Three-dimensional computed tomographic reconstruction of the orbital fracture.
Screen shot of the computed planning of the orbital reconstruction. The red portion represents the area of the injured orbit that differs from the uninjured orbit when the mirroring function of the software is applied; this is the defect area of the orbital floor that needs to be bridged with an implant.
Stereolithographic rapid prototype repair model with a thin layer of plaster to correct small irregularities.
Pressure unit with the counterdie inserted in the base ring.
Finished custom-made orbital floor implant with extension over the infraorbital rim for optimal orientation and fixation.
Insertion of the implant via a lower lid incision. The fixation screw is visible on the lateral aspect.
Routine postoperative occipitomental radiograph.
Lieger O, Richards R, Liu M, Lloyd T. Computer-Assisted Design and Manufacture of Implants in the Late Reconstruction of Extensive Orbital Fractures. Arch Facial Plast Surg. 2010;12(3):186-191. doi:10.1001/archfacial.2010.26
Copyright 2010 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2010
To evaluate the use of computer-assisted designed and manufactured (CAD/CAM) orbital wall and floor implants for late reconstruction of extensive orbital fractures.
We performed a retrospective data review on 29 patients treated for extensive orbital fractures from January 1, 1997, through December 31, 2007, at the University College London Hospitals. The use of a CAD/CAM technique based on cross-sectional computed tomographic scans, generating an accurate stereolithographic model, enabled surgeons and technicians to plan and create the best dimension and position of the implant. Sheet titanium was then pressed to shape from a design outlined on a counterdie of the new reconstructed model.
Twenty-nine patients with late enophthalmos due to complex orbital fractures underwent successful reconstruction surgery. Enophthalmos was corrected in all patients. Diplopia was improved in 14 patients, and extraocular movement was improved in 13.
The CAD/CAM implants represent a financially viable method for secondary reconstruction of the orbit. This method enables the surgeon to plan the operation in detail, facilitates the surgical procedure, and can help to improve the outcome.
Orbital fractures are reported in more than 40% of maxillofacial injuries and therefore represent the most common trauma to the midface.1 Posttraumatic orbital deformities caused by missed diagnosis or incorrect reconstruction of the preinjury anatomy of the orbit account for serious complications, including enophthalmos, diplopia, and visual acuity disturbance. The goal of the secondary operative reconstruction is to repair the orbit in an anatomically correct order and to mobilize and free periorbital tissue to achieve optimal positioning of the globe and to improve eye movement.2 Fat atrophy was long believed to be a central problem in the cause of enophthalmos. However, different studies showed that there is no evidence for this assumption.3- 7
To date, forward movement of the globe has often been achieved by volume reduction. The criterion standard implants for this procedure have been rigid calvarial bone grafts or large titanium mesh. Precise reconstruction, especially of the retrobulbar bulge, often was impossible. Today, advances in imaging techniques and navigation systems enable the surgeon to perform preoperative planning, which allows more accurate intraoperative placement of the implants. However, even with the help of these modern tools, bulky calvarial bone grafts or free hand-bent titanium mesh do not seem to be the optimal implants in precise anatomical reconstructions of the orbit.8,9
Custom-made titanium implants created using computer-assisted design and manufacturing (CAD/CAM) are already common in reconstructive cranioplasty surgery.10- 13 With the help of this experience and knowledge, we developed custom-made orbital implants. They enabled us to achieve optimal reconstruction in areas of limited visibility and nearby important anatomical structures such as the optic nerve.
The purpose of this study was to assess the outcome of secondary orbital repair using CAD/CAM implants in all patients treated during a 10-year period.
From January 1, 1997, through December 31, 2007, CAD/CAM implants for orbital reconstruction were produced at the Department of Medical Physics and Bioengineering at the University College London Hospitals for a total of 71 patients. These operations were mainly performed in 3 different centers. Thirty-two of these patients were treated at the University College London Hospitals for unilateral posttraumatic enophthalmos with or without diplopia. A side difference of at least 2 mm, measured by Hertel exophthalmometry, was considered to be enophthalmos.6,14 The medical records of these patients were scrutinized. Three patients were excluded from the study because of insufficient follow-up. Twenty-nine patients were examined preoperatively and postoperatively by ophthalmologists and maxillofacial surgeons (T.L.). The extent of the fracture was assessed with the help of preoperative computed tomographic (CT) scans and classified according to Jaquiéry et al15 (Table 1). The extent of enophthalmos was quantified with a Hertel exophthalmometer. Diplopia was determined with the Harms wall (an orthoptic method for analyzing the field of binocular vision), and ocular motility was evaluated by assessing 8 fields of gaze. In all patients, standard follow-up at 2 and 4 weeks and 3 and 6 months was performed by the maxillofacial surgeon (T.L.). Routine postoperative ophthalmological follow-up was performed between 5 and 7 months. For the present analysis, the most recent orthoptic and ophthalmological examination results were used.
To generate a 3-dimensional (3D) image, a thin-section CT scan (≤1 mm) of the patients' orbits and periorbital region is obtained. The data are then transferred by local network to the Medical Graphics and Imaging workstation at University College London. A computer 3D model is reconstructed via in-house–developed 3D imaging software (Figure 1).16 The imaging software allows the user to assess the extent of the defect and make measurements to determine suitable attachment sites. By selecting the appropriate Hounsfield threshold level, the defective orbit floor can be examined closely. The orbital repair is achieved by mirroring the good orbit of the patient and aligning this with intact parts of the damaged orbit (Figure 2). Screw-fixing hole locations are determined by the thickness of the bones around the defect. Stereolithography files of the repaired structure and the damaged structure are exported from the software.
Two stereolithographic models, an untouched defect model and a repair model, are built using a rapid prototype machine. With older CT scans, the surface of the repair model may be slightly terraced because of the finite thickness of each CT section. To correct any irregularity, the surface of the reconstructed orbit may be smoothened with an ultrathin layer of plaster (Figure 3).
The plate is normally designed to have a small flange that fits to the contour of the middle of the infraorbital margin and extends far enough posteriorly to rest on intact bone. A notch is put on the flange, which lines up with the infraorbital foramen. This helps to achieve correct placement of the implant. The implant is secured anteriorly into the orbital rim by means of screws through holes in the flange. The required outline is defined on the repair model, which is transferred to a counterdie that is then inserted into the base ring of the pressure unit (Figure 4). A 25-cm flat, round sheet of commercially pure annealed titanium, 0.5 mm thick (American Society for Testing and Materials grade 1; Imperial Metal Industries alloy No. 115), is pressure formed between the upper and lower chambers of the unit. The plate is then trimmed to remove excess metal, and the holes are drilled and the notch is cut appropriately. Finally, the plate is cleaned, polished, etched, and anodized (Figure 5).
All patients received perioperative prophylactic antibiotic therapy. Surgery was performed with the patient under general anesthesia. Subciliary or lower eyelid/midtarsal incisions were used to expose the infraorbital rim and the orbital floor. Subperiosteal dissection was used to identify stable bony structures around the defect. Subsequently, the sterilized implant was inserted and screwed to the bone using the predrilled holes (Figure 6). Postoperative occipitomental (Waters view) radiographs were obtained routinely (Figure 7).
This study was a retrospective analysis of 29 patients (26 male; average age, 33.4 [range, 16.2-56.3] years) with posttraumatic enophthalmos. The average interval between trauma and surgical enophthalmos correction was 1.6 years (range, 2 months to 4.1 years). Pure blow-out fractures were recorded in 16 patients. Ten patients sustained associated zygomaticomaxillary fractures and 3 had orbitonasal-ethmoidal fractures. Primary orbital revision was performed in 3 patients only, with the use of polydioxanone implants. All 3 patients had sustained a zygomaticomaxillary fracture and showed delayed postoperative enophthalmos. In all other patients, no primary orbital reconstruction was performed. The defect size was classified according to Jaquiéry et al15 using the preoperative CT scans. According to that classification method, 20 patients sustained a class III defect and 9 patients sustained a class IV defect.
Preoperatively impaired ocular mobility and diplopia were seen in 24 patients (Table 2). Three patients had normal binocular vision, and 2 patients were blind on the affected side. Hertel exophthalmometer assessment showed a mean preoperative difference of 3.1 mm (range, 2.0-7.0 mm). Routine postoperative ophthalmological follow-up was performed between 5 and 7 months after surgery. The mean follow-up was 14 (range, 5.5-29.0) months. Diplopia was cured in 5 patients, with improvement in 9 others (Table 3); ocular motility was normalized in 4 patients and improved in 9. In 1 patient, diplopia resolved although the motility of the eye globe remained unchanged. No worsening of double vision or motility was recorded.
The last Hertel assessment showed an average improvement of 2.9 mm (range, 1.5-5.0 mm) with a postoperative mean difference of 0.2 mm (range, −0.5-2.0 mm). No difference in the Hertel assessment was found in 17 patients. In 12 patients, an important improvement was achieved. Only 2 of the patients showed a difference of 2 mm, which by definition is still classified as enophthalmos. In 1 of these patients, the postoperative results at 7 months showed an excellent result but worsened as a result of posttraumatic globe shrinking. Clinical follow-up examination at 2 and 4 weeks and 3 and 6 months by the maxillofacial surgeon showed good wound healing in 28 patients, with no signs of infection. One patient had an infection in the region of the implant, which was successfully treated with antibiotics. None of the implants had to be removed. No additional orbital correction surgery was performed.
Posttraumatic enophthalmos occurs when no primary reconstruction of orbital injuries was performed or when insufficient outcome was obtained after primary reconstruction caused by poor mesh placement or contouring and resorption or displacement of bone grafts. In the treatment of enophthalmos and persisting diplopia, 3D orbital reconstruction remains the primary goal. It is generally agreed that the main reason for enophthalmos is an increase in volume of the posterior segment of the orbit and changes in the deep orbital cone area.7,17 Several methods and implant materials have been described for the reconstruction of bony orbital defects.8,18- 20 Depending on the size of the defect, various materials are preferred. Some authors suggest using synthetic absorbable implants such as polyglactin and polydioxanone sheets to cover small defects.21,22 In larger orbital defects, however, the stability of these materials seems insufficient. In these cases titanium mesh, autogenous bone grafts, and hydroxyapatite blocks are widely used.23
In secondary orbital reconstruction, autogenous calvarial bone grafts are still accepted as the criterion standard.24,25 However, there are important disadvantages of these grafts, such as minimal moldability, donor site defects, and potential resorption over time.26- 30 Titanium is a widely used allogenic implant material in maxillofacial reconstruction because of its availability, excellent biocompatibility, and osseointegration. It has proved to be ideal to cover larger defects in the repair of enophthalmos.31,32 Nevertheless, it is a great challenge for the surgeon to restore the correct 3D shape of the orbit by bending the implant. Especially in secondary repair, anatomical landmarks are often lacking because of extensive defects or bone remodeling. Modern preformed titanium mesh may facilitate the bending process but often leaves the surgeon with a feeling of insecurity after insertion. Unless a navigation system is used, it is very difficult to verify the correct anatomical 3D reconstruction. Especially when deep structures such as the retrobulbar bulge, the deep orbital cone, or the conjunction between floor and medial orbital wall in the posterior segments need to be reconstructed, every surgeon is pushed to the limit of his or her skills.17 In addition, intraoperative bending and repetitive insertion and removal of the mesh for orbital reconstruction may cause damage to the orbital soft tissue.
In all 29 patients in this study, enophthalmos was treated successfully. The last Hertel assessment showed an average improvement of 2.9 mm. Diplopia was improved or cured in 14 of 24 patients. Unfortunately, few studies on this topic exist in the English literature. The largest group of patients undergoing late repair of enophthalmos, to our knowledge, was assessed by Fan et al.33 In that study, 17 patients were treated using osteotomies and polyethylene implants. Another smaller study was performed by Klein and Glatzer,8 with custom-made glass-bioceramic implants used in 8 patients. However, only Fan et al33 published a detailed list of the outcome measurements. When we compared the results, the preoperative degree of enophthalmos was more pronounced in the Chinese group (4.8 mm) than in our study (3.1 mm). Fan et al33 measured mean preoperative to postoperative changes of 4.6 mm, whereas we found a difference of 2.9 mm. The results in the follow-up Hertel assessment between the 2 previously published groups are similar. However, the Chinese patients showed a much better outcome with regard to diplopia. Fifteen of those 17 patients (88%) had complete correction or improvement of the impaired double vision compared with 14 of 24 (58%) in our study. An explanation for this important discrepancy may be the difference in time delay between trauma and operation. In the Chinese study the mean delay was 2 months, whereas in our study it was 14 months. Finally, it remains unclear to what extent the patients in both studies sustained any injuries to the orbital content, particularly the muscles.
Limitations of custom-made orbital implants may be their cost and production time. One implant costs approximately £1000 ($1640, as of August 2009). At our center and at Cavendish Implants Ltd, this cost covers the planning, the models, and the production of a single custom-made implant. In our experience, the use of preformed implants, however, drastically reduces the operation time and therefore the costs because fewer manipulations of the soft tissue and less trial fitting into the orbital cavity are required.
Secondary reconstruction of orbital defects is still very challenging. This study shows that CAD/CAM orbital implants can be used successfully in the late repair of extensive orbital fractures. Although the costs may still be rather high, the use of custom-made implants will become more and more important because of their undeniable advantages over other methods.
Correspondence: Olivier Lieger, MD, DMD, Department of Cranio-Maxillofacial Surgery, Cantonal Hospital Lucerne, 6000 Lucerne 16, Switzerland (firstname.lastname@example.org).
Accepted for Publication: December 4, 2009.
Author Contributions:Study concept and design: Lieger, Richards, Liu, and Lloyd. Acquisition of data: Lieger and Lloyd. Analysis and interpretation of data: Lieger and Lloyd. Drafting of the manuscript: Lieger and Lloyd. Critical revision of the manuscript for important intellectual content: Richards, Liu, and Lloyd. Statistical analysis: Lieger and Lloyd. Obtained funding: Lieger and Lloyd. Administrative, technical, and material support: Lieger, Richards, Liu, and Lloyd. Study supervision: Lloyd.
Financial Disclosure: None reported.