Journal Logo


Development of a New Large-Animal Model for Composite Face and Whole-Eye Transplantation: A Novel Application for Anatomical Mapping Using Indocyanine Green and Liquid Latex

Bravo, Miguel G. M.D.; Granoff, Melisa D. B.A.; Johnson, Anna Rose M.P.H.; Lee, Bernard T. M.D., M.B.A., M.P.H.

Author Information
Plastic and Reconstructive Surgery: January 2020 - Volume 145 - Issue 1 - p 67e-75e
doi: 10.1097/PRS.0000000000006322
  • Free
  • SDC
  • Associated Video


It has only recently been possible to correct severe deformities of the face with reconstructive facial transplantation using vascularized composite allotransplants. This innovation catalyzed a paradigm shift for reconstruction of complex defects, including the face and orbital region.1 As of 2018, 404 partial or full-face transplants have been performed in 43 patients.2–5 However, no facial transplants to date have included vascularized composite allotransplantation of the globe. Of patients who have undergone facial transplantation, five have been blind in at least one eye.6–10 Significant vision loss caused by trauma is often accompanied by severe facial injury, with 58 percent of U.S. wartime globe injuries including severe facial injury. In patients who suffer facial trauma that affects the extraocular muscles or eyelids even without globe trauma, vision loss and blindness can still occur because of corneal exposure, resulting in ulceration.11 However, despite technological and surgical advancements, an approach to allow blind patients to recover this functional deficit has not yet emerged. Therefore, there is great therapeutic potential for the development of a technique for full facial transplant with whole-eye transplantation.

Although vascularized composite allotransplantation of the eye and orbit has been attempted in multiple mammalian and cold-blooded models, concerns regarding the feasibility of restoring vision remain. A review by Ellenberg et al. reported that of 173 mammalian eye transplants recorded in the literature, only two described recovery of visual function, one of which was later called into question.12 More recent studies have established the possibility of axon recovery of the optic nerve in mice, with regeneration of retinal ganglion cells imaged along the entire length of the visual pathway into the lateral geniculate nucleus and superior colliculus.13 Although these results show great promise, they are not clinically applicable in their current form because of the use of a genetically modified mouse (phosphatase and tensin homolog knockout). Therefore, further studies in mammals more similar to humans using clinically applicable methods are needed.

Selection of an appropriate animal model for face and whole-eye transplantation that recapitulates human anatomy has proved challenging. Identification of a structurally similar model would provide opportunity for further anatomical study and procedural refinement for eventual human translation. The porcine model has been used for research and medical training because of its anatomical and physiologic similarities to humans, making it an ideal model with which to study surgical technique, nerve regeneration, vision restoration, and immunosuppression.14,15 The porcine orbit is larger than the human orbit, and lacks a lateral wall, which allows for a broader field of dissection, optimal for experimenting with different approaches to the orbital contents before attempts in humans.16 In addition, the porcine eye is largely perfused by the external ophthalmic artery extending off the external carotid. This not only facilitates dissection because of ease of access of the external carotid artery, but also allows for dissection of a variety of face and eye flaps that are supplied by a common pedicle (Fig. 1).16 Part of the difficulty of establishing recovery of visual function in the animal model is the objective assessment of sight. Gizewski et al. proposed a novel method for assessing visual cortex function in swine using functional magnetic resonance imaging with light flash stimulus.17 This further establishes the swine as an ideal large-animal model for this technique. Before any surgical technique can be attempted in a living animal, however, the approach must be tested for feasibility and reproducibility in a cadaveric model.

Fig. 1.
Fig. 1.:
Vasculature of the porcine face.

In a human cadaveric study of composite eyeball-periorbital transplantation flaps, a number of anatomical difficulties were identified. For example, the inclusion of the bony orbit would require an intracranial approach, but excluding the bony orbit would complicate restoration of the lacrimal system and muscle/tendon attachments.18 A proposed solution to this anatomical limitation in humans is an endoscopic approach using a combined transorbital, endonasal, and transcranial approach and exenteration as described by Davidson et al.19 These and further anatomical concerns require further research in animal and cadaveric models before translation into patients can be attempted.

The hemifacial flap has been previously described by our laboratory and others in a swine model.20–22 It has also been performed in a small-animal model that assessed functional outcomes and sensation after composite tissue allotransplantation.23 In this study, we aim to modify the previously described hemifacial swine flap to include the orbit because the vascular pedicle supplies both of these structures.20 In other cadaveric models, latex has been used to assist with identification of cerebral vasculature for resident teaching in specialties including neurosurgery.24 We aimed to expand on this by adding indocyanine green to facilitate assessment of flap perfusion using near-infrared Fluorescence-Assisted Resection and Exploration (FLARE; Curadel, Natick, Mass.). Use of indocyanine green has been described with success in this model.20 The combination of the latex mixture and imaging technology will provide the opportunity to delineate the vascular territory of three different hemifacial flaps. In addition, we will attempt to provide a preliminary assessment of perfusion patterns of the orbit and face in each flap design. The aim of this study is to describe three different techniques for harvesting composite face and whole-eye transplantation flaps using cadaveric Yorkshire pigs and to establish a novel approach for anatomical identification and visualization of flap perfusion patterns based on fresh porcine cadaver specimens and colored latex and indocyanine green injection.


Protocol for Pig Dissection

Six surgical dissections of a composite face and whole-eye transplantation were performed using Yorkshire pig cadavers (Parsons EM & Sons, Inc., Hadley, Mass.) by a single surgeon between July of 2018 and September of 2018. Institutional animal care and use committee approval was not required for this study, as all experiments were performed on discarded carcasses that were used in nonsurvival studies by other research groups. No pig cadaver had a history of head or craniofacial trauma. All pigs weighed between 40 and 50 kg. Tissue harvest occurred immediately after the pigs were euthanized. To facilitate our dissection, the pig head was decapitated and both external carotid arteries were ligated. The internal and external jugular veins were clamped. Both carotid arteries were then cannulated using two 16-gauge angiocatheters. The carotids were flushed using 200 cc of 0.9% saline and 1500 IU of heparin after release of the venous clamps. Then, 100 cc of dyed latex (FX Latex, Beverly, Mass.) was injected into the external carotid arteries to provide distinct visualization of pertinent vasculature during dissection.

The liquid rubber latex was mixed with two teaspoons of colored powder pigment (Stardust Micas; KSSMC LLC, Titusville, N.J.) and 0.3 cc of indocyanine green (Akom, Inc., Lake Forest, Ill.). Indocyanine green facilitated visualization of the vascular tree around the eye and confirmed globe and retinal perfusion. Images were acquired using the near-infrared fluorescence imaging system (Fluorescence-Assisted Resection and Exploration).

Surgical Technique

All surgical dissections were performed by the first author (M.G.B.). We developed and performed three approaches to harvest the globe, periorbital tissue, and orbit attached to the facial flap and perfused by the external carotid pedicle. Our surgical approach to harvesting the orbit mirrors that described by Kyllar et al.16 The adjacent orbital skin island was marked as follows: medially 1 cm rostral to the lacrimal duct, 2 cm below the inferior orbital rim of the zygomatic bone, and 2 cm cephalad to the superior orbital rim of the frontal bone (Fig. 2). The lateral borders differed according to flap design.

Fig. 2.
Fig. 2.:
Eye skin island. Borders of the eye skin island: medial, 1 cm rostral to the lacrimal duct; inferior, 2 cm below the inferior orbital rim (zygomatic bone); lateral, equidistant between the ear and eye; and superior, 2 cm above the superior orbital rim (frontal bone).

Chimeric Whole-Eye and Hemifacial Flap

Our chimeric flap was designed using two skin islands (Fig. 3). The first was based on the anatomical boundaries described previously in the hemifacial swine flap.20 The second skin island can be seen in detail in Figure 2. The lateral landmark for the eye skin island was determined from an anatomical point equidistant from the anterior ear and lateral eye. The eye, orbital tissue, and face were harvested as part of the chimera. The external carotid artery and external jugular vein were the single vascular pedicle for this composite flap.

Fig. 3.
Fig. 3.:
Chimeric whole-eye and hemifacial flap: a, external carotid artery; b, external ophthalmic artery and vein; c, temporal artery; d, frontal bone; e, optic nerve; f, zygomatic bone.

After harvesting the hemifacial flap, further exposure of the maxillary artery was achieved by performing a partial mandibulectomy. After elevating the masseter muscle off the mandible, its periosteum was stripped in preparation for the osteotomy. In addition, periosteal stripping of the posterior aspect of the mandible was performed to minimize inadvertent damage to the maxillary artery. The osteotomy was performed along the angle of the mandible, and both coronoid and condylar processes were released from their insertion point using a periosteal elevator while holding the mandibular ramus with a bone clamp. This allowed further visualization of the maxillary artery and facilitated a view of its external ophthalmic branch, which entered the inferior orbital wall (Fig. 4).

Fig. 4.
Fig. 4.:
View after resection of the mandible ramus: a, external carotid artery; b, maxillary artery; c, external ophthalmic artery; d, mandible; e, zygoma; f, temporomandibular joint

The posterior aspect of the orbit could then be accessed by means of its lateral edge. The ligamentum orbitale that extends between the zygomatic process of the frontal bone and the frontal process of the zygomatic bone was excised. After retracting the temporalis muscle, the length of the extraocular muscles and optic nerve were visible. The optic nerve was traced and severed at the point where it exits the optic canal. The external ophthalmic artery and vein were then tagged to avoid damage during subsequent osteotomies. Osteotomies of the frontal bone, os lacrimale, and zygoma were then performed. The frontal osteotomy was performed 2 cm ventral to the supraorbital rim, including the foramen supraorbitale. The os lacrimale osteotomy was performed 1 cm anterior to the foramen lacrimale accessing the sinus lacrimalis and labyrinthus ethmoidalis. The zygomatic osteotomy was performed rostral to the processus zygomaticus ossis temporalis.

Monobloc Whole-Eye and Hemifacial Flap

In the monobloc flap, the hemifacial flap, eyeball, orbit, and periorbital tissue were harvested en bloc (Fig. 5). The lateral margin of the orbital portion of the flap was the posterior edge of the ear, based off the previously described hemifacial flap.

Fig. 5.
Fig. 5.:
Monobloc whole-eye and hemifacial flap: a, frontal bone; b, optic nerve; c, lacrimal gland; d, zygomatic bone.

Access to the condylar and coronoid processes for partial mandibulectomy was then developed inferiorly. A partial mandibulectomy, similar to the above, was then performed. However, during the inferior approach, careful retraction of the hemifacial flap was required. It is important to avoid placement of excess tension on the pedicle during retraction. The external ophthalmic artery and vein were visualized entering the orbit following partial mandibulectomy. These vessels should be protected by a malleable retractor before the zygomatic osteotomies are performed.

In this approach, access to the posterior aspect of the orbit cannot be achieved solely by release of the ligamentum orbitale. Instead, a frontal bone osteotomy is required. The frontal bone osteotomy was performed 2 cm superior to the supraorbital rim and included the foramen supraorbitale. This provided access to the posterior edge of the orbit and exposure of the extraocular muscles and optic nerve. Special care at this point in the dissection to elevate the extraocular muscles is needed to cut the optic nerve at the anatomical point where it enters the orbit from the optic canal.

Bipedicled Whole-Eye and Facial Flap

In the bipedicled flap, both sides of the face were harvested using the margins described in the monobloc technique, with the flaps conjoined by means of the frontal bone and overlying soft tissue (Fig. 6). The superior and inferior markings of the skin island connecting the hemifacial flaps were marked using the supraorbital foramen superiorly and lacrimal duct inferiorly.

Fig. 6.
Fig. 6.:
Bipedicled whole-eye and facial flap: a, external carotid artery; b, external jugular vein; c, frontal bone; d, optic nerves; e, external ophthalmic arteries.

The inferior frontal bone osteotomy is performed in a straight line connecting the point 3 cm anterior to the lacrimal duct of each eye. This osteotomy must be performed parallel to the axis of the snout to remain parallel to the orbital floor. This is paramount, as modifying the alignment of the osteotomy can lead to orbital floor injury, including rupture and inadvertent puncture of intraorbital contents. The superior frontal bone osteotomy is then performed perpendicular to the cranium until the dura mater is reached. It should be noted that the frontal bone of the pig is especially thick and pneumatized, which could increase the difficulty of performing the osteotomy.


All three flaps were harvested using six cadaveric pigs. All flaps were based on the external carotid system. No anatomical variations regarding the location or presence of the pig orbit and external ophthalmic artery were identified.

The mixture of liquid latex and indocyanine green was injected into the external carotid artery of all six pigs. No dye leakage was noted. On gross inspection, the dyed latex mixture was observed in the external ophthalmic artery entering the orbit (Fig. 4). Fluorescence was observed using the near-infrared Fluorescence-Assisted Resection and Exploration imaging system. All flap vascular territories had consistent, well-defined perfusion patterns (Fig. 7).

Fig. 7.
Fig. 7.:
Perfusion of the external ophthalmic artery.

Following flap dissection, the eyeball was enucleated, and the anterior surface was removed to assess retinal perfusion using near-infrared imaging. [See Figure, Supplemental Digital Content 1, which shows retinal perfusion using near-infrared imaging. (Above, left) Enucleation of the eye, with visualization of the external ophthalmic artery surrounding the optic nerve. (Above, center and above, right) Near-infrared imaging illustrating indocyanine green–latex mixture. (Below) Assessment of retinal perfusion with near-infrared imaging after removal of the anterior surface of the eye,]


Trauma resulting in blindness and severe facial deformity warrants a better reconstructive approach for this complex problem that incorporates vascularized composite allotransplantation of the eye.6 Any degree of sight restoration that can be reestablished in these patients has the potential to vastly improve quality of life and function. Previous studies have accomplished functional whole-eye transplantation in cold-blooded animals.12 The only successful mammalian whole-eye transplantation with documented retinal ganglion cell regeneration was performed in a murine model using research methods inapplicable to human translation.13 In this study, we describe a novel approach for anatomical identification and visualization of flap perfusion patterns based on fresh porcine cadaver specimens and colored latex and indocyanine green injection.

Our approach provides an anatomical assessment of face and whole-eye transplantation vasculature patterns. It also provides real-time visualization using near-infrared light fluorescence imaging without the extravascular contamination often seen when indocyanine green is used in cadaveric tissue. The distribution of the material was confirmed with gross inspection of the latex and near-infrared visualization of the indocyanine green, establishing this method as a novel, useful approach to vascular visualization and testing in this model. In future studies, we hope to develop this technique further for application in a live porcine model.

The three face and whole-eye transplantation flaps described in this study illustrate novel surgical approaches for correcting severe facial deformities and accompanying bilateral or unilateral blindness in a porcine model. Benefits of the chimeric flap include improved exposure of the external ophthalmic artery arising from the maxillary artery using a secondary incision. This approach may be limited because of an increased risk of skeletonizing or inadvertently injuring the vasculature during separation of the eye and periorbital tissue from their attachments. The monobloc is an improvement on the chimeric flap because complete dissection of the ophthalmic vessels is not required, reducing their risk of damage. However, this limits exposure of orbital contents and increases the technical difficulty of the dissection. The bipedicled approach is the most clinically translatable for patients who have lost vision in both eyes accompanied by severe facial damage. We aim to better evaluate the bipedicled flap in live animals for study of optic nerve function after transplantation in future studies.

Limitations of this study remain. Although the porcine cadaver provides an excellent research model for planning and executing different surgical approaches, our results do not translate directly into humans. This study does not assess flap survival or functionality. As we transition to a live porcine model, we will be able to evaluate and further study ischemia time of the mammalian eye and its accompanying structures. In a pilot, live-animal model, we were able to illustrate perfusion to the chimeric flap and orbit using our novel indocyanine green near-infrared imaging technology. [See Video (online), which demonstrates indocyanine green imaging of the orbit in the chimeric flap.] We plan to continue to broaden our approach in a live model to be able to study eye reperfusion and ischemic time as part of the chimeric flap. Furthermore, although we were unable to address these parameters in this model, we acknowledge that assessment of ischemia time of the eye is of central importance to determine the feasibility of eye vascularized composite allotransplantation in living tissue.

{"href":"Single Video Player","role":"media-player-id","content-type":"play-in-place","position":"float","orientation":"portrait","label":"Video.","caption":"This video demonstrates indocyanine green imaging of the orbit in the chimeric flap.","object-id":[{"pub-id-type":"doi","id":""},{"pub-id-type":"other","content-type":"media-stream-id","id":"1_61bq5ron"},{"pub-id-type":"other","content-type":"media-source","id":"Kaltura"}]}

Moreover, although anatomical variations in the porcine model provide relative ease of dissection, this does not directly translate to an understanding of relevant human anatomy and variations. For example, the surgical approach to the orbit differs and requires special consideration. The thickness of the frontal, maxillary, and zygomatic bones are different in the porcine model than in the human model, requiring a different osteotomy approach. The porcine orbit is open on its posterolateral aspect, whereas the human orbit is enclosed by bone circumferentially. This facilitates easier surgical access to the porcine orbit.16 This relative ease of access is important for our study, which establishes a model that can be used to study optic nerve regeneration and perfusion in a mammalian species. Although other large mammals such as primates may be more anatomically comparable to humans, the ethical and logistical considerations of performing these transplants in primates are not compensatory for the added anatomical equivalence. Instead, our immediate follow-up studies focus on optic nerve regeneration in a porcine model using our described surgical technique, and thus will be studied as an autotransplantation. In addition, next steps from this technique include study of the human anatomy in a human cadaver using our described surgical technique and novel indocyanine green–latex vascular imaging. Furthermore, venous structures are poorly defined in a cadaveric model. We anticipate improvements in venous anatomical delineation as we transition to a live model.

This study has presented the design and surgical approach of three novel composite face and whole-eye transplantation models in a large mammal cadaver and the use of a novel indocyanine green–latex mixture that allows for superior vascular visualization in cadaveric studies. We believe this will serve as a foundation for translation into a live porcine model for further study of optic nerve regeneration in the large mammal.


The authors would like to formally acknowledge and thank the Beth Israel Deaconess Animal Research Facilities staff for assistance and support of this project.


1. Sosin M, Mundinger GS, Dorafshar AH, et al. Eyelid transplantation: Lessons from a total face transplant and the importance of blink. Plast Reconstr Surg. 2015;135:167e–175e.
2. Caterson EJ, McCarty JC. Face transplant: Status of current supporting technology to plan and perform the operation and monitor the graft in the postoperative period. J Craniofac Surg. 2018;29:820–822.
3. Lindford AJ, Mäkisalo H, Jalanko H, et al. The Helsinki approach to face transplantation. J Plast Reconstr Aesthet Surg. 2019;72:173–180.
4. Lemieux D, Bernstien J. The inside story of Canada’s first face transplant. Available at: Accessed December 12, 2018.
5. Brigham and Women’s Hospital. Mitch Hunter: Second full face transplant in US. Available at: Accessed December 12, 2018.
6. Carty MJ, Bueno EM, Lehmann LS, Pomahac B. A position paper in support of face transplantation in the blind. Plast Reconstr Surg. 2012;130:319–324.
7. Kollar B, Pomahac B. Facial restoration by transplantation. Surgeon 2018;16:245–249.
8. Lemmens GM, Poppe C, Hendrickx H, et al. Facial transplantation in a blind patient: Psychologic, marital, and family outcomes at 15 months follow-up. Psychosomatics 2015;56:362–370.
9. Gastman B, Hashem AM, Djohan R, et al. Malignant pyoderma associated with granulomatosis with polyangiitis (Wegener granulomatosis) as a unique indication for facial vascularized composite allotransplantation: Part I. Plast Reconstr Surg. 2016;137:1007e–1015e.
10. Lassus P, Lindford A, Vuola J, et al. The Helsinki face transplantation: Surgical aspects and 1-year outcome. J Plast Reconstr Aesthet Surg. 2018;71:132–139.
11. Liu K, Gu B, Chiang CA, Zhang DS, Li Q. Rescue of vision in burn patients with total eyelid loss. Burns 2012;38:269–273.
12. Ellenberg D, Shi J, Jain S, et al. Impediments to eye transplantation: Ocular viability following optic-nerve transection or enucleation. Br J Ophthalmol. 2009;93:1134–1140.
13. de Lima S, Koriyama Y, Kurimoto T, et al. Full-length axon regeneration in the adult mouse optic nerve and partial recovery of simple visual behaviors. Proc Natl Acad Sci USA. 2012;109:9149–9154.
14. Farkas LG, Munro IR, Vanderby BM. Quantitative assessment of the morphology of the pig’s head used as a model in surgical experimentation: Part 1. Methods of measurements. Can J Comp Med. 1976;40:397–403.
15. Farkas LG, James JS, Vanderby BM. Quantitative assessment of the morphology of the pig’s head used as a model in surgical experimentation part 2: Results of measurements. Can J Comp Med. 1977;41:455–459.
16. Kyllar M, Štembírek J, Danek Z, et al. A porcine model: Surgical anatomy of the orbit for maxillofacial surgery. Lab Anim. 2016;50:125–136.
17. Gizewski ER, Schanze T, Bolle I, de Greiff A, Forsting M, Laube T. Visualization of the visual cortex in minipigs using fMRI. Res Vet Sci. 2007;82:281–286.
18. Siemionow M, Bozkurt M, Zor F, et al. A new composite eyeball-periorbital transplantation model in humans: An anatomical study in preparation for eyeball transplantation. Plast Reconstr Surg. 2018;141:1011–1018.
19. Davidson EH, Wang EW, Yu JY, et al. Total human eye allotransplantation: Developing surgical protocols for donor and recipient procedures. Plast Reconstr Surg. 2016;138:1297–1308.
20. Nguyen JT, Ashitate Y, Buchanan IA, et al. Face transplant perfusion assessment using near-infrared fluorescence imaging. J Surg Res. 2012;177:e83–e88.
21. Kuo YR, Shih HS, Lin CC, et al. Swine hemi-facial composite tissue allotransplantation: A model to study immune rejection. J Surg Res. 2009;153:268–273.
22. Park J, Yim S, Eun SC. Experimental design for composite face transplantation. J Craniofac Surg. 2016;27:843–845.
23. Washington KM, Solari MG, Sacks JM, et al. A model for functional recovery and cortical reintegration after hemifacial composite tissue allotransplantation. Plast Reconstr Surg. 2009;123(Suppl):26S–33S.
24. Alvernia JE, Pradilla G, Mertens P, Lanzino G, Tamargo RJ. Latex injection of cadaver heads: Technical note. Neurosurgery 2010;67(Suppl Operative):362–367.

Supplemental Digital Content

Copyright © 2019 by the American Society of Plastic Surgeons