*Review* **Personalized Medicine Workflow in Post-Traumatic Orbital Reconstruction**

**Juliana F. Sabelis 1,\* , Ruud Schreurs 1,2, Harald Essig <sup>3</sup> , Alfred G. Becking <sup>1</sup> and Leander Dubois <sup>1</sup>**

	- Geert Grooteplein Zuid 10, 6525 GA Nijmegen, The Netherlands

**Abstract:** Restoration of the orbit is the first and most predictable step in the surgical treatment of orbital fractures. Orbital reconstruction is keyhole surgery performed in a confined space. A technology-supported workflow called computer-assisted surgery (CAS) has become the standard for complex orbital traumatology in many hospitals. CAS technology has catalyzed the incorporation of personalized medicine in orbital reconstruction. The complete workflow consists of diagnostics, planning, surgery and evaluation. Advanced diagnostics and virtual surgical planning are techniques utilized in the preoperative phase to optimally prepare for surgery and adapt the treatment to the patient. Further personalization of the treatment is possible if reconstruction is performed with a patient-specific implant and several design options are available to tailor the implant to individual needs. Intraoperatively, visual appraisal is used to assess the obtained implant position. Surgical navigation, intraoperative imaging, and specific PSI design options are able to enhance feedback in the CAS workflow. Evaluation of the surgical result can be performed both qualitatively and quantitatively. Throughout the entire workflow, the concepts of CAS and personalized medicine are intertwined. A combination of the techniques may be applied in order to achieve the most optimal clinical outcome. The goal of this article is to provide a complete overview of the workflow for post-traumatic orbital reconstruction, with an in-depth description of the available personalization and CAS options.

**Keywords:** patient-specific implants; orbital reconstruction; computer-assisted surgery; surgical navigation; additive manufacturing

## **1. Introduction**

The orbit is an inward-projecting bony structure in the shape of a cone (or pyramid) at the transition between midface and skull base [1–3]. The base of the orbit, the orbital rim, is composed of thick bone; in contrast, the orbit's inner walls are thin bony structures. The orbit provides the casing for the soft-tissue structures associated with the visual (motor) system: neurovascular structures, connective tissue, ocular muscles, and the globe [4,5].

With its central position and thin bony walls, the orbit is probably the most vulnerable part of the facial skeleton [1,2]. Two possible theories of orbital fracture pathogenesis have been suggested. The buckling theory suggests that the energy of a traumatic impact on the orbital rim after blunt-force trauma is propagated to the thin inner walls and leads to a fracture in these weaker structures [6–8]. Hydraulic theory considers increased pressure after impact on the globe and orbital contents as the main reason for orbital wall fractures. The exact nature and specifics of a fracture may be explained by a combination of both mechanisms [7].

The orbital volume may be increased, and soft tissue may be displaced into the adjacent sinuses due to the impact or the dislocation of supporting bony structures. The

**Citation:** Sabelis, J.F.; Schreurs, R.; Essig, H.; Becking, A.G.; Dubois, L. Personalized Medicine Workflow in Post-Traumatic Orbital Reconstruction. *J. Pers. Med.* **2022**, *12*, 1366. https://doi.org/10.3390/ jpm12091366

Academic Editors: Joep Kraeima, Sebastiaan de Visscher and Max J.H. Witjes

Received: 4 July 2022 Accepted: 22 August 2022 Published: 24 August 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

globe's position may be displaced after trauma, for instance, with inward displacement (enophthalmos) or inferior displacement (hypoglobus). The orbital soft tissue may be affected by the traumatic impact as well. The structural integrity and functional capacities of connective tissue or extraocular muscles may be disrupted, resulting in a disturbance of ocular motility and double vision (diplopia). The location and type of the impact, in combination with the amount of energy transferred to the orbit's bony structures and orbital soft tissue, are responsible for the heterogeneity in clinical presentation.

There is an ongoing debate on the indication of surgical reconstruction, and systematic reviews have not been able to provide evidence-supported guidelines [9–11]. Some advocate a radical approach to prevent clinical symptoms [12], while others choose a more conservative approach with a delayed surgery if clinical symptoms develop [13]. Indication for reconstruction remains a subjective decision in most cases, depending on the surgeon and patient characteristics. The surgical management of orbital fractures focuses on the repositioning the orbital contents and the globe and reinstating the structural support to recover ocular function. Restoration of the orbit is the first and most predictable step in the surgical treatment of orbital fractures [14,15].

Nowadays, titanium mesh implants have become the preferred biomaterial for the surgical reconstruction of the orbit. Titanium implants can be categorized into flat implants, preformed implants, and patient-specific implants (PSIs). Flat implants are manually shaped and trimmed by the surgeon. A generic or individual model of the (mirrored) orbit may aid in the molding process. Preformed implants have a predefined shape, based on a model of the average orbit [10,16]. Patient-specific implants (PSIs) are designed on an individual basis for the patient and are subsequently produced through additive manufacturing.

The soft tissue's intricate architecture and the proximity to vital structures pose surgical challenges in orbital reconstruction [17,18]. Orbital reconstruction is keyhole surgery performed in a confined space. This contributes to limited visualization, which is further enhanced by protruding fat. The margin of error is small: an incorrectly positioned implant may have significant implications for the clinical outcome and the patient's quality of life, and it is considered a ground for revision surgery in the literature [19,20]. Medical technology has been incorporated in the clinical workflow of orbital reconstructions to reduce the risk of implant malpositioning [21].

This technology-supported workflow, called computer-assisted surgery (CAS), has become the standard for complex orbital traumatology in many hospitals [22]. The introduction of CAS has also enabled personalization of the treatment: treatment planning is customized to fit the options and needs of the patient, and intraoperative guidance is adjusted to the anatomical possibilities. The main aim of this article is to provide a complete overview of the (CAS) workflow for orbital reconstruction, with an in-depth description of the techniques embedded in the workflow and with a special focus on treatment personalization through patient-specific implant design.

#### **2. Post-Traumatic Orbital Reconstruction Workflow**

The conventional workflow of post-traumatic orbital reconstruction and possible CAS techniques are illustrated in Figure 1. The individual phases are explained in detail in the following paragraphs.

*J. Pers. Med.* **2022**, *12*, 1366 3 of 17

**Figure 1.** Schematic overview of the technological possibilities within the orbital reconstruction workflow. Chronologically, a clinical workflow consists of diagnostics, planning, surgery, and evaluation. In the conventional workflow, the surgeon is dependent on clinical assessment and preoperative imaging, visual appraisal during surgery and postoperative imaging for evaluation. CAS consists of several techniques that may be combined and affect one or multiple workflow stages. Several CAS technologies may be combined. *2.1. Diagnostics*  A thorough clinical and radiographic evaluation of the patient is essential to deter-**Figure 1.** Schematic overview of the technological possibilities within the orbital reconstruction workflow. Chronologically, a clinical workflow consists of diagnostics, planning, surgery, and evaluation. In the conventional workflow, the surgeon is dependent on clinical assessment and preoperative imaging, visual appraisal during surgery and postoperative imaging for evaluation. CAS consists of several techniques that may be combined and affect one or multiple workflow stages. Several CAS technologies may be combined.

#### mine the optimal treatment. Clinical evaluation should at least assess the amount of globe displacement and the degree of double vision. The Hertel exophthalmometer is the sim-*2.1. Diagnostics*

plest tool to quantitatively measure the relative ventrodorsal globe position [23]. It is the current gold standard despite the known limitations, such as the asymmetry of lateral orbital rims, the compression of soft tissue, and the lack of a uniform technique [23,24]. There are no readily available reproducible tools for measuring the relative craniocaudal globe position [25]. It is assessed by the Hirschberg test, which evaluates the light reflex centered on each pupil to reveal vertical asymmetry [26]. Alternative methods have been proposed to quantify globe position differences based on imaging, but these have not been broadly implemented [23,27]. One of the difficulties in clinical decision-making is to address the most common complaint in orbital fractures: diplopia. It is challenging to find the actual cause of diplopia; in most cases, it is caused by a restriction of ocular motility. Ocular motility can be disturbed by impingement or the entrapment of the ocular muscles and surrounding soft tissue, but also by muscle edema, muscle injury, hemorrhage, emphysema, or motor nerve A thorough clinical and radiographic evaluation of the patient is essential to determine the optimal treatment. Clinical evaluation should at least assess the amount of globe displacement and the degree of double vision. The Hertel exophthalmometer is the simplest tool to quantitatively measure the relative ventrodorsal globe position [23]. It is the current gold standard despite the known limitations, such as the asymmetry of lateral orbital rims, the compression of soft tissue, and the lack of a uniform technique [23,24]. There are no readily available reproducible tools for measuring the relative craniocaudal globe position [25]. It is assessed by the Hirschberg test, which evaluates the light reflex centered on each pupil to reveal vertical asymmetry [26]. Alternative methods have been proposed to quantify globe position differences based on imaging, but these have not been broadly implemented [23,27].

palsy. In a trauma setting, orthoptic measurements may be challenging to perform due to logistics, limited mobility of the patient, or considerable periorbital swelling. Absolute restrictions, as seen in trapdoor fractures, need to be treated shortly after the trauma. Ocular motility deficits with different etiology may improve or resolve over time, and surgery might not be indicated. In these cases, it is advisable to perform several orthoptic assessments over time to monitor spontaneous improvement. Moreover, the orthoptist One of the difficulties in clinical decision-making is to address the most common complaint in orbital fractures: diplopia. It is challenging to find the actual cause of diplopia; in most cases, it is caused by a restriction of ocular motility. Ocular motility can be disturbed by impingement or the entrapment of the ocular muscles and surrounding soft tissue, but also by muscle edema, muscle injury, hemorrhage, emphysema, or motor nerve palsy. In a trauma setting, orthoptic measurements may be challenging to perform due to logistics, limited mobility of the patient, or considerable periorbital swelling. Absolute restrictions, as seen in trapdoor fractures, need to be treated shortly after the trauma. Ocular motility deficits with different etiology may improve or resolve over time, and surgery might not be indicated. In these cases, it is advisable to perform several orthoptic assessments over time to monitor spontaneous improvement. Moreover, the orthoptist may be able to differentiate between possible causes of double vision through repeated measurements [9,28].

Computed tomography (CT) is the modality of choice for radiographic evaluation because of the superior visualization of bony structures. The size and extent of the fracture may be estimated or measured in the coronal, sagittal, or axial plane. Considering that the bone is paper-thin in certain areas, a maximum slice thickness of 1.0 mm is essential for evaluation. In individual cases, the evaluation of soft-tissue changes may become important. Shape alterations of the inferior rectus muscle have been reported to affect delayed or postoperative enophthalmos [29–31] and may affect treatment decisions [10]. In addition, herniation of the orbital soft tissue might be an indication for surgical reconstruction. Magnetic resonance imaging (MRI) provides superior soft-tissue contrast compared to CT and is more sensitive for identifying extraocular muscle or periorbital fat entrapment [32,33]. Nevertheless, the acquisition of an MRI is not part of the standard imaging protocol for orbital trauma [13]. This may change in the future, considering that all subsequent treatment steps benefit from optimizing the information collection in the diagnostics phase.

#### *2.2. Advanced Diagnostics*

Advanced diagnostics aim to maximize the information extracted from the available image data. For this purpose, the CT scan is imported into the virtual surgical planning software. The CT scan is subdivided into voxels (3D pixels), each with a grayscale value corresponding to the X-ray absorption within that volume. These voxels may be segmented (grouped) based on the tissue type or anatomical structure they belong to. Anatomical structures of interest in orbital trauma are the orbit, orbital cavity, and possibly surrounding bony structures such as the zygomatic complex. The segmentation is visualized as an overlay in the multi-planar view and as a 3D model. Additional information may be collected through quantification (e.g., volume measurement), or manipulation (e.g., mirroring) of the segmented anatomy (illustrated in Figure 2). The unaffected contralateral orbit and orbital cavity can provide a reference for the affected orbit in unilateral fractures, which provides insight into the extent of the fracture and displacement of orbital walls or surrounding bony structures. The volume of the affected orbit can be compared to the unaffected healthy side to determine the relative volume change, since it has been proven that the orbits are highly symmetrical [33]. These volumetric changes can be incorporated into the treatment plan [34]. Information may also be extracted from multiple image sets. Image fusion allows aligning multiple datasets of the same modality over time or image sets from different modalities. The image sets can be simultaneously visualized and evaluated after image fusion. The segmentation process can also be based on information from multiple fused modalities. *J. Pers. Med.* **2022**, *12*, 1366 5 of 17

**Figure 2.** Advanced diagnostics for two cases. I: Solitary orbital reconstruction (**A**–**D**). (**A**) Visualization of the 3D bone surface model. (**B**) Segmentation of the unaffected orbit. (**C**) Mirroring of the segmented orbit to the affected, contralateral side. (**D**) Visualization of additional structures, such as the globe and eye muscles. II: Zygomatic complex fracture (**E**–**H**). (**E**) Visualization of the 3D bone surface model. (**F**) Segmentation of the unaffected side. (**G**) Mirroring of the segmentation to the affected, contralateral side. (**H**) Visualization of the zygomatic complex displacement. *2.3. Virtual Surgical Planning*  **Figure 2.** Advanced diagnostics for two cases. I: Solitary orbital reconstruction (**A**–**D**). (**A**) Visualization of the 3D bone surface model. (**B**) Segmentation of the unaffected orbit. (**C**) Mirroring of the segmented orbit to the affected, contralateral side. (**D**) Visualization of additional structures, such as the globe and eye muscles. II: Zygomatic complex fracture (**E**–**H**). (**E**) Visualization of the 3D bone surface model. (**F**) Segmentation of the unaffected side. (**G**) Mirroring of the segmentation to the affected, contralateral side. (**H**) Visualization of the zygomatic complex displacement.

dorsal ledge and fixation possibility on the infraorbital rim.

ing time and extent of manipulation inside the orbit during surgery [36].

Virtual surgical planning (VSP) is a simulation of the actual surgery on the imaging data [35]. It is based on information gathered in the previous treatment stages. The exact content of the virtual surgical planning depends on the type of implant. If a flat mesh plate

In the preformed implant setting, the stereolithographic model (STL) of a preformed implant is imported into the planning environment to perform a virtual reconstruction of the affected orbit. The implant's fitting potential is evaluated and its optimal position for an accurate reconstruction of the pretraumatized anatomy is simulated. The potential of VSP in the preformed implant setting is thus highly dependent on the willingness of the implant manufacturers to provide STLs of their preformed implants. In modern planning software, the implant may be automatically aligned to another virtual model, for instance, the mirrored orbit. Manual adjustments could be necessary to prevent interferences with the bone and ensure the orbital defect is covered, with adequate implant support on the

The implant may be virtually trimmed to simulate the cutting of medial or posterior parts of the implant. The surgery can be simulated multiple times in the virtual surgical planning, with different implant types and sizes (Figure 3). This enables comparison between preformed implant options and substantiated decision making before surgery. The number of try-ins in virtual planning is limitless without consequences for the patient, in contrast to try-ins during actual surgery. Establishing the optimal position in virtual planning provides the surgeon with intraoperative feedback, which could reduce the operat-

3D printed to serve as individual bending template(s) for molding the flat mesh.

#### *2.3. Virtual Surgical Planning*

Virtual surgical planning (VSP) is a simulation of the actual surgery on the imaging data [35]. It is based on information gathered in the previous treatment stages. The exact content of the virtual surgical planning depends on the type of implant. If a flat mesh plate is used, the virtual models of the mirrored orbit and affected orbit can be exported and 3D printed to serve as individual bending template(s) for molding the flat mesh.

In the preformed implant setting, the stereolithographic model (STL) of a preformed implant is imported into the planning environment to perform a virtual reconstruction of the affected orbit. The implant's fitting potential is evaluated and its optimal position for an accurate reconstruction of the pretraumatized anatomy is simulated. The potential of VSP in the preformed implant setting is thus highly dependent on the willingness of the implant manufacturers to provide STLs of their preformed implants. In modern planning software, the implant may be automatically aligned to another virtual model, for instance, the mirrored orbit. Manual adjustments could be necessary to prevent interferences with the bone and ensure the orbital defect is covered, with adequate implant support on the dorsal ledge and fixation possibility on the infraorbital rim.

The implant may be virtually trimmed to simulate the cutting of medial or posterior parts of the implant. The surgery can be simulated multiple times in the virtual surgical planning, with different implant types and sizes (Figure 3). This enables comparison between preformed implant options and substantiated decision making before surgery. The number of try-ins in virtual planning is limitless without consequences for the patient, in contrast to try-ins during actual surgery. Establishing the optimal position in virtual planning provides the surgeon with intraoperative feedback, which could reduce the operating time and extent of manipulation inside the orbit during surgery [36]. *J. Pers. Med.* **2022**, *12*, 1366 6 of 17

**Figure 3.** Virtual fitting of different preformed implants in a solitary orbital fracture. Three-dimensional models of the preformed implants of KLS Martin (**A**), Synthes (**B**) and Stryker (**C**) are visualized with potential cutting lines (black lines) in the first column. The implants are virtually positioned (red contour) and the fit is evaluated in the coronal, sagittal, and axial slices. Important considerations are adequate support (on the posterior ledge, on the medial wall and on the infraorbital rim) and a lack of interference with bone*. 2.4. Patient-Specific Implant Design*  **Figure 3.** Virtual fitting of different preformed implants in a solitary orbital fracture. Threedimensional models of the preformed implants of KLS Martin (**A**), Synthes (**B**) and Stryker (**C**) are visualized with potential cutting lines (black lines) in the first column. The implants are virtually positioned (red contour) and the fit is evaluated in the coronal, sagittal, and axial slices. Important considerations are adequate support (on the posterior ledge, on the medial wall and on the infraorbital rim) and a lack of interference with bone.

Reconstructing the orbit with a PSI is the ultimate step of individualization for orbital reconstruction. A PSI is virtually modelled from scratch, using information from the (advanced) diagnostics phase and exported virtual models. In dedicated design software, a

surgical planning to improve the fit, but the design of the prototype is adjusted and the novel prototype is reimported. Even though the design of a PSI is not set in stone by protocols, various design considerations have been described in the literature. An overview of options is summarized in Table 1. This overview is not comprehensive, and novel de-

0.3 mm [22,37,38]

Present [37,39,40]

Absent [38,41,42]

Porous [42,45,46]

ledge

Horizontal [22,37,40,43]

Three points [22] Infraorbital rim, medial wall, posterior

Squares [38,39,41,44]

[38] Anteromedial, anterolateral, posterior

sign options are regularly introduced in the literature.

Thickness Positioning,

Atraumatic cord Positioning,

Grid Clinical symp-

Support Stability, accu-

stability

toms

racy

stability

**Design Consideration Effect on Options References Notes** 

**Table 1.** List of different design considerations as reported in the literature.
