ESTHESIONEUROBLASTOMA Richard E. Davis, MD, and Mark C. Weissler, MD
Esthesioneuroblastoma (ENB) is a rare malignant neoplasm arising from the cribriform olfactory epithelium. Its aggressive biologic behavior is characterized by inapparent submucosal spread, local recurrence, atypical distant metastases, and poor long-term prognosis. Historically, the suspicion of neck metastasis has not been associated with ENB. A retrospective review of all ENB patients treated at the University of North Carolina Hospitals since 1972 has revealed an unusually high incidence of cervical metastases (four of four patients). This prompted a review of the ENB surgical literature. All English-literatureseries published since the widespread use of craniofacial surgery with accruals of eight or more patients were examined for reports of neck metastasis. Although the cumulative cervical metastatic rate reached 27% (55 of 207 patients), mosi authors failed to recognize the potential risk of cervical metastasis in ENB. Furthermore, in patients suffering from Kadish stage C disease, the cervical metastatic rate climbed to 44% (25 of 57 Kadish stage C patients). This further emphasizes the importance of the neck as a high-risk metastatic site. These findings suggest that cervical metastases may be an imponant consideration in the evalu-
From the Division of Otolaryngology-Head and Neck Surgery, University of North Carolina School of Medicine, Chapel Hill, North Carolina. Acknowledgments: The authors express their appreciation to William W. Shockley, MD, for his insightful editorial comments, and to Ms. Hope Merritt for her assistance in the preparation of this manuscript. Presented at the Southern Section Meeting of the American Laryngological, Rhinological and Otological Society, Sea Island, Georgia, January 17, 1992. Address all correspondence to Dr. Weissler at the Division of Otolaryngology-Head and Neck Surgery, CB X7070, 610 Burnett-Womack Building, University of North Carolina, School of Medicine, Chapel Hill, NC 27599-7070. Accepted for publication May 18, 1992, CCC 0148~6403/92/060477-06 @ 1992 John Wiley & Sons, Inc.
Esthesioneuroblastomaand Neck Metastasis
ation and treatment of patients suffering from ENB. Strategies for evaluation and management of the neck are discussed. HEAD & NECK 1992;14:477-482 0 1992 John Wiley & Sons, Inc.
Esthesioneuroblastoma (ENB)is a rare, aggressive malignancy of neural crest origin arising from the olfactory epitheli~rn.’-~With more than 200 cases reported by 1986,4ENB remains poorly understood, owing to both its infrequent occurrence and the lack of consensus regarding treatment approaches. The clinical diagnosis of ENB is complicated by vague and indistinct early symptoms often leading to delayed patient presentation. Eventually, the development of nasal obstruction and epistaxis, or less frequently, anosmia, epiphora, proptosis, or diplopia results in the discovery of a vascular, pinkish-gray, polypoid mass in the nasal ~ a u l tTumor .~~~ staging is then performed according to the Kadish criteria.6 In addition to delays in presentation, the diagnosis of ENB may also be complicated by the tumor’s nondescript microscopic appearance. As a result of its undifferentiated histology, ENB has been mistaken for various malignancies, including lymphoma, melanoma, undifferentiated carcinoma, and s a r ~ o m a . ~ This ’ ~ ’ ~frustrating ’~ feature prompted Ogura to refer to ENB as the “great imp ~ s t e r . ”Consequently, ~ ultrastructural analysis or special histologic stains are often required to establish the correct histopathologic diagn~sis.~ Although ENB continues to be histologically subclassified according to the presence or absence of
HEAD & NECK
November/Decernber 1992
477
rosette formation, the prognosis of ENB appears to be independent of this histologic Clinically, the behavior of ENB is typified by an aggressive and persistent clinical course. Tugrowth is characterized by inapparent submucosal spread, and local recurrence is common regardless of the treatment modality chosen. In one of the largest published ENB series, Olsen and DeSanto noted local recurrence in 12 of 21 patients (57%) ranging from 7 months to 10 years following initial treatment.3 Local recurrence was observed in all treatment groups including surgery, radiotherapy, and combined therapy. In the Cantrell series at the University of Virginia, eight of 12 patients received combined surgery and radiotherapy, yet local recurrence remained the most common cause of death.' In addition to frequent local recurrence, ENB has also been shown to metastasize aggressively with documented sites, including bone, lung, liver, parotid, skin, spinal cord, eye, face, scalp, aorta, spleen, adrenal, ovary, vagina, and nodes of the thorax, abdomen, or In the Olsen and DeSanto series, metastases occurred in 13 of 21 patients (62%).3Although two patients were found t o have metastases at the time of diagnosis, metastases were also discovered as many as 10 years following initial presentation. Several patients also experienced multiple metastatic episodes. Although the site of metastasis is widely variable and often atypical, Olsen and DeSanto reported cervical nodes to be the single most common site, developing in 10 of 21 ENB patients (48%).3Beitler et al. found cervical metastases to be as common as local recurrenceg and Cantrell et al. found cervical spread in two of three patients suffering from metastatic ENB.2 Levine et al. reported nine patients with ENB extending beyond the nasal vault and paranasal sinuses [Kadish group C (Table l)I.* Two of these patients (22%) were found to have cervical metastasis a t the time of initial presentation. In a literature review of 110 patients by Bailey and Barton, 24 patients (22%) were found to have metastatic disease, with cervical lymph nodes being the most common site for metastatic spread (eight patients).' Other authors have also found cervical nodes to be the most common site of m e t a s t a s i ~ . ~ , ~ J ~ This report summarizes the ENB experience at the University of North Carolina Hospitals since 1972. A retrospective review of the UNC Hospitals tumor registry since 1972 revealed
478
Esthesioneuroblastoma and Neck Metastasis
Table 1. Kadish staging system for ENB. Group A Group B Group C
Tumor confined to the nasal cavity. Tumor involving the nasal cavity and paranasal sinuses. Tumor spread beyond the nasal cavity and aaranasal sinuses.
only four patients diagnosed with ENB. All four patients suffered from metastatic neck disease at some point in their treatment and three of these patients presented with palpable neck disease. In this report we explore the rationale for treatment of the neck in ENB and suggest guidelines for the use of prophylactic neck therapy. MATERIALS AND METHODS
A retrospective review was conducted on all patients with a histopathologic diagnosis of ENB treated at the University of North Carolina Hospitals during the craniofacial era. Four patients were staged according to the criteria established by Kadish (Table 1) using surgical biopsy, surgical pathology, and computed tomographic (CT) or magnetic resonance imaging (MRI) data. Statistical evaluation of neck metastasis versus Kadish stage was conducted using contingency table analysis. RESULTS
A retrospective review of the University of North Carolina Hospitals Tumor Registry since 1972 revealed only four patients treated for ENB. All four patients were staged as Kadish stage C at the time of initial presentation by virtue of extensive local disease (Table 2). Three patients had direct invasion of the frontal lobe and one patient had extradural extension into the anterior cranial fossa. The patient with extradural anterior fossa invasion and two patients with frontal lobe invasion presented with palpable cervical metastatic disease. The fourth patient later developed cervical metastases when he recurred in both necks. One patient died during therapy and two patients are presently without evidence of disease after follow-up periods of 1 and 4 years. The fourth patient has recently completed combined chemotherapy and radiotherapy for a recurrence in the left parotid and is presently without evidence of disease. A literature search was conducted on the English surgical literature published since the widespread use of craniofacial surgery. All ENB
HEAD & NECK
NovembedDecember 1992
Table 2. UNC ENB experience since 1972 Patient
Kadish group at diagnosis
Neck metastases at diagnosis
Neck failure
Status
Yes
No
NED after XRT/chemo
Yes
No
NED after XRTkhemo
Yes
No
Deceased after XRT/chemo
No
Yes
NED after surgery chemo/XRT for third recurrence
C (bilateral ethmoid sinus, frontal lobes and neck) C (left paranasal sinuses, frontal lobe and neck) C (bilateral ethmoid and frontal sinuses, nasopharynx, orbit, extradural anterior fossa, and right neck) C (bilateral paranasal sinuses and left frontal lobe)
1 2 3
4
NED, no evidence of disease; XRT, radiotherapy; chemo, chemotherapy.
series with accruals of eight or more patients were evaluated. Those series which provided data for the site and frequency of metastasis were used to determine the cumulative cervical metastasis rate. A total of 207 patients were identified out of eight published series. Fifty-five of these 207 patients developed cervical metastasis during their disease course, producing a cumulative cervical metastatic rate of 27% (Table 3). For those series in which data for Kadish staging was provided, the cumulative cervical metastatic rate was determined according to each Kadish stage. Statistical analysis of Kadish stage versus incidence of cervical metastases revealed a statistically significant relationship. Cervical metastases occurred in 14% of Kadish group A patients (five of 36), 11%of Kadish
Table 4. Contingency table of neck metastases vs Kadish stage. Neck status
Kadish stage
Neck mets
No neck mets
Totals
A B C
5 6 25
31 51 32
36 57 57
Totals
36
114
150
p = 0.0001, contingency table analysis; mets = metastases
group B patients (six of 57), and 44% of Kadish group C patients (25 of 57) ( p = 0.0001 contingency table analysis, Figure 1, Table 4). CASEREPORT
Table 3. Percentage of reported neck metastases during craniofacial era. Total patients
Author
Year
Schenck7 Bailey'
1972 1975
Kadish' Djalilian' Cantrell* Elkon=
1976 1977 1977 7979
Olsen3 Levine4 Beitlerg This study
1983 7986 1991 1991
8 110 (combined series) 19 19 12 79 (combined series) 21 16 14 4
Total
207
27%
% patients with neck
metastases
A 37-year-old man was initially seen at the University of North Carolina Hospitals Neurosurgical service with a 3-week history of severe headache, nausea, vomiting, and blurred vision in the
37 (3/8) 22 (241110) 21 (4/19) 21 (4119) 17 (2112) Not specified* 48 (10121) Not specified* 28 (4/14) 100 (414)
P C
e n t
. . . . . . . ..... . . .
p
Esthesioneuroblastoma (ENB) is a rare malignant neoplasm arising from the cribriform olfactory epithelium. Its aggressive biologic behavior is characterized by inapparent submucosal spread, local recurrence, atypical distant metastases, and poor long-term prognosis. Historically, the suspicion of neck metastasis has not been associated with ENB. A retrospective review of all ENB patients treated at the University of North Carolina Hospitals since 1972 has revealed an unusually high incidence of cervical metastases (four of four patients). This prompted a review of the ENB surgical literature. All English-literatureseries published since the widespread use of craniofacial surgery with accruals of eight or more patients were examined for reports of neck metastasis. Although the cumulative cervical metastatic rate reached 27% (55 of 207 patients), mosi authors failed to recognize the potential risk of cervical metastasis in ENB. Furthermore, in patients suffering from Kadish stage C disease, the cervical metastatic rate climbed to 44% (25 of 57 Kadish stage C patients). This further emphasizes the importance of the neck as a high-risk metastatic site. These findings suggest that cervical metastases may be an imponant consideration in the evalu-
From the Division of Otolaryngology-Head and Neck Surgery, University of North Carolina School of Medicine, Chapel Hill, North Carolina. Acknowledgments: The authors express their appreciation to William W. Shockley, MD, for his insightful editorial comments, and to Ms. Hope Merritt for her assistance in the preparation of this manuscript. Presented at the Southern Section Meeting of the American Laryngological, Rhinological and Otological Society, Sea Island, Georgia, January 17, 1992. Address all correspondence to Dr. Weissler at the Division of Otolaryngology-Head and Neck Surgery, CB X7070, 610 Burnett-Womack Building, University of North Carolina, School of Medicine, Chapel Hill, NC 27599-7070. Accepted for publication May 18, 1992, CCC 0148~6403/92/060477-06 @ 1992 John Wiley & Sons, Inc.
Esthesioneuroblastomaand Neck Metastasis
ation and treatment of patients suffering from ENB. Strategies for evaluation and management of the neck are discussed. HEAD & NECK 1992;14:477-482 0 1992 John Wiley & Sons, Inc.
Esthesioneuroblastoma (ENB)is a rare, aggressive malignancy of neural crest origin arising from the olfactory epitheli~rn.’-~With more than 200 cases reported by 1986,4ENB remains poorly understood, owing to both its infrequent occurrence and the lack of consensus regarding treatment approaches. The clinical diagnosis of ENB is complicated by vague and indistinct early symptoms often leading to delayed patient presentation. Eventually, the development of nasal obstruction and epistaxis, or less frequently, anosmia, epiphora, proptosis, or diplopia results in the discovery of a vascular, pinkish-gray, polypoid mass in the nasal ~ a u l tTumor .~~~ staging is then performed according to the Kadish criteria.6 In addition to delays in presentation, the diagnosis of ENB may also be complicated by the tumor’s nondescript microscopic appearance. As a result of its undifferentiated histology, ENB has been mistaken for various malignancies, including lymphoma, melanoma, undifferentiated carcinoma, and s a r ~ o m a . ~ This ’ ~ ’ ~frustrating ’~ feature prompted Ogura to refer to ENB as the “great imp ~ s t e r . ”Consequently, ~ ultrastructural analysis or special histologic stains are often required to establish the correct histopathologic diagn~sis.~ Although ENB continues to be histologically subclassified according to the presence or absence of
HEAD & NECK
November/Decernber 1992
477
rosette formation, the prognosis of ENB appears to be independent of this histologic Clinically, the behavior of ENB is typified by an aggressive and persistent clinical course. Tugrowth is characterized by inapparent submucosal spread, and local recurrence is common regardless of the treatment modality chosen. In one of the largest published ENB series, Olsen and DeSanto noted local recurrence in 12 of 21 patients (57%) ranging from 7 months to 10 years following initial treatment.3 Local recurrence was observed in all treatment groups including surgery, radiotherapy, and combined therapy. In the Cantrell series at the University of Virginia, eight of 12 patients received combined surgery and radiotherapy, yet local recurrence remained the most common cause of death.' In addition to frequent local recurrence, ENB has also been shown to metastasize aggressively with documented sites, including bone, lung, liver, parotid, skin, spinal cord, eye, face, scalp, aorta, spleen, adrenal, ovary, vagina, and nodes of the thorax, abdomen, or In the Olsen and DeSanto series, metastases occurred in 13 of 21 patients (62%).3Although two patients were found t o have metastases at the time of diagnosis, metastases were also discovered as many as 10 years following initial presentation. Several patients also experienced multiple metastatic episodes. Although the site of metastasis is widely variable and often atypical, Olsen and DeSanto reported cervical nodes to be the single most common site, developing in 10 of 21 ENB patients (48%).3Beitler et al. found cervical metastases to be as common as local recurrenceg and Cantrell et al. found cervical spread in two of three patients suffering from metastatic ENB.2 Levine et al. reported nine patients with ENB extending beyond the nasal vault and paranasal sinuses [Kadish group C (Table l)I.* Two of these patients (22%) were found to have cervical metastasis a t the time of initial presentation. In a literature review of 110 patients by Bailey and Barton, 24 patients (22%) were found to have metastatic disease, with cervical lymph nodes being the most common site for metastatic spread (eight patients).' Other authors have also found cervical nodes to be the most common site of m e t a s t a s i ~ . ~ , ~ J ~ This report summarizes the ENB experience at the University of North Carolina Hospitals since 1972. A retrospective review of the UNC Hospitals tumor registry since 1972 revealed
478
Esthesioneuroblastoma and Neck Metastasis
Table 1. Kadish staging system for ENB. Group A Group B Group C
Tumor confined to the nasal cavity. Tumor involving the nasal cavity and paranasal sinuses. Tumor spread beyond the nasal cavity and aaranasal sinuses.
only four patients diagnosed with ENB. All four patients suffered from metastatic neck disease at some point in their treatment and three of these patients presented with palpable neck disease. In this report we explore the rationale for treatment of the neck in ENB and suggest guidelines for the use of prophylactic neck therapy. MATERIALS AND METHODS
A retrospective review was conducted on all patients with a histopathologic diagnosis of ENB treated at the University of North Carolina Hospitals during the craniofacial era. Four patients were staged according to the criteria established by Kadish (Table 1) using surgical biopsy, surgical pathology, and computed tomographic (CT) or magnetic resonance imaging (MRI) data. Statistical evaluation of neck metastasis versus Kadish stage was conducted using contingency table analysis. RESULTS
A retrospective review of the University of North Carolina Hospitals Tumor Registry since 1972 revealed only four patients treated for ENB. All four patients were staged as Kadish stage C at the time of initial presentation by virtue of extensive local disease (Table 2). Three patients had direct invasion of the frontal lobe and one patient had extradural extension into the anterior cranial fossa. The patient with extradural anterior fossa invasion and two patients with frontal lobe invasion presented with palpable cervical metastatic disease. The fourth patient later developed cervical metastases when he recurred in both necks. One patient died during therapy and two patients are presently without evidence of disease after follow-up periods of 1 and 4 years. The fourth patient has recently completed combined chemotherapy and radiotherapy for a recurrence in the left parotid and is presently without evidence of disease. A literature search was conducted on the English surgical literature published since the widespread use of craniofacial surgery. All ENB
HEAD & NECK
NovembedDecember 1992
Table 2. UNC ENB experience since 1972 Patient
Kadish group at diagnosis
Neck metastases at diagnosis
Neck failure
Status
Yes
No
NED after XRT/chemo
Yes
No
NED after XRTkhemo
Yes
No
Deceased after XRT/chemo
No
Yes
NED after surgery chemo/XRT for third recurrence
C (bilateral ethmoid sinus, frontal lobes and neck) C (left paranasal sinuses, frontal lobe and neck) C (bilateral ethmoid and frontal sinuses, nasopharynx, orbit, extradural anterior fossa, and right neck) C (bilateral paranasal sinuses and left frontal lobe)
1 2 3
4
NED, no evidence of disease; XRT, radiotherapy; chemo, chemotherapy.
series with accruals of eight or more patients were evaluated. Those series which provided data for the site and frequency of metastasis were used to determine the cumulative cervical metastasis rate. A total of 207 patients were identified out of eight published series. Fifty-five of these 207 patients developed cervical metastasis during their disease course, producing a cumulative cervical metastatic rate of 27% (Table 3). For those series in which data for Kadish staging was provided, the cumulative cervical metastatic rate was determined according to each Kadish stage. Statistical analysis of Kadish stage versus incidence of cervical metastases revealed a statistically significant relationship. Cervical metastases occurred in 14% of Kadish group A patients (five of 36), 11%of Kadish
Table 4. Contingency table of neck metastases vs Kadish stage. Neck status
Kadish stage
Neck mets
No neck mets
Totals
A B C
5 6 25
31 51 32
36 57 57
Totals
36
114
150
p = 0.0001, contingency table analysis; mets = metastases
group B patients (six of 57), and 44% of Kadish group C patients (25 of 57) ( p = 0.0001 contingency table analysis, Figure 1, Table 4). CASEREPORT
Table 3. Percentage of reported neck metastases during craniofacial era. Total patients
Author
Year
Schenck7 Bailey'
1972 1975
Kadish' Djalilian' Cantrell* Elkon=
1976 1977 1977 7979
Olsen3 Levine4 Beitlerg This study
1983 7986 1991 1991
8 110 (combined series) 19 19 12 79 (combined series) 21 16 14 4
Total
207
27%
% patients with neck
metastases
A 37-year-old man was initially seen at the University of North Carolina Hospitals Neurosurgical service with a 3-week history of severe headache, nausea, vomiting, and blurred vision in the
37 (3/8) 22 (241110) 21 (4/19) 21 (4119) 17 (2112) Not specified* 48 (10121) Not specified* 28 (4/14) 100 (414)
P C
e n t
. . . . . . . ..... . . .
p
