Radiotherapy of esthesioneuroblastoma

Int. J. Radiation Oncology Biol. Phys., Vol. 49, No. 1, pp. 155–160, 2001 Copyright © 2001 Elsevier Science Inc. Printed in the USA. All rights reserved 0360-3016/01/$–see front matter

PII S0360-3016(00)00811-7

CLINICAL INVESTIGATION

Head and Neck

RADIOTHERAPY OF ESTHESIONEUROBLASTOMA HANS TH. EICH, M.D.,* SUSANNE STAAR, M.D.,* OLIVER MICKE, M.D.,† PETER D. EICH, M.D.,‡ HARTMUT STU¨ TZER, PH.D.,§ AND ROLF-PETER MU¨ LLER, M.D.* Departments of *Radiotherapy, and §Medical Statistics, Biometrics, and Epidemiology, University of Cologne, Germany; Departments of †Radiotherapy and ‡Radiology, University of Muenster, Germany Purpose: Only 3% of all malignant intranasal tumors are esthesioneuroblastomas (ENB). As the tumor is very rare, the number of ENB treated in individual departments is small. In order to evaluate the efficacy of radiotherapy (RT), patients’ data of 2 centres were analysed with reference to new reports in literature. Methods and Materials: From 1981 to 1998, 17 patients with ENB, 8 men and 9 women aged between 6 and 81 years, were treated in the departments of radiotherapy of the universities of Cologne and Muenster. The tumors were Kadish Stage B in 4/17 patients and Stage C in 13/17 patients. Treatment included incomplete surgery and irradiation in 2/17 patients, adjuvant RT postoperatively in 6/17 patients, definitive RT in 7/17 patients and RT after incomplete surgery of recurrent tumors in 2/17 patients. Postoperatively, the median target dose of EBRT was 56 (range 50 – 60) Gy; for definitive RT it was 58 (range 40 –70) Gy. Results: After a median follow-up period of 86 (range 2–208) months 10/17 patients showed no evidence of disease (NED). There were 6 patients treated with radical complete surgery plus postoperative irradiation and 5 of them were NED. There were 7 patients treated with only irradiation and 3 of those patients were NED. Of 2 patients with incomplete surgery and irradiation there was one patient NED. Of 2 patients with incomplete resection of recurrent tumor who received irradiation, there was one patient NED. 2 of the patients with NED died after 22 and 94 months respectively. 4/17 patients died as a result of local recurrence and 2/17 patients as a result of distant metastases (liver, brain). One patient with a recurrent tumor is alive. Median survival of all 17 patients was 94 months. Progressive disease after definitive RT occurred after a median of 11 months. Conclusions: Esthesioneuroblastomas are radiocurable tumors. In correlation to literature a primarily complete tumor resection followed by adjuvant RT (50 – 60 Gy) offers the best disease free survival. © 2001 Elsevier Science Inc. Esthesioneuroblastoma, Radiotherapy.

neuroepithelial tumors in mice (5). Mice from a transgenic line that expressed the human adenovirus type 12E1A and E1B genes tend to develop ENB at approximately 6 month of age. In most cases type C retrovirus particles were seen in the tumor rosettes (6). In the same study three cases of feline C-particle positive ENB’s were reported. Also in cats spontaneous olfactory neuroblastomas type C retroviral particles have been seen. These particles have been classified as Feline Leukaemia Virus (FeLV) by polymerase chain reaction and immunhistochemistry. It may be interesting that these cats did not show any other evidence of neoplasms or leukaemia (7). In the histopathologic diagnosis of ENB immunhistochemistry and electron microscopy are helpful (8). ENB may be confused with lymphoma, extramedullary plasmocytoma, undifferentiated carcinoma, malignant melanoma and rhabdomyosarcoma (9). The positive reaction for S-100 protein and the presence of NSE in combination with a negative stain for epithelial- and lymphoma marker espe-

INTRODUCTION Esthesioneuroblastoma (ENB) is a rare and uncommon tumor arising from the olfactory epithelium in the upper nasal cavity and often shows an intracranial extension. This tumor constitutes 3% of all intranasal neoplasms. ENB can be seen in all ages, with peaks in the second and sixth decades of life (1, 2). The distribution between the genders is roughly equal. As in all intranasal tumors, initial symptoms are non-specific and include nasal obstruction, epistaxis, cephalgia, hyposmia, exophthalmos and amaurosis in correlation to the tumor extension. Tumor etiology and pathogenesis is not yet clear (3). Vollrath et al. (4) induced a tumor in the regio olfactoria in rats by applying nitrosamine. These tumors were morphologically and histobiochemically equivalent to ENB. Further examinations showed aberrations in chromosome 1. These chromosome aberrations were interpreted as the basis for tumor induction in this animal model (4). Other authors discuss a virus genesis. Polyoma virus has been seen in Reprint requests to: Hans Theodor Eich, M.D., Department of Radiotherapy, University of Cologne, Joseph-Stelzmann Str. 9, D-50924 Ko¨ln, Germany; Tel.: ⫹49 221 478-5449; Fax: ⫺6158.

Accepted for publication 21 July 2000.

155

75 32 34 81 39 65 58 6 43 10 73 65 65 58 68 80 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Abbreviations: NED ⫽ no evidence of disease; Incomplete-Rec ⫽ incomplete surgery of recurrent tumor; mo ⫽ month; RT ⫽ radiation therapy.

14 63 84 18 85 7 12 155 8 6 8 5 10 123 – 93 36 15 65 208 22 86 8 57 161 33 7 8 6 11 125 2 94 39 – – – – – – brain (50 mo) – – – lymph node (2 mo) liver (6 mo) – – liver (0 mo) – lymph node (36 mo)

Metastases Chemotherapy Surgery Stage

– – – – – – – preoperative – – – palliative – – – – –

Disease-free survival (months) Overall survival (months) Gender

Between 1981 and 1998, seventeen patients with histologically confirmed ENB were treated in the departments of radiotherapy of the universities of Cologne and Muenster. Patients included nine women and eight men aged between 6 and 81 years (median 58 years). Clinical records and radiographic studies of the patients were reviewed. Documented data of the initial presenting symptoms, ENT examination, surgical intervention, radiation treatment protocols and treatment planning as well as the results and imaging of radiographs were analysed. The diagnosis of ENB was based on histopathological verification and the radiographic findings of the tumor localization. All patients had computed tomography (CT) initially and in recent years magnetic resonance imaging (MRI) (5/17) additionally. One patient showed liver metastases at the time of primary therapy. All patients were staged according the Kadish system (27), (Table 1, Table 2) retrospectively. This staging system has also been used by most other authors and a comparison with reported studies is possible. The follow-up period ranged from 2 to 208 months (median 86 months). Treatment included incomplete surgery and irradiation in 2/17 patients, adjuvant radiotherapy postoperatively in 6/17 patients, definitive radiotherapy in 7/17 patients and radio-

Patient

METHODS AND MATERIALS

Year of RT

cially suggest the classification as a neurogenic tumor (10, 11, 12). Since the first description of ENB by Berger and Luc (13) in 1924, about 945 cases have been reported in the world literature (2). As the tumor is rare the number of ENB treated in individual departments is small. Consequently large co-ordinated follow-up studies are difficult if not impossible. Almost all studies in literature are retrospective reports, often case reports (2, 9, 14, 15, 16, 17). Randomized treatment trials are not practical for such a rare tumor. This may also be a reason for the lack of generally accepted treatment recommendations for ENB. Optimal treatment is still controversial. The treatment of ENB reported in literature includes primary surgery, primary radiotherapy, combination of both modalities with pre- and postoperative radiotherapy, chemotherapy and high doses of chemotherapy followed by autologous marrow transplant (2, 15, 18 – 25). In order to evaluate the efficacy of radiotherapy, especially in multimodality therapy, patients’ data of two centres were analysed retrospectively with reference to new reports in literature.

No. of daily fractions

Tumor limited to the nasal cavity Tumor involving the nasal and paranasal cavities Tumor extends beyond the nasal and paranasal cavities

Total dose (Gy)

A B C

Age (years)

Extension

Table 2. Patient characteristics, treatment, and outcome

Type

Outcome

Table 1. Tumor stage definition according to Kadish et al. (Ref. 27)

NED NED Alive, Rec Died, NED NED NED Dead of metastases NED Dead of disease NED Dead of disease Dead of metastases Dead of disease NED Dead of disease Died, NED NED

Volume 49, Number 1, 2001

1998 1994 1982 1993 1991 1998 1990 1986 1981 1998 1985 1985 1986 1989 1993 1991 1996

Physics

33 25 30 31 25 30 33 22 25 34 29 35 33 31 16 25 30

●

59 50 60 56 50 60 60 44 50 61 58 70 66 56 40 50 55

Biology

Radical Radical Radical Radical Radical Radical Incomplete Incomplete Incomplete-Rec Incomplete-Rec – – – – – – –

●

B C B B B C C C C C C C C C C C C

I. J. Radiation Oncology

M M F F M M F M M F F M F F F F M

156

Esthesioneuroblastoma

Table 3. Presenting symptoms of esthesioneuroblastoma in 17 patients Symptoms

No. of patients

Nasal obstruction Epistaxis Headache Loss of sense of smell Face pain Diplopia Visual dysfunction Exophthalmos Mental change Facial mass Neck mass

9 7 6 3 2 2 2 1 1 1 1

therapy after incomplete surgery of recurrent tumors in 2/17 patients. In 6/17 patients a complete tumor resection could be achieved: 4 patients received a craniotomy and lateral rhinotomy (craniofacial resection), 2 patients a nasal tumor resection. These 6 patients received adjuvant radiotherapy postoperatively with a median dose to the target volume of 56 Gy (range 50 – 60 Gy). Dose per fraction was 1,8 or 2 Gy, median number of fractions 30. For the 11/17 patients who received definitive radiotherapy after biopsy (7/17) or incomplete surgery of primary and recurrent tumors (4/17), the median dose to the target volume was 58 Gy (range 40 –70 Gy), median dose per fraction 2 Gy (range 1,8 –3Gy) with a median number of 29 fractions (range 16 –35). Three locally advanced tumors received radiotherapy with palliative intention. 6/17 patients were irradiated with telecobalt, 11/17 patients with 5–10 MV photons of a linear accelerator. None of the patients had elective lymph node irradiation. Radiotherapy was delivered after complex individual treatment planning. In 14/17 patients a 3-field technique was used: two lateral wedged fields and one anterior field. 3/17 patients were treated with two fields: 2/3 patients with oblique lateral fields, 1/3 patients with two lateral opposed fields. The irradiated volume varied upon the tumor extension and localization including a safety margin of 1–2 cm. The planning target volume was covered by the 90% isodose. Response to therapy was defined by ENT examination and CT or MRI, and classified as complete response, partial response, stable disease and progressive disease. Survival was calculated according to Kaplan-Meier (26) from the date of primary therapy, whereas the disease free survival times were calculated from the last day of the given radiotherapy.

● H. T. EICH et al.

157

evidence of disease (NED). There were 6 patients treated with radical complete resection plus postoperative irradiation and 5 of them were NED. There were 7 patients treated with only irradiation and 3 of those patients were NED. Of 2 patients treated with incomplete surgery and irradiation there was one patient NED. Of 2 patients with incomplete resection of recurrent tumor who also received irradiation, there was one patient NED. 2 of the patients with NED died after 22 and 94 months respectively (patients no. 4 and 16) (Figure 1a, 1b); 4/17 patients died as a result of a local recurrence, and 2/17 patients as a result of distant metastases (patients no. 7 and 12). One patient with recurrence is still alive since 208 months. Median survival calculated according to Kaplan-Meier of all 17 patients was 94 months (95% confidence interval 9,8; 178,2) (Figure 2). During follow-up, 4/17 patients (23%) developed nodal or distant metastases (liver, brain). Progressive disease after definitive radiotherapy occured after a median of 11 months, whereas in some cases a long lasting remission could be observed. One 58-year old woman with a tumor Kadish Stage C is in complete remission 125 months after radiotherapy with a total dose of 56 Gy (patient no. 14). 2/17 patients received chemotherapy (Table 2). One patient (no. 12) developed fulminant liver metastases 6 months after the end of radiotherapy and received palliative chemotherapy with a combination of cisplatin, endoxan, adriamycin and methotrexate. This patient died of progressive liver metastases but without recurrence at the primary site. A 6-year old boy (patient no. 8) received polychemotherapy preoperatively. 2 cycles of ACVD (adriamycin, cisplatin, vincristine, DTIC), 2 cycles of PCV (DDP, cisplatin, VM 16) and 3 cycles of IVP (ifosphamide, VP 16). After incomplete surgery he was irradiated with a total dose of 44 Gy, 5 times per week with 2 Gy daily. Due to tumor extension into the orbit and residual tumor postoperatively, it was not possible to spare the orbits. After a follow-up period of 155 months the patient is alive with no evidence of disease. Patient no. 8 showed remarkable side effects of treatment. Acute reactions included erythema, edema of the left eye and reversible cilia loss followed later by cataract and radiation induced retinopathia combined with an impaired vision. 13 years after multimodality therapy he showed a slight face asymmetry. Further sequelae of radiotherapy complained of by other patients were a reversible impared sense of taste, xerostomia and irreversible hyposmia. Optic nerve damage, blindness secondary to radiotherapy, osteoradionecrosis and necrosis of the frontal brain could not be observed. DISCUSSION

RESULTS Patients’ characteristics, treatment modalities and radiotherapy data are summarized in Table 2 in detail. Presenting symptoms at the time of diagnosis are listened in Table 3. The median interval from the onset of symptoms until diagnosis was 5 months (range 1–35 months). After a median follow-up of 86 months (range 2–208 months) 10/17 patients showed no

The use of radiotherapy as the only treatment modality for ENB has already been described by Berger in 1924 in the first documented case of ENB (13). In the following years surgical methods of treatment were progressively used as an alternative to radiotherapy (2). However, an epidemiological study at the Mayo Clinic between 1951 and 1990 showed no significant difference in survival between patients treated with surgery

158

I. J. Radiation Oncology

●

Biology

●

Physics

Volume 49, Number 1, 2001

Fig. 1a. Esthesioneuroblastoma in Kadish stage C at diagnosis (patient no. 16, Table 2) Fig. 1b. Follow-up computed tomography 7 months after definitive radiotherapy (50 Gy)

alone and those receiving radiotherapy (20). An increase in local tumor control of ENB was achieved by combining surgical techniques and radiotherapy, which was used pre- as well as postoperatively. After surgery only the rates of local recurrences quoted in the literature are between 44 – 86%, whereas surgery plus radiotherapy resulted in reduced local failures (0 – 40%) (Table 4) (1, 8, 20, 27, 28). Eden et al. did not find a significant difference in survival comparing pre- and postoperative radiotherapy. However, they were able to show improved local tumor control following preoperative radiotherapy (19). Spaulding et al. reported an improvement in survival in a group of 34 patients after preoperative radiotherapy. The 5-/10-year survival rates were 81% and 54.5%. Through the introduction of more sophisticated surgical techniques such as craniofacial resection and more complex beam arrangements (3-field techniques) using megavoltage radiation in the middle of the seventies, Spaulding showed an improvement in the 2-year survival rate from 70% to 87% in 30 patients when comparing the treatment periods of 1969 –1975 and 1976 – 1990 (29). At present there is neither a generally accepted primary concept for treatment of ENB nor there are generally accepted recommendations for radiotherapy. However, when

choosing the most adequate therapy for the individual patient, the staging classification by Kadish, which is proven as the most important prognostic factor, should be taken into account (Table 1), (1, 27). The Kadish staging system underwent multiple modifications, the advantages and disadvantages of which will not be discussed here (2, 22, 28, 30). Since in the literature the original Kadish classification is predominantly employed, we have deliberately chosen this classification for reasons of comparability. Skolnik recommend the use of radiotherapy in Kadish stage A and B, and in stage C the combination of surgery and radiotherapy (2, 23). Urdaneta (31) and Broich (2) advocate a combined approach for all stages. In general, total reference doses of 45 Gy preoperatively and 50 – 60 Gy postoperatively are recommended. For definitive radiotherapy, doses of 60 –70 Gy should be applied (23). A second essential prognostic factor defining the individual radiotherapy approach is the histopathological grading according to the Hyam’s system (32, 33). Morita et al. (22) showed in a retrospective analysis of 49 patients with ENB that the 5-year survival rate for patients with low-grade tumors (grade I, II) was 80% compared to only 40% for patients with high-grade tumors (grade III, IV). The 5-year

Esthesioneuroblastoma

● H. T. EICH et al.

159

Table 4. Local recurrence First author (Ref. no.) Elkon (1) Kadish (27) Foote (20) O’Connor (8) Dulgherow (28) Eich

Fig. 2. Overall survival in esthesioneuroblastoma (n ⫽ 17) according to Kaplan-Meier

survival of all patients was 69%. The authors recommend to adjust therapy according to grading: Surgery alone is considered effective for low-grade tumors, if there are tumor free resection margins. Radiotherapy is recommended for low-grade tumors borderline resected, residual or recurrent tumors, and for all high-grade tumors. High-grade tumors should be treated with additional chemotherapy due to their adverse prognosis (22). Presently, chemotherapy is not considered as first line treatment (2). It is used increasingly in patients with unresectable primary tumors, recurrences and distant metastases. The agents used most frequently are cisplatin, etoposide, adriamycin, cyclophosphamide, vincristin, 5-fluorouracil, doxorubicin, and thiothepa. In individual cases there are data about the successful use of preoperative chemotherapy followed by surgery and radiotherapy in patients with stage C disease (2). Patient no. 8 also showed benefits from the use of preoperative chemotherapy. Furthermore there have been reports about the use of high dose chemotherapy followed by autologous bone marrow transplant (2, 25). Since the incidence of cervical lymph node metastases in early stage of disease is less than 10%, elective neck irradiation or neck dissection is generally not recommended (23). In

Surgery alone

Surgery and radiotherapy

44% 50% 59% 75% 86% –

0% 0% 8% 29% 40% 17%

contrast, patients with Kadish stage C disease developed cervical lymph node metastases in up to 44% (2, 23) and in the presented group of Stage C patients 15% (2/13). Therefore, we suggest that in advanced disease stages radiotherapy of the cervical lymph nodes, radical neck dissection or a combination of both should be considered. Contrast agent enhanced CT and/or MRI currently are the imaging modalities used to classify tumor stage and choose the therapy of ENB (34, 35). Based on CT-scan or MRI computerized treatment planning, inhomogeneities between bony structures and the air containing nose and paranasal sinuses, as well as the topographic vicinity of the target volume to the olfactorial area, eyes, optic nerve, optic chiasm, pituitary and frontal lobes, have to be taken into account. In the presented group of patients we were able to achieve a homogenous dose distribution within the target volume in 13/17 patients with advanced tumors using conventional 3-field techniques, i.e. two lateral opposing wedged fields and an anterior field. An alternative to conventional 3-field techniques has been the introduction of 3-D radiation treatment planning. It allows an individual dose adjustment to the irregularly shaped target volume using (micro-) multileaf collimators and noncoplanar beam arrangements. The target volume can be homogeneously encircled by the 90% isodose. Whether the use of 3-D planning can lead to an improvement of treatment results of ENB or to a further reduction of side effects remains to be seen. The retrospective analysis of our results verify that ENB is a radiocurable tumor (2, 13, 23). According to literature and our data, complete surgical resection should be attempted followed by adjuvant radiotherapy with 50 – 60 Gy. Furthermore we would recommend the use of radiotherapy alone especially for palliative treatment. Using this approach, long lasting remissions up to 125 months could be approached. Careful irradiation treatment planning is necessary to avoid severe complications in the regions of nose, orbitae and the frontal brain. Even with multimodality treatments ENB have to be considered with caution in view of their potential for metastases.

REFERENCES 1. Elkon D, Hightower SI, Lim ML, et al. Esthesioneuroblastoma. Cancer 1979;44:1087–1094. 2. Broich G, Pagliari A, Ottaviani F. Esthesioneuroblastomas: a general review of the cases published since the discovery of the tumor in 1924. Anticancer-Res 1997;17:2683–2706.

3. Herrold KM. Induction of olfactory neuroepithelial tumors and syrian hamsters by diethylnitrosamine. Cancer 1962;17: 114 –121. 4. Vollrath M. Chemically induced esthesioneuroblastoma: a cytogenic, cell culture and biochemical investigation with im-

160

5.

6. 7. 8. 9. 10. 11.

12.

13. 14.

15. 16. 17. 18. 19. 20.

I. J. Radiation Oncology

●

Biology

●

Physics

plications for tumor histogenesis. Arch Otorhinolaryngol 1989;246:71–78. Dawe CJ. Neoplasms induced by Polyomavirus virus in the upper respiratory Tract of mice. In: Muir CS, Shanmugaratnam K editors. Cancer of the Rhinopharynx: Copenhagen, Ejanar Munksgaard; 1967. p. 179 –196. Koike K, Jay G, Hartley JW, et al. Activation of retrovirus in transgenic mice: association with development of olfactory neuroblastomas. J Virology 1990;64:3988 –3991. Schrenzel MD, Higgins RJ, Hinrichs SH, et al. Type C retroviral expression in sponataneous feline olfactory neuroblastomas. Acta Neuropathol 1990;80:547–553. O’Connor TA, McLean P, Juillard GJF, et al. Olfactory neuroblastoma. Cancer 1989;63:2426 –2428. Slevin NJ, Irwin CJ, Banerjee SS, et al. Olfactory neural tumors—the role of external beam radiotherapy. J Laryngol Otol 1996;110:1012–1016. Hirose T, Scheithauer BW, Lopes MB, et al. Olfactory neuroblastoma. An immunohistochemical, ultrastructural, and flow cytometric study. Cancer 1996;77:1957–1959. Taxy JB, Bharani NK, Mills SE, et al. The spectrum of olfactory neural tumors: a light-microscopic immunohistochemical and ultrastructural analysis. Am J Surg Pathol 1986; 10:687– 695. Wick MR, Stamley SJ, Swanson PE. Immunohistochemical diagnosis of sinonasal melanoma, carcinoma, and neuroblastoma with monoclonal antibodies HMB-45 and anti-synaptophysin. Arch Pathol Lab Med 1988;112:616 – 620. Berger RL, Luc R. L’esthesioneuroepitheliome olfactif. Bull Assoc Franc Pour L’Etude Cancer 1924;13:410 – 420. Bhattacharyya N, Thornten AF, Joseph MP, et al. Successful treatment of esthesioneuroblastoma and neurocrine carcinoma with combined chemotherapy and proton radiation. Arch Otolaryngol Head Neck Surg 1997;123:34 – 40. Guedea F, Van Limbergen E, Van den Bogaert W. High dose level radiation therapy for local tumor control in esthesioneuroblastoma. Eur J Cancer 1994;30:1757–1760. Herzog J, Schmidt B, Petersen D, et al. Neuroblastoma of the olfactory nerve—Radiotherapeutic experiences with six patients. Strahlenther Onkol 1988;164:515–526. Jehunen AP, Kairemo KJ, Lehtonen HP, et al. Treatment of olfactory neuroblastoma. A report of 11 cases. Am J Clin Oncol 1996;19:375–378. Bilsky MH, Kraus DH, Strong EW, et al. Extended anterior craniofacial resection for intracranial extension of malignant tumors. Am J Surg 1997;174:565–568. Eden BV, Debo RF, Larner JM, et al. Esthesioneuroblastoma. Long-term outcome and patterns of failure—the University of Virginia experience. Cancer 1994;73:2556 –2562. Foote RL, Morita A, Ebersold MJ, et al. Esthesioneuroblastoma: the role of adjuvant radiation therapy. Int J Radiat Oncol Biol Phys 1993;27:835– 842.

Volume 49, Number 1, 2001

21. Mc Elroy EA, Buckner JC, Lewis JE. Chemotherapy for advanced esthesioneuroblastoma: the Mayo Clinic experience. Neurosurgery 1998;42:1023–1027. 22. Morita A, Ebersold, MJ, Olsen KD, et al. Esthesioneuroblastoma: prognosis and management. Neurosurgery 1993;32:706 –714. 23. Perez CA, Clifford Chao KS. Unusual nonepithelial tumors of the head and neck. In: Perez CA, Brady LW editors. Principles and Practice of Radiation Oncology, 3rd edition. Philadelphia: Lippincott-Raven Publishers; 1998. p. 1111–1116. 24. Polin RS, Sheehan JP, Chenelle AG, et al. The role of adjuvant treatment in the management of esthesioneuroblastoma: the University of Virginia experience. Neurosurgery 1998;42: 1029 –1037. 25. Stewart FM, Lazarus HM, Levine PA, et al. High-dose chemotherapy and autologous marrow transplantation for esthesioneuroblastoma and sinonasal undifferentiated carcinoma. Am J Clin Oncol 1989;12:217–221. 26. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457– 484. 27. Kadish S, Godman M, Wang CC. Olfactory neuroblastoma: a clinical analysis of 17 cases. Cancer 1976;37:1571–1576. 28. Dulguerow P, Calcaterra T. Esthesioneuroblastoma: the UCLA experience 1979 –1990. Laryngoscope 1992;102:843– 848. 29. Spaulding CA, Kranyak Ms, Constable WC, et al. Esthesioneuroblastoma: a comparison of two treatment eras. Int J Radiat Oncol Biol Phys 1988;15:581–590. 30. Biller HF, Lawson W, Lachdev V, et al. Esthesioneuroblastoma. Surgical treatment without radiation. Laryngoscope 1992;102:843– 848. 31. Urdaneta N, Fischer JJ, Knowlton A. Olfactory neuroblastoma: Observations on seven patients treated with radiation therapy and review of the literature. Am J Clin Oncol 1988; 11:672– 678. 32. Hyams VJ: Olfactory neuroblastoma (case 6), In: Batsakis JG, Hyams VJ, Morales AR editors. Special Tumors of the Head and Neck. Chicago, American Society of Clinical Pathologists; 1983. p. 24 –29. 33. Hyams VJ. Tumors of the upper respiratory tract and ear. In: Hyams VJ, Batsakis JG, Michaels L editors. Atlas of Tumor Pathology, Second Series, Fascicle 25. Washington, DC, Armed Forces Institute of Pathology; 1988. p. 240 –248. 34. Hurst RW, Erickson S, Cail WS, et al. Computed tomographic features of esthesioneuroblastoma. Neuroradiology 1989;31: 253–257. 35. Schuster JJ, Phillips CD, Levine PA. MR of esthesioneuroblastoma (olfactory neuroblastoma) and appearance after craniofacial resection. Am J Neuroradiol 1994;15:1169 – 1177.