Esthesioneuroblastoma: Clinical Presentation, Radiological, and Pathological Features, Treatment, Review of the Literature, and the University of Virginia Experience

Neurosurg Focus 12 (5):Article 4, 2002, Click here to return to Table of Contents

Esthesioneuroblastoma: clinical presentation, radiological, and pathological features, treatment, review of the literature, and the University of Virginia experience

ROD J. OSKOUIAN, JR., M.D., JOHN A. JANE, SR., M.D., PH.D., AARON S. DUMONT, M.D., JONAS M. SHEEHAN, M.D., JEFFREY J. LAURENT, M.D., AND PAUL A. LEVINE, M.D. Departments of Neurological Surgery and Otolaryngology–Head and Neck Surgery, University of Virginia School of Medicine, Charlottesville, Virginia

Esthesioneuroblastoma is a rare and malignant upper nasal cavity neoplasm involving the anterior skull base. Treatment includes surgery, radiotherapy, chemotherapy, or a combination. The ideal treatment modality has yet to be determined. Esthesioneuroblastoma often lies in proximity to the optic nerves, optic chiasm, and the orbit. Resection risks damaging these critical structures, and radiotherapeutic techniques, similar to those applied for paranasal sinus tumors, may damage these vital structures and result in late sequelae such as blindness and cortical necrosis. Management strategies for this neoplasm lack uniformity, and there is no universally accepted staging system. In this paper the authors discuss the clinical presentation, radiological and pathological features, and treatment of this rare, malignant skull base neoplasm, as well as review the literature. They also present their results and treatment regimen, which includes preoperative radio- and chemotherapy or 1) craniofacial resection if the lesion has a significant intrac- erebral component, or 2) frontal sinus resection if little intracranial extension exists.

KEY WORDS • esthesioneuroblastoma • radiation therapy • chemotherapy • skull base tumor

L’esthesioneuroepitheliome olfactif, a term coined by Since the original papers,2,3 numerous studies have been Berger and Luc3 in 1924, was originally used to describe published, in each of which the patient population was a nasal mass that exhibited cellular rosettes and fibrils. limited and the treatment regimens multiple.29,62 Although The astonishing observations made by these authors were the data reported in these studies have been of great inter- the recognition of the superior location of the mass in the est over the years, the clinical and scientific merits have anterior nasal cavity and realization that it developed from been diminished because of the small population sizes and the olfactory placode. Two years later Berger and Cou- the varied treatment protocols. tard2 reported a similar but slightly different case, in which the tumor was termed esthesioneurocytome olfactif, and in 1929, a third variant was named esthesioneuroblastome.55 These two initial reports were followed by a slew CLINICAL PRESENTATION of others in which this most elusive tumor of the superior Because sinonasal malignant lesions are rare and diffi- nasal vault was characterized. Based on the classification cult to distinguish from their benign counterparts, their of McCormack and Harris,45 some authors initally be- diagnosis is challenging. These similarities between be- lieved that there were three types of esthesioneuroblas- nign and malignant lesions at initial presentation may lead tomas. McCormack and Harris classified the tumor as an to a significant delay in the diagnosis of malignancy. esthesioneuroepithelioma if it exhibited rosettes, an esthe- Given the absence of classic symptoms, patients are often sioneuroblastoma if it had pseudorosettes, or an esthesio- not examined until they present with nasal obstruction, neurocytoma if neither rosettes nor pseudorosettes were headache, nasal pain, excessive lacrimation, epistaxis, an- present. osmia, and visual changes.16 The most common complaint is longstanding unilateral nasal obstruction that has been Abbreviations used in this paper: CSF = cerebrospinal fluid; present for months, sometimes years, before presentation. CT = computerized tomography; MR = magnetic resonance; The disease is initially demonstrated with a bimodal age PNET = primitive neuroectodermal tumor; pPNET = peripheral distribution; it peaks first between age 10 and 20 years and PNET; TNM = tumor node metastasis; UVA = University of later between age 50 and 60 years, and is slightly more Virginia. common in females.7 The tumor can spread submucosally

Neurosurg. Focus / Volume 12 / May, 2002 1

Unauthenticated | Downloaded 09/25/21 01:13 AM UTC R. J. Oskouian, et al.

in all directions, thereby involving the paranasal sinuses, perintense to gray matter on T2-weighted images. Mul- 1 nasal cavity, and surrounding structures. These lesions tiplanar T2-weighted MR images provide excellent con- can cross the cribriform plate and invade the brain or seed trast between the very high–intensity signal present after the cerebrospinal fluid.61 Such lesions often require resec- sinus fluid or mucosa obstruction and the intermediate- tion via a craniofacial approach. This uncommon neo- intensity signal of the adjacent tumor. plasm is locally aggressive and can metastasize by lym- In assessing CT and MR imaging studies, it is some- phatic and hematogenic routes. Esthesioneuroblastoma is times difficult to determine whether the dura is invaded by an aggressive malignant tumor derived from the special- tumor. Because of the thin fovea and cribriform complex, ized neuroepithelium of the upper nasal cavity; it accounts the dura is often the significant barrier to tumor penetra- for 3% of all endonasal neoplasms.7 In cases of metastatic tion into the brain. Notwithstanding both neuroimaging disease, 20 to 30% will involve the CNS. In these cases studies, it may not be certain that the dura is involved until the neoplasm invades the CNS by growing through the surgery. As a result, the team must be prepared to resect cribriform plate and invading the anterior cranial base. dura and replace it with autogenous fascia, temporalis, or When it has entered the CNS, esthesioneuroblastoma fascia lata, for example, when performing craniofacial re- metastasizes to either the brain parenchyma or lepto- section. Computerized tomography and MR imaging are meninges. complementary examinations performed preoperatively to define the extent of tumor, staging, and surgical approach. The primary advantage of CT studies is that they allow the RADIOLOGICAL FEATURES team to evaluate osseous involvement around the nasal 16 Contrast-enhanced axial and direct coronal CT scan- septum, orbit, and anterior skull base. ning is performed in all patients to obtain all necessary preoperative information, such as proximity to the adjacent sinuses, in Kadish Stage A or Stage B tumors31 (Table PATHOLOGICAL CHARACTERISTICS 1). The protocols for CT scanning must include axial and Since its first descriptions,2,3 the histogenesis of this coronal scans of 1- to 5-mm-thick slices with intravenous neoplasm has been of debate. This controversy has given contrast. Esthesioneuroblastomas are solid and enhancing rise to a variety of names including esthesioneuro- nasal cavity masses that may manifest erosion into nearby blastoma, olfactory neuroepithelioma, olfactory neuro- osseous structures of the orbital plate of the ethmoid bone, blastoma, and others. Several anatomical origins of esthe- cribriform plate, and fovea ethmoidalis (Fig. 1). In cases sioneuroblastoma have been suggested, including the in which CT scanning reveals nasal roof erosion evidence sympathetic fibers of the anterior nasal cavity, the neu- of intracranial extension, we also recommend acquiring roectodermal cells of the sphenopalatine ganglion, the MR images because they provide a more sensitive assess- 10,12,13,49, 60 olfactory placode, and the vomeronasal organ. ment for unrecognized intracranial lesions. Intralesional 56,58,64 It was not until Obert, et al.,50 demonstrated that the calcification observed on neuroimaging can be considered 42 tumor originated from the superior nasal vault that agree- pathognomonic for esthesioneuroblastoma, and the pres- ment was needed and it was concluded that the tumor was ence of cysts along the intracranial margins of Kadish of neuroectodermal origin and more from the olfactory Stage C lesions in which there is intracranial extension 67 epithelium. Early attempts to categorize these neoplasms also yields a definitive diagnosis. based on gene expression at a molecular level have not In Kadish Stage C disease, it is important to perform T1- correlated with clinical outcomes. weighted gadolinium-enhanced MR imaging to obtain es- Esthesioneuroblastomas are thought to arise from cells sential additional information, particularly when walls in the olfactory epithelium. These malignant cells are of the bones are thinned or eroded. Fat-suppression T1- mitotically active and are precursor cells that give rise to weighted imaging visualizes the tumor more clearly. On sustentacular and neuronal cells. On high-power micro- T1-weighted MR images esthesioneuroblastomas present scopic examination they are shown to have a vascular as homogeneously enhancing tumors with intermediate stroma with fibrovascular components and pseudorosettes signal intensity, whereas on T2-weighted images the orig- (Fig. 2). The tumor is histologically similar to malignant inal intensity is increased. Occasionally in cases in which lesions of the sympathetic ganglia, adrenal medulla, and there is intracranial extension tumor, cysts at the tumor the retina. Common features are small, round neuroep- margins are also present. The main advantage of an MR ithelial cells arranged in rosette or pseudorosette patterns, imaging is its ability to delineate intraorbital and intrac- separated by fibrous elements. Rosettes consist of a cen- erebral extension. The tumor appears as hypointense to tral space ringed by columnar cells with radially oriented 16 gray matter on T1-weighted images and isointense or hy- nuclei. In the past considerable effort was expended to define histological subtypes. The authors of numerous studies have confirmed, however, that the histological 7 TABLE 1 subtype has no bearing on clinical behavior or prognosis. Immunohistochemical and histological examinations Kadish esthesioneuroblastoma staging system show that esthesioneuroblastoma is of neuroendocrine cell Stage Description origin and that a broad panel of antibodic agents is necessary to verify the diagnosis.14 A neural derivation is indi- A tumor limited to the nasal cavity cated ultrastructurally by the presence of dense neuro- B tumor involves the nasal & paranasal cavities secretory granules and neuritic processes within tumor C tumor extends beyond the nasal and paranasal cavities cells.16 In summary, a diagnosis of esthesioneuroblasto-

2 Neurosurg. Focus / Volume 12 / May, 2002

Unauthenticated | Downloaded 09/25/21 01:13 AM UTC Review of esthesioneuroblastoma

Fig. 1. Axial and coronal postcontrast CT scans. Left and Center: Scans obtained in a patient with a homogeneous soft-tissue mass with relatively uniform enhancement centered in the superior nasal cavity (typical feature). Bone erosion is frequent and usually accompanied by molding of bone. Right: Postopertive scan demonstrating results of enbloc tumor resection following craniofacial resection. Craniofacial reconstruction as well as coronal and axial CT scans allow visualization of any intracranial extension. The resection cavity shows that the entire cribriform plate and crista galli have been removed. mas is determined using a combination of factors, includ- a histopathological diagnosis is essential before initiating ing light microscopic appearance, and a battery of immu- treatment for esthesioneuroblastoma. A biopsy specimen nohistochemical stains, (chromogranin, synaptophysin, should not be obtained until all neuroimaging studies have S-100 protein, and neuron-specific enolase).23 been performed to avoid swelling effects and the inadver- tent biopsy sampling of other nasal tumors of neurogenic Genetic and Molecular Analysis origin. Developmental biologists have shown that the human Multimodality treatments are the most frequently advo- equivalent of HASH-1, a marker expressed in the imma- cated interventions, and craniofacial resection is the most ture Drosophila fly olfactory neurons, was expressed in common therapy. Craniofacial resection has evolved into the best surgical procedure for achieving safe, en bloc six primary and one metastatic specimens of esthesio- 35,37–40 neuroblastoma; the olfactory master protein, a protein ex- resection of disease. In 1959 with treatment focused pressed in mature olfactory neurons only, was not. Based on surgery as the primary modality and with the knowl- edge that the incidence of tumor recurrence was signifi- on the results of this study, the authors suggested that es- 41 thesioneuroblastoma might develop from immature, cant, Malecki described for the first time the concept of rather than mature, olfactory neurons.8 the en bloc resection, rather than fragmented excision, of The authors of genetic and molecular studies have tumors of the superior nasal vault with adequate margins. Despite this report, however, it was not until the 1970s shown that esthesioneuroblastoma is a member of the Ew- 33 9 ing sarcoma and PNET family. The translocation com- that Ketcham, et al., and then Clifford popularized cran- monly seen in Ewing sarcoma and PNETs (EWS/FLI1 iofacial resection in which the cribriform plate was re- genes) has been confirmed in esthesioneuroblastoma moved to prevent the perineural extension of the tumor to cell lines.68 Alterations of chromosome 1 in pPNETs have the frontal lobes along the olfactory rootlets. In the 1970s been described. Ninety-five percent of pPNETs show at our institution, the senior author, (J.A.J.) and his col- chromosomal translocations that are considered specific league G. S. Fitz-Hugh, M.D. in the Department of Oto- for this class of tumor [t(11;22)(q24;q12) or t(21;22) laryngology–Head and Neck Surgery, began to advocate a (q22;q12)], and these translocations fuse a portion of the multidisciplinary craniofacial approach and resection; in this procedure the goal was to obtain an en bloc resection, EWS gene on chromosome 22 to either of the two Ets fam- 19,35,37–40 ily transcription factors [FLI1 on 11q24 (5, 6) or ERG on and the emphasis was on a team approach. In one 21q22].68 Consequently, these fusion genes express chi- study the disease-free survival rate improved from 37.5 to 82% when extracranial excision and craniofacial resection meric proteins capable of cell transformation and act as 37–39 aberrant transcription factors.27,51 Identification of fu- were compared. sion transcripts [EWSyFLI1 and EWSyERG] with re- verse-transcription polymerase chain reaction has formed SURGERY-RELATED NUANCES AND the basis a sensitive and specific diagnostic test for COMPLICATIONS pPNETs. Endocranial extension and the close relationship to the ethmoid roof and cribriform plate require a combined TREATMENT OF ESTHESIONEUROBLASTOMA transfacial and neurosurgical approach. A craniotomy is The importance of staging the extent of disease in ma- probably not justified for tumors classified as T1 by Biller, lignant neoplasms has long been recognized because the et al., (Table 2) in which there is clear radiological evi- stage usually correlates with prognosis.16,17,26 Establishing dence of a normal cribriform plate and upper ethmoid

Neurosurg. Focus / Volume 12 / May, 2002 3

Unauthenticated | Downloaded 09/25/21 01:13 AM UTC R. J. Oskouian, et al.

TABLE 2 Biller esthesioneuroblastoma staging system

Stage Description

T1 tumor in the nasal cavity & sinuses (except sphenoid) w/ or w/out erosion of anterior cranial fossa bone T2 periorbital or anterior cranial fossa extension T3 brain involvement w/ resectable margins T4 unresectable tumor

tracerebral bleeding or internal carotid artery injury, are exceptional but serious complications. Radiotherapy-related complications also are rare and include necrosis of the bone flap, retinopathy, postirradiation cataracts, and vision loss. Delayed complications include necrosis of the frontal bone flap, lacrimal drainage obstruction, and frontal sinus mucoceles. Some surgeons have prevented pneumocephalus in the immediate postoperative period by using a nasal trumpet in the ipsilateral nasal fossa and nasal packing to divert pressurized air.15 The cranial floor is repaired using vari- ous techniques, such as a pericranial flap, temporalis mus- cle, and fascia transposition, or a layer of fascia lata held with fibrin glue.30 These procedures have prevented the herniation of cranial contents into the nasal cavity and minimized the incidence of CSF leakage. In a series recently reported by our group,37 nine (25.7%) of 35 patients suffered central nervous system complications. These included one case of elevated intracranial pressure, two of pneumocephalus that required de- compressive surgery, one of a cerebrovascular accident that caused a temporary left hemiplegia, which complete- ly resolved, and five of CSF leaks, four of which were treated satisfactorily with lumbar peritoneal shunts and one of which required no therapy. In eight (22.9%) of the 35 patients orbital complications Fig. 2. Photomicrographs depicting characteristic cytologic were secondary to therapy. Three developed epiphora, two features of this uncommon neoplasm, including uniform, small, radiation-induced cataracts, one radiation-induced ker- round-to-oval cells with coarsely granular chromatin, multiple small nucleoli, prominent nuclear membranes, and scant cytoplasm atopathy, and two developed transient diplopia secondary with occasional Homer–Wright rosettes. H & E, original magnifi- to fourth cranial nerve dysfunction. Seven (20%) of the 35 cations 10 (upper) 100 (center), and 200 (lower). patients suffered systemic complications that were not life threatening (hyponatremia, diabetes insipidus, wound ser- oma, pulmonary embolus, hypothyroidism, amenorrhea, and prolactinoma) and for which they received treatment. cells exist. All other patients should undergo a transfacial In 22 cases the tumors were Stage C, and the patients approach combined with a bifrontal craniotomy.19,28,37–41, therefore underwent chemotherapy; chemotoxicity result- 43,44,53,59,63 A craniofacial resection permits en bloc resec- ed in (18.2%): one patient who suffered a myocardial in- tion of the tumor, better visualization of any intracranial farction, one who developed bilateral vocal cord palsies, extension, and protection of the brain and optic nerves. one who suffered peripheral neuropathy, and one who The resection should include the entire cribriform plate developed herpes zoster during his chemotherapy treat- and crista galli (Fig. 3). ment. Only three patients (8.6%) developed infectious The postoperative complications are infections, includ- complications postoperatively. In two patients infected ing abscess surrounding the flap and, less commonly, sub- bone flaps occurred, requiring removal, and in one patient dural or epidural abscesses.57 The closure of the anterior an epidural abscess developed, necessitating surgical skull base is difficult and not always watertight. As a re- drainage. There were no cases of meningitis. Cosmetic sult, meningitis can develop in some patients if there is a complications developed in five patients (14.3%): a sad- skull base defect and resultant CSF leak with the need for dle-nose deformity that necessitated a calvarial bone graft repair.30 Postoperative development of pneumocephalus in one case, a nasocutaneous fistula in one case, resorbed can also occur with resultant brain compression. Blind- frontal bone flaps in two cases, and enophthalmos sec- ness secondary to optic nerve injury as well as either in- ondary to radiotherapy in the fifth case.

4 Neurosurg. Focus / Volume 12 / May, 2002

Unauthenticated | Downloaded 09/25/21 01:13 AM UTC Review of esthesioneuroblastoma

Mayo Clinic the authors reported that there was no differ- ence in survival rates between patients treated with surgery and those with radiotherapy.20 In other studies authors have also shown that an increase in local tumor control was achieved by combining pre- and postoperative radiotherapy.69,70 The local tumor control rates were between 44 and 86% for surgery alone, whereas surgery combined with radiotherapy resulted in reduced local fail- ures of 0 to 40%.15,18,20,31 Adjuvant radiotherapy is thought to augment surgical 5-year cure rates from 10 to 15%; the 5-year survival rate for radiotherapy alone is reported to be 23% whereas for surgery combined with radiotherapy it is 44%.52 In light of these results, our treatment protocol has evolved from one of excision to multimodality treatment. The use of palliative high-dose irradiation alone for unresectable disease yields a 5-year survival rate of 10 to 15%.66 Therefore, radiotherapy alone as a primary treatment modality is no longer a recommended option except for palliation in patients with advanced disease.46 In 1998 Polin and colleagues54 reported that patients at the UVA exhibited a significant response to adjuvant therapy with or without chemotherapy. Moreover, they found that in patients in whom a response to adjuvant therapy was demonstrated preoperatively there was a statistically significant chance of being disease free.

CHEMOTHERAPEUTIC TREATMENT The role of chemotherapy in the management of patients with esthesioneuroblastoma is continually evolv- ing.48 Our philosophy in treating patients with Kadish Stage C tumors is based on the principle that these tumors share certain biological characteristics with other chemosensitive tumors of neural crest origin (for example, neuroblastomas, high-grade neuroendocrine carcinomas, and PNETs).4 In 1990 Goldsweig and Sundaresan22 reviewed Fig. 3. Coronal CT scan and photomicrographs. Upper: Co- 25 cases reported in the literature and found improvement ronal CT scan obtained in a 47-year-old man who presented with a in 19 of the 20 patients in whom chemotherapy alone was 2-year history of nasal obstruction and epistaxis. The CT scan 22 revealed a mass and extension into the ipsilateral cribiform plate undertaken for recurrence or metastatic disease. The most with intracranial extension. A direct coronal fine-cut 3-mm CT commonly used regimen includes cyclophosphamide with scan is the initial radiological study of choice. Esthesioneuroblas- vincristine. Since 1981, patients with Kadish Stage C tu- toma does not have a specific radiological appearance and is visu- mors have undergone preoperative chemotherapy at our alized as a homogeneous soft-tissue mass with uniform and mod- institution. Cyclophosphamide (650 mg/m2) and vincris- erate contrast enhancement. Computerized tomography studies are tine (1.5 mg/m2 with a maximum dose of 2 mg) were giv- essential for correct staging and should be evaluated carefully to en every 3 weeks for a total of six cycles. Esthesioneu- assess for erosion of the lamina papyracea, cribriform plate, and roblastomas are chemosensitive tumors. Treatment of fovea ethmoidalis. Obstruction of the sinus draining ostia results in an accumulation of nasal secretions, which tend to be difficult to Stage C disease is warranted, and multimodality therapy is differentiate from tumor tissue when viewed on a CT scan. Cen- reasonable in patients with metastatic disease. ter and Lower: Photomicrographs. Morphologically and ultrastructurally, esthesioneuroblastoma cells resemble those of sympathetic neuroblastomas. Both tumors show rosette formation and minimal REVIEW OF PUBLISHED SERIES cellular pleomorphism; on electron microscopy, dense-core cy- Skolnick, et al.,65 reported in a series of 50 patients in toplasmic granules can be identified. H & E, original magnifica- whom 5-year follow up was obtained. The 5-year survival tion 100 (center) and 200 (lower). rates were 64% for patients treated with surgery alone, 38% for those who underwent radiotherapy alone, and 50% for those who received combined therapy; the local ADJUVANT THERAPY recurrence rate was 48%. The other report of significance The use of radiotherapy as the only intervention in was published by Elkon, et al.,18 in 1979. Theirs was both cases of esthesioneuroblastoma was originally described a follow up to the report by Skolnick, et al., and a review by Berger, et al.3 In the subsequent years surgical treat- of an additional 76 cases described in the literature be- ment became more prevalent and was often undertaken as tween 1967 and 1977. They concluded that there was no the sole treatment for these neoplasms. In a study from the advantage to multimodality therapy in the management of

Neurosurg. Focus / Volume 12 / May, 2002 5

Unauthenticated | Downloaded 09/25/21 01:13 AM UTC R. J. Oskouian, et al. early disease and that although treatment of advanced dis- TABLE 4 ease provided palliation a cure did not occur. Homzie and Hyams histopathological grading system* Elkon24 surveyed the world literature and excluded some cases because they lacked the necessary information to LA Nuclear Preser- Mitotic Poly- Fibrillary evaluate their prognostic variables and 11 cases because Grade vation Index morphism Matrix Rosettes Necrosis adequate follow-up data were not available. The authors determined that the cure rate for patients treated after 1966 I + zero none prominent HW none was 64%, whereas that for those treated before 1966 was II + low low present HW none 23%. These data suggested that the changes in treatment III +/– mod mod low FW rare patterns ultimately affected cure rate and that those IV +/– high high absent none frequent patients who survived for at least 36 months had an excel- * FW = Flexner–Wintersteiner; HW = Homer–Wright; LA = lobular ar- lent chance of reaching the 5-year survival mark. chitecture; mod = moderate; + = present; +/– = present/absent. Traditionally, esthesioneuroblastoma is staged, as suggested by Kadish.31 These authors divided the lesions into icant method of predicting outcome compared with the three stages: Stage A in which the tumor is limited to the 31 nasal fossa; Stage B in which the tumor extends to the Kadish staging system. paranasal sinuses; and Stage C in which the tumor extends Treatment results reported in the past were most likely beyond the paranasal sinuses (Table 1). Currently, there flawed, probably by the inclusion of cases of sinonasal undifferentiated and neuroendocrine carcinoma (aggres- are no epidemiological, molecular, or pathological prog- 36 nostic indicators that are more valuable in predicting re- sive diseases associated with poor survival). In recent currance of disease or biological behavior than the Kadish series, the 5-year survival rate has varied from 50 to 80%, staging system.37 The main shortcoming of the Kadish with most studies indicating survival rates of greater than classification is that the language of defining Stage C is 70%. The Kadish system has a definitive prognostic too broad. Recognizing these inadequacies, a proposed value, and in patients harboring Stages A and B tumors, classification is based on the TNM system, which is pred- the 5-year survival rate appears to be 86%, whereas in those with Stage C tumors the 5-year survival rate appears icated on CT and MR imaging findings. The study of es- 39 thesioneuroblastoma has clearly suffered from a uniform, to be 72%. The most frequent recurrence is local, with effective staging system. Two other staging systems, the rates ranging from 20 to 40%. Although distant metastasis Biller method5 and the Dulguerov method,15 have also is not frequent (8%), it carries a very poor prognosis, and been described. In 1992 Dulguerov and Calcaterra15 criti- the presence of palpable lymph nodes is the most impor- cized the Biller staging system (Table 2) because cribri- tant prognostic factor for survival (for example, the 5-year survival is 0% in cases in which nodes are present and form plate involvement was assumed in all cases and be- 34 cause craniotomy was required for accurate staging. As a 65% in those in which they are not). In studies from the result, they proposed a modified TNM system based on Mayo Clinic the authors emphasize that the Hyams histo- CT and/or MR imaging findings (Table 3). pathological grade is the most important prognostic factor, 47 even though the results of other studies as well ours have Morita, et al., published a series of 49 patients treated 46 at the Mayo Clinic over a 39-year period; they performed failed to confirm this (Table 4). numerous treatment techniques that evolved over this long evaluation period. In seeking to determine prognostic fac- THE UVA MULTIDISCIPLINARY APPROACH tors, they found that the pathological grading system pro- AND RESULTS posed by Hyams25 (Table 4) provided a statistically signif- The treatment protocol undertaken in our patient population was the same as that previously reported by Levine, et al., in 198638 with use of the Kadish staging system. For TABLE 3 all stages of disease (Kadish Stages A–C), a preoperative 50-Gy dose of radiation was delivered because esthesio- Modified TNM staging system neuroblastomas are proven to be radiosensitive31,38 and be- Classi- cause preservation of the eye in sinonasal malignancies 44,53 fication Definition has been shown. In addition, patients harboring Kadish Stage C lesions, in whom the tumor extended into the or- tumor bit, or intracranially, or in whom cervical metastases were T1 tumor involving the nasal cavity &/or paranasal sinuses (ex- cluding sphenoid), sparing the most superior ethmoidal cells present, but not those with distant metastatic disease, T2 tumor involving the nasal cavity &/or paranasal sinuses (in- chemotherapy was undertaken (cyclophosphamide and cluding the sphenoid), w/ extension to or erosion of cribi- vincristine) prior to the radiotherapy, as previously de- form plate scribed.38,71 As in other institutions in which superior nasal T3 tumor extending into orbit or protruding into anterior cranial vault tumors are treated, a team including the otolaryngol- fossa, w/out dural invasion ogist–head and neck surgeon, neurosurgeon, neuropathol- T4 tumor involving the brain node ogist, radiation oncologist, and hematologist–oncologist N0 no cervical lymph node metastasis evaluated each patient and was instrumental in determin- N1 any form of cervical lymph node metastasis ing and delivering treatment. metastasis Our belief is that the majority of Grade IV esthesioneu- M0 no metastasis roblastomas, as described by Hyams,25 can also be classi- M1 distant metastasis fied as sinonasal undifferentiated carcinomas.21 Because

6 Neurosurg. Focus / Volume 12 / May, 2002

Unauthenticated | Downloaded 09/25/21 01:13 AM UTC Review of esthesioneuroblastoma of the considerably more aggressive clinical behavior of mean of 72 months (median 60 months) following defini- the sinonasal undifferentiated carcinoma,36 the outcomes tive diagnosis, whereas in the other 10 (28.6%) delayed of this grading analysis can be affected. cervical or distant metastatic disease developed at one or A retrospective analysis of the medical records obtained more sites during their lives. In 37% of the patients at least in patients treated at our institution between September one episode of metastatic disease eventually occurred. 1976 and May 1998 was performed. Although additional The longest duration before presentation with this first patients were evaluated and treated at our institution prior recurrence was 160 months. to this date, September 1976 reflects the time at which we Of those patients in whom delayed cervical metastases introduced craniofacial resection as the standard surgical developed, the mean time following diagnosis of this oc- therapy for superior nasal vault tumors. In a cohort of 45 currence was 84 months (range 42–124 months). Kaplan– patients, 10 were excluded because surgery had not been Meier survival curves were estimated for the percentage performed at our institution. The remaining 35 patients of disease-free survival by month of follow up. The dis- were found to have completed definitive curative therapy ease-free survival was 80.4% at 8 years, and the total at our institution and data were deemed appropriate for survival (inclusive of all diseases for this patient group) this analysis. The overall disease-free survival and total was 75.2%.39 survival rates for these patients were calculated. Kaplan– There were five episodes of distant metastasis in four Meier survival curves32 were estimated for the percentage patients, and these occurred a mean of 33.3 months fol- of disease-free survival and total survival, which included lowing diagnosis. In order of frequency the involved sites cause-specific death, as well as death of nonesthesioneu- were bone, lung, and prostate. In one patient a radiation- roblastoma causes. induced primary squamous cell carcinoma developed 153 Of 19 patients with Grade II tumors, seven developed months after treatment, occurring in the contralateral infe- recurrent disease, and of five with Grade III tumors, four rior turbinate.39 developed recurrences. Of the 19 Grade II tumors, there The mean follow-up period in these 35 patients was were six Stage B tumors and 13 Stage C; in two of the six 95.7 months. The 80.4% 8-year disease-free survival rate patients harboring Grade II/Stage B tumors local recur- proves the value of aggressive surgical therapy in con- rence/metastases developed, and in five of the 13 patients junction with adjuvant therapy (radiotherapy with or with- harboring Grade 2/Stage C lesions local recurrence/metas- out chemotherapy) in treating this malignant neoplasm. Of tases developed. One of the Grade III/Stage B lesions re- note 22 (62.9%) of the 35 patients presented with ad- curred, and two of the four Grade III/Stage C lesions vanced Stage C disease. This 80.4% 8-year disease-free recurred as well. The one patient with a Grade I/Stage C survival rate also includes 13 patients (37.1%) who devel- lesion developed no additional disease. We were unable to oped metastatic disease (two at presentation and 11 in a correlate any clinical significance with this pathological delayed fashion). grading system; however, we did find that the Kadish sys- The first metastases following completion of therapy tem was more clinically applicable, because the Stage C was 74.8 months (range 10–160 months). In light of this, disease in both Grades II and III lesions was more clini- these patients require a longer follow-up period. In the cally aggressive (Table 5). studies published by Kadish and colleagues in 1976, only Only two (6%) of the 35 patients in this series present- two patients with Stage C disease survived for 3 years or ed with cervical metastases, although in nine (25.7%) of more. In our series, 22 (62.9 %) of 35 patients presented 35 at least one episode of cervical metastasis occurred with Stage C disease. either prior to or after the completion of therapy. Of the seven in whom cervical disease developed following pri- CONCLUSIONS mary therapy, four sustained two to four episodes of cervical recurrences, despite surgical and adjuvant therapeu- We believe that esthesioneuroblastoma can be a highly tic intervention. This outcome is consistent with the data curable sinonasal malignancy, even in cases in which pa- published by Davis and Weissler,11 who showed, by cal- tients when present with advanced disease. Our results culating the cumulative published data derived from the and those of others indicate that surgery as well adjuvant studies of 10 authors (total number of patients 207), that therapy should be recommended for the treatment of the ultimate cervical metastasis rate was 27% in this com- esthesioneuroblastoma. We also believe that preoperative posite group. Also in this group of seven patients with radiotherapy can reduce the tumor burden, allowing a delayed recurrent cervical disease, four (57%) of seven greater chance of gross-total resection.37–39,54 Adjuvant developed additional metastatic disease at a later date. therapy should be considered in patients with advanced In five patients (14.3%) local recurrence developed a disease and unresectable disease. An extended follow-up period is required for these patients.

TABLE 5 References Initial Kadish stage determined in 35 patients at UVA 1. Bailey BJ, Barton S: Olfactory neuroblastoma. Management and prognosis. Arch Otolaryngol 101:1–5, 1975 Kadish Stage No. of Patients 2. Berger L, Coutard H: L’esthesioneurocytome olfactif. Bull Assoc Fr Etude Cancer 15:404–414, 1926 (Reference unver- A3ified) B10 C223. Berger L, Luc H, Richard R: L’esthesioneuro epitheliome olfactif. Bull Assoc Fr Etude Cancer 13:410–420, 1924

Neurosurg. Focus / Volume 12 / May, 2002 7

Unauthenticated | Downloaded 09/25/21 01:13 AM UTC R. J. Oskouian, et al.

4. Bhattacharyya N, Thornton AF, Joseph MP, et al: Successful 27. Jackson PK, Chevalier S, Philippe M, et al: Early events in treatment of esthesioneuroblastoma and neuroendocrine carci- DNA replication require cyclin E and are blocked by p21CIP1. noma with combined chemotherapy and proton radiation. Re- J Cell Biol 130:755–769, 1995 sults in 9 cases. Arch Otolaryngol Head Neck Surg 123: 28. Jane JA, Park TS, Pobereskin LH, et al: The supraorbital rim 34–40, 1997 approach: technical note. Neurosurgery 11:537–542, 1982 5. Biller HF, Lawson W, Sachdev VP, et al: Esthesioneuroblasto- 29. Jekunen AP, Kairemo KJ, Lehtonen HP, et al: Treatment of ma: surgical treatment without radiation. Laryngoscope 100: olfactory neuroblastoma. A report of 11 cases. Am J Clin On- 1199–1201, 1990 col 19:375–378, 1996 6. Boysen M, Lovdal O, Tausjo J, et al: The value of follow-up in 30. Johns ME, Winn HR, McLean WC, et al: Pericranial flap for the patients treated for squamous cell carcinoma of the head and closure defects of craniofacial resection. Laryngoscope 91: neck. Eur J Cancer 28:426–430, 1992 952–959, 1981 7. Broich G, Pagliari A, Ottaviani F: Esthesioneuroblastoma: a 31. Kadish S, Goodman M, Wang CC: Olfactory neuroblastoma. A general review of the cases published since the discovery of the clinical analysis of 17 cases. Cancer 37:1571–1576, 1976 tumor in 1924. Anticancer Res 17:2683–2706, 1997 32. Kaplan E, Meier P: Nonparametric estimation from incomplete 8. Carney ME, O’Reilly RC, Sholevar B, et al: Expression of the observations. J Am Stat Assoc 53:457–481, 1958 human Achaete-scute 1 gene in olfactory neuroblastoma (esthe- 33. Ketcham AS, Chretien PB, Van Buren JM, et al: The ethmoid sioneuroblastoma). J Neurooncology 26:35–43, 1995 sinuses: a re-evaluation of surgical resection. Am J Surg 126: 9. Clifford P: Transcranial approach to cancer of the antroeth- 469–476, 1973 moidal area. Clin Otolaryngol 2:115–130, 1977 34. Koka VN, Julieron M, Bourhis J, et al: Aesthesioneuroblas- 10. Cullen MJ, Cusack DA, O’Briain DS, et al: Neurosecretion of toma. J Laryngol Otol 112:628–633, 1998 arginine vasopressin by an olfactory neuroblastoma causing re- 35. Levine PA, Debo RF, Meredith SD, et al: Craniofacial resection versible syndrome of antidiuresis. Am J Med 81:911–916, 1986 at the University of Virginia (1976–1992): survival analysis. 11. Davis RE, Weissler MC: Esthesioneuroblastoma and neck me- Head Neck 16:574–577, 1994 tastasis. Head Neck 14:477–482, 1992 36. Levine PA, Frierson HF Jr, Stewart FM, et al: Sinonasal undif- 12. Devaney K, Wenig BM, Abbondanzo SL: Olfactory neuroblas- ferentiated carcinoma: a distinctive and highly aggressive neo- toma and other round cell lesions of the sinonasal region. Mod plasm. Laryngoscope 97:905–908, 1987 Pathol 9:658–663, 1996 37. Levine PA, Gallagher R, Cantrell RW: Esthesioneuroblasto- 13. Djalilian M, Zujko RD, Weiland LH, et al: Olfactory neuro- ma: reflections of a 21-year experience. Laryngoscope 109: blastoma. Surg Clin North Am 57:751–762, 1977 1539–1543, 1999 14. Du ZM, Li YS, Wang BF: Electron microscopic and immuno- 38. Levine PA, McLean WC, Cantrell RW: Esthesioneuro- histochemical findings in a case of olfactory neuroblastoma. J blastoma: the University of Virginia experience 1960–1985. Clin Pathol 46:83–85, 1993 Laryngoscope 96:742–746, 1986 15. Dulguerov P, Calcaterra T: Esthesioneuroblastoma: the UCLA 39. Levine PA, Scher RL, Jane JA, et al: The craniofacial resec- experience 1970–1990. Laryngoscope 102:843–849, 1992 tion—eleven-year experience at the University of Virginia: 16. Ebersold MJ, Olsen KD, Foote RL, et al: Esthesioneuroblasto- problems and solutions. Otolaryngol Head Neck Surg 101: ma, in Kaye AH, Laws ER Jr (eds): Brian Tumors: An Ency- 665–669, 1989 clopedic Approach. London: Churchill Livingstone, 1995, pp 40. Lund VJ, Howard DJ, Wei WI, et al: Craniofacial resection for 825–838 tumors of the nasal cavity and paranasal sinuses—a 17-year ex- 17. Eden BV, Debo RF, Larner JM, et al: Esthesioneuroblastoma. perience. Head Neck 20:97–105, 1998 Long-term outcome and patterns of failure—the University of 41. Malecki J: New trends in frontal sinus surgery. Acta Otolar- Virginia experience. Cancer 73:2556–2562, 1994 yngol 50:137–140, 1959 (Reference unverified) 18. Elkon D, Hightower SI, Lim ML, et al: Esthesioneuroblastoma. 42. Manelfe C, Bonafe A, Fabre P, et al: Computed tomography in Cancer 44:1087–1094, 1979 olfactory neuroblastoma: one case of esthesioneuroepithelioma 19. Fitz-Hugh GS, Allen MS Jr, Rucker TN, et al: Olfactory neu- and four cases of esthesioneuroblastoma. J Comput Assist To- roblastoma (esthesioneuroepithelioma). Arch Otolaryngol 81: mogr 2:412–420, 1978 161–168, 1965 43. Maniglia AJ: Indications and techniques of midfacial deglov- 20. Foote RL, Morita A, Ebersold MJ, et al: Esthesioneuroblasto- ing. A 15-year experience. Arch Otolaryngol Head Neck ma: the role of adjuvant radiation therapy. Int J Radiat Oncol Surg 112:750–752, 1986 Biol Phys 27:835–842, 1993 44. McCary WS, Levine PA, Cantrell RW: Preservation of the eye 21. Frierson HF Jr, Mills SE, Fechner RE, et al: Sinonasal undif- in the treatment of sinonasal malignant neoplasms with orbital ferentiated carcinoma. An aggressive neoplasm derived from involvement. A confirmation of the original treatise. Arch Oto- schneiderian epithelium and distinct from olfactory neuroblas- laryngol Head Neck Surg 122:657–659, 1996 toma. Am J Surg Pathol 10:771–779, 1986 45. McCormack LJ, Harris HE: Neurogenic tumors of the nasal 22. Goldsweig HG, Sundaresan N: Chemotherapy of recurrent fossa. JAMA 157:318–321, 1955 esthesioneuroblastoma. Case report and review of the literature. 46. McElroy EA Jr, Buckner JC, Lewis JE: Chemotherapy for Am J Clin Oncol 13:139–143, 1990 advanced esthesioneuroblastoma: the Mayo Clinic experience. 23. Hirose T, Scheithauer BW, Lopes MB, et al: Olfactory neuro- Neurosurgery 42:1023–1028, 1998 blastoma. An immunohistochemical, ultrastructural, and flow 47. Morita A, Ebersold MJ, Olsen KD, et al: Esthesioneuroblasto- cytometric study. Cancer 76:4–19, 1995 ma: prognosis and management. Neurosurgery 32:706–715, 24. Homzie MJ, Elkon D: Olfactory esthesioneuroblastoma—vari- 1993 ables predictive of tumor control and recurrence. Cancer 46: 48. Nguyen QA, Villablanca JG, Siegel SE, et al: Esthesioneuro- 2509–2513, 1980 blastoma in the pediatric age-group: the role of chemotherapy 25. Hyams VJ, Batsakis JG, Michaels L: Tumors of the Upper and autologous bone marrow transplantation. Int J Pediatr Respiratory Tract and Ear. Washington, DC: Air Force In- Otorhinolaryngol 37:45–52, 1996 stitute of Pathology, 1988, pp 240–248 49. Oberman HA, Rice DH: Olfactory neuroblastomas: a clinico- 26. Irish J, Dasgupta R, Freeman J, et al: Outcome and analysis of pathologic study. Cancer 38:2494–2502, 1976 the surgical management of esthesioneuroblastoma. J Otolar- 50. Obert GJ, Devine KD, McDonald JR: Olfactory neuroblastoma. yngol 26:1–7, 1997 Cancer 13:205–215, 1960 (Reference unverified)

8 Neurosurg. Focus / Volume 12 / May, 2002

Unauthenticated | Downloaded 09/25/21 01:13 AM UTC Review of esthesioneuroblastoma

51. Papadaki H, Kounelis S, Kapadia SB, et al: Relationship of p53 63. Shah JP, Galicich JH: Craniofacial resection for malignant tu- gene alterations with tumor progression and recurrence in ol- mors of ethmoid and anterior skull base. Arch Otolaryngol factory neuroblastoma. Am J Surg Pathol 20:715–721, 1996 103:514–517, 1977 52. Parsons JT, Mendenhall WM, Mancuso AA, et al: Malignant 64. Singh W, Ramage C, Best P, et al: Nasal neuroblastoma secret- tumors of the nasal cavity and ethmoid and sphenoid sinuses. ing vasopressin. A case report. Cancer 45:961–966, 1980 Int J Radiat Oncol Biol Phys 14:11–22, 1988 65. Skolnick EM, Massari FS, Tenta LT: Olfactory neuroepithe- 53. Perry C, Levine PA, Williamson BR, et al: Preservation of the lioma. Review of the world literature and presentation of eye in paranasal sinus cancer surgery. Arch Otolaryngol Head two cases. Arch Otolaryngol 84:84–93, 1966 (Reference un- Neck Surg 114:632–634, 1988 verified) 54. Polin RS, Sheehan JP, Chenelle AG, et al: The role of preoper- 66. Slevin NJ, Irwin CJ, Banerjee SS, et al: Olfactory neural ative adjuvant treatment in the management of esthesioneuro- tumours—the role of external beam radiotherapy. J Laryngol blastoma: the University of Virginia experience. Neurosurgery Otol 110:1012–1016, 1996 42:1029–1037, 1998 67. Som P, Lidov M, Brandwein M, et al: Sinonasal esthesioneu- 55. Portmann G, Bonnard M: Sur un cas de tumeur nerveuse des roblastoma with intracranial extension: marginal tumor cysts as foisse nasales (esthesioneuroblastome). Arch Otolaryngol 13: a diagnostic MR finding. AJNR 15:1259–1262, 1994 52–56, 1929 (Reference unverified) 68. Sorensen PH, Wu JK, Berean KW, et al: Olfactory neuroblas- 56. Ramsay HA, Kairemo KJ, Jekunen AP: Somatostatin receptor toma is a peripheral primitive neuroectodermal tumor related to imaging of olfactory neuroblastoma. J Laryngol Otol 110: Ewing sarcoma. Proc Natl Acad Sci USA 93:1038–1043, 1161–1163, 1996 1996 57. Richtsmeier WJ, Briggs RJ, Koch WM, et al: Complications 69. Terz JJ, Young HF, Lawrence W Jr: Combined craniofacial and early outcome of anterior craniofacial resection. Arch Oto- resection for locally advanced carcinoma of the head and neck laryngol Head Neck Surg 118:913–917, 1992 I. Tumors of the skin and soft tissues. Am J Surg 140:613–617, 58. Schall LA, Lineback M: Primary intranasal neuroblastoma: re- 1980 port of 3 cases. Ann Otol Rhinol Laryngol 60:221–229, 1951 70. Terz JJ, Young HF, Lawrence W Jr: Combined craniofacial re- (Reference unverified) section for locally advanced carcinoma of the head and neck II. 59. Schramm VL Jr, Myers EN, Maroon JC: Anterior skull base Carcinoma of the paranasal sinuses. Am J Surg 140:618–624, surgery for benign and malignant disease. Laryngoscope 89: 1980 1077–1091, 1979 71. Wade PM Jr, Smith RE, Johns ME: Response of esthesioneu- 60. Schuster JJ, Phillips CD, Levine PA: MR of esthesioneuroblas- roblastoma to chemotherapy. Report of five cases and review of toma (olfactory neuroblastoma) and appearance after craniofa- the literature. Cancer 53:1036–1041, 1984 cial resection. AJNR 15:1169–1177, 1994 61. Shaari CM, Catalano PJ, Sen C, et al: Central nervous system Manuscript received March 25, 2002. metastases from esthesioneuroblastoma. Otolaryngol Head Accepted in final form April 30, 2002. Neck Surg 114:808–812, 1996 Address reprint requests to: Rod J. Oskouian Jr., M.D., Depart- 62. Shah JP, Feghali J: Esthesioneuroblastoma. Am J Surg 142: ment of Neurological Surgery, University of Virginia, Box 800212, 456–458, 1981 Charlottesville, Virginia 22908. email: [email protected].

Neurosurg. Focus / Volume 12 / May, 2002 9

Unauthenticated | Downloaded 09/25/21 01:13 AM UTC