Guidelines for Orbital Defect Assessment and Patient-Specific Implant Design: Introducing OA 2 (Orbital Assessment Algorithm)

Abstract

Study Design

This study presents a review of the evolutionary development in reconstructive orbital surgery over the past 3 decades. Additionally, it proposes the Orbital Assessment Algorithm (OA²) to enhance decision-making for intraorbital reconstruction of post-traumatic orbital deformities.

Objective

The objective of this paper is to provide insights into modern post-traumatic orbital reconstruction from a surgeon’s perspective, with a specific focus on adult patients. It aims to highlight the advancements in computer-aided design and manufacturing techniques, particularly in the field of reconstructive orbital surgery, and to introduce the OA² as a tool for improved decision-making in this context.

Methods

The study conducts a comprehensive review of the evolution of reconstructive orbital surgery, focusing on the integration of 3D technology into surgical practices. It also outlines the development and rationale behind the proposed OA2, emphasizing its potential to enhance the accuracy and efficacy of intraorbital reconstruction procedures for post-traumatic deformities.

Results

The review demonstrates the significant progress made in reconstructive orbital surgery, particularly in leveraging 3D technology for virtual modeling, navigation, and the design and manufacturing of patient-specific implants. The introduction of the OA² provides a structured approach to assessing and addressing post-traumatic orbital deformities, offering potential benefits in decision-making and surgical outcomes.

Conclusions

In conclusion, this paper underscores the pivotal role of computer-aided design and manufacturing in advancing reconstructive orbital surgery. It highlights the importance of integrating innovative design concepts into implant manufacturing processes and emphasizes the potential of the OA² to guide surgeons in the management of post-traumatic orbital deformities, ultimately contributing to improved patient outcomes.

Introduction

Orbital deformities may be either acquired (post-traumatic, postablative, or post-inflammatory) or result from congenital deformities (e.g. hemifacial microsomia, or Crouzon’s disease). Although significant, orbital deformities might not necessarily lead to functional losses like loss of stereoscopic vision despite globe dystopia. Six out of twelve cranial nerves contribute to orbital and periorbital function, where this function might not be related to orbital deformity but to impaired nerve dysfunction only. The craniomaxillofacial reconstructive surgeon must identify orbital deformities and malfunctions first clinically, and then according to clinical patient needs. ¹ This may warrant the use of appropriate imaging technologies, such as cone-beam computerized tomography (CT), helical CT, and magnetic resonance imaging (MRI).^2,3 Plain radiographic films like Water’s view, lateral cephalogram, and posterior anterior (PA)-views are historical and no longer have any diagnostic significance; diagnostic imaging for orbital reconstruction today should be 3D-based. Although there is still little evidence that surgically achieving normal orbital volume and shape in unilateral trauma or in congenital deformities (in comparison to reference data from unaffected individuals) is mandatory, it is consented today, that post-traumatic primary deformity repair should attempt to achieve premorbid volume and shape. ⁴ One of the major confusions surrounding the topic of post-traumatic orbital reconstruction is that the literature usually erroneously refers to orbital fractures and their treatment. In most cases, however, this is not a fracture reduction, but the restoration of post-traumatic shape and volume defects. The international AOCMF Orbita3 study ⁵ had one of the first prospective quantitative study approaches after the pioneering publications of Paul Manson^2,6,7 on the correlation of orbital shape and volume regarding enophthalmos and hypoglobus. Unfortunately, the study was conducted at a time when only few clinics were starting to use computer-aided design (CAD)/computer-aided manufacturing (CAM) manufactured implants using the selective laser melting (SLM) technique. Most patient-specific implants were still preformed manually on patient-individual biomodels. This might even today be a fast and cheap method to acquire an individually preshaped orbital implant on an inhouse printed biomodel. ⁸ The result was that the more patient-specific the shape of a titanium implant, the more accurate the reconstruction. Additionally, the use of intra-operative navigation was statistically significant in terms of volume restoration of post-traumatic orbital deformity. ⁵ Several other studies demonstrate the value of intra-operative navigation for more precise post-traumatic orbital reconstruction.^9-11

Secondary orbital reconstructions depend even more on the underlying pathologies or previously performed treatment protocols. ¹² Reconstruction of these defects needs to consider previous treatments, initial surgical interventions, and the original trauma. This is even more relevant in cases where (neo-) adjuvant therapies like radio-, chemo-, and immuno-therapy, have been performed. Although many different types of biomaterials have been used for intraorbital reconstruction, it is generally agreed that materials that are biologically inert and do not disrupt the orbital soft tissues are best for allowing proper orbital function.^13-18 The hope to get clinically relevant input on the question of which implant material serves best for post-traumatic orbital reconstruction, is unfortunately reduced to absurdity when appraising existing meta-analyses; not even the evaluation methods for recording post-traumatic orbital defects are clear.^19,20 More recent publications on post-traumatic orbital floor restoration, which compare titanium mesh with bioresorbable based reconstruction, are misleading, because they do not refer to time points of post-operative evaluation, nor the extent and location of the post-traumatic defect in the addressed orbit, nor is it clear as to what the indication is for using a bioresorbable or a titanium implant. ²¹

The purpose of this whitepaper is to describe and review the evolutionary changes that have occurred in diagnostics, planning, approaches, biomaterials, and quality control to allow for best possible and predictable reconstructive results within the orbit, focusing on adult patients. Most of these aspects account for all types of indications of orbital reconstruction, however, this publication focuses mainly on post-traumatic orbital defect reconstruction.

Workflow

The clinical assessment of post-traumatic orbital pathology ranges from acute emergency assessment in the field to controlled—in hospital—assessment with appropriate instruments and assessment tools. The first step is always to determine appropriate information on orbital function including visual acuity per side, extra ocular movements, globe position and presence of diplopia. ²² The bipupillary line should be assessed as well as corneal sagittal projection, where 2 mm of under or over projection of a corneal surface is considered as normal. Pupillary width, form and function has to be checked as well as gross motility testing within the visual field in all directions. The use of a vision card is helpful in both adults and children to document visual acuity for each eye. Appropriate orbital examination should also include an assessment of the palpebral and lacrimal drainage systems and their function, including documenting any lacerations. To document clinical findings, it can be helpful to draw a simple sketch of the position of the medial and lateral canthi, scleral and corneal show, and globe position. Detection of change in quality of periorbital soft tissues, e.g. elasticity of eyelids via snap testing, has gained importance, especially in secondary orbital reconstruction during treatment planning especially. ¹

Imaging

Today, the gold standard in imaging is acquiring a detailed 3D-volume data set, ¹ because 2D-imaging is not sufficient for the proper assessment of orbital deformities (congenital or acquired) and presurgical planning. For assessment of the hard tissues, cone-beam CT or helical CT should be considered. Additionally, a contrast-enhanced helical CT scan would aid in demonstrating intraorbital soft tissue masses, the position and shape of the eye muscles, the globe, and the optic nerve.^6,23

MRI volume data is beneficial to further assess the soft tissues or to investigate the presence of osteomyelitis of the orbital bones. However, imaging is not only important for diagnostics but also to aid in digital workflow. Such a pathway must fulfill certain technical standards that will allow proper segmentation, virtual modelling, and designing of a digital blueprint for either printing out a biomodel or even to allow for computer-aided design of patient-specific implants.

The technical standards for imaging are as follows:

1. Cone-beam CT: The resolution (100-600 μm voxel size) should ideally aim for a higher resolution-mode, and the field of view should always allow a proper view of both the orbit and neighboring cranial and midfacial structures. In case of landmark-based navigation, a data set must scan a more extensive field, which might need the inclusion of the frontal bones, calvaria, or lateral skull base. In cases where the data set must be referenced with the real patient during surgery while using navigation, the final reference method must be known and considered during the initial scanning procedure, e.g. whether or not a referencing splint should be used in the maxilla during the scanning process.

2. Helical CT: A 0.6-1 mm thick slice should be used. Like the determination of the appropriate field of view in cone beam CT, the same decision must be performed for helical CT scanning. If soft tissue information is needed, a decision must be made on whether contrast is required.

3. MRI: Depending on the clinical questions, the appropriate MRI sequence can be chosen. The field of view must be determined in the same way as used in the cone beam CT or helical CT case. In case of navigation the combined use of different imaging modalities (i.e. CT and MRI) should be considered.

All three imaging technologies should first allow for appropriate diagnosis and then be robust enough for the creation of a digital workflow. ²⁴ However, to allow the 3D-data set to be part of a digital workflow, it needs to be exported in a Digital Imaging and Communications in Medicine (DICOM)-format to allow for free use on professional and medically licensed imaging analyzing platforms, independent of certain viewers (often combined with the imaging machine itself).^25,26 Exporting the data set in DICOM-format enables a qualified start of a digital workflow (first, you must be able to “see it”). Clinical experience shows that about 50% of referred 3D-data sets from other institutions do not fulfill these above-mentioned basic requirements such that the data set is only limited for use in diagnostics. This again results in unnecessary and repetitive imaging with additional radiation exposure to patients and increases the use of resources and costs.

Planning Software

For more than 20 years, software like Voxim® (IVS solutions, Chemnitz, Germany) set the standard for orbital vs craniomaxillofacial interactive imaging analysis (Figure 1). This software already combined the idea of a robust DICOM viewer together with a potent toolbox for interactive analysis, patient-specific planning, virtual modelling, and stereolithography (STL)-export of virtual designs. Unfortunately, it is currently not available and none of the national and international medical academic teaching platforms have managed to replicate it. These institutions have not included similar platforms into their educational and teaching schedules but have rather left it to individual surgeons worldwide to find and implement available digital technologies in clinical assessment. The biomedical industry does not yet see a benefit in the development of such software unless it can be somehow linked to the sales of medical products.

Figure 1.

Coronal view (Voxim®, IVS Solutions, Chemnitz, Germany); aligned dataset with a centered calibrated metric grid (1 box = 1 cm) for efficient measurement.

The digital process starts with the first interaction between a patient and surgeon: “see it” then “say it” (communication between patient and surgeon and surgeon to biomedical engineer) followed by “solve it” (creation of the medical device). Appropriate quality control can only be accomplished if the surgeon fully understands the underlying problem, is able to fully assess the deformity, reconstruct it and then post-operatively ensure that the pre-operative plan and digital workflow were fully met. This can only be achieved in a predictable way if a robust DICOM viewer is created and made widely available (Figure 1). Voxim® enabled deformity assessment with alignment of the dataset and the use of a calibrated grid and coordinate system into the appropriate multiplanar and 3D-view. However, Voxim® must now be regarded as an outstanding, historical tool, which is no longer available.

Today, Visage7.1® (Visage Imaging, San Diego, CA, USA; Richmond VIC, Australia; Berlin, Germany) is one of the most robust DICOM-viewers with very fast data processing. This platform enables very quick alignment and assessment of any cone beam, helical CT, and MRI data set. However, it is still driven from a diagnostic perspective (radiologist view) and is not focused on virtual design or creating an interface for 3D-printing technology or navigation, like other available software, e.g. Elements® (Brainlab, Munich, Germany). Visage® is user-friendly for early morning rounds, teaching, investigation of patients in clinics or during surgery, or even restaging evaluations in post-tumor follow-up (Figures 2 and 3). One of the main strengths is the easy alignment of the data set, measurement of distances and angles, e.g. between the medial wall and the orbital floor.

Figure 2.

Visage® is a robust DICOM viewer, that efficiently allows data set alignment and identification of key features of a post-trauma volume dataset. The red line in the axial view showing the impressed medial orbital wall (A) is set to acquire the corresponding paramedian oblique sagittal view (B), where the whole length of the orbital floor is depressed (this proves that it is the post-traumatic defect which is the issue and not the fracture). The coronal view displays the down-fracture of the orbital floor including the transition zone to the medial wall, the inferior rectus muscle shows rounding (C). The 3D view (D) only serves as an overview, with dislocation of the left nasal bones.

Figure 3.

Example of a clinical case with a post-traumatic right orbital defect, on which the OA² is applied to check the below-mentioned ten key signs (Figures 6 and 7) using Visage®. The medial posterior bulge on the right side is missing (A); the anterior-posterior projection of the orbital floor shows a huge post-traumatic defect (B); the coronal view in hard (C) and soft-tissue (D) window shows the involvement of the orbital floor over the full width starting medially from the transition zone, which itself is in original position. Rounding of the inferior rectus muscle is demonstrated in (D). In such a case, a patient-specific implant helps to provide a safe post-traumatic orbital defect reconstruction.

The software Elements® not only includes a robust DICOM-viewer, but enables any kind of virtual modelling, atlas-based segmentation, and export of virtual designs. It also allows for import of STL-implants and is directly connected to the intra-operative interfaces (navigation and imaging) in the operating room. If it comes to navigation or even the fusion of an intra-operatively acquired cone-beam CT with a pre-operatively planned orbital implant, the intra-operative superimposition of 2 data sets can be performed quickly on site (Figures 4–6).

Figure 4.

Pointer-based navigation for position control during surgery (Elements®, Brainlab, Munich, Germany) with the digital object of the left patient-specific 2 wall orbital implant; in the upper row a multiplanar (axial, sagittal, coronal) and 3D-view show the pointer-tip at the key area at the beginning of the medial posterior bulge at the anterior part of the posterior third of the orbit.

Figure 5.

Trajectory-based navigation for position control of the patient-specific implant (purple) during surgery (Elements®, Brainlab, Munich/Germany) of the same patient with the digital object of the left patient-specific 2 wall orbital implant; the left picture shows navigation of the medial (A) and the right picture of the lateral trajectory in 3D (B).

Figure 6.

Superimposition of the virtual planning (purple) with the post-operative CT scan of the left orbit showing the adequate result of the implant position.

Elements® is a cutting-edge software, which now firstly is used from deformity assessment up to computer-assisted planning with STL-export of a virtual blueprint and later implementation of the computer-aided manufactured patient-specific implant in STL-format (Figure 4). This allows for intra-operative navigation, either trajectory-based or pointer-based. Easily, metric or volumetric measurements are possible.

Trumatch® is the planning platform from DePuy Synthes (West Chester, PA, USA), which is the digital interface for the interaction between surgeon and engineer to guide the digital construction plan during the design process up to design freeze and manufacturing of a patient-specific implant (Figure 7).

Figure 7.

Three sagittal views in a CT scan of a right orbit demonstrate the extent of the post-traumatic orbital floor defect (23,8 mm) in the paramedian oblique sagittal view (A); a proposed implant design is displayed in the equivalent view (B); due to the flat designed posterior projection of the implant in red a correction is implemented by the surgeon for the engineer to change the implant design (C). The posterior projection of the implant was designed straight, i.e. without the preventive design feature of an “inverted snow shovel”; the final virtual orbital implant design is shown (D) in the en face 3D view (Materialise, Leuven, Belgium). (note: the final 3-D design significantly lacks important features, a modern 3D-printed orbital implant should have: implemented landmarks (trajectories, points), metric information on the implant, visible insertion control information on the implant).

3D-Modelling of the Relevant Bone Structures and the Orbital Assessment Algorithm (OA²)

The exact number of different bony structures that contribute to the orbit does not help the surgeon to address all the above-mentioned aspects or later described design features. The relevant bony structures of the orbit are part of our proposed Orbital Assessment Algorithm (OA²) summarized in Figure 8. These include the often-reliable lateral orbital wall with a strong and thick zygomatic and sphenoid bone. The inferior orbital fissure plays an important role separating the orbital floor from the lateral wall and is the landmark that often must be dissected—especially in secondary orbital reconstructions—and allows, after dissection, a clear view to the projection of the lower border of the lateral orbital wall. If the latter is followed posteriorly, the posterior sturdy point of the orbital floor can be safely reached below the optic nerve, i.e. the posterior ledge formed by the orbital process of the palatine bone. The medial orbital wall formed by the ethmoid and lacrimal bone is even thinner than the orbital floor. A clinically non-visible buttress is formed at the junction of the medial orbital wall with the floor, the so-called transition zone. It often must be bridged from the orbital floor to the medial orbital wall in cases of complete loss of the medial to lateral projection of the orbital floor. The entry to the orbit is formed by the very strong cortical orbital rim, which encircles the full orbit and is referred to as the orbital frame. Laterally, the zygomatic bone contributes to the orbital rim, medially it is the maxillary bone. The center of the inferior orbital rim with the above-mentioned posterior ledge (located below the optic nerve canal) resembles the paramedian oblique sagittal projection and view, which allows assessment of the anterior-posterior integrity of the orbital floor (Figure 8). This view shows the importance of the first 8-10 mm from the inferior orbital rim, which contribute to the vertical position of the globe, including posteriorly to the inferior orbital rim the so-called post-entry-zone, which shows a dip and then an ascending line, which is S-shaped and finishes pointing-down at the posterior ledge. This is why this projection shows a typical “lazy S”-shape (Figure 8).

Figure 8.

Representative views and key signs (1-10) for orbital deformity analysis. The figure shows a summary of the schematic views of a left sided intact bony orbit in coronal, (paramedian oblique) sagittal and axial view; accordingly, the ten important key signs are marked, listed and implemented into the relevant views within the Orbital Assessment Algorithm (OA²). Axial view: The dashed line corresponds to the paramedian oblique vector - connecting the center of the inferior orbital rim and the entrance of the bony optic nerve canal (each marked with a large cross). Coronal view: The large cross marks the transition zone (1) between medial wall and orbital floor; the oval shape (10) symbolizes the regular intact shape of the inferior rectus muscle. Paramedian oblique sagittal view: the large cross marks the posterior ledge of the orbit.

In the axial view of the aligned data set, another key area in the posterior third of the orbit can be easily determined, i.e. the medial posterior bulge, formed by the ethmoidal and palatine and maxillary bone (Figures 4 and 8).

The S-shape of the orbital floor together with the medial posterior bulge and an intact lateral orbital wall result in the appropriate orbital volume control with proper projection of the globe. If these structures are significantly displaced, then typically intraorbital widening occurs with consecutive enophthalmos and possible hypoglobus.^7,27 A typical paralleling radiographic sign could be the rounded inferior rectus muscle in the coronal view (Figure 9), ²⁸ which is an indicator for an enlarged orbital volume with relevant opening of the periorbita.

Figure 9.

In the schematic coronal view the inferior rectus muscle (10) is rounded due to a post-traumatic deformity associated with breach of the periorbita; the muscle is rounded in a coronal view on a real patient helical CT scan (soft-tissue window) shown below.

The goal of 3D-modelling is to reshape these bony structures in case of congenital or acquired deformities. Brainlab has simplified 3D-modelling significantly by replacing semi- or automatic segmentation algorithms of data sets by an atlas-based segmentation. This means that the deformity can be perfectly displayed and segmented, and a virtual reconstruction to the original can be automatically designed (Figures 10 and 11). These designs are limited to bony reconstructions only, thus intraorbital or periorbital soft tissue reconstruction has to follow. One promising approach is the introduction of surgical simulators for reconstructive orbital surgery to internalize the complex anatomy for targeted reconstruction. ²⁹

Figure 10.

Due to the atlas function implemented in Brainlab software, a fast algorithm-based segmentation is possible. The fully segmented 3D-data set is displayed for the 3D-, axial, sagittal and coronal view.

Figure 11.

A virtually modelled orbit with thickened thin orbital walls can be exported as an STL-file and serve for a print-out biomodel with autoclavable resin, so that intra-operative individualized adaptation on top of the biomodel is feasible even with non-preformed or preformed orbital meshes (Brainlab, Munich, Germany).

Planning the Surgery

The pre-operative plan has to be adjusted depending on the timing of surgery. During emergency orbital reconstruction, there is no time for patient-specific implant design and manufacturing, thus either non-preformed or preformed orbital implants have to be used (Figure 12). In non-emergency cases, patient-specific design and implant manufacturing can be applied. An adequate volume data set assessment should be performed in either case, during which the OA² check list is applied (Figures 8 and 9). For intra-operative referencing, the diagnostic dataset should be combined with an upper dental arch-borne navigation-splint mounted with four 2.0 screws (each with a cruciform screwhead).

Figure 12.

Different orbital implants are displayed: non preformed fan-plate (A), with preshaping on a patient-specific biomodel (B); a preformed two-wall orbital plate (DePuys Synthes, West Chester, PA, USA) - based on statistical mean values of individuals - for the left orbital floor and medial orbital wall (C) and a first generation patient-specific orbital floor implant manufactured in SLM-technique (KLS Martin Group, Tuttlingen, Germany) (D).

It is advisable to include real time navigation into reconstructive orbital surgery; however, this requires the appropriate setup of a volume data set that can be used for navigation together with a virtually corrected and designed final orbital-outcome of the intended orbital reconstruction.^25,30-32 Where available, intra-operative imaging allows for acquisition of the post-operative 3D-scan; this is useful if revision is needed as it can then be performed immediately during the same surgery (Figure 13).

Figure 13.

Before the end of surgery, an intra-operative cone-beam CT was acquired and the dataset aligned; the orbital implant position is shown in coronal, axial and paramedian oblique sagittal view (A-C) (the same patient addressed in Figures 14–16). (Note: the implant lacks the “inverted snow shovel”-design at the posterior ledge in (C)).

With these described steps, it is not difficult to realize that orbital reconstructive surgery is headed towards a quantifiable quality control process, such as that which has been practiced in orthognathic surgery over the past several decades. Whereas in orthognathic surgery, this analogous workflow has been successfully transferred into a digital workflow, it was orbital surgery that spearheaded this transition into a digital workflow.

The only analogous way of preforming implants and biomaterials historically was on a patient-specific or non-patient-specific biomodel that was used to, for example, pre-shape a thin mesh-based orbital (fan) plate according to the individual reconstructive needs before or during surgery. ⁷ This modelling process can be seen in Figure 12, whereas Figure 7 shows a digital blueprint in a primary orbital reconstructive case with the true-to-original reconstruction of an orbital floor and medial wall repair.

Implant Design

Depending on the biomaterial used for orbital reconstruction, reliable and stable preforming can be done. The more stability required, the more stable the biomaterial has to be. This is why the authors recommend looking for an adequate biomaterial that matches the needs of a specific orbital reconstruction. Furthermore, any implant used should allow for post-operative or intra-operative assessment of position, i.e. it should be radiopaque (“You have to prove that you have been there!”, quote from Prof. Paul Manson, Baltimore, Maryland, US; AOCMF Course on Orbital Reconstruction, Thessaloniki, Greece, 2008) (Figure 14). This is why bioresorbable materials for an orbital implant cannot be checked for correct positioning; this is a major deficit in addition to their inability to be complexly formed, and the fact that they resorb per definition (Figure 15B). At present, alternative biomaterials such as PEEK cannot meet these requirements because they are not radiopaque and, at least at present, are not comparable with titanium-based implants in terms of shape and, in particular, thickness. However, these materials may well play a role in the future.

Figure 14.

Image fusion of 2 volume data sets from the same patient: paramedian oblique sagittal view pre- (A) and post-reconstruction (B) of a post-traumatic orbital floor defect. The post-traumatic orbital floor defect starts from the post-entry zone to the posterior ledge, i.e. the full length of the orbital floor is affected and depressed in the anterior to posterior projection. A non-preformed fan-plate (DePuys Synthes, West Chester, PA, USA) was manually shaped on a patient-specific biomodel for post-traumatic orbital defect reconstruction (note: other than SLM-manufactured patient-specific orbital implants these fan-plate based orbital implants might easily be deformed during the insertion process and they often show sharp edges see Figure 9A.

Figure 15.

Comparison of the upper side of a patient-specific orbital implant in SLM-technique and a manually bent and shaped/cut fan-plate (A): the sharp edges and non-preventive-design of the fan-plate-based orbital implant is obvious. Other than the form-stable implants in (A) an original polydioxanone (PDS) foil is shown in (B), the shaped implant prior to insertion is shown in (C). (note: the important key areas cannot be adequately addressed in the implant design (C)).

Today, a modern orbital implant is not only defined by its biomaterial or original form, but also by its ability to fulfill other design needs such as implemented functionalization and prevention of potential errors (Figure 15A). The modern orbital implant itself should indicate directly (via anatomical landmarks and extensions), or indirectly (via navigational actors, i.e. trajectories) the correct position of the implant within the addressed orbital space (Figures 5 and 16).

Figure 16.

State-of-the-art left large two-wall orbital implant with two navigationable lateral orbital wall extensions (implemented holes) and a lateral on top of the orbital rim-based fixation (KLS Martin Group, Tuttlingen, Germany). The shepherd´s crook shape to the medial orbital wall is intended for the volume correction need (A and B). The design of the orbital implant allows for self-centering in the dissected orbit. (A) biomodel helps the surgeon to have the physical plan of the orbital implant position and the corrected orbit (A); two implemented trajectories allow for navigational control of the correct implant position in the left orbit.

It should be designed in a preventive way so that no negative interaction with sharp edges and orbital soft tissues is possible. Such interactions might lead to damage of the optic nerve or exposed eye muscles. To avoid major adverse effects of post-traumatic orbital reconstruction, an intended deviation in design from rebuilding original anatomy should be implemented into new generation orbital implants. ³³ The orbital implant should be as thin as possible, but as reinforced as needed to allow for form stability while inserting (i.e. no deformation should be possible). The overall design of the implant should be as small as possible, so that even transconjunctival placement without lateral canthotomy is feasible as this provides favorable aesthetic outcomes. ¹² Fixation of an orbital implant should be reduced to 1 or 2 fixation screws, which can be done inside, on top, or outside of the orbital rim. According to the needs for reconstruction of an inner orbital deformity, an implant should be designed as wide as needed so that the implant edges rest securely on reliable bony ledges. Figures 12 and 15A show the evolution in orbital implant design from non-preformed to preformed to individually preshaped and patient-specific implants with and without additional functions implemented into the design.

Design of Guides

Drill-guides or positioning guides may play a role in orbital reconstruction for positioning or screw angulation (Figure 17); however, they are not part of standard procedures. They might play a role in oncology surgery, where defined ablation must be combined with guides in certain cases and where predrilling might be helpful.

Figure 17.

The implant inserted for the patient in Figure 7 via a retroseptal transconjunctival approach, where the incision is placed around 1 cm posterior to the lower eyelid margin (A); a brain or orbital retractor keeps the orbital tissues out of the field of vision (B); the patient-specific implant and an extra drill guide with a separate screw fixation to the inferior orbital rim are shown in (C) (Materialise, Leuven, Belgium); following insertion and single screw fixation on the inferior orbital rim the orbital implant is shown from above in (D).

Surgical Technique

The decision to reconstruct the inner orbital structures should be guided by thorough pre-operative imaging analysis of the aligned volume data set, in conjunction with a comprehensive clinical assessment of the patient; additionally, the patient’s preferences and desires for correction must be taken into account. This will determine the adequate surgical technique for orbital reconstruction. Historically, transcutaneous incisions were used to reach the inner orbit. Thirty years ago, a lateral eyebrow incision was used to reach the lateral orbit and fronto-zygomatical suture. Lower eyelid approaches were the subciliary, midpalpebral or inferior rim incisions. The medial orbital wall was addressed via a coronal approach from the top. Nowadays, transcutaneous incisions should be limited to the upper blepharoplasty approach to reach the lateral orbital roof or lateral orbital wall including the zygomatico-frontal suture, unless pre-existing scars or lacerations exist. Upper lateral eyebrow incisions should remain historical.

The orbital floor and transition zone to the medial wall and the lower part of the lateral orbital wall can be reached via the pre- or retroseptal lower transconjunctival approach, with or without canthotomy and cantholysis ³⁴ (Figure 18). Pre-, trans- or retrocaruncular medial transconjunctival approaches might be indicated when the medial orbital wall needs to be adequately explored to allow implants to be positioned in the area of the medial orbital wall towards the inferior anterior skull base. According to the recommendation of oculoplastic surgeons, transpalpabral lower eyelid approaches should be avoided. The best orbital approach is the one that does not interfere with palpabral function. Special care must be taken to stay in between bone and the periorbita during dissection, which might be difficult in secondary orbital reconstruction cases.^35,36 In cases where the periorbita is significantly injured, scarred, or even disrupted, special care must be taken to avoid harming the inferior orbital nerve or the inferior oblique muscle and to appropriately retract any prolapsing fat.

Figure 18.

3D and multiplanar view (screen capture) from the Curve® navigation system (Brainlab, Munich, Germany) during intra-operative pointer-based real time navigation (the same patient addressed in Figures 7 and 17).

Any surgical technique is dependent on the appropriate use of the surgical instruments. Eyelid and orbital retractors with or without printed measuring scales or preshaped tips can play a significant role. Proper retraction is key in making orbital reconstruction a safe procedure allowing easy implant insertion and final positioning. Any approach and orbital dissection has to be adequate enough to allow for proper visualization of all of the areas of interest, as well as allowing enough space to safely insert and position an orbital implant. The authors do not intentionally take any additional measures to decrease intra-ocular pressure for the purpose of improving visualization.

Rehabilitation and Follow-Up

During any orbital reconstruction, great care should be taken during soft tissue closure ensuring meticulous suspension of the conjunctiva and/or skin wound margins to reduce post-operative lid complications. Additionally, the pre-operative forced duction test should always be compared with a forced duction test at the end of surgery, after implant placement but before wound closure. Of course, it is always advisable to check for any Optic nerve interference, for example by testing the pupil reaction before, during and after the operation.

A recommended follow-up of patients after reconstructive orbital surgery should include visits after 1, 4 and 12 weeks, and after 1 year. Corrective surgery might be indicated (eyelid position, lacrimal drainage system or eye muscles), depending on functional impairments. Patients should be followed for at least 3 months following orbital reconstruction. This timeline allows for most of the swelling to resolve and for the soft tissues to return to normal. The final assessment of both function and appearance should therefore wait for at least 3 months; some authors even suggest waiting for 6 months before deciding on the need for further interventions. A final judgement on the sequelae of a non-treated post-traumatic orbital deformity is only possible 3 months after the trauma. Typically, an earlier decision could be expected 2 weeks after the trauma, when the peri- and intraorbital swelling has decreased. The same accounts for the post-operative follow-up after primary orbital reconstruction and even more after secondary orbital reconstruction, where sometimes months are needed to see the result.

A post-operative assessment of visual acuity should be performed as early as possible and should even be considered in the recovery room when the patient is awake enough to follow instructions. In case of any onset of significant intraorbital swelling (e.g. retrobulbar hematoma) or loss of function (e.g. visual acuity decrease) early revisional surgery is indicated. The recommended 3D-scan is required showing the post-operative implant position, and depending on the individual situation, removal of any intraorbital hardware may need to be considered. Detection of optic nerve trauma requires an emergency treatment.^37-39 In case of double vision, occlusion or prism foils might be helpful at an early time point under the care of an Ophthalmologist. If eye-muscle surgery is considered for correction of double vision, the surgery should be planned after 3 months, though it is better to wait for up to 6 months.

Guidelines for the 3D-Modelling

As shown in Figure 19, screenshots of the imaging analysis of the orbital deformity, with measurements of the addressed defect areas and—if needed—angle measurements in representative 2D-views allow for easy communication between the surgeon and biomedical engineer (Figure 19). The engineers can use available techniques that help virtually reconstruct a deformed CT-data set. In cases of unilateral deformities, mirroring helps to create an individualized virtual ideal design for the affected orbit. The key, however, is to define the correct mirroring plane. The use of the atlas function within the software might be an easy alternative that can be used in either uni- or bilateral orbital deformities to virtually correct the deformed individual data set. Artificial intelligence might contribute to improvements in virtual planning in the future.

Figure 19.

Right paramedian oblique sagittal view of a post-traumatic secondary orbital deformity prior to reconstruction, after primary repair with PDS foil. The curved thick line shows the rough implant design with the post-entry zone (dip) and the “Lazy S”-shape levering on the posterior ledge area with the posterior margin of the implant pointed down.

If both the orbit and the lateral midfacial frame is affected in primary cases or in revision cases (secondary reconstruction), OA² has to be applied. Both the correction of the outer frame and the simultaneous correction of the shape and volume of the orbit must be planned simultaneously (Figure 20): For a patient-specific implant involving both the inner orbit and the outer lateral midfacial frame, each implant should be designed independently. Together, they should independently define the position of a re-osteotomized malar bone, if necessary. In secondary cases, additional measures to decrease orbital volume are required. This is where spacers (Figure 21) are utilized; their number and placement depend entirely on clinical judgment, which must be carefully considered prior to surgery.

Figure 20.

For the patient in Figure 19 the aligned cone-beam CT dataset is shown in the coronal (A), paramedian oblique sagittal (B) and axial (C) view prior to secondary orbital reconstruction. (Note: the right orbital volume is significantly enlarged due to primary reconstruction elsewhere with a bioresorbable implant for the post-traumatic orbital floor defect; the outer frame with the zygoma and zygomatic arch is under corrected).

Figure 21.

Post-operative cone-beam (A, B, D) and helical CT (C) planes show the post-traumatic orbital reconstruction due to the radiopacity of the patient-specific implant (A) in comparison to the unaffected orbital floor (B) in the paramedian oblique sagittal view; in addition to the implant three radiopaque spacers ⁴¹ - made out of the same titanium grade V like the SLM-manufactured orbital implant - are shown to additionally downsize the orbital volume that has been even increased by for- and outwarding the impacted malar bone (C, D).

Post-traumatic secondary corrections with enopthalmos and hypoglobus often require volumetric over-correction in addition to true-to-original reconstruction—the gold standard for primary post-traumatic orbital reconstructions—and must be based on the clinical findings (Figure 21).^31,40

Guidelines for Patient-Specific Implant Modelling

The guidelines for patient-specific implant modelling include the following:

• An implant thickness of least dimension with adequate stiffness and stability, so that there is no secondary deformation during insertion and fixation

• A fully smooth outer edge

• A minimum of 2 fixation holes in case a screw doesn’t tighten

• Avoidance of medial fixation due to bone density and lacrimal apparatus

• An adequate width and length to be safely seated on reliable bony structures

• Additional phalanges or extensions to have position control towards individual unharmed orbital structures (self-centering implant)

• An open or meshed designed surface to enable potential draining of a hematoma into the adjacent sinuses

• Over-correction of the down-directed posterior implant area towards the posterior ledge like an “inverted snow shovel” design, so that no harm occurs to the area close to the optic canal (preventive design element)

• Ideally anatomic landmarks or trajectories should be implemented into the orbital implant to allow for pointer-based and trajectory-based navigational control (functionalized design)

• The biomaterial should be radiopaque to allow for radiographic control of implant position.

The application of the above guidelines can be seen in the cases shown in Figures 22–24. These cases show how an extended orbital floor defect was restored with a state-of-the-art orbital implant. Position control was already achieved intra-operatively with a C-arm cone-beam CT. By respecting the guidelines, excellent restoration of the medial posterior bulge, transition zone, posterior ledge and post-entry zone was performed.^26,42

Figure 22.

Two axial planes of a left corrected post-traumatic orbital defect are shown with a state-of-the-art patient-specific orbital implant recontouring the medial posterior bulge and the medial orbital wall (A and B); the paramedian oblique sagittal cone-beam CT slices show post-operatively, that the orbital implant is spanning the full sagittal projection of the orbital floor up to the posterior ledge (C); for comparison the equivalent right unaffected orbit is displayed (D).

Figure 23.

Axial spiral CT slice showing an intact outer frame (A); the coronal views of the same patient show the hard tissue window (B) and the soft tissue window (C), where significant rounding of the inferior rectus muscle is obivous (with a flat, lentil shaped unharmed contralateral inferior rectus muscle on the right orbit) as a typical feature of significant orbital volume enlargement, resulting from rupture of the periorbita together with significant orbital floor displacement.

Figure 24.

Intra-operatively acquired cone-beam CT with axial (A), coronal (B) views and adjustment for the paramedian oblique sagittal plane in the axial view (C) with the corresponding sagittal view (D) of the same patient from Figure 23. A perfect fit of the orbital implant reconstructing the full orbital floor is demonstrated, extended over the transition zone to the medial wall and seated with an extra extension to the lateral orbital wall.

Guidelines for the Implant Design – Material Properties and Structural Strength

Today, evidence suggests that fewer reoperations are needed when careful imaging analysis is performed before surgery or even during surgery (i.e. analysis of intra-operative 3D cone beam CT after reduction of the lateral orbital frame secondary to zygomaticomaxillary complex fractures). Historically, soft materials were used as orbital implants, which after getting moist, do not keep their form (e.g. lyodura or ethisorb®) (Figures 25 and 26).

Figure 25.

Typical example of an inadequate primary repair of a post-traumatic left orbital defect with PDS foil and consecutive need for an iatrogenic secondary orbital reconstruction due to inappropriate orbital recontouring over time. Coronal view (A) and paramedian oblique sagittal view (B) showing the sagging of the orbital contents. (Note: these primarily inadequate treatments of extended post-traumatic orbital defects with bioresorbable materials predictably result in clinically relevant orbital deformities (enophthalmos, hypoglobus) requiring more complex secondary orbital reconstruction).

Figure 26.

Coronal view of a secondary orbital reconstruction needed after midfacial fracture repair elsewhere including primary PDS-based reconstruction of the inner orbit (A). The large arrows highlight the enlargement of the left orbital volume. TZ marks the transition zone between medial wall and floor. A corresponding coronal view post-operatively is shown on the Brainlab platform (B) with a patient-specific orbital implant super-imposed onto the virtual planning (interrupted line). In addition, three simultaneously trapezoid spacers are shown: they were needed to allow for a balanced volume control in addition to the orbital implant and to correct globe dystopia: coming from an outwards gaze of the left eye luckily no further eye-muscle correction was needed, because regular stereoscopic vision was achieved after secondary orbital reconstruction.

Even today, PDS® foil is often used without a clear and quantifiable indication. A recent publication by Taxis et al ²¹ describes this method, which, however, cannot predictably achieve the desired and required shape and volume restoration, particularly in the case of extensive post-traumatic orbital defects. This applies to both adult and pediatric patients. According to the authors’ experience, pediatric patients can be treated like adults for orbital reconstruction starting from the age of seven. Even when metallic implants are used, growth is not compromised, as fixation involves only 1 or 2 screws to maintain the position, rather than extensive craniofacial plating. This highlights the fact that there is still a lot of educational work to be done in the recognition and treatment of post-traumatic deformities. Applying the proposed OA² would help to firstly, address the reconstruction needs more adequately for these post-traumatic orbital defects and secondly, even avoid unnecessary surgeries.

Today, experts agree that a strong, radiopaque, form-stable orbital implant should be the most common choice for surgeons performing orbital reconstruction. These biomaterials range from autogenous bone to metallic or ceramic implants, and a more patient-specific implant approach is recommended the larger the defect is. When choosing biomaterials, the size of the defect may be a deciding factor, i.e. autogenous bone cannot be designed below 1-2 mm thickness, whereas SLM technology allows manufacturing at thicknesses of around 0.3 mm.^43,44 However, as previously mentioned, mechanical interaction during the insertion, positioning, and fixation process should allow for an implant that does not change its design during the procedure. Such deformation often occurred with the use of meshes that had only 1 thickness and was documented in the AOCMF Orbital3 Study.^5,45 Additionally, the orbit-facing side of the implant should be as smooth as possible. In the future, hybrid implants with coatings such as silicone could be beneficial. Based on the authors’ experience, unpredictable cases occasionally occur where unwanted fibrous attachment of the eye muscles to the perforated titanium surface takes place. A non-adherent surface might help to prevent this issue. Existing hybrid implants where a core metallic implant (mesh-based) is encased with e.g. medpore® are reported to be applicable; however, in cases where secondary intervention is needed they have shown a massive fibrous ingrowth making dissection or any removal very difficult. By today’s standard the most reliable orbital implants are the 0.3 mm thick SLM-orbital implants surrounded by a 0.5 mm circular rounded border. Apart from radiopacity, the chosen implant design needs to be open for additional features like phalanges or extensions that enable the implant to lever on intact and reliable anatomic structures, so that self-positioning information is included into the implant design (Figure 16). Furthermore, the structural strength exhibited by the implant results from the biomechanical forces exerted during the insertion process, because the biomechanical load onto the implant is very low. Finally, navigational control during surgery should still be regarded as an important tool for intra-operative quality control to avoid pitfalls.^46,47 The decision to use 2 implants instead of a single 1 for patient-specific reconstruction depends on the extent of the reconstruction, the surgical approach, and the surgeon’s skills. According to the authors’ experience, a two-piece implant is rarely necessary.

Safe Zones for Screw Fixation

As previously noted, 1 to 2 mini-screw fixations are sufficient to allow for adequate stability of an orbital implant. As a basic rule, more screws might be needed when the orbital implant is less patient specific, and when there are fewer additional design features added to the implant to allow it to lever onto reliable anatomic structures. Safe zones for screw fixation include the inferior orbital rim (inside, on top, or outside), and the lateral orbit (inside, on top, or outside). Any screw fixation needs to respect important anatomical structures that should not be compromised, e.g. inferior orbital nerve or lateral canthus. Screws with a thin screw-shaft diameter with a low thread-pitch, e.g. 1.2, 1.3, 1.5, 1.55 mm, and a length of around 4-6 mm are recommended.

Conclusion

There have been significant advances and improvements in orbital surgery due to the implementation of 3D-technologies for a better understanding of the underlying morphological problem causing the orbital deformity. By taking a comprehensive 3D-data set which exceeds a purely diagnostic level and making this the starting point of a digital workflow, while being aware of functional aspects and under the perspective of clinical relevance, orbital surgery has pioneered in all medical specialties in terms of CAD/CAM technologies and innovative concepts and implant design. Thus, orbital reconstructive surgery has advanced from an early trial-and-error procedure with a lack of quality control with inappropriate 2D radiographs towards a quantifiable and usually predictable result that can be monitored through all phases of treatment. To successfully implement modern orbital reconstruction, it is necessary understand that it is not a matter of post-traumatic fracture treatment, but rather of managing defects. This can only be achieved by consistently dealing with volume dataset information using a robust DICOM viewer. However, despite all modern technologies, biological rules and principles, especially of the soft tissues, still prevail as they do in any area of craniomaxillofacial reconstruction.