Predictability in orbital reconstruction: A human cadaver study. Part II: Navigation-assisted orbital reconstruction

Accepted Manuscript Predictability in orbital reconstruction: a human cadaver study. Part II: navigationassisted orbital reconstruction Leander Dubois, Ruud Schreurs, Jesper Jansen, Thomas J.J. Maal, Harald Essig, Peter J.J. Gooris, Alfred G. Becking PII:

S1010-5182(15)00246-2

DOI:

10.1016/j.jcms.2015.07.020

Reference:

YJCMS 2143

To appear in:

Journal of Cranio-Maxillo-Facial Surgery

Received Date: 5 May 2015 Revised Date:

6 July 2015

Accepted Date: 21 July 2015

Please cite this article as: Dubois L, Schreurs R, Jansen J, Maal TJJ, Essig H, Gooris PJJ, Becking AG, Predictability in orbital reconstruction: a human cadaver study. Part II: navigation-assisted orbital reconstruction, Journal of Cranio-Maxillofacial Surgery (2015), doi: 10.1016/j.jcms.2015.07.020. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Predictability in orbital reconstruction: a human cadaver study. Part II: navigation-assisted orbital reconstruction

RI PT

Leander Duboisa, Ruud Schreursb, Jesper Jansena, Thomas J.J. Maalb, Harald Essigc, Peter J.J. Goorisa, Alfred G. Beckinga

of Oral and Maxillofacial Surgery, Orbital Unit, Academic Medical Centre of

SC

aDepartment

Amsterdam, University of Amsterdam, Academic Centre for Dentistry (ACTA), Meibergdreef 9,

b3D

M AN U

1105 AZ Amsterdam ZO, The Netherlands (Head: prof.dr. J. de Lange).

Laboratory Oral and Maxillofacial Surgery, University of Amsterdam, Meibergdreef 9, 1105

AZ, Amsterdam ZO, The Netherlands (Head: prof.dr. J. de Lange). c

Department of Oral and Maxillofacial Surgery, University Hospital of Zürich, Frauenklinikstrasse 24,

TE D

CH-8091 Zürich, Switzerland (Head: prof.dr. dr. M. Rücker).

Leander Dubois

EP

Send correspondence and reprint requests to:

AC C

Department of Oral and Maxillofacial Surgery Academic Medical Centre, Academic Centre for Dentistry Amsterdam, University of Amsterdam Meibergdreef 9

1105 AZ Amsterdam ZO The Netherlands T: +31 20 5661364 F: +31 20 5669032

E-mail: [email protected]

ACCEPTED MANUSCRIPT

Funding: This study was supported by the KLS Martin Group (Tuttlingen, Germany) and BrainLAB AG (Feldkirchen, Germany). All orbital implants for this study were donated and their stereolithographic files provided by KLS Martin. The navigation equipment suitable for cadaver experiments was

RI PT

provided by BrainLAB AG. Neither organization had any involvement in the study design, collection and interpretation of data, writing of the manuscript, and/or decision to submit the manuscript for

AC C

EP

TE D

M AN U

SC

publication.

ACCEPTED MANUSCRIPT Predictability in orbital reconstruction: a human cadaver study. Part II: navigation-assisted orbital reconstruction SUMMARY

Preformed orbital reconstruction plates are useful for treating orbital defects. However, intraoperative

RI PT

errors can lead to misplaced implants and poor outcomes. Navigation-assisted surgery may help optimize orbital reconstruction. We aimed to explore whether navigation-assisted surgery is more predictable than traditional orbital reconstruction for optimal implant placement. Pre-injury computed tomography scans

SC

were obtained for 10 cadaver heads (20 orbits). Complex orbital fractures (Class III–IV) were created in all orbits, which were reconstructed using a transconjunctival approach with and without navigation. The best

M AN U

possible fit of the stereolithographic file of a preformed orbital mesh plate was used as the optimal position for reconstruction. The accuracy of the implant positions was evaluated using iPlan software. The consistency of orbital reconstruction was lower in the traditional reconstructions than in the navigation group in the parameters of translation and rotation. Implant position also differed significantly in the

TE D

parameters of translation (p = 0.002) and rotation (pitch: p = 0.77; yaw: p <0.001; roll: p = 0.001). Compared with traditional orbital reconstruction, navigation-assisted reconstruction provides more predictable

EP

anatomical reconstruction of complex orbital defects and significantly improves orbital implant position.

Keywords: navigation; orbital fractures; orbital implants; reconstructive surgical procedures; surgery,

AC C

computer-assisted; treatment outcome

ACCEPTED MANUSCRIPT INTRODUCTION

The orbit is often affected by traumatic injuries, which can result in aesthetic deficits and functional ocular impairment, especially when the reconstruction is suboptimal (Wilde and Schramm, 2014). Precise reconstruction of the orbit is the primary step in restoring normal function and aesthetics of the orbit;

RI PT

however, it is difficult to accomplish (Dubois et al., 2015a). The complexity of orbital reconstruction in posttraumatic and post-ablative defects is well described in the published literature (Hammer, 1995, Burnstine, 2002, Gellrich et al., 2002, Ewers et al., 2005, Schramm et al., 2009, Rana et al., 2012, Essig et al., 2013b,

SC

Dubois et al., 2015a, b, and c).

M AN U

Reconstruction outcome is unpredictable because of the difficulty of optimally reconstructing the complex orbital contour (Ellis and Tan, 2003, Markiewicz et al., 2012, Dubois et al., 2015c). Anatomical orbital landmarks function as important surgical guides and may be helpful during reconstruction of the orbit. However, it may be difficult to locate the posterior ledge, which is the most important dorsal anatomical landmark and provides essential support for reconstruction material (Manson et al., 1986, Hammer, 1995,

TE D

Kakibuchi et al., 2004). This is especially the case in a traumatized orbit with the combination of a comminuted fracture of the thin orbital floor and disrupted orbital soft tissue.

EP

Limited surgical exposure hinders the view of the orbital defect and the verification of implant position during surgery (Markiewicz et al., 2012), which may be an important reason for suboptimal placement of

AC C

the implant and for unsatisfactory outcomes such as enophthalmos and/or diplopia. Manson et al. (1986) were the first to associate improper implant positioning with inadequate restoration of orbital volume, resulting in enophthalmos.

Recently, preoperative computer-assisted planning with virtual implant placement has been combined with intraoperative navigation in an attempt to reconstruct the bony orbit more accurately and optimize treatment outcome. The first step in computer-assisted surgery for orbital reconstruction is segmenting the orbital walls and orbital volume. By using a mirroring technique (Bruneau et al., 2013, Essig et al., 2013b), the unaffected side can be copied to the deformed side, creating a template for a custom-made ideal orbit.

ACCEPTED MANUSCRIPT The outcome of the surgical correction depends on the shape and positioning of the orbital implants (Metzger et al., 2007, Rana et al., 2012, Essig et al., 2013b). One advantage of using a preformed implant is that a stereolithographic (STL) software file of the implant can be used preoperatively to find the optimal fit and position in a digital environment (Fig. 1). However, an optimally formed implant does not automatically

RI PT

result in optimal implant placement (Fig. 2).

A computer model can be used intraoperatively as a virtual template to navigate the preplanned bony contours and assess implant fit. Intraoperative navigation assists the surgeon in optimal reconstruction

SC

(Gellrich et al., 2002, Schmelzeisen et al., 2004, Bell and Markiewicz, 2009, Yu et al., 2010, Andrews et al., 2013, Yu et al., 2013). Wilde et al. (2014) suggested that computer-assisted surgery helps achieve

M AN U

predictable outcomes in reconstruction. In the largest clinical cohort study published to date (n = 104), Yu et al. (2013) demonstrated that navigation-assisted surgery (NAS) provided promising and accurate results for the treatment of midfacial deformities. As described by Zhang et al. (2012), NAS probably increases predictability and could become an essential part of the workflow for complex orbital reconstructions.

TE D

Markewietz et al. (2012) concurred with the authors of previous reports and concluded that future studies should explore whether NAS is preferable to conventional techniques for true-to-origin orbital reconstruction.

EP

The aim of this study was to assess the predictability of navigation-assisted orbital reconstruction for

AC C

implant positioning in complex defects in a human cadaver model (Class III–IV) (Dubois et al., 2015c).

MATERIALS AND METHODS

Materials

The human specimens used for this study had been previously used for a study of endoscopic-assisted orbital reconstruction (Dubois et al., unpublished results). Ten human cadaver heads were obtained from the Department of Anatomy of the Academic Medical Centre of the University of Amsterdam. One orbit was excluded because of sinus pathology (osteoma), resulting in a total of 19 orbits eligible for this study.

ACCEPTED MANUSCRIPT The orbital floor and medial wall were fully exposed through a standard transconjunctival incision and retroseptal preparation. Following the Jaquiéry classification (Jaquiéry et al., 2007, Kunz et al., 2013), complex orbital defects (Class III–IV) were created with piezoelectric surgery (Mectron, Carasco, Italy) Computed tomographic scans of the cadaver heads were performed at baseline (with intact orbits, T0), after

RI PT

creation of the orbital defects (T1), and postoperatively after implant placement (T2) (Sensation 64, Siemens Medical Solutions, Forchheim, Germany). Scan parameters included collimation of 20 × 0.6 mm, 120 kV, 350 mAs, pitch 0.85, FOV 30 cm, matrix 512 × 512, reconstruction slice thickness of 0.75 mm with

SC

overlapping increments of 0.4 mm, bone kernel H70s, and bone window W1600 L400.

M AN U

Validation study

The poor consistency of traditional orbital reconstruction, as has been described in an endoscope study (Dubois et al., unpublished results), is shown in Table 1. A validation study was performed to investigate the consistency of navigation-assisted orbital reconstruction. Two oral and maxillofacial surgeons (LD and PS), experienced in the field of orbital reconstruction, performed 10 orbital reconstructions on the cadaver

TE D

heads using preformed orbital titanium mesh plates (KLS Martin, Tuttlingen, Germany) and a navigationassisted workflow (Curve, BrainLAB AG, Feldkirchen, Germany).

EP

Methods

In the first surgical session, all 19 orbits were reconstructed with a transconjunctival approach (traditional

AC C

group) by both surgeons (LD, PG). In a second session, the orbits were reconstructed with a transconjunctival approach combined with navigation by the same surgeons. The implant was placed in the correct position according to the navigation outcomes. The implants were fixed with one bone screw. The drill holes were covered and camouflaged between the two sessions by filling with DuraLay (Reliance Dental Mfg. Co., Worth, Illinois, USA). For consistency of measurement, one surgeon (LD) performed the reconstructions twice for both groups. After each reconstruction, the surgeon filled out a questionnaire regarding the perceived predictability and quality of reconstruction. These methods were similar to those

ACCEPTED MANUSCRIPT used by the same research group in a previous study on endoscopic-assisted orbital reconstruction (Dubois et al., unpublished results).

Contour analysis

RI PT

The quality of the reconstruction was evaluated using iPlan software (version 3.05, BrainLAB AG, Feldkirchen, Germany). The optimal implant position was determined by information from the T0 and T1 scans with two surgeons (LD, PG) in agreement. In both the traditional and navigation groups, the

postoperative (T2) scan was superimposed on the T0 scan with the image fusion modality, available in the

SC

iPlan software. A threshold segmentation with a threshold of ≥1200 Hounsfield units was performed to

M AN U

segment the resulting implant in the T2 scan. The osteosynthesis screw and excess bony tissue were removed from the segmentation result because these structures were not present in the STL model of the planned implant. The three proximal osteosynthesis rings were excluded because their position could have been altered during the surgery by bending. An STL model of the planned implant and the segmentation of the resulting implant were exported from the iPlan software. The orbital implant dislocation frame (OIDF),

TE D

described in a previous study (Schreurs et al., unpublished results) was used to quantify rotational differences (roll, pitch, and yaw) and translational differences between the STL models of the planned implant and the final implant (Fig. 3). Translational differences were expressed as total displacement, which

AC C

Statistical analysis

EP

resulted in the Euclidean distances of the translation given in the x, y and z directions.

To assess inter- and intraobserver variability between the reconstructions, the interclass correlation coefficient (ICC) was calculated for the pitch, yaw, roll, and translation for the reconstructions in the validation study. A paired t test was used to compare the translation and rotation (pitch, yaw, and roll) of the preformed orbital plate in relation to the planned ideal implant position between the traditional and navigation groups. Statistical data analysis was performed using SPSS software (IOS X, version 22.0; SPSS, Inc., Chicago, IL, USA). P <0.05 was considered statistically significant.

RESULTS

ACCEPTED MANUSCRIPT Validation study

The consistency of the orbital implant position in terms of translation (ICC 0.76, 95% confidence interval [CI] 0.25–0.94) and rotation (pitch: ICC 0.78, 95% CI 0.30–0.95; yaw: ICC 0.87, 95% CI 0.53–0.97; roll: ICC

lower in the traditional group than in the navigation group (Table 1).

Comparison of traditional and navigation groups

RI PT

0.51, 95% CI 0.18–0.86) was high in the navigation group. The consistency in orbital reconstruction was

SC

On the CT images, none of the orbital implants were positioned below the ledge in the traditional or navigation groups.

M AN U

Full-defect contours were exposed in 47.4% of the reconstructions in the traditional group. However, defect boundaries were visualised by navigation in 100% of the reconstructions in the navigation group. The surgeons’ satisfaction rate was lower in the traditional group than in the navigation group (Wilcoxon signed

TE D

rank test, p <0.001), which was in accordance with the actual postoperative results.

The difference between the two groups in implant position (translation, yaw, and roll) was statistically significant, with the navigation group showing more favourable results (paired t test, p <0.05; Table 2).

EP

Even in cases of navigation-assisted orbital reconstruction, medial wall involvement resulted in significantly higher yaw than reconstructions in orbits with intact medial walls (independent sample t test, p = 0.008).

AC C

However, medial wall involvement was not associated with significant differences in translation, pitch, or roll (Table 3). More complex defects (Class III vs. IV) were not associated with a higher degree of implant dislocation (independent sample t test, p >0.05; Table 4).

Plots of the implant positions and the most frequent translations and rotations are shown in Fig. 4 and 5. The 95% confidence interval has been marked for both groups.

DISCUSSION

Advanced diagnostic techniques, such as mirroring the unaffected orbit to set anatomical boundaries and

ACCEPTED MANUSCRIPT fitting a preformed or patient-specific implant in a digital environment, have proven to be viable tools for true-to-origin orbital reconstruction (Gellrich et al., 2002, Schmelzeisen et al., 2004, Markiewicz et al., 2011 and 2012, Cai et al., 2012, Rana et al., 2012, Essig et al., 2013b). With preoperative planning, the surgeon is able to set a clear target for ideal implant position. However, the relationship between the final position

RI PT

and the planned position must be verified intraoperatively. Navigation-assisted surgery provides the surgeon with an intraoperative tool for orientation and comparison of the actual implant location with the target location, presumably preventing improper implant placement.

SC

Several reports have described the use of intraoperative navigation for orbital reconstruction of posttraumatic defects (Gellrich et al., 2002, Schmelzeisen et al., 2004, Bell and Markiewicz, 2009, Yu et al., 2010,

M AN U

Zhang et al., 2012, Rana et al., 2012, Yu et al., 2013, Gander et al., 2015). Some of these studies showed that navigation-assisted orbital reconstruction is effective in restoring orbital volume and globe dimensions in complex defects (Gellrich et al., 2002, Markiewicz et al., 2011 and 2012, Zhang et al., 2012, Rana et al., 2012). However, none of these studies compared the effects of orbital reconstruction with and without the

TE D

use of intraoperative navigation. The authors believe that this is the first study to directly compare the effect of navigation on orbital reconstructive surgery within the same specimen.

This study clearly indicates that NAS is a reliable tool that enables significantly better and more consistent

EP

orbital implant placement. It may also reduce the risk of improper placement for experienced orbital

AC C

surgeons who have no experience with navigation (PG). A common cause of inaccurate implant placement is the inability to define the posterior orbital ledge (Manson et al., 1986, Kakibuchi et al., 2004). In our study, none of the implants were positioned below the ledge, even in the traditional group. This study demonstrates that when navigation is not used, suboptimal implant placements by experienced surgeons are most frequently caused by translation in the x-axis, in combination with yaw (Fig. 4 and 5). Navigationaided surgery significantly reduces these translation and rotation errors. Perfect placement was defined as the difference between the planned position and the actual position within a calibration error of 1–2 mm (Metzger et al., 2007, Zhang et al., 2012, Essig et al., 2013a), which corresponds to a deviation of less than 1 mm per axis and a yaw of less than 8°. Although navigation enables significantly better placement than

ACCEPTED MANUSCRIPT other techniques while retaining a similar level of precision (Markiewicz et al., 2011 and 2012, Cai et al., 2012, Essig et al., 2013b), it is still not perfect. More predictable true-to-origin reconstructions may be possible with further development of the technology. Given the current dissimilarities, our group promotes intraoperative imaging as an additional tool to check the orbital implant position.

RI PT

Although NAS can be used for all orbital reconstructions, previous studies have suggested that it is primarily beneficial for complex defects (Gellrich et al., 2002, Markiewicz et al., 2012, Essig et al., 2013b). Complex defects include those with medial wall involvement, loss of the transition zone, or orbital defects extending

SC

to the posterior third of the orbital floor. Through the combination of medial and inferior-posterior orbital fat bulging, it may often be difficult to place the implant in the optimal position. In this study, NAS

M AN U

significantly reduced improper implant placements in these complex defects, as compared with conventional implant placement performed without the aid of computer-assisted navigation.

This study was performed with cadavers to directly compare the quality of orbital reconstruction with and without intraoperative navigation. In our opinion, a cadaver study has many advantages for verifying the

TE D

quality of treatment options. Treating the same fracture twice under comparable circumstances makes it possible to compare both methods on the same cadaver and to evaluate intra- and intersurgeon variability. The laboratory setting is also free of the stress and distraction of a clinical situation, where time constraints,

EP

psychological stress, and fatigue may come into play. However, cadaver tissues are of a different consistency

AC C

than those of living patients, and in a cadaver model important clinical problems such as enophthalmos and diplopia are not displayed. Because this study focused on lost anatomical boundaries and optimal reconstruction, the texture of the human cadaver specimen was acceptable. Peri-orbital tissues are difficult to assess in cadavers because of the firmer consistency of the orbital fat, but the main focus of this study was reconstruction of the bones, which is of fundamental importance.

Most studies use orbital volume and reduction of volume as tools to describe the effect of the reconstruction (Andrades et al., 2009, Markiewicz et al., 2011 and 2012, Strong et al., 2013b, Novelli et al., 2014). As described by Strong et al. (2013a), orbital volume is an important predictor of enophthalmos. However, critically assessing the actual implant position is also an important factor. As shown by Schreurs

ACCEPTED MANUSCRIPT et al. (2015), even a malpositioned orbital implant can lead to a significant reduction of orbital volume and a good clinical outcome, although it is not the desired result (Fig. 6).

By using the OIDF, the actual result can be instantly compared with the desired result, and the differences between planned and actual results may be quantified for all rotational and translational parameters

RI PT

(Schreurs et al., 2015). In NAS, an important surgical goal is created in the preoperative planning phase; NAS is regarded as target surgery. The additional value of navigation assistance in orbital reconstructive surgery can therefore be quantified only if the surgical outcome is compared with the surgical target. Comparing the

SC

location of the final implant with that of the planned implant makes it possible to compare accuracy and predictability between traditional and navigation-assisted reconstructions (Schreurs et al., 2015). In our

M AN U

study, additional analysis with the OIDF enabled comparisons of a higher level than simply making the conclusion that navigation assistance improved implant placement. The rotational and translational parameters also provided details on how the implant positioning was improved.

TE D

CONCLUSION

Although several previous studies have reported promising results, this human cadaver study clearly proves that for true-to-origin orbital reconstructions, the results of navigation-assisted reconstruction are more

AC C

ACKNOWLEDGEMENTS

EP

consistent and predictable than those of traditional reconstruction techniques.

We thank Nick H.J. Lobé and Ludo Beenen, from the Department of Radiology, and Eric J. Lichtenberg and Petra E.M.H. Habets, from the Department of Anatomy, Embryology and Physiology, for their assistance in logistics and cadaveric scanning. We also thank Irene H. A. Aartman for her advice on statistical analysis.

ACCEPTED MANUSCRIPT

RI PT

REFERENCES

Andrades P, Hernandez D, Falguera MI, Millan JM, Heredero S, Gutierrez R, et al.: Degrees of tolerance in post-traumatic orbital volume correction: the role of prefabricated mesh. J Oral Maxillofac Surg 67(11):2404 2411, 2009. doi:10.1016/j.joms.2008.11.024.

2.

Andrews BT, Surek CC, Tanna N, Bradley JP: Utilization of computed tomography image-guided navigation in orbit fracture repair. Laryngoscope 123(6):1389 1393, 2013. doi:10.1002/lary.23729.

3.

Bell RB, Markiewicz MR: Computer-assisted planning, stereolithographic modeling, and intraoperative navigation for complex orbital reconstruction: a descriptive study in a preliminary cohort. J Oral Maxillofac Surg 67(12):2559 2570, 2009. doi:10.1016/j.joms.2009.07.098.

4.

Bruneau M, Schoovaerts F, Kamouni R, Dache S, De Witte O, de Fontaine S: The mirroring technique: a navigation-based method for reconstructing a symmetrical orbit and cranial vault. Neurosurgery 73(1 Suppl Operative):24 29, 2013. doi:10.1227/NEU.0b013e318282a4e3.

5.

Burnstine MA: Clinical recommendations for repair of isolated orbital floor fractures: an evidencebased analysis. Ophthalmology 109(7):1207 1213, 2002.

6.

Cai EZ, Koh YP, Hing ECH, Low JR, Shen JY, Wong HC, et al.: Computer-assisted navigational surgery improves outcomes in orbital reconstructive surgery. J Craniofac Surg 23(5):1567 1573, 2012. doi:10.1097/SCS.0b013e318260ef33.

7.

Dubois L, Jansen J, Schreurs R, Saeed P, Beenen, L, Gooris PJJ, et al.: Predictability in orbital reconstruction: a human cadaver study. Part I: Endoscopic-assisted orbital reconstruction. J Craniomaxillofac Surg (unpublished results).

8.

Dubois L, Steenen SA, Gooris PJJ, Mourits MP, Becking AG: Controversies in orbital reconstruction-I. Defect-driven orbital reconstruction: A systematic review. Int J Oral Maxillofac Surg 44(3):308 315, 2015a. doi:10.1016/j.ijom.2014.12.002.

9.

Dubois L, Steenen SA, Gooris PJJ, Mourits MP, Becking AG: Controversies in orbital reconstruction-II. Timing of post-traumatic orbital reconstruction: A systematic review. Int J Oral Maxillofac Surg 44(4):433 440, 2015b. doi:10.1016/j.ijom.2014.12.003.

10.

Dubois L, Steenen SA, Gooris PJJ, Bos RRM, Becking AG: Controversies in orbital reconstruction-III. Biomaterials for orbital reconstruction: A review with clinical recommendations. Int J Oral Maxillofac Surg, 2015c. (in press)

11.

Ellis E, Tan Y: Assessment of internal orbital reconstructions for pure blowout fractures: cranial bone grafts versus titanium mesh. J Oral Maxillofac Surg 61(4):442 453, 2003. doi:10.1053/joms.2003.50085.

AC C

EP

TE D

M AN U

SC

1.

ACCEPTED MANUSCRIPT Essig H, Dressel L, Rana M, Rana M, Kokemueller H, Ruecker M, et al.: Precision of posttraumatic primary orbital reconstruction using individually bent titanium mesh with and without navigation: a retrospective study. Head Face Med 9(1):18, 2013a. doi:10.1186/1746-160X-9-18.

13.

Essig H, Rana M, Kokemueller H, Zizelmann C, See C, Ruecker M, et al.: Referencing of markerless CT data sets with cone beam subvolume including registration markers to ease computer-assisted surgery-A clinical and technical research. Int J Med Robot 9(3):e39 e45, 2013b. doi:10.1002/rcs.1444.

14.

Ewers R, Schicho K, Undt G, Wanschitz F, Truppe M, Seemann R, et al.: Basic research and 12 years of clinical experience in computer-assisted navigation technology: a review. Int J Oral Maxillofac Surg 34(1):1 8, 2005. doi:10.1016/j.ijom.2004.03.018.

15.

Gander T, Essig H, Metzler P, Lindhorst D, Dubois L, Rücker M, et al.: Patient specific implants (PSI) in reconstruction of orbital floor and wall fractures. J Craniomaxillofac Surg 43(1):126 130, 2015. doi:10.1016/j.jcms.2014.10.024.

16.

Gellrich NC, Schramm A, Hammer B, Rojas S, Cufi D, Lagrèze W, et al.: Computer-assisted secondary reconstruction of unilateral posttraumatic orbital deformity. Plast Reconstr Surg 110(6):1417 1429, 2002. doi:10.1097/01.PRS.0000029807.35391.E5.

17.

Hammer B: Orbital Fractures: Diagnosis, Operative Treatment, Secondary Corrections. Seattle: Hogrefe and Huber, 1995.

18.

Jaquiéry C, Aeppli C, Cornelius P, Palmowsky A, Kunz C, Hammer B: Reconstruction of orbital wall defects: critical review of 72 patients. Int J Oral Maxillofac Surg 36(3):193 199, 2007. doi:10.1016/j.ijom.2006.11.002.

19.

Kakibuchi M, Fukazawa K, Fukuda K, Yamada N, Matsuda K, Kawai K, et al.: Combination of transconjunctival and endonasal-transantral approach in the repair of blowout fractures involving the orbital floor. Br J Plast Surg 57(1):37 44, 2004.

20.

Kunz C, Sigron GR, Jaquiéry C: Functional outcome after non-surgical management of orbital fractures--the bias of decision-making according to size of defect: critical review of 48 patients. Br J Oral Maxillofac Surg 51(6):486 492, 2013. doi:10.1016/j.bjoms.2012.09.016.

21.

Manson PN, Grivas A, Rosenbaum A, Vannier M, Zinreich J, Iliff N: Studies on enophthalmos: II. The measurement of orbital injuries and their treatment by quantitative computed tomography. Plast Reconstr Surg 77:203 214, 1986.

22.

Markiewicz MR, Dierks EJ, Bell RB: Does intraoperative navigation restore orbital dimensions in traumatic and post-ablative defects? J Craniomaxillofac Surg 40(2):142 148, 2012. doi:10.1016/j.jcms.2011.03.008.

23.

Markiewicz MR, Dierks EJ, Potter BE, Bell RB: Reliability of intraoperative navigation in restoring normal orbital dimensions. J Oral Maxillofac Surg 69(11):2833 2840, 2011. doi:10.1016/j.joms.2010.12.043.

24.

Metzger MC, Hohlweg-Majert B, Schön R, Teschner M, Gellrich NC, Schmelzeisen R, et al.: Verification of clinical precision after computer-aided reconstruction in craniomaxillofacial surgery. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 104(4):e1 e10, 2007.

25.

Novelli G, Tonellini G, Mazzoleni F, Bozzetti A, Sozzi D: Virtual surgery simulation in orbital wall reconstruction: integration of surgical navigation and stereolithographic models. J Craniomaxillofac Surg 42(8):2025 2034, 2014. doi:10.1016/j.jcms.2014.09.009.

26.

Rana M, Essig H, Rücker M, Ruecker M, Gellrich NC: Development and demonstration of a novel computer planning solution for predefined correction of enophthalmos in anophthalmic patients

AC C

EP

TE D

M AN U

SC

RI PT

12.

ACCEPTED MANUSCRIPT using prebended 3D titanium-meshes--a technical note. J Oral Maxillofac Surg 70(11):e631 638, 2012. doi:10.1016/j.joms.2012.07.028. Schmelzeisen R, Gellrich NC, Schoen R, Gutwald R, Zizelmann C, Schramm A: Navigation-aided reconstruction of medial orbital wall and floor contour in cranio-maxillofacial reconstruction. Injury 35(10):955 962, 2004. doi:10.1016/j.injury.2004.06.005.

28.

Schramm A, Suarez-Cunqueiro MM, Rücker M, Kokemueller H, Bormann KH, Metzger MC, et al.: Computer-assisted therapy in orbital and mid-facial reconstructions. Int J Med Robot 5(2):111 124, 2009. doi:10.1002/rcs.245.

29.

Schreurs R, Dubois L, Becking AG, Maal TJJ: Measuring predictability in orbital reconstruction: Quantitative assessment of orbital implant position. PLOS ONE (unpublished results).

31.

Strong EB, Fuller SC, Chahal HS. Computer-aided analysis of orbital volume: a novel technique. Ophthal Plast Reconstr Surg. 29(1):1 5, 2013a. doi:10.1097/IOP.0b013e31826a24ea.

32.

Strong EB, Fuller SC, Wiley DF, Zumbansen J, Wilson MD, Metzger MC: Preformed vs intraoperative bending of titanium mesh for orbital reconstruction. Otolaryngol Head Neck Surg 149(1):60 66, 2013b. doi:10.1177/0194599813481430.

33.

Wilde F, Schramm A: Intraoperative imaging in orbital and midface reconstruction. Facial Plast Surg 30(5):545 553, 2014. doi:10.1055/s-0034-1393700.

34.

Yu H, Shen G, Wang X, Zhang S: Navigation-guided reduction and orbital floor reconstruction in the treatment of zygomatic-orbital-maxillary complex fractures. J Oral Maxillofac Surg 68(1):28 34, 2010. doi:10.1016/j.joms.2009.07.058.

35.

Yu H, Shen SG, Wang X, Zhang L, Zhang S: The indication and application of computer-assisted navigation in oral and maxillofacial surgery-Shanghai’s experience based on 104 cases. J Craniomaxillofac Surg 41(8):770 774, 2013. doi:10.1016/j.jcms.2013.01.016.

36.

Zhang S, Gui H, Lin Y, Shen G, Xu B: Navigation-guided correction of midfacial post-traumatic deformities (Shanghai experience with 40 cases). J Oral Maxillofac Surg 70(6):1426 1433, 2012. doi:10.1016/j.joms.2011.03.068.

AC C

EP

TE D

M AN U

SC

RI PT

27.

ACCEPTED MANUSCRIPT TABLES Table 1 Interclass correlation coefficients (ICC) for inter- and intra-surgeon variability. Table 2 Evaluation of planned vs. realized orbital implant position by using the orbital implant dislocation frame. *Significant values, p <0.05.

RI PT

Table 3 Effect of medial wall involvement on planned vs. realized orbital implant position by using the orbital implant dislocation frame (navigation group). * Significant values p <0.05.

AC C

EP

TE D

M AN U

SC

Table 4 Effect of complexity of the defect on planned vs. realized orbital implant position by using the orbital implant dislocation frame (navigation group). * Significant value, p <0.05.

ACCEPTED MANUSCRIPT FIG. LEGENDS Fig. 1 BrainLab planning: ideal implant fit: a) three-dimensional view; b) coronal view, anterior part; c) coronal view, posterior part; d) sagittal view.

Fig. 3. Degrees of freedom in translation and rotation.

RI PT

Fig. 2 Discrepancies in implant positions based on the three-dimensional reconstruction of pre- and postoperative computed tomography images: a) coronal view, anterior part; b) coronal view, posterior part; c) sagittal view; d) three-dimensional view.

Fig. 4 a) Cranial view, orbital implant; b) plot in cranial view of the navigation group; 95% confidence interval, conventional group (dotted line), navigation group (straight line).

SC

Fig. 5 a) Frontal view, orbital implant; b) plot in frontal view of the navigation group; 95% confidence interval, conventional group (dotted line), navigation group (straight line).

AC C

EP

TE D

M AN U

Fig. 6 Example of misplaced implant, with a volume decrease of 4 cm3 and poor implant positioning.

ACCEPTED MANUSCRIPT Table 1 Interclass correlation coefficients (ICC) for inter- and intra-surgeon variability Conventional

.691

.322

.760

.665

95% CI[.106-.921]

95% CI [-.348-.774]

95% CI[.246-.940]]

95% CI[.314-.856]

.254

.067

.783

95% CI[-.451-763]

95% CI[-.557-.643]

95% CI[.300-.947]

95% CI[.330-.861]

.533

.385

.871

.301

95% CI[-.150-.871]

95% CI[-.283-.801]

95% CI[.533-.969]

95% CI[-.165-.657]

.742

.599

.512

.283

95% CI[.208-.935]

95% CI[-.005-.883]

95% CI[-.177-864)

RI PT

inter

TE D

.675

SC

Roll

intra

EP

Yaw

inter

AC C

Pitch

intra

95% CI[-.184-.646]

M AN U

Translation

Navigation

ACCEPTED MANUSCRIPT Table 2 Evaluation of planned vs. realized orbital implant position by using the OIDF. * Significant values, p <0.05.

Navigation

Paired t test

Mean

SD

Mean

SD

Translation

4.98mm

2.19mm

3.29mm

1.64mm

Pitch

-1.29°

3.05°

-1.13°

2.15°

Yaw

17.81°

10.91°

8.82°

8.05°

Roll

-9.81°

9.09°

-2.3°

4.79°

Translation

4.90mm

1.35mm

Pitch

0.43°

6.26°

Yaw

21.99°

Roll

-11.91°

t

M AN U

(PG)

p

3.60

18

0.002*

-0.30

18

0.77

4.51

18

<0.001*

-3.81

18

0.001*

SC

(LD)

df

RI PT

Conventional

1.35mm

2.27

9

0.05*

-2.10°

3.46°

1.58

9

0.15

13.17°

7.94°

10.5°

2.28

9

0.05*

6.13°

-9.60°

5.13°

-1.56

9

0.15

AC C

EP

TE D

4.15 mm

ACCEPTED MANUSCRIPT Table 3 Effect of medial wall involvement on planned vs. realized orbital implant position by using the orbital implant dislocation frame (navigation group). Significant values p <0.05.

Translation

No medial wall fracture Mean SD 3.24 1.86

Medial wall fracture Mean SD 3.32 1.51

p 0.91

Pitch Yaw Roll

-0.57 3.97 -4.12

-1.63 13.19 -.68

0.30 0.008* 0.17

1.63 6.95 3.75

AC C

EP

TE D

M AN U

SC

2.60 6.40 5.37

RI PT

*

ACCEPTED MANUSCRIPT Table 4 Effect of complexity of the defect on planned vs. realized orbital implant position by using the orbital implant dislocation frame (navigation group). Significant value, p <0.05.

Translation

Class III Mean SD 3.15 1.69

Class IV Mean SD 3.47 1.64

p 0.68

Pitch Yaw Roll

-1.54 10.71 -2.06

-0.57 6.25 -2.65

0.35 0.24 0.80

2.2 10.24 4.45

AC C

EP

TE D

M AN U

SC

2.12 5.81 5.22

RI PT

*

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT