Introduction

Reconstruction of the paranasal sinuses presents unique challenges restoring optimal cosmesis, structure, and function of the adjacent anatomy including the skull base, orbital floor, hard palate, and nasal cavity. There are a wide variety of pathologies that can present in the paranasal sinuses that may lead to surgical defects that necessitate reconstruction. The most common benign lesions in the sinonasal cavity are inverted papilloma and osteoma [1]. Resection of these benign tumors is generally limited; however, lesions which are recurrent or particularly extensive may require reconstruction. Malignant tumors include squamous cell carcinoma, adenoid cystic carcinoma, adenocarcinoma, sarcomas, and esthesioneuroblastomas. Malignancies of the sinuses are rare, making up less than 5% of all head and neck cancers, and often present with advanced stage at diagnosis [1]. In addition to sinonasal masses, both blunt and penetrating traumatic injuries to the face can lead to defects involving the paranasal sinuses necessitating reconstructive surgery. Many systemic diseases such as granulomatosis with polyangiitis affect the sinonasal cavities [2, 3], though reconstructive surgeries for these patients more frequently focus on addressing saddle nose deformity and septal perforation as opposed to reconstruction of the paranasal sinuses [4, 5].

Goals for reconstruction include restoring facial appearance, providing a water-tight dural seal, separation of the intracranial space from the paranasal sinuses, and supporting the orbital contents. Most reconstruction involving the skull base will require soft tissue flap reconstruction, often involving free tissue transfer, to obliterate dead space and provide a barrier against the dura. To avoid enophthalmos, exophthalmos, or dystopia, the orbital walls need to be precisely reconstructed with care to avoid injury to the optic and oculomotor nerves. When planning for reconstructive options, a patient’s comorbidities and history should be carefully reviewed, as medical conditions, prior surgery, or prior radiation may limit reconstructive options. A selection of published articles assessing paranasal sinus reconstruction is summarized in Table 1. While this is not a comprehensive list of the available data, it provides an overview of reconstructive outcomes.

Table 1 Review of select studies investigating outcomes after paranasal sinus reconstruction
Full size table

In this review, we aim to describe approaches to reconstruction of the paranasal sinuses by highlighting options which restore form and function to the anatomic structures that may be affected (Fig. 1).

Fig. 1
figure 1

Algorithm for reconstruction of paranasal sinus defects

Full size image

Maxillary Sinus

The maxillary sinus is the site of 50–70% of paranasal sinus malignancies [1]. Anatomically, the maxilla forms the structures of the orbital floor, lateral nasal wall, hard palate, and contour of the midface. Defects secondary to maxillary sinus tumors or trauma should consider restoration to the absent portions. Various classification schemas have been proposed to describe maxillary defects and guide reconstructive options [6,7,8,9,10]. These classification systems may have limited application in clinical practice, and we present a framework for reconstruction considering the anatomic subunits of the maxilla.

Hard Palate and Alveolus

Infrastructure maxillectomy involves removal of the lower portion of the maxilla including the hard palate and alveolus. These defects introduce a connection between the oral and nasal cavity, affecting a patient’s speech and swallowing. Additionally, removal of the tooth-bearing segments of the maxilla alters mastication. Non-surgical reconstructive options include an obturator or prosthesis. Many studies describe satisfactory outcomes with regard to dental restoration, speech, and swallowing when using an obturator [8, 11]. The benefit of placing an obturator or prosthesis lies in its efficiency and simplicity. In patients with malignant tumors, removal of prosthesis facilitates visual monitoring of recurrence over time. This option may be considered for patients who are not optimal candidates for extensive surgeries, as this option can shorten operative times and hospitalization. Additionally, certain pathologies cannot be accurately diagnosed on intraoperative frozen section, and a staged reconstruction allows for confirmation of negative margins prior to definitive reconstruction. A temporary obturator may also be utilized to restore function while awaiting reconstructive surgery. Patient compliance with usage, cleaning and maintaining the obturator, and routine follow-up with prosthodontist will maximize benefit.

For defects lateral to the canine, isolated soft tissue reconstruction may be performed without need for bony reconstruction [12]. Local flaps, such as temporalis muscle, temporoparietal fascia (TPF), or pectoralis muscle, may be appropriate. Pedicled buccal fat pad, facial artery musculomucosal (FAMM), and submental island flaps have also been described in the literature [13,14,15]. For larger defects, we recommend free tissue transfer such as radial forearm free flap (RFFF), anterolateral thigh (ALT), or rectus abdominis myofascial free flap. The RFFF provides a pliable skin paddle which can be harvested up to 30 × 15 cm in size [16]. The ALT flap can be versatile in its design, as a variable amount of muscle, fat, and skin can be included for the defect size. The soft tissue volume may be useful for obliteration of dead space.

Free nonvascularized bone grafting may be used for small bony defects on the alveolus in conjunction with local soft tissue reconstruction as described above. These grafts may be harvested from the anterior iliac crest, posterior iliac crest, calvarium, or tibia [17, 18]. The advantage for these grafts includes faster harvest time compared to free flaps and low donor site morbidity. However, nonvascularized bone grafting is associated with higher rates of wound breakdown in the setting of postoperative radiation [18, 19]. There may also be limitations to the length of free bone grafts. Historic studies have suggested the need for vascularized bone grafts for mandibular defects over 6–9 cm in length [20, 21]. More recently, nonvascularized bone grafts up to 22 cm have been used for mandibular reconstruction [22]. Extrapolating these studies to maxillary reconstruction may be limited, as many of these studies describe an extraoral approach without any soft tissue reconstruction. Thus, we recommend using vascularized osseous free flaps for segments longer than 3 cm and for patients who are undergoing postoperative radiation. For patients who are poor free flap candidates, local vascularized bone flap options include the parietal osteofascial flap (supplied by the superficial temporal artery) and the temporalis myo-osseus flap (supplied by the deep temporal artery).

For large bony defects requiring reconstruction, the next option is an osteocutaneous or myo-osseous free flap. These flaps can provide bone stock for future dental implants, and should be considered for defects which extend medial to or including the canine. The separation of the oral cavity from the sinonasal cavity must also be restored. The selection of free flap is influenced by the volume of soft tissue and bone needing reconstruction, pedicle length and donor vessel availability, and surgeon preference or comfort. The fibula free flap has a good bone stock and long pedicle length, but can present challenges with geometry of the skin paddle and require multiple osteotomies for complex defects. Advantages for this flap include the length of bone available, long pedicle length, and ability to use a two-team approach for simultaneous flap harvest during the resection. The scapula and subscapular system provides flexibility in the volume of bone (lateral scapula as well as the scapular tip) and soft tissue harvested for reconstruction. However, this flap cannot be harvested at the same time as the ablative surgery which can increase operative times. Deep circumflex iliac artery (DCIA) bone flap can provide significant soft tissue volume with the internal oblique muscle and a large segment of iliac crest bone [23, 24]. Interestingly, this flap has higher rates of reported donor site morbidity and flap failure when used for head and neck reconstruction compared to its use for extremity reconstruction [25]. It is not frequently used as a first-line option for head and neck reconstruction, perhaps related to surgeon preference.

Orbital Floor and Orbital Rim

Failure to repair defects of the orbital floor put patients at risk of developing hypoglobus and diplopia. Limited defects of the bony orbital floor < 50% of its width and medial to the course of the infraorbital nerve do not typically require rigid reconstruction. Similarly, isolated lamina papyracea defects do not require reconstruction, as the periorbita is able to support the orbital contents [26]. Periorbital defects may be supported by silastic sheets, TPF, fascia lata, free mucosal grafts, or acellular dermal matrix [26].

For larger defects of the orbital floor involving > 50%, rigid reconstruction should be considered. Reconstruction of the orbital floor may be completed with titanium mesh or porous polyethylene implants [27]. These implants may be bent or trimmed intraoperatively to fit the defect, or custom-made (i.e., video surgical planning) to precisely replicate a patient’s anatomy based on preoperative imaging. Similar to palate and alveolar ridge defects, small defects of the orbital rim and floor may be repaired with free nonvascularized bone grafts (anterior iliac crest, posterior iliac crest, calvarium, or tibia). For more extensive orbital defects or for patients planning to undergo postoperative radiation, vascularized free tissue transfer is utilized for orbital reconstruction. Free flap options include fibula, osteocutaneous radial forearm, DCIA, serratus anterior with rib, and scapula tip. Additionally, an osseous free flap can be used reconstruct the orbital rim or zygoma while utilizing titanium or porous polyethylene implant to address the orbital floor defect. Restoring the zygoma will ensure adequate facial projection and width. The osteocutaneous radial forearm flaps include a partial-thickness segment of the radius harvested with the pliable skin paddle. The volume of harvested bone is limited to reduce the risk of pathologic donor site fracture, and cannot support dental implantation. However, the size of bone harvested can be optimal for reconstruction of the orbital rim. If both the orbital floor and alveolus have been resected, scapula or DCIA free flaps can provide a large portion of bone to reconstruct the defect. For complex defects involving the maxilla and orbit, fibula free flap can be fashioned with multiple segments for each anatomic structure. When doing so, carefully planning must be taken with attention to maintaining sufficient pedicle length as well as ensuring sufficient length of each bone segment (at least 3 cm).

For tumor resections involving orbital exenteration or orbitomaxillectomy, the orbital cavity reconstruction is frequently reconstructed with regional or free tissue transfer. Regional flaps may include temporalis muscle, TPF, or pericranial flap [28••]. Free tissue transfer options commonly used include ALT free flap with tailoring of included vastus lateralis, fat, or skin. If additional pedicle length is required, identifying distal perforators along the anterolateral thigh and perforator dissection should be performed until sufficient length is achieved. Furthermore, use of multiple perforators allows for chimeric ALT flaps, which provides further options when addressing multiple adjacent reconstructive sites. For example, the ALT free flap can also provide multiple skin paddles for both external and mucosal defects. Rectus abdominis free flap is an alternate option. For patients planning to use an orbital prosthesis, excessive soft tissue bulk should be avoided in the reconstruction to allow for placement of the orbital prosthesis. In patients with isolated orbital exenteration, RFFF is a good option. A folded thin, pliable RFFF, “pacman flap” provides for an effective orbital reconstruction maintaining the natural concavity, facilitating ease of prosthesis placement [29•].

Ethmoid and Sphenoid Sinuses

Key considerations in reconstructing ethmoid and sphenoid sinus defects largely focus on restoration of the skull base, dural reconstruction with a water-tight seal to prevent cerebrospinal fluid (CSF) leak, and minimizing risk of meningitis. During the advent of endoscopic skull base surgery, rates of CSF leak after endoscopic endonasal approaches (EEA) to the skull base have been as high as 65% [30••]. This rate has decreased significantly as surgical techniques have improved. Depending on the location and extent of the primary lesion, tumors of the anterior skull base can be approached endoscopically or via craniofacial resection [31]. For any involvement or exposure of the intracranial space, involving neurosurgical colleagues for a multi-team approach can optimize outcomes and patient care [32].

Free mucosal grafts may be harvested from the inferior turbinate, middle turbinate, septum, or nasal floor. Vascularized tissue in the form of nasoseptal flaps based off the posterior septal artery is another option. Both nasoseptal and free mucosal grafts avoid a separate donor site, and studies have demonstrated similar outcomes between the two after uncomplicated pituitary surgery [33, 34]. A septal floor rotational flap pedicled off the ethmoidal arteries is an alternative low-morbidity local flap which can be considered if the nasoseptal flap pedicle is compromised or affected by tumor resection [35]. Defects with high-flow CSF leaks may require additional reconstructive measures [36]. Additional autologous tissue reconstructive options including fat graft, fascia lata flap, TPF flap, or pericranial flap may be used. TPF flaps can be transposed via the supraorbital or infratemporal corridors to reconstruct the skull base [37]. For tumors resected via endoscopic approach, these additional reconstructive options require additional incisions, adding surgical morbidity. However, for patients who have previously received radiation, the advantage of these options is to provide non-radiated tissue into the surgical site for reconstruction. Fibrin sealant and collagen matrix may also be used to bolster the skull base repair. Absorbable and nonabsorbable nasal packing can be used to further support the grafts during healing.

For large defects or for patients receiving postoperative radiation, free flap reconstruction should be considered for soft tissue volume. This can be combined with any local flaps for multilayer reconstruction of the dura as needed [38]. The open approach often provides direct access to many of these commonly used local options, as well as unique variations. The “helmet-visor” pericranial flap has been described, pedicled from bilateral superficial temporal arteries, as an alternative when the supraorbital/supratrochlear vascular pedicle has been compromised [39]. The reverse temporalis muscle flap, perfused by the superficial temporal artery, is another similar option [40].

Due to the distance between the anterior skull base and neck vasculature, pedicle length is an important consideration of free flap selection. The superficial temporal vessels, traditionally considered a “back-up” for patients with vessel-depleted necks, can be favorable recipient vessels for many skull base reconstructions [41, 42]. Reconstruction of large anterior skull base defects requires sufficient soft tissue volume to obliterate dead space. Reliable options include ALT, latissimus dorsi, or rectus abdominis myofascial free flap [43]. As described above, the ALT can be designed with variable amount of muscle, fat, or skin depending on the defect size—the latissimus dorsi flap also provides similar benefits. However, the latissimus dorsi flap is harvested with the patient in the lateral decubitus position which limits the ability to simultaneously harvest the flap during tumor ablation.

Frontal Sinus

The anterior wall of the frontal sinus provides contour to the glabella, while the posterior wall of the frontal sinus separates the cranial vault from the sinonasal cavity. Reconstruction of defects involving the frontal sinus should aim to address the affected components. Open craniofacial resections as described in the previous section may or may not involve the frontal sinus. If the craniotomy bone flap is preserved, this will restore the contour of the glabella. Any defects which involve the anterior wall of the frontal sinus can be reconstructed with titanium mesh, porous polyethylene, methylmethacrylate, or plastic polymers. Free bone grafts or split calvarial bone grafts may also be used, and favored for patients planning to undergo postoperative radiation. Calvarial grafts can be harvested from the craniotomy bone flap or from the parietal bone. Pericranial flaps can be used to provide the inner lining of reconstructed anterior frontal sinus walls to preserve function of the frontal sinus [44]. Osteocutaneous free tissue transfer is another option. Bony reconstruction of the frontal bar, orbital rim, or walls can be completed with osteocutaneous radial forearm, FFF, or scapula free flap. Depending on the size of the soft tissue defect, reconstruction may be better suited with a larger soft tissue-only flap, as the soft tissue volume with fibula and osteocutaneous radial forearm is limited.

For posterior table defects, the frontal sinus cranialization or obliteration can be pursued [45]. The mucosa of the frontal sinus and outflow tracts are removed to prevent mucocele formation, and the outflow tract can be plugged with fat or temporalis fascia. The anterior skull base defects can also be supported with pericranial flaps and TPF flaps if the vascular supply (supraorbital/supratrochlear artery and superficial temporal artery, respectively) is intact. For any concurrent defects in the dura, collagen matrix can be used to decrease the risk of postoperative CSF leak. Large, anterior skull base defects which involve the posterior table of the frontal sinus are often reconstructed with soft tissue free flaps, described in the previous section.

Concurrent cutaneous defects of the forehead or scalp can be addressed with a variety of reconstructive options. Acellular dermal matrix, skin graft, tissue expansion, local rotational flap, and free flap can all be considered depending on the size of the defect [46].

Conclusion

In summary, approach to reconstruction of the paranasal sinuses can be guided by restoring the function of affected anatomic structures. There are a wide variety of local, regional, and free flaps in addition to non-surgical options that can lead to excellent functional outcomes.