Further to squamous cell carcinoma, which is the most frequent histology worldwide, other epithelial cancers may arise into the nasal cavity and paranasal sinuses, namely nonsquamous cell carcinomas, which originate with mucous membranes, minor salivary glands, and seromucinous glands. Essentially, these cancers are classified into the following subtypes: adenocarcinoma (intestinal and nonintestinal types), salivary gland carcinoma (adenoid cystic and non–adenoid cystic types), and poorly differentiated tumors such as sinonasal undifferentiated carcinoma (SNUC) and neuroendocrine carcinoma. A multidisciplinary approach is required for management of these diseases, because poorly differentiated tumors may benefit from nonsurgical treatments such as induction chemotherapy and/or exclusive radiochemotherapy, which should be tailored to the specific histology and stage of the tumor, taking also into account the age and general condition of the patient. Endoscopic endonasal surgery is emerging as a pivotal treatment for the surgical resection of these tumors, even if traditional craniofacial and transfacial approaches should be still considered in select cases of locally advanced diseases.
30 Nonsquamous Cell Carcinoma of the Nasal Cavity and Paranasal Sinuses
Nonsquamous cell carcinomas of the nasal cavity and paranasal sinuses originate from mucous membranes, minor salivary glands, and seromucinous glands and may be broadly classified into the following subtypes: adenocarcinoma (ADC; intestinal and nonintestinal types), salivary gland carcinoma (adenoid cystic and non–adenoid cystic types), and poorly differentiated tumors such as SNUC and neuroendocrine carcinoma.
Although squamous cell carcinoma of the nasal cavity and paranasal sinuses represents the most frequent histology worldwide, the incidence, site, and histological type of sinonasal cancers can vary in different geographical areas as a result of occupational, social, and genetic factors. In this regard, selected subtypes of nonsquamous cell carcinoma can represent the most frequent histology in specific areas, such as ADC in some European countries and undifferentiated carcinoma in Chinese endemic areas (e.g., Hong Kong).1
30.2.1 Incidence and Epidemiology
ADC represents the most common mucosal epithelial malignancy in Europe, although it comprises only the 20 to 30% of all sinonasal primary malignant tumors worldwide.2 ADC usually presents in the fifth and sixth decades (mean age 58 years). It predominantly occurs in the ethmoid sinuses (85%) and olfactory region (13%), whereas a primary origin in the maxillary sinus is a rare event.2 , 3 Because of their occupational exposure to wood and leather dusts, males are affected four times more frequently than women. Hard woods such as ebony, oak, and beech confer the highest risk of developing sinonasal ADCs, which is further increased by inhalation of formaldehyde or other substances normally used in the wood industry.4 The strong correlation between intestinal-type ADC and exposure to wood and leather dusts makes this disease almost exclusive to carpenters and furniture makers. Accordingly, in many countries (Italy, Germany, Great Britain, Belgium, France, etc.), the tumor is considered an employment-related disease. Recent studies have shown that even short periods of exposure (< 5 years) can lead to an increased risk of ADC. In general, there is an average latency period of 40 years, although it can range between 20 and 70 years.4 Despite this clear correlation, the molecular mechanisms leading to the development of sinonasal ADC are still unknown. Because wood dust does not have mutagenic properties, it is hypothesized that prolonged exposure to and irritation by wood dust particles stimulate cellular turnover by inflammatory pathways.5
Sinonasal ADCs can be broadly divided into intestinal-type (ITAC) and non–intestinal-type (NITAC).
ITAC resembles ADC of the intestinal tract and shows a distinctive immunohistochemical phenotype, with all cases expressing CK20, CDX-2, and villin. These staining characteristics can also be found in precancerous lesions, as ITAC seems to be preceded by intestinal metaplasia of the respiratory mucosa, which is accompanied by a switch to an intestinal phenotype.5
Traditionally, two classifications of ITACs have been proposed. Barnes divided these tumors into five categories: papillary (18%), colonic (40%), solid (20%), mucinous (14%), and mixed (8%).6 Kleinsasser and Schroeder divided ITACs into four categories: papillary tubular cylinder cell types I to III (I = well-differentiated, II = moderately differentiated, III = poorly differentiated), alveolar goblet type, signet-ring type, and transitional type.7 Either classification is acceptable and related to prognosis, but because of its simplicity the Barnes classification is generally the most used.
NITACs are even rarer than ITACs and are merely divided into low- and high-grade subtypes, with the low grade presenting mostly in the ethmoid cells and the high grade in the maxillary sinus.3 In this subset of ADC, occupational exposure has been rarely observed, and the grade of differentiation of tumor cells has been described as the most important prognostic factor.
30.2.3 Clinical Features
ADCs present with nasal obstruction, rhinorrhea, and epistaxis, which are often unilateral. Anosmia is generally the earliest symptom reported by patients but is generally underrated. Given the paucity of symptoms in early stages and the indolent growth of the tumor, diagnosis is usually late, with patients presenting for medical attention when they have advanced-stage tumors. Because the large majority of cases originate in the nasoethmoidal complex, the most frequent pattern of spread of disease follows the olfactory phyla toward the anterior cranial fossa.8 Extension into the cranial cavity may result in neurological symptoms, such as headache. Extension may also occur into the pterygopalatine and infratemporal fossae with numbness or pain in the maxillary region. Cervical lymph node involvement at presentation can be seen in less than 6 to 8% of cases, with an added risk of neck recurrences that has been estimated to be around 4 to 6%.9 Distant metastases are infrequent at presentation (1–3%) and generally involve the lung and liver. Notably, leptomeningeal metastasis has been described for advanced-stage ADC in a nonnegligible percentage of cases (5.4% of cases).10 On endoscopic examination, ITAC commonly appears as an exophytic mass filling the olfactory cleft, usually at the level of the tail of the superior turbinate and bulging into the nasal cavity, often with a gray, necrotic, and friable appearance.
Surgery is the mainstay for the treatment of sinonasal ADC. Endoscopic endonasal surgery is tailored to the extension of disease and may range from purely endonasal resection to an expanded resection including the ethmoidal roof and dura of the anterior skull base, from the posterior wall of the frontal sinus back to the planum sphenoidal (endoscopic resection with transnasal craniectomy).8 Transnasal skull base reconstruction is performed using pedicled local flap such as nasoseptal flap11 and septal flip flap12 whenever possible; however, in the large majority of cases, the nasal septum is involved by tumor and a multilayer skull base reconstruction using autologous materials such as fascia lata or iliotibial tract is recommended (Fig. 30.1).13 In cases of massive involvement of the dura over the orbital roof or brain parenchyma infiltration, the endoscopic endonasal technique is combined with an external transcranial approach (cranioendoscopic resection).
In patients who have ITAC, in light of its association with occupational exposure, the entire ethmoid box has been traditionally considered to be at risk of harboring preneoplastic changes or developing metachronous lesions. Accordingly, a bilateral ethmoid labyrinth resection, including involved and uninvolved sides, has been generally suggested as prudent (Fig. 30.2).5 However, this policy has been recently challenged, with some Belgian authors having proposed unilateral surgical resection for unilateral extended ADCs, for which they have reported similar overall survival (OS) rates, with a slight increase in recurrence rate.14 , 15 However, currently available data preclude any definitive conclusions, and additional studies of larger cohorts of patients, with longer follow-up, will be needed to adequately address this issue.
Endoscopic endonasal surgery is effective as a single treatment modality for early-stage (T1–T2) low-grade lesions that have been radically resected with negative margins.16 In contrast, postoperative intensity-modulated radiotherapy (IMRT) improves survival rates for high-grade sinonasal ADCs (G3, signet-ring variant, solid type) regardless of the stage of disease at presentation. The role of adjuvant IMRT is also widely accepted for advanced-stage lesions (T3–T4) and in the presence of positive surgical margins.10 Given the risk of leptomeningeal spread at diagnosis or late during follow-up, prophylactic brain irradiation can be considered in high-grade lesions with intracranial invasion.10 Conversely, elective treatment of N0 neck lymph nodes is not recommended because of the low risk of regional metastases (6–8%) as reported in a large-scale review published in 2010.1 Recently, proton therapy following surgical resection has shown promising results compared with conventional radiotherapy or IMRT, with superior local control rates for ADC (80 vs. 50–60%).17
Notably, in the presence of advanced-stage ITAC (T3–T4), a neoadjuvant chemotherapy regimen based on cisplatin, fluorouracil, and leucovorin followed by surgery and radiation has been proposed for tumors with wild-type or functional p53 protein that has been highly effective, showing promising results in terms of disease-free survival (DFS).18 In addition, a subset of ITACs, mostly found in woodworkers, showed high expression of epithelial growth factor receptor (EGFR) on immunophenotyping, suggesting the possibility of using anti-EGFR therapy (Video 30.1).19
30.2.5 Outcome and Prognosis
The most important factors affecting survival in ADC are tumor stage and histological grade. A multicenter study from France analyzed a large cohort of 418 patients who had sinonasal ADC and found that stage of disease (T4), lymph node involvement, and intracranial extension were significantly associated with poorer outcomes.20
Recently, a large case series by Nicolai et al on 169 consecutive patients affected by ITAC reported DFS of 85.2%, 73.3%, and 71.7% at 1, 3, and 5 years, respectively, with OS of 93.0%, 80.5%, and 68.8% at 1, 3, and 5 years, respectively. OS and DFS were negatively affected by histological grade, advanced pT stage, dural and brain involvement, and positive surgical margins.10 Recurrences were observed in 21.3% of cases: mainly local relapses (16%), leptomeningeal spread (5.4%), regional failure (1.8%), and distant metastases (6.5%). Comparable values of 5-year OS and recurrence rates were also reported by Camp et al (68% from a series of 123 patients)21 and Vergez et al (62% from a series of 159 patients).9 Overall, these data strongly support a definitive paradigm shift in the management of ITAC toward endoscopic surgery with or without adjuvant IMRT instead of external surgical techniques, which still play a role for only a small subset of patients.
Notably, in sinonasal ADC the follow-up period has a crucial impact on OS, in particular when unilateral resection was performed—as demonstrated by the Belgian experience, in which 5-year OS dropped from 83% on a midterm follow-up (median 42 months)14 to 63% 3 years later (median follow-up 61 months).15 This supports the need for strict postoperative endoscopic and radiological surveillance of patients, who should be regularly followed for at least 10 years to promptly detect late toxicities and local or distant recurrences.
30.3 Salivary Gland Carcinomas
30.3.1 Adenoid Cystic Carcinoma
Incidence and Epidemiology
Adenoid cystic carcinoma (AdCC) represents only 10% of salivary gland malignancies of any site, but it is the most frequent tumor arising in the minor salivary glands, accounting for about 60% of cases.22 , 23 , 24 , 25 , 26 , 27 Consequently, it is by far the most frequent salivary tumor occurring in the nasal cavity and paranasal sinuses. The most involved subsites are the maxillary sinus (up to 45–50%), nasal cavity (20–30%), ethmoid (10%), and sphenoid sinus (4–5%).23 , 27 , 28 , 29 , 30
No etiologic factor has been recognized to date. The male: female ratio is variable in different series; the incidence in both genders is likely to be roughly equal, with a slight prevalence in women.30 The peak age incidence is between the fifth and sixth decades of life. However, AdCC can arise at any age, and its occurrence in young adults is not rare.23 , 28 , 29 It occurs mostly in Caucasians (over 70% of cases), followed by blacks, Asians, and Pacific Islanders (about 10% each).30
AdCC has been classified into three different histologic subtypes: tubular, cribriform, and solid. Although several historical reports have stated that no correlation between histologic subtype and outcome is present,31 , 32 Spiro et al and Batsakis et al suggested that the solid type is the most aggressive form and that its presence can negatively affect prognosis.33 , 34 Accordingly, tumor grade has been correlated to evaluation of the prevalent histologic subtype. Two different scores are currently used:
Predominantly tubular, no solid
Mostly tubular, occasional solid
Mixed with substantial solid (> 50%)
Predominantly tubular, no solid
Predominantly cribriform, < 30% solid
Solid > 30%
Recently, van Weert et al retrospectively analyzed 81 patients surgically treated for head and neck AdCC to define the prognostic value of the two currently used grading systems and to propose a simplified scheme.36 According to the authors, the presence of any solid component, irrespective of the proportion, is a negative prognosticator and can effectively identify a subgroup of tumors with aggressive behavior and poor outcome.
This simplified score (presence vs. absence of solid form) is very promising due to its high interpathologist reliability, high reproducibility, and remarkable predictive value, comparable with traditional grading systems.36
The clinician should also be aware of a new pathologic entity called “high-grade transformation of AdCC.” This rare and aggressive variant features dedifferentiated areas and a typical chromosomal translocation leading to MYB–NFIB fusion protein.37
AdCC of the sinonasal tract is usually characterized by slow and indolent growth, with a distinctive tendency to spread along nerves (perineural and intraneural invasion) and along subperiosteal and submucosal planes (Fig. 30.3).
Symptoms at presentation are usually scarce and subtle, and this feature is probably responsible for the high proportion of tumors diagnosed in the late stage (up to 76% of T3–T4 tumors and up to 80% of tumors in stage IV).23 , 28 , 38 , 39 The most frequent symptoms are nasal obstruction, facial pain, and epistaxis.23 Other complaints specifically related to nerve infiltration have been reported, such as paraesthesia of the cheek or trigeminal neuralgia. Other signs (diplopia, epiphora, proptosis, dental instability, maxillary swelling, loss of visual acuity) are less common and suggest advanced disease.
Endoscopic examination usually shows a lobulated lesion covered by normal mucosa (Fig. 30.4). Due to the almost omnipresent perineural spread of AdCC, clinical examination is often of little value to estimate tumor extension, and radiological studies such as MRI with contrast enhancement are essential.
Site-Specific Routes of Spread
The most critical means of spread of AdCC is along nerves, which can be regarded as the highway through which the tumor can overcome anatomical barriers and reach regions distant from the site of origin. Herein, the most important patterns of spread are briefly summarized.
AdCC arising in the maxillary sinus can infiltrate the infraorbital nerve: by anterograde spread it can reach the premaxillary soft tissues, while by retrograde growth it can involve the maxillary nerve and Meckel’s cave. Then, by anterograde route, the tumor can spread along any branch of the trigeminal nerve. For example, it can involve the supraorbital fissure and the orbit via the ophthalmic nerve, reaching other ocular nerves and/or the optic nerve; it can spread to the hard palate through the palatine nerves or reach the parotid region through the auriculotemporal nerve (branch of the mandibular nerve).
Another critical area is the pterygopalatine fossa (PPF), which can be involved either by direct tumor extension or by perineural spread. The PPF is a crossway of nerves (autonomic nerves arising from the pterygopalatine ganglion and anastomosing with branches of the trigeminal nerve) and vessels (branches of the maxillary artery, such as the infraorbital, posterosuperior alveolar, sphenopalatine, and vidian arteries, and so forth). Thus, from this fossa the tumor can reach infratemporal fossa and masticatory space, inferior orbital fissure and orbit, cavernous sinus and Meckel’s cave, and middle cranial fossa (Fig. 30.5).
Moreover, the tumor can extend along the carotid artery through the vidian nerve or by direct osseous permeation of the basisphenoid (Fig. 30.6).
Finally, AdCC can also reach the facial nerve through either the aforementioned auriculotemporal nerve or the greater superficial petrosal nerve (which originates from the geniculate ganglion and gives origin to the vidian nerve by joining the deep petrosal nerve; Fig. 30.7).
Nasoethmoid tumors can easily involve the anterior cranial fossa, either along olfactory nerves or by osseous permeation. Moreover, the orbit may be invaded via the ethmoidal vascular–nervous bundles. Finally, the tumor can reach the PPF via direct extension or vascular/neural spread and from there can spread to all the aforementioned structures.
Lymph-node metastases are rare. In fact, AdCC per se displays little tendency toward lymphatic spread, and sinonasal tumors are usually at low risk of nodal metastasis owing to the scarcity of lymphatic vessels in this localization (with the exception of the hard palate and infratemporal fossa). The combination of these two aspects can explain the low incidence of nodal metastases (3.6–5.3%).30 , 40
Conversely, systemic spread is a frequent finding. Lungs are by far the most involved site, followed by liver, bone, and brain. Up to 5 to 7% of patients have metastases at presentation, whereas recurrence at distant sites can occur in about 40% of cases.23 , 30
Surgery still plays a pivotal role in the treatment of sinonasal AdCC. The study of preoperative radiological imaging is of the utmost importance to estimating tumor extension, and surgical resection should be guided by meticulous examination of all possible avenues of spread of AdCC. All the named nerves adjacent to the tumor should be carefully inspected and perhaps also biopsied; if involved by tumor, they should be resected until clear margins are obtained. Critical areas, such as the PPF, should be always explored and possibly included in the resection. In addition, when the tumor is localized in proximity to bone, subperiosteal spread has to be ruled out by frozen section. Full-thickness resection of bone adjacent to the lesion is recommended in AdCC, not only when clear signs of permeation/erosion are detected, but also when the subperiosteal plane is invaded.41
In view of the insidious pattern of growth of AdCC and the complexity of sinonasal region, negative margins can be obtained in only a minority of cases (about 40%) whatever the aggressiveness of the surgical plan.23 A matter of debate is whether surgery is the ideal primary treatment when involvement of critical anatomical area (i.e., cavernous sinus) is clearly present. In view of the emerging role of new radiotherapies, as described later in this chapter, aggressive surgical resections should be balanced against their real effect on prognosis and morbidity, which can significantly worsen the patient’s residual quality of life. In selected cases, a therapeutic plan combining surgical debulking with a minimal residual disease and subsequent radiotherapy can be considered to decrease overall morbidity. However, further clinical experience is needed to explore the validity of this paradigm. Treatment planning should always be discussed by an experienced multidisciplinary team.
Site-Specific Surgical Issues
Endoscopic medial maxillectomy is indicated only for tumors limited to the medial wall of the maxillary sinus. Otherwise, depending on the extension of the tumor, more aggressive surgical procedures such as subtotal, total, and radical extended maxillectomies are deemed necessary. Orbital clearance is indicated when the tumor transgresses the periorbita with macroscopic invasion of the orbital contents. If the tumor is in close contact with the anterior skull base, then the corresponding bony layer of the skull base should be removed. In case of involvement of olfactory phyla and/or (trans)dural invasion, resection of the dura and olfactory bulbs is mandatory.
Whenever indicated, endoscopic endonasal surgery is the first choice for resection of nasoethmoid AdCC. In this setting, the advantages related to the endoscopic view (greater magnification of the surgical field, closer point of view, easier dissection of small anatomical structure) are particularly valuable to explore the possible pattern of spread of AdCC (submucosal/subperiosteal, invasion of PPF, etc.).
The main contraindications for a purely endoscopic approach are involvement of the anterior plate of the frontal sinus, massive involvement of the orbit, dural invasion over the orbit (lateral to the vertical line passing through the pupil), massive infiltration of the brain, massive involvement of the lacrimal pathway, and involvement of the hard palate or nasal bones. In these cases, combined cranioendoscopic approach or craniofacial resection are indicated.
The goals of the reconstruction are to separate the sinonasal from the intracranial space and the oral cavity, allow dental prosthetic rehabilitation, and restore facial contour. Several options can be used to reconstruct maxillary defects: prosthesis (palatal obturator), pedicled flaps, and free flaps. According to Okay’s classification,42 for small defects (class Ia and Ib) an obturator or a fat pad pedicled flap can be adequate. For larger and more complex defects (class II and III z, f, zf), the use of osteomuscular/osteomuscular–fasciocutaneous free flaps (fibula, iliac crest, anterolateral thigh, and tip of the scapula) ensures better functional and aesthetic results.43 , 44 , 45 , 46 In case of orbital exenteration, the reconstruction can be accomplished using an osteomuscolar/osteomuscolar–cutaneous free flap associated with orbital prosthesis; alternatively, a fasciocutaneous, myocutaneous, or fasciomyocutaneous free flap (rectus abdominis, anterolateral thigh) can be used to fill the orbital cavity.44 , 46
In case of dural resection, vascularized flaps (pericranial flap, nasoseptal flap) are the optimal choice for safer and faster healing. Multilayer reconstruction using autologous graft (i.e., iliotibial tract and fat) is an effective and safe alternative option when pedicled flaps are unavailable for oncological reasons or if their harvesting would result in unnecessary morbidity (i.e., harvesting of pericranial flap after pure endoscopic resections). Allogenic materials can be used when the reconstruction is completed by a vascularized flap. Finally, a free flap is almost never required to reconstruct an isolated anterior skull base defect.8 , 47 , 48
In view of the high frequency of positive microscopic margins, adjuvant radiotherapy is an essential treatment for locally advanced AdCC of the sinonasal tract.41 , 49 , 50 Much as with surgery, fields of irradiation should be planned to take into account the possible pathways of spread of AdCC. For example, in case of perineural spread, all cranial nerves at risk should be irradiated up to their entry point in the skull base.
Promising data on the efficacy of heavy-ion therapy (carbon and proton) have been recently published, both as single-modality treatment for large inoperable AdCC and in an adjuvant setting.51 , 52 Critical extensions of the tumor, predictable efficacy and morbidity of each treatment, expertise of the surgeon and radiation oncologist, and patient’s willingness are all important factors to consider in treatment choice and planning.41
Chemotherapy is reserved for palliative treatment of metastatic disease; of note, the role of systemic therapies remains controversial, with no real benefit in primary, adjuvant, and palliative settings reported.53
Outcome and Prognosis
The natural history of AdCC is characterized by high risk of local relapse (Fig. 30.8) and/or distant metastasis, typically many years after primary treatment. Disease-specific survival (DSS) is quite high at 5 years (70%) but it drops at 10 years (40%) and further decreases at 20 years (15%).23 , 30
Recurrence rate is about 60% (range, 40–77%) with a slight prevalence of distant metastasis (40%) over local relapse (30%). Generally distant metastasis causes a remarkable decrease in survival rates (from about 65% to 20% at 5 years). Interestingly, lung and liver metastases are usually associated with a better prognosis (with a mortality rate of 50% at 4 years) than those of bone and brain (mortality rate ranging from 70% up to 100% at 1 year).26 , 54 , 55 , 56 , 57 , 58 AdCC arising in the nasal cavity tends to have a better prognosis, whereas tumors originating in the sphenoid have the worst.
The most important negative prognostic factors are invasion of the skull base, pT4 classification, IV stage, solid subtype, and positive margins.23 , 26 , 27 , 54 , 57 Perineural spread also has a negative effect on survival; in particular, intraneural invasion was a strong negative prognosticator in a large multicentric study on head and neck AdCC published in 2015.58 The negative impact of the involvement of the orbit and infratemporal fossa is still unclear, as it did not reach statistical significance in the series from MD Anderson.23
Finally, it is worth remembering that the indolent and slow growth usually attributed to AdCC is typical of low- and intermediate-grade tumors. Conversely, high-grade variants (predominantly solid tumors, high-grade transformation of AdCC) can display a very aggressive biological behavior, resulting in advanced local disease and early regional and distant spread. Accordingly, the prognosis can be poor (5-year DSS ranging from 30 to 40%).30
30.3.2 Non–Adenoid Cystic Carcinoma Salivary Tumors
Non-AdCC salivary tumors (NAdCC) of the sinonasal tract are heterogeneous and very rare. Consequently, published evidence is confined to small case series and case reports. In general, NAdCC are diagnosed at a relatively earlier stage than AdCC.59 Because of the paucity of data, biological behavior, treatment strategy, and prognostic factors must be inferred by NAdCC of other subsites. Tumor grade is probably the most important parameter affecting biologic behavior and prognosis. Low- and intermediate-grade lesions display indolent growth, low tendency to spread along nerves and subperiosteal/perichondral planes, and favorable prognosis.60 , 61 , 62 , 63 , 64 Generally, a less aggressive policy of surgical resection is advised in these cases, with adjuvant radiotherapy usually recommended only in locally advanced tumors.41
Conversely, high-grade tumors can display a more aggressive and rapid pattern of growth with perineural and lymphovascular spread. Accordingly, the surgical approach should be more aggressive and adjuvant radiotherapy is often required. Moreover, these tumors are associated with a high rate of distant metastasis. In fact, despite a reasonably good rate of local control, prognosis is still poor due to distant failure. The efficacy of chemotherapy is largely unsatisfactory; new regimens and target therapies are currently under investigation.64