Maxillary reconstruction is a complex part of head and neck surgery.
Considering the defects using the Brown classification and using the concept of midfacial buttresses helps guide surgeons to choosing the appropriate bony reconstruction.
A variety of techniques using 3-dimensional computer-assisted planning can be used ranging from printing biomodels, to 3-dimensional printing models, cutting guides, plates, and fixation guides.
Maxillary reconstruction is a challenging area of head and neck reconstructive surgery. Since the advent of microvascular reconstruction, the surgeon has a large armamentarium of options to consider after ablative surgery to the maxilla and midface. With the development of computer-assisted surgical planning, the reconstructive surgeon can now accurately design and plan bony reconstruction of the midface to within millimeter accuracy, and this has changed the way we approach this area. The maxillectomy defect can produce a complex defect often involving a variety of structures, including the tooth bearing alveolus, the palate, paranasal sinuses, nasal cavity, and orbital cavity. Several bones are often involved, including the maxilla, the palatine bone, the ethmoid bone, the zygomatic bone, and less frequently the nasal bones. Loss of these anatomic structures leads to complex functional and aesthetic consequences and, when considering reconstruction in these patients, the surgeon should determine their goals for each individual patient, because these will help to choose the ideal reconstructive option. These goals can include closure of an oronasal communication, achieving midfacial soft tissue support and symmetry, avoiding velopharyngeal insufficiency, maintaining eye position, dental rehabilitation, and bony support for a facial prosthesis.
In this article, we present an overview of the use of 3-D computer-assisted surgical planning to achieve accurate reconstructions of the complex maxillary defect. We briefly discuss the Brown classification system and use it to guide the reader through the various reconstructive options for the complex maxillary defect.
Three-dimensional computer-assisted surgical planning and manufacturing
Significant progress into the development of 3-dimensional (3D) printing and computer-aided design and manufacture has led to a rapid growth in virtual surgical planning options. The main choices for any hospital or department are whether to purchase a 3D printer and the planning software and design everything in house or to outsource to a company with a proprietary workflow. The extent of virtual planning for every case can range from printing of a stereolithographic model alone that allows a 3D analysis of the case and potential for prebending plates, to completely custom planning with 3D printing of models, guides, and plates. In our unit, as an additional workflow of our maxillofacial prosthetics laboratory, 3D printers have been purchased, and Materialize™ software used for surgical planning. The advantages of an in-house system includes rapid prototyping and planning to be performed, allows the surgeons to easily communicate with the technicians and biomedical engineers to make small modifications when necessary, less expensive for the institution, and requires less time for planning and manufacturing ( Table 1 ). In our unit, the planning is all completed virtually, 3D biomodels and cutting guides are printed, and the plates are prebent on the biomodels and sterilized. Prebending reconstruction plates allows for a more cost-effective option than printing in hospital grade titanium. Several cases will be shown to highlight how 3D-assisted surgical planning has revolutionized our practice and improved surgical outcomes in maxillary reconstruction.
|Stage||Action||Average time required||Additional steps for proprietary planning|
|3D planning (PC workup||
|3D printing and postprocessing of model||
|Laboratory process of plate||
|Completion||–||27 h||Variable delivery between 2 and 4 wk|
Classification of midface defects
Several classifications exist for midfacial defects; however, the Brown classification will be used in this article. The Brown classification was described in 2010, provides a framework for considering potential reconstructive options for maxillary and midface defects, and will be used in this report ( Fig. 1 ). The classification consists of 6 vertical and 4 horizontal maxillectomy defect patterns. The vertical classification involves (1) maxillectomy not causing an oronasal fistula, (2) not involving the orbit, (3) involving the orbital floor and potentially the orbital adnexae but the eye remains in situ, (4) with orbital enucleation or exenteration, (5) orbitomaxillary defect (ie, without involving the alveolus), and (6) nasomaxillary defect. The horizontal classification involves (a) a palatal defect only, not involving the dental alveolus, (b) less than or one-half unilateral, (c) less than or equal to one-half bilateral transverse anterior, and (d) greater than one-half maxillectomy. In addition to using this classification, it is also valuable to consider the bony buttresses of the midface. The key bony buttresses to consider in midface reconstruction as per Yamamoto and colleagues are the horizontal–zygomaticomaxillary (infraorbital) and vertical nasomaxillary buttress, and the oblique pterygomaxillary (zygomatic buttress) ( Fig. 2 ). Having an understanding of these buttresses helps the reconstructive surgeon to choose the ideal reconstructive option.
The Birmingham approach to midfacial defect reconstruction
In this section, we discuss our approach to complex maxillary and midfacial reconstruction and demonstrate various cases where this approach was used ( Fig. 3 ).
Class I defects
In class I and some class II defects (IIa and low-level and posterior IIb), the goal of reconstruction is closure of an oroantral or oronasal fistula and dental rehabilitation. When the patient is dentate, very good results are achieved with either prosthetic obturation or soft tissue reconstruction alone or in combination with zygomatic implants. When the class IIb defect is posterior to the canine or first premolar, there is no advantage to bony reconstruction in these cases.
Class II defects
In anterior or large class IIb, II2, or IId defects, there is significant loss of cheek and upper lip support, which requires an underlying bony framework to preserve facial form and symmetry. In these cases, our preference is the fibula free flap. The fibula is ideal in these cases because it can provide a substantial length of bone, has a long pedicle, has a reliable soft tissue skin paddle, and provides a good bony framework for dental implants. In hemimaxillectomy (class IIb/IIc) defects, the fibula can either be used to recreate the shape of the alveolar arch ( Figs. 4 and 5 ) or it can be reliably used to recreate the pterygomaxillary buttress. In class II defects, the reconstruction can often be done with a single straight strut of fibula ( Fig. 6 ), but it may need to incorporate multiple osteotomies if the anterior maxilla is too prognathic in relationship to the zygoma or if the defect crosses the midline. The fibula used in this way can often provide adequate soft tissue support for the cheek and upper lip and it provides a framework for implant rehabilitation for at least the anterior teeth. Even in a class IId defect, the fibula provides adequate bone length for complete maxillary low-level reconstruction.
An alternative option includes the thoracodorsal angular artery flap (scapula tip) for these defects. The scapula tip can be an effective alternative for maxillary reconstruction and it can be oriented to reconstruct the palatal surface ( Fig. 7 ) or the anterior maxillary wall. The use of the thoracodorsal angular artery scapular tip flap is particularly useful when there is a significant amount of facial skin loss requiring reconstruction. The bone stock of scapular tip has been criticized for consideration of dental rehabilitation; however, implant rehabilitation is still possible in scapular tip flaps, although onlay bone grafting may be required. The deep circumflex iliac artery (DCIA) is rarely used for these defects, but it can be used for a high class II defect when the height of the resection is at least 3 cm to allow for the DCIA height to adequately capture the DCIA vessels. The natural curve of the iliac creates a very good platform for implants and good facial soft tissue support for class II high defects ( Fig. 8 ).
Class III defects
With class III defects, in addition to the alveolar bone loss, there is also loss of bony support for the cheek skin, orbital floor, and rim. This defect is the most difficult to reconstruct, and failed reconstruction has a devastating consequence of midfacial soft tissue collapse and loss of orbital bony support, resulting in hypoglobus and enophthalmous. Any reconstructive option being considered should at least aim to address these key features. Historically, the majority of cases reported in the literature have been reconstructed with soft tissue flaps only, mainly the rectus abdominis or the anterolateral thigh flap. Nonvascularized bone grafts, such as iliac crest or calvarium, were often used in conjunction for orbital floor reconstruction. However, this option provides no bony replacement of the alveolus and hence dental rehabilitative options are limited. There is also a risk of bone graft loss or infection, particularly after radiotherapy. Composite microvascular reconstructive options, when possible, are ideal in these cases. In our experience, the ideal microvascular option is the DCIA flap, because its shape and contour allow for restoration of the zygomaticomaxillary, pterygomaxillary, and the nasomaxillary buttresses simultaneously. With the use of 3D computer-aided planning, the patient’s own iliac crest can be scanned and used to determine the exact part and shape of the iliac crest to harvest. We prepare custom cutting guides for the iliac crest, and this allows the surgeon to harvest the exact shape and amount of bone planned to reconstruct the defect ( Fig. 9 ). The bone is secured with either a prebent reconstruction plate or miniplates. An orbital floor titanium mesh is bent on a stereolithic model or a custom orbital floor plate is printed and secured to the DCIA with miniscrews. The thickness and shape of the bone is very good for implant insertion. The internal oblique muscle is used to close the oral communication and this epithelializes to provide an excellent soft tissue platform for future dental rehabilitation (see Fig. 9 ). Pedicle length is often a concern and, to increase its length, the area of bony harvest on the iliac crest can be moved posteriorly to a maximum of 5 cm from the anterior superior iliac spine. This maneuver allows additional length of pedicle without compromising bony vascularity.
The scapula free flap, using the lateral surface of the scapula, can also be used for class III defects. It is becoming an increasingly popular reconstruction option ( Fig. 10 ). Also, the ability to use it as a chimeric flap with a second bony segment, skin, and muscle options provides the reconstructive surgeon with the ability to reconstruct complex defects involving midfacial bone and skin. It can even be osteotomized to achieve the ideal orientation required. The scapula tip can also be used for these defects. The scapula tip has a significant advantage over the DCIA in that it has a longer pedicle; however, this advantage is only achieved if no scapular skin is required and in this situation, the teres major can be used to close the oral mucosal defect. Virtual planning can be used to isolate the scapular shape and determine the ideal orientation for reconstruction of the defect in each individual case. Where possible, having the thickest part of the scapula tip facing the alveolus is ideal for implant provision, but this strategy does shorten the pedicle length. Morphologically, the scapula tip has excellent morphologic consistency with the palate and orbital floor.
The fibula free flap has a limited role in class III defects. Although it can be used to reconstruct all 3 bony buttresses, this would involve 2 osteotomies and very sharp angles between them, leaving a central area with no bony support mid cheek. Another option is double barreling the fibula, orientating it horizontally while removing a segment in between (on the nasal side). These techniques lead to significant challenges with regard to orientation of the skin island and inset. Also, the need for multiple osteotomies compromises the pedicle length, which is a crucial concern, especially in midface defects. Hence, most authors feel that the fibula free flap is inadequate for class III and IV defects. ,
Class IV defects
When the maxillectomy is associated with orbital exenteration, the need to reconstruct the orbital floor is no longer necessary to support the eye position. The goals of reconstruction should be to reestablish midfacial projection, close the oroantral communication, and provide a platform for prosthetic replacement of the eye or obliteration of the defect. In our experience, similar to a class III defect, a DCIA with internal oblique muscle has been found to appropriately replace the bony framework of the midface ( Fig. 11 ). The inferior aspect of the reconstruction can be used to support implants for dental rehabilitation, and the superior aspect can be used to support implants for an orbital prosthesis ( Fig. 12 ). The scapula free flap can also be used here with or without the latissmus dorsi muscle, which can be used to close the oral communication and obliterate the orbital cavity if needed as well. The use of muscle to obturate the orbit allows for favorable contraction to provide space for an orbital prosthesis. This is particularly necessary when the eyelids have been resected as part of the ablation. When a significant amount of skin is resected in a class IV defect, then a scapula free flap is the ideal option, although over time, the skin sags with gravity. Secondary procedures are often necessary to remove excess skin, and maintain tone.
Class V and class VI defects
Class V defects include the maxilla, orbital rim, and occasionally orbital exenteration without involving alveolar resection. These resections often involve skin as well. In these cases, the role of a composite reconstruction is to reconstruct the zygomaticomaxillary buttress for facial support, to allow for application of an orbital floor prosthesis, or to provide a bony framework for implants to be inserted for orbital prosthetic replacement. Class VI defects are rare and often involve a significant aspect of nasal resection. When nasal bone or nasomaxillary defects are created, these often require composite reconstruction to create a framework to support the soft tissue.
Owing to the limited amount of bone needed, a composite radial forearm free flap is often used to reconstruct the infraorbital rim or the nasal bone framework. Three-dimensional computer planning allows for construction of a 3D-printed stereolithographic model of the radius, calculation of 40% circumference of the radius and preparation of a cutting guide. The shape of radius bone needed can be planned, and osteotomies are preplanned to guide the surgeon to achieve the most accurate result. The 3D-printed radius model can also be used to prebend a distal radius plate to ensure accurate adaptation at the time of surgery and reduce the risk of radius fracture ( Fig. 13 ).
When total rhinectomy defects are created and the plan is for autologous reconstruction, a composite radial forearm can be used to provide bony support and the radial skin used for nasal line, and a paramedian forehead flap can be used to replace the overlying skin for a better color match.
Complex multiclass defects
Computer-assisted 3D surgical planning has also significantly improved the way we reconstruct complex midface defects, such as combined rhinectomy and maxillectomy defects. With 3D computer-assisted planning, the planned reconstructions can be designed to facilitate reconstruction of the maxilla first, followed by provision of implants to support a dental and a nasal prosthesis ( Fig. 14 ). Care needs to be taken to ensure adequate upper lip length to create a natural appearing nasolabial fold between skin and nasal prosthesis, and to maintain lip competence. In cases where there is significant tissue loss, using computer-aided planning, midface structures can be superimposed to facilitate planning of the reconstruction. Such advances in virtual planning software have greatly improved the ability to reconstruct complex defects such as gunshot wounds with great precision ( Fig. 15 ).