Benign Tumors of the Sinonasal Tract

Inverted Papilloma

Inverted papilloma is the second most frequent benign tumor of the sinonasal tract after osteoma and is the most common form of sinonasal papilloma. The lesion is estimated to represent 0.4% to 4.7% of all surgically removed nasal tumors, with an incidence ranging from 0.74 to 2.3 new cases per 100,000 inhabitants per year. Men are more commonly affected than women (2 to 3 : 1 male to female ratio), and the lesion is most often seen in the fifth and sixth decades of life.

Inverted papilloma predominantly arises from the lateral nasal wall and maxillary sinus: maxillary medial wall, in particular in the region of the fontanelles, is the most common site of origin. Frontal and sphenoid sinuses are rarely involved primarily. Often, the lesion extensively involves multiple sites (30% of cases), making it difficult to assess precisely its origin. Bilateral papillomas of any type are exceptional.

With regard to histologic appearance, inverted papilloma is composed exclusively or almost exclusively of hyperplastic ribbons of basement membrane–enclosed epithelium that grow downward into the underlying stroma. The epithelium is multilayered and formed of squamous or ciliated columnar cells mixed with mucocytes and transmigrating neutrophils. Inverted papilloma may be occasionally found in association with sinonasal hamartomas.

Malignant transformation is observed in 5% to 15% of cases; synchronous occurrence is more frequent than metachronous. The vast majority of malignant tumors arising from inverted papillomas are squamous cell carcinomas, but other tumors such as sinonasal undifferentiated carcinomas, mucoepidermoid carcinoma, and verrucous carcinoma may rarely occur.

Different etiologic factors have been investigated. Exposure to organic solvent is significantly related to a higher risk of developing inverted papilloma with a dose–response relationship, while no association has been demonstrated with smoke or alcohol. Instead, smoke has been recently recognized as a risk factor for recurrence and malignant transformation. Hong et al. analyzed a retrospective cohort of 162 inverted papillomas and focused on 17 cases (10.5%) of malignant degeneration. Fourteen squamous cell carcinomas occurred among 53 smokers (26.4%), while only three arose in 109 nonsmokers (2.8%). Smoking conferred a 12-fold higher risk of malignant transformation ( P < .001).

The presence of human papillomavirus (HPV) in inverted papilloma has been extensively investigated, but its role remains ambiguous. In fact, whether HPV positivity represents a pure colonization or indicates that the virus is an etiologic factor is still a matter of debate. A large meta-analysis by Lawson et al. in 2008 found an overall HPV incidence of 22% to 26%, with an increased rate of detection of high-risk HPV subtypes in cases of dysplasia (56%) and carcinoma ex-inverted papilloma (55%). This finding appeared to support an etiologic role for HPV, especially in cases of malignant transformation. However, this meta-analysis included several old studies in which the techniques used to detect HPV were unable to distinguish between active and “passenger” virus. Despite the introduction of excellent methods to identify transcriptionally active HPV (i.e., viral mRNA in situ hybridization), published evidence still remains controversial. Rooper et al. retrospectively analyzed 52 inverted papillomas: 30 benign, 6 with dysplasia, and 16 with malignant transformation. Transcriptionally active HPV was not detected in any of them, which led the authors to conclude that in most cases carcinomatous transformation of inverted papillomas is driven by non-HPV mechanisms. Similar conclusions were reached by Mohajeri et al., who looked for all known HPV subtypes in 76 inverted papillomas by DNA extraction, amplification, and sequencing. HPV DNA was found in 10 specimens (13%) including nononcogenic types 6 and 11 (8%) and oncogenic types 16, 18, 45, and 56 (5%), while it was not isolated in ex-inverted papilloma carcinomas. p16 cell staining was a poor surrogate for HPV positivity. Conversely, in other research using RNA in situ hybridization technique, the analysis of 19 inverted papillomas showed HPV positivity in 100% of cases, although in 58% of tumors, less than 1% of cells exhibited transcriptional activity. In conclusion, the possible role of HPV as an etiologic factor in the development of inverted papilloma and its malignant transformation is currently unclear (although apparently unlikely) and needs further investigation.

On a molecular basis, inverted papilloma is characterized by activating epidermal growth factor receptor (EGFR) mutations, which are preserved even in cases of malignant transformation. Conversely, oncocytic papillomas and related carcinomas constantly harbor activating KRAS mutations, which are never found in inverted papillomas. These findings suggest that inverted and oncocytic papillomas represent two distinctive entities, with peculiar mutations representing a trait that is also maintained in cases of malignant degeneration.

Unilateral nasal obstruction with watery rhinorrhea is the most common symptom prompting the patient to seek an otorhinolaryngologic consultation; unilateral rhinosinusitis with related symptoms, such as headache and facial pain, may be caused by mechanic obstruction of sinus drainage. Epiphora, proptosis, and diplopia may be associated with advanced lesions involving the orbit and should raise the suspicion of malignant transformation. Nasal endoscopy typically shows a pale, polypoid mass with a papillary or cerebriform appearance protruding from the middle meatus ( Fig. 50.1 ), which strongly suggests the nature of the lesion. Sometimes, the coexistence of inflammatory polyps may make endoscopic diagnosis trickier. In any case, a biopsy performed under endoscopic guidance is indicated to establish definitive histology.

Imaging studies are required to provide a tentative diagnosis, assess the extent and site of origin of the tumor, disclose its relationship with surrounding structures (i.e., orbit, skull base, optic nerve, and internal carotid artery), and assess possible malignancy. In our experience, these goals are best achieved by MRI with gadolinium enhancement. In fact, this examination can provide better differentiation between tumor and inflammatory mucosal changes and more precise contouring and pattern of growth of the lesion. Moreover, the so-called cerebriform-columnar pattern, which is highly predictive of inverted papilloma, can be easily detected by MRI ( Figs. 50.2 and 50.3 ). This radiologic finding reflects the histologic arrangement of inverted papilloma characterized by the alternation of regular parallel folds made of a highly cellular metaplastic epithelium and of an underlying less cellular stroma. However, even MRI has limitations: when the lesion completely fills a sinus, it may be difficult to distinguish a tumor growing from a small area of insertion from one extensively involving the mucosa. The site of origin of the lesion can be reliably identified when associated with a bony strut, focal hyperostosis, and osteitic changes on CT but may also be identified on MRI (see Fig. 50.3 ). Finally, MRI may also suggest a malignant transformation of the lesion on the basis of the loss of the typical cerebriform-columnar pattern and the evidence of aggressive and infiltrative tumor growth ( Fig. 50.4 ).

Endoscopic endonasal surgery is the gold standard of treatment for inverted papilloma. A meta-analysis of 2006 comparing endoscopic and external approaches in contemporary (1992–2004) and historical (1970–1995) cohorts clearly showed that patients treated with endoscopic surgery had a significantly lower recurrence rate in the contemporary series, and external approaches in recent cohorts were associated with a risk of recurrence similar to that of historical series. This suggests that the improvement in local control may be mostly attributed to the introduction of endoscopic resections. Two more recent meta-analysis confirmed this finding and showed that endoscopic resection is associated with a reduction of the risk of recurrence by between one-third and one-half. Moreover, unquestionable advantages of endoscopic surgery over external approaches are the absence of a facial incision, negligible facial swelling, shorter in-patient time, and a reduction in postoperative pain and paraesthesia. Therefore, endoscopic surgery is preferred whenever possible. However, an exclusive endoscopic approach may be contraindicated in the following situations: (1) concomitant presence of a malignancy involving critical areas (see contraindications for endoscopic surgery in sinonasal malignant tumors); (2) site of origin on the anterior wall or lateral recess of the frontal sinus, or massive involvement of the mucosa of the frontal sinus and/or of a supraorbital cell; and (3) orbital involvement, which is a very uncommon situation usually found in patients who have already undergone one or more surgical procedures.

Whatever the approach, the key concept for a radical resection is to dissect the involved mucosa in the subperiosteal plane and drill the underlying bone. Thus, the extent of the operation is dictated by the site of origin and area of mucosa involved by the lesion.

Recently, a paradigm shift toward a more conservative, pedicle-oriented resection has been observed. This approach was first described by Landsberg in 2008 and popularized more recently by Kamel et al. and Pagella et al. Briefly, the procedure starts with debulking of the intranasal part of the tumor with the aim of detecting the site of origin of the lesion. Subsequently, both the subperiosteal dissection of the mucosa and drilling of the underlying bone are limited to the point of insertion of the tumor. The key concept of this technique is that the extent of surgery is not predetermined but modulated according to the site of attachment of the lesion. In their study, Pagella et al. compared a historical cohort undergoing complete demucolization of the involved sinus and extensive drilling with a group of patients who underwent the pedicle-oriented procedure. No difference in terms of recurrence rate was observed, while a decrease in operating times and postoperative complications was experienced with the new technique.

In accordance with this trend toward a more conservative resection, greater attention has been paid to minimize the morbidity related to the surgical corridor. We have recently proposed a modular classification of endoscopic medial maxillectomies, which progressively widens the size of the targeted area exposed. The classification ranges from a middle antrostomy (type A) to an endoscopic Sturmann–Canfield procedure (type D), which entails the removal of the medial portion of the anterior wall of the maxillary sinus to enable access to all the antrum walls and recesses. The clear advantage is the possibility of minimizing the corridor on the basis of the surgical need, with the option of intraoperative shift to a larger approach if there is inadequate exposure.

The “TuNa-saving” technique is a modified medial maxillectomy in which the head of the inferior turbinate (Tu) and the nasolacrimal duct (Na) are preserved, while the inferior meatus is drilled anteriorly up to the piriform aperture. It can provide access to almost the whole maxillary sinus and is indicated in cases of inverted papillomas originating from the inferior portion of the lateral and/or anterior walls of the maxillary sinus. Preliminary results have shown good oncologic control and a significant decrease in epiphora and dacryocystitis compared with classical extended medial maxillectomies.

The prelacrimal approach (with or without medial displacement of inferior turbinate and nasolacrimal duct) provides direct access inside the maxillary sinus by removing the bone between the lacrimal pathway and the piriform aperture (with possible extension to the medial part of the anterior maxillary wall). The mucosal incision is sutured at the end of the procedure, so that the lateral nasal wall is completely preserved. Preliminary results are promising with reported good disease control, no epiphora or dry nose, and minor transitory sensitive impairment of the upper lip. However, a possible drawback may be the impossibility to have full visualization of the maxillary sinus during nasal endoscopy at follow-up.

As a final remark, it should be noticed that the Sturmann–Canfield procedure may be associated with relevant morbidity, in terms of epiphora (12%), premaxillary paraesthesia (24%), and hypoesthesia of ipsilateral superior incisive and canine teeth (29%). Accordingly, its use should be limited, and thorough preoperative counseling of the patient should be undertaken.

Different degrees of frontal sinus involvement can be observed, ranging from a limited lesion marginally growing from the ethmoid into the frontal recess ( Fig. 50.5 ) to very complex cases in which most or all the sinus mucosa is diseased. Because the key point for ultimate success is removal of the lesion along the subperiosteal plane with the possibility of drilling the underlying bone, the surgical choice may involve a Draf IIB or Draf III endoscopic sinusotomy or a combination of an endoscopic approach with external sinusotomy through an osteoplastic flap depending on the extent of frontal sinus mucosa involvement, a finding that can usually only be definitively assessed at the time of surgery. Even the involvement of an extensively pneumatized supraorbital cell is frequently challenging. Exposure through the nose may be increased by coagulation and sectioning of the anterior ethmoid artery and by displacement of the orbit after drilling of the upper portion of the lamina papyracea. However, whenever the lesion involves a cell that extends far posteriorly and/or laterally over the orbit and complete resection cannot be achieved transnasally, the surgeon should resort to a frontal osteoplastic flap.

Postoperative surveillance is almost exclusively clinical. Regardless of the access selected, resection of inverted papillomas should lead to the creation of a largely marsupialized cavity that will allow wide access for endoscopic inspection during follow-up. It is our policy to perform this every 4 months during the first 2 years, every 6 months until the fifth, and then yearly until the 10th year. Postoperatively, MRI or CT is indicated only when a sinus that was originally involved by the lesion is not accessible for exploration owing to scar closure, the patient is symptomatic, or a residual or recurrent lesion has been histologically documented.

Recurrence rates after transnasal resection of inverted papillomas dramatically dropped after the introduction of microscopic and endoscopic techniques. Therefore, some traits that were historically attributed to this lesion (i.e., high tendency to local recurrence and multifocality) were more likely the consequence of inadequate treatment rather than the expression of intrinsic features of the tumor, and they should now be regarded as myths to dispel.

In a recent meta-analysis, the recurrence rates after endoscopic resection and external and combined approaches were 13.8%, 18.7%, and 12.9%, respectively. However, in the largest mono­institutional series recently published, the recurrence rate was well below 10%.

Incomplete or inadequate resection, high-stage, recurrent tumor, and smoking are the most relevant risk factors for relapse. In a recent meta-analysis, Krouse stage 3 (involvement of any wall of the maxillary sinus except the medial, frontal sinus, or sphenoid sinus) yielded a significant increase in the risk of recurrence (51%) when compared with stage 2 (involvement of ostiomeatal complex region, ethmoid, or medial wall of the maxillary sinus). This trend reflects how the extension toward paranasal sinuses may jeopardize a successful resection. Finally, dysplasia is emerging as an effective predictor of recurrence. In the study of Safadi et al. on 92 patients, the presence of dysplasia conferred a 10-fold higher risk of recurrence. In our experience, the risk is tripled (14.6% vs 4.8%, OR: 3.1, P = .048; unpublished data).

A staging system would facilitate the comparison of results among different institutions, but unfortunately, none of the numerous classifications specific for inverted papilloma has been widely accepted as a standard of reference.