31 Esthesioneuroblastoma

Valerie J. Lund and David J. Howard


Esthesioneuroblastoma is a rare and unique tumor arising from olfactory epithelium, which prior to the advent of craniofacial surgery was associated with a poor prognosis. Conventional craniofacial resection doubled survival compared with previous treatments, and more recent advances in diagnosis and the combination of surgery and radiotherapy have further improved survival. The use of endoscopic surgery in selected cases has also reduced morbidity with similar outcomes. Genomic analysis of the tumor suggests new possibilities in targeted treatments in the future. However, lifelong follow-up is required irrespective of the treatment, because the natural history of this tumor extends over a lifetime.

31 Esthesioneuroblastoma

31.1 Incidence and Epidemiology

In common with all tumors of the anterior skull base, esthesioneuroblastoma (ENB) is comparatively rare. This malignant neuroendocrine neoplasm classically arises from the olfactory mucosa and was first recognized by Berger et al in 1924, who coined the term esthesioneuroepitheliome olfactif.1 However, a wide range of other terms has been used, including esthesioneurocytoma, esthesioneuroma, intranasal neuroblastoma, olfactory neuroepithelial tumor, and olfactory neuroblastoma. In 1966, Skolnik et al found only 97 cases reported in 42 papers in the English literature, with most authors having treated only two or three cases, and by 1989 O’Connor estimated that ≤ 300 cases had been published, which represented 1 to 5% of all malignant tumors of the nasal cavity.2 ,​ 3 However, this number had risen to 945 by 1997, not including a large series from the Armed Forces Institute of Pathology or another from the Institut Gustave-Roussy.4 ,​ 5 ,​ 6 In 2000, the National Cancer Database included 664 cases from more than 500 U.S. hospitals over a 10-year period (1985–1995). In the last 40 years, this tumor is being described in increasing numbers, almost certainly due to increased awareness and improved histological techniques for diagnosis, and likely occurs in ~ 0.1/100,000 people/year. Even so, the difficulty of accruing large individual series compromises statistical analysis of outcomes. The authors, who work in a tertiary referral center, have managed 125 cases since 1970.

Hitherto, unlike for adenocarcinoma, occupational factors in the development of olfactory neuroblastoma have not been identified in men other than a single case report in a woodworker.7 However, in rodents the administration of N-nitroso compounds has been reported to produce esthesioneuroepitheliomas when administered parenterally, orally, or topically, as has administration of bischloromethyl ether.8 ,​ 9 ,​ 10 ,​ 11

There appears to be a slight male preponderance in the literature, and the tumor may occur over a wide age range (3–90 years), with a reported bimodal peak in the second/third decades and sixth/seventh decades.3 ,​ 12 ,​ 13 ,​ 14 ,​ 15 In our own series, this male preponderance is more pronounced, with a 1.3:1 ratio, and the age range is 12 to 88 years (mean 48 years), though occurrence in children under 10 has occasionally been reported.12 ,​ 16

Recent genomic analyses suggest that there may be some specific genetic variations that could predispose to the development of this tumor and that might provide therapeutic opportunities in the future.17 ,​ 18 ,​ 19

31.2 Pathology

Olfactory neuroblastoma generally arises in the nasal roof, corresponding to the anatomical distribution of the olfactory epithelium, which extends from the olfactory niche onto the upper nasal septum and superior turbinates on the lateral wall. Evidence for the origin of olfactory neuroblastoma from specialized olfactory epithelium, however, is somewhat circumstantial, though tumors occasionally found outside this distribution have been ascribed to ectopic olfactory epithelium.5 Tumors arising in the cribriform niche can easily spread superiorly along olfactory fibers into the anterior cranial fossa to affect the olfactory bulb and tracts. The superior septum is often involved, and from thence the tumor may spread to the contralateral side and into the ethmoids and adjacent orbit. Earlier histological studies suggest that there is microscopic intracranial involvement in many patients even when not suggested by imaging and macroscopic appearance at surgery, but this has not been supported by more recent analysis.20 ,​ 21

Macroscopically, the tumor is characteristically a polypoid reddish gray mass that bleeds readily. Microscopically, the tumor typically forms clusters of cells arranged in patterns, which vary from small nests surrounded by a fibrillary stroma to diffuse areas separated by fibrovascular septa. The cells may palisade around blood vessels, and occasionally true rosettes form. Hyam proposed a grading based on tumor differentiation, which is widely used.5

It had been suggested that ENB was part of the Ewing’s sarcoma/peripheral neuroectodermal group of tumors, but this has not been supported by immunohistochemical studies.22 However, it can present some difficulties in diagnosis, even when using modern techniques, and can be confused with a host of other small-cell tumors, such as lymphoma, malignant melanoma, peripheral neuroectodermal tumors, and sinonasal undifferentiated carcinoma, by those unfamiliar with sinonasal malignancy. This prompted Ogura and Schenck to describe ENB as the “great imposter.”23

Immunohistochemistry using a broad panel of antibodies is usually employed to confirm the diagnosis. These include general neuroendocrine markers such as neuron-specific enolase (NSE), synaptophysin, chromogranin, and protein gene product–9.5 (PGP-9.5), which are usually positive.24 S100 positivity can be demonstrated at the periphery of the tumor nests, and some tumors are also positive using MNF 116 and CAM 5.2, both of which are stains for certain cytokeratins. Conversely, LP 34, a high–molecular weight cytokeratin stain; epithelial membrane antigen (EMA); carcinoembryonic antigen (CEA); and glial fibrillary acidic protein (GFAP) are generally negative.

31.3 Staging

A number of staging systems have been proposed. Kadish et al’s is a somewhat crude system that divides tumors into three stages25:

  • Stage A: lesions confined to the nasal cavity

  • Stage B: involvement of nasal cavity plus one or more of the paranasal sinuses

  • Stage C: involvement beyond the nasal cavity, including the orbit, skull base, intracranial cavity, cervical lymph nodes, or systemic metastases

This can now be considerably refined by the use of modern imaging protocols validated by craniofacial resection.26 ,​ 27 However, neither the Kadish system nor subsequent modifications have proved entirely successful, largely owing to the late presentation of most patients and despite efforts to refine advanced disease by creating a stage D for metastases.25 ,​ 28 However, attempts to correlate both staging (Kadish/Morita) and histological grade (Hyam) have had variable success.29 ,​ 30 The staging system proposed by Dulgerov in 2001 is probably the most often used (Table 31.1).13

Table 31.1 Olfactory neuroblastoma: staging system after Dulguerov et al13




Tumor involving the nasal cavity and/or paranasal sinuses (excluding the sphenoid sinus), sparing the most superior ethmoidal cells


Tumor involving the nasal cavity and/or paranasal sinuses (including the sphenoid sinus), with extension to or erosion of the cribriform plate


Tumor extending into the orbit or protruding into the anterior cranial fossa, without dural invasion


Tumor involving the brain


No cervical lymph node metastases


Any form of cervical lymph node metastases


No metastases


Any distant metastases

31.4 Treatment

31.4.1 Clinical Features

The usual site of origin results in fairly innocuous symptoms initially, remarkable only for their sudden onset and unilaterality. As a consequence, there is often considerable delay in diagnosis, with some patients waiting for more than a year (24% of 40 cases reported by Schwabb et al).6 There is little specific to this particular tumor, whose symptoms are common to all nasal cavity lesions: blockage, discharge, some epistaxis as a result of the tumor’s vascularity, and hyposmia. In a series of 42 patients, unilateral obstruction occurred in 93%, epistaxis or serosanguineous discharge in 55%, and rhinorrhea in 30%.15 Anosmia was reported rarely (5%) in our patients, and invasion of the anterior cranial fossa is otherwise generally silent. As the tumor spreads to the orbit, patients may develop epiphora, displacement of the eye, diplopia, and eventually visual loss, though this last is usually a late phenomenon. Ocular symptoms occurred in 11% of our patients.15 Curiously, in this series the left side was more often affected (62%) than the right (29%), with both sides affected at presentation in 9%. Occasionally involvement of the Eustachian orifice may result in otalgia and conductive hearing loss as a result of serous otitis media.

The incidence of cervical metastases varies considerably from report to report, compounded by the generally small numbers in each series. Rinaldo et al reported a lymph node metastatic frequency of 23.4% (range 5–100%), though only eight studies validated the diagnosis of ENB with immunohistochemistry.14 A recent report by Nalavenkata et al found 7% at presentation and 9% during a mean follow-up of 41 months in 113 patients.31

Because ENB is a neuroendocrine tumor, it can be associated with paraneoplastic syndromes caused by inappropriate hormone production, such as antidiuretic hormone secretion (SIADH), first described by Bouche et al in 1967.32 This has been estimated to occur in ~ 2% of cases, which has been our own experience, and may precede diagnosis of the tumor by some years (mean 3.5 years).33 If suspected, immunohistochemical staining for arginine vasopressin confirms the diagnosis. Normalization of sodium and osmolality can be anticipated after tumor removal in most cases, and because disease recurrence can also manifest as SIADH, some authors consider sodium levels to be a form of tumor marker that could be monitored as part of routine follow-up.33 ,​ 34 ,​ 35 ,​ 36

31.4.2 Imaging

Ideally, all patients are submitted to a preoperative imaging protocol, which employs a combination of high-resolution contrast-enhanced CT (coronal, axial, and sagittal planes) combined with multiplanar MRI enhanced with gadolinium diethylenetriamine penta-acetic acid (DTPA), which should include the neck. (Fig. 31.1).37 If available, ultrasound of the neck combined with fine-needle aspiration cytology is a useful screening technique having a high degree of accuracy.38

Fig. 31.1 Imaging of olfactory neuroblastoma. (a) Coronal CT showing a mass of olfactory neuroblastoma filling the right nasal cavity without obvious erosion of the skull base. (b) Coronal MRI (T1-weighted unenhanced) in the same patient, showing tumor confined to the nasal cavity and confirming suitability for an endoscopic resection. (c) Coronal CT in another patient showing olfactory neuroblastoma filling upper right nasal cavity and ethmoids with extension through the lamina papyracea and associated with new bone formation on the anterior skull base. (d) Coronal postgadolinium MRI scan in another patient showing olfactory neuroblastoma extending into anterior cranial cavity with an associated peritumoral cyst (arrow), requiring combined endoscopic and conventional craniofacial approach. ([a] and [b] reproduced with permission from Lund VJ, Howard DJ, Wei WI, Tumors of the Nose, Sinuses, and Nasopharynx, New York, NY: Thieme; 2012.)

Many centers also undertake imaging for systemic metastases using PET-CT or MRI, although the incidence of systemic disease at presentation is extremely low.39 Some clinicians have advocated 68Ga DOTATATE PET/CT because of its greater specificity for ENB.40

After surgery, all patients should also undergo a rigorous follow-up protocol of outpatient/office endonasal endoscopy together with MRI of the head and neck every 3 to 4 months for the first 2 years and then every 6 months thereafter.26 Formal biopsy of any suspicious areas should be performed as required. Patients may develop recurrence many years after initial treatment (22 years from ostensibly curative treatment in one of our patients), which can occur anywhere in relation to the surgical field, orbit, or intracranial cavity, suggesting some local embolic phenomenon. Distant en plaque dural deposits are not uncommon, so the entire intracranial compartment should be visualized on MRI (Fig. 31.2).

Fig. 31.2 Coronal MRI (T1 with gadolinium) showing dural deposits distant from the craniofacial cavity, found 11 years after treatment.

Initial imaging often suggests the diagnosis but also, more important, indicates extent.37 ,​ 41 No features are specific to ENB, but the position of the mass and associated bone erosion indicates a malignant nasal tumor and can be associated with intracranial peritumoral cysts. Coronal CT remains the most accurate method of demonstrating early anterior skull base erosion, whereas the addition of contrast enhancement and MRI shows extent of intracranial and orbital spread. Typical features are an intense signal on precontrast T2-weighted spin echo sequences and strong enhancement after gadolinium on T1-weighted sequences. However, even the most sophisticated imaging cannot be absolutely relied on to demonstrate involvement of the dura and orbital periosteum, which can be determined only by surgery with histological confirmation.

31.4.3 Treatment

The advent of craniofacial resection in the 1970s revolutionized the treatment of ENB, doubling survival figures, and rapidly became the “gold standard.” In a 2001 meta-analysis, Dulgerov et al confirmed that this procedure, combined with radiotherapy, was the treatment of choice.13

31.4.4 Surgery

Craniofacial resection was introduced in the 1970s by Ketcham and others, providing the combination of an en bloc oncologic resection having low morbidity and excellent cosmesis.42 ,​ 43 ,​ 44 By approaching the tumor from the nose and anterior cranial fossa, the operation directly addresses the origin and local spread of this tumor, allowing resection of dura and the olfactory system, including the olfactory epithelium, cribriform plate, olfactory bulb, and tracts. This directly deals with macro- and microscopic spread of disease, reducing local recurrence.45 ,​ 46

There are many variations on the technique, but essentially all involve some form of craniotomy together with a nasal approach using various incisions and forms of repair. Use of a coronal incision in the scalp and a sublabial incision for a midfacial degloving can hide these scars, though the use of an extended lateral rhinotomy or a supraorbital spectacle incision heals well. The skull base repair may be effected using a pericranial flap or fascia lata and split skin. In our craniofacial series, which extended over 35 years, postoperative hospital stay has been, on average, 14 days, and major complications have been low.46

Prior to craniofacial resection, conventional wisdom dictated that the orbit should be sacrificed if tumor either was adjacent to or had transgressed the periosteum. However, it became clear that a significant proportion of these eyes could be salvaged without compromising survival. If the tumor has not penetrated the full thickness of orbital periosteum on frozen section, it is possible to resect it widely and skin graft the area. Nonetheless, the orbit should be cleared if there is full-thickness periosteal penetration or frank infiltration of orbital contents when cure would otherwise be possible.

In recent years, an endoscopic resection with curative intent has been used for selected cases, usually for those who are without significant intradural extension or who are a poor anesthetic risk. The endoscopic approach should not be regarded as a limited procedure, because it can encompass most of what was previously achieved by formal craniofacial approaches, including a wide-field resection of tissue, and also allows determination of the exact site of origin of the tumor.21 ,​ 26 ,​ 47 ,​ 48 ,​ 49 ,​ 50 Some authors have also suggested that olfactory function may be salvaged by preservation of one or both olfactory bulbs in selected cases.51 If there is significant intradural or frontal lobe involvement, endoscopic and external craniotomy approaches can be combined.52 ,​ 53 ,​ 54 Endoscopic surgery can also have a role in the management of localized recurrence.

There is insufficient evidence to support prophylactic treatment of the N0 neck, although a selective neck dissection is undertaken in the presence of disease.55

31.4.5 Radiotherapy

Radiation in this area must be carefully administered to deliver the maximum dose while preserving the adjacent brain and optic nerves. An external megavoltage beam and three-field technique has generally been used. An anterior port combined with wedged lateral fields delivers a dose of 60 to 70 Gy in 30 to 35 sessions over 6 to 7 weeks.56

Because of the proximity of the optic chiasm, intensity-modulated radiotherapy (IMRT) is preferred if available.57 Generally, radiotherapy has been used as an adjunct to surgery, and postoperative delivery is preferred.15 ,​ 58

Stereotactically guided radiotherapy has been used in combination with surgery, which may offer improved local control with minimal collateral damage, although numbers and follow-up preclude definitive evidence.59 Similarly, the advantages of proton beam therapy remain to be determined owing to the present lack of numbers and follow-up.56

31.4.6 Chemotherapy

The use of concomitant chemotherapy has not been fully evaluated, though chemosensitivity has been found in retrospective series. ENB has been shown to respond to platinum-based regimes.60 ,​ 61 ,​ 62 ,​ 63 ,​ 64 Cyclophosphamide, doxorubicin, docetaxel, irinotecan, and etoposide have also been used in more advanced disease.65 ,​ 66 ,​ 67

Since 1999 it has been our practice to give two courses of this adjunctive cisplatin at the beginning and halfway through the postoperative course of radiotherapy, based on data suggesting that doing so reduces recurrence.64

31.5 Outcome and Prognosis

Prior to the advent of conventional craniofacial resection (cCFR), the use of lateral rhinotomy and radiotherapy provided poor results of ≤ 40% at 5 years, largely due to its inability to deal with intracranial spread.68 ,​ 69 ,​ 70 Craniofacial resection specifically addressed this area and allows removal of the olfactory bulbs and tracts, where microscopic disease might be residing undetected. As a result, when large series with long-term follow-up after craniofacial are considered, the 5-year overall survival is seen to have improved significantly or even to have doubled, as in our own series to 77%, or to 89% in that of Diaz et al.15 ,​ 71 ,​ 72 ,​ 73 ,​ 74 ,​ 75 ,​ 76 However, there is continued loss over time, as well as local recurrence many years after treatment (range 12–144 months, mean 37 months, longest interval 22 years). In an earlier analysis of our cCFR series of 42 cases, disease-free survival dropped from 77% at 5 years to 53% at 10 years. In a further study of this cohort enlarged to 56 individuals, 15-year survival fell to 40%, underscoring the importance of long-term follow-up.46

The most frequent recurrence is local and occurred in 17% of our series, in keeping with other published series using craniofacial resection and radiotherapy. Local recurrence has been shown to be significantly decreased by the addition of radiotherapy (28% vs. 4%), but it does not seem to matter whether this is given before or after surgery, even when the therapeutic interval differs between pre- and postoperative administration.15 This applies to both survival (p = 0.52) and complications (p = 0.07). Interestingly, previous treatment did not seem to affect 5-year actuarial survival, either, but in patients who developed local recurrence, 5-year survival after further salvage treatment was 54%. However, it should again be noted that the site of “local” recurrence can be anywhere and on either side of the nose, sinuses, orbit, or intracranial cavity, so follow-up must be especially vigilant if this disease is to be detected early.

A recent study of 95 of our ENB patients allowed comparison of 30 who underwent endoscopic sinus surgery (ESS), with 65 undergoing cCFR.27 This showed significantly higher survival in those treated endoscopically (100% 5- and 97% 10-year overall survival), reflecting the selection of patients who have more limited disease but nonetheless confirming that this approach is at least as good as craniofacial resection, bringing much lower complication and morbidity rates (Fig. 31.3).27 ,​ 77 ,​ 78 ,​ 79 ,​ 80 ,​ 81 ,​ 82 ,​ 83 ,​ 84 ,​ 85 As a consequence, our present approach for ENB treatment is as follows, as supported by a recent systematic review86:

Fig. 31.3 Kaplan-Meier graph of overall survival comparing endoscopic sinus surgery (ESS) to cCFR. (Reproduced with permission from Lund VJ, Stammberger H, Nicolai P, et al, European position paper on endoscopic management of the nose, paranasal sinuses and skull base, Rhinology 2010; (Suppl 22), 105:86–100.)

  • T1, T2: ESS plus radiotherapy, with or without adjuvant chemotherapy

  • T3: ESS or ESS/conventional craniofacial resection (cCFR) plus radiotherapy, with or without adjuvant chemotherapy

  • T4: cCFR plus radiotherapy, with or without adjuvant chemotherapy

A Cox regression analysis of the type of surgery, stage, orbital and/or brain involvement showed both orbital extension (p = 0.002) and intracranial involvement (p = 0.022) to be significant independent factors affecting outcome. Without orbital involvement, 5-year DFS was 88.3%, a figure that fell to 55.6% if periosteum was involved and to 25% with frank involvement of the globe (Mantel-Cox p = 0.00) (Fig. 31.4). Similarly, disease-free survival (DFS) was significantly higher if there was no intracranial involvement (92.3% 5 years and 76.3% 10 years). If dura was involved, then 5-year survival was 68.9% and 10-year survival 56.8%, figures that fell to 41.7% and 10.4%, respectively, if there was intracerebral disease (Mantel-Cox p = 0.000; Fig. 31.5).

Cervical lymphadenopathy also constitutes an important prognostic factor. Koka et al showed a 29% survival with nodes versus 64% without nodes, which was supported by two subsequent meta-analyses.13 ,​ 14 ,​ 87 Distant metastases with locoregional control are relatively rare (≤ 10%) and carry a poor prognosis.15 ,​ 62

Fig. 31.4 Kaplan-Meier graph of disease-free survival according to orbital involvement. (Reproduced with permission from Lund VJ, Stammberger H, Nicolai P, et al, European position paper on endoscopic management of the nose, paranasal sinuses and skull base, Rhinology 2010; [Suppl 22], 105:86–100.)
Fig. 31.5 Kaplan-Meier graph of disease-free survival according to intracranial involvement. (Reproduced with permission from Lund VJ, Stammberger H, Nicolai P, et al, European position paper on endoscopic management of the nose, paranasal sinuses and skull base, Rhinology 2010; [Suppl 22], 105:86–100.)

31.6 Conclusion

ENB is a rare nasal tumor that has a unique natural history. Its management requires experience in histopathology, radiology, sinonasal surgery, and medical oncology. It is optimally treated by a combination of surgery and radiotherapy. Conventional craniofacial resection doubled survival compared with previous surgical treatments. However, endoscopic resection now offers an appropriate alternative in selected cases with minimal complications and morbidity. The role of chemotherapy remains to be determined, but genomic analysis may open up new avenues for treatment.88 Lifelong follow-up is required irrespective of the treatment, for the natural history of this tumor extends over a lifetime.

31.7 References

  • 5 Hyams VJ. Olfactory neuroblastoma. In: Hyams VJ, Baksakis JG, Michaels L, eds. Tumours of the Upper Respiratory Tract and Ear. Washington, DC: Armed Forces Institute of Pathology; 1998:240–248

  • 8 Magee PN, Montesano R, Preussmann R. N-nitroso compounds and related carcinogens. In: Searle CE, ed. Chemical Carcinogens. Washington, DC: USA American Chemical Society; 1976:491–625

  • 12 Lund V, Howard D, Wei W. Tumors of the nose, sinuses and nasopharynx. New York: Thieme; 2014

  • 26 Lund VJ, Stammberger H, Nicolai P, et al. European position paper on endoscopic management of tumours of the nose, paranasal sinuses and skull base. Rhinol Suppl. 2010;22:46–51

  • 29 Saade R, Roberts D, Ow Tj, et al. Prognosis and biology in esthesioneuroblastoma: staging versus grading dilemma—the MDACC experience. J Neurol Surg, Part B: Skull Base. Conference: 24th Annual Meeting North American Skull Base Society. Abstract 75, 2014

  • 34 Kunc M, Gabrych A, Czapiewski P, Sworczak K. Paraneoplastic syndromes in olfactory neuroblastoma. Contemp Oncol (Pozn). 2015;19(1):6–16

  • 37 Madani G, Beale T, Lund V. Imaging of sinonasal tumors.. Seminars in Ultrasound. CT and MRI 2009; 30 (1) 25-38 NOT_FOUND;INVALID_JOURNAL

  • 46 Howard DJ, Lund VJ, Wei WI. Craniofacial resection for tumors of the nasal cavity and paranasal sinuses: a 25-year experience. Head Neck 2006; 28(10):867–873

  • 53 Draf W, Schick B, Weber R, et al. Endoscopic micro-endoscopic surgery of nasal and paranasal sinus tumours. In: Stamm AC, Draf W, eds. Micro-Endoscopic Surgery of the Paranasal Sinuses and the Skull Base. Berlin, Germany: Springer; 2000:481–488

  • 88 Lechner M, Wells G, Steele C, et al. Clinical and mutation profiling of olfactory neuroblastoma, establishment of novel olfactory neuroblastoma cell culture model and results from drug screening ERS London 2018 Abstract

Only gold members can continue reading. Log In or Register to continue

Feb 8, 2021 | Posted by in NEUROSURGERY | Comments Off on 31 Esthesioneuroblastoma
Premium Wordpress Themes by UFO Themes