibular reconstruction: com-parison of planned and final results. J Oral Maxillofac Surg 2010;68: 2824–2832.

19. Kokemueller H, Tavassol F, Ruecker M, et al. Complex midfacialreconstruction: a combined technique of computer-assistedsurgery and microvascular tissue transfer. J Oral Maxillofac Surg2008; 66: 2398–2406.

20. Ewers R, Schicho K, Undt G, et al. Basic research and 12 years ofclinical experience in computer-assisted navigation technology:a review. Int J Oral Maxillofac Surg 2005; 34: 1–8.

Figure 8. (A) Pre-operative left facial gunshot defect with affected malar and zygoma arch. (B) Postoperative facial symmetry aftercomputer-assisted zygoma reconstruction with vascularized iliac crest bone graft



Computerassisted zygoma reconstruction with vascularized iliac …€¦ · a surgical guide was...

Documents

Transcript of Computerassisted zygoma reconstruction with vascularized iliac …€¦ · a surgical guide was...