new 45 glomus tumors 02

Esthesioneuroblastoma

Esthesioneuroblastomas (ENB), first described by Berger and Luc (20), are rare tumors thought to arise in the olfactory receptors in the nasal mucosa or the cribriform plate of the ethmoid bone. The olfactory nerves perforate grooves in the ethmoid bone in the cribriform plate and continue into the subarachnoid spaces, accounting for the high incidence of intracranial extension (20,75).

Epidemiology

ENB constitutes 3% of all endonasal neoplasms. In the United States, according to the data from the Surveillance, Epidemiology, and the End Results Program, 84 cases of ENB were registered from 1978 to 1990 (72). About 945 cases have been reported in the world literature (31). The review authors' cases accounted for 198 and collaborative efforts accounted for 747 cases. Sex distribution was 53.6% male and 46.64% female. Kadish classification was applied to 563 cases; 103 (18.3%) class A, 182 (32.2%) class B, and 278 (49.4%) class C cases.

Herrold (117) induced olfactory neuroblastoma in the nasal cavity of hamsters by injection of dimethylnitrosamine and other nitroso derivatives. What role these agents may play in humans is unknown. No other risk factors have been identified. There appears to be a slight male predominance. The age incidence has a bimodal distribution, with peaks at 11 to 20 years and 40 to 60 years, the highest incidence at 51 to 60 years (75,143).

Natural History

Although others thought that EBN were of ectodermal origin, most observers believe the tumor to be of neuroectodermal origin in the olfactory epithelium (11,97). Most of these tumors occur high in the nasal cavity or in the lateral wall adjacent to the ethmoid. The tumor may spread to the opposite ethmoid bone, superiorly to the frontal sinus and anterior cranial fossa, posteriorly to the sphenoid sinus, nasopharynx, and base of skull, laterally to the orbits, forward to the frontonasal angle, or inferiorly to the nasal cavity and antrum (Fig. 45.9). Lymphatic spread may be to the subdigastric, posterior cervical, submaxillary, or preauricular nodes, as well as to the nodes of Rouviere. The exact incidence of distant metastases is uncertain; it has been quoted to be as high as 50%, but this rate is influenced by the use of chemotherapy in high-risk patients.

Figure 45.9. Coronal magnetic resonance imaging scan showing a large soft tissue mass and bone destruction in the right ethmoidal maxillary sinuses and nasal cavity secondary to extensive (Kadish stage C) esthesioneuroblastoma.

Clinical Presentation

These tumors tend to be friable and bleed easily. The most common clinical symptoms are epistaxis and nasal blockage. Patients also may have local pain or headache, visual disturbances, rhinorrhea, tearing, proptosis, or swelling in the cheek (134). The symptoms may be associated with a mass in the neck.

Diagnostic Work-Up and Staging

Physical examination may show the inferior aspect of a polypoid friable mass in the nasal cavity. Ocular findings or a mass in the nasopharynx may be present.With early lesions, radiographs or CT or MRI may show only nonspecific opacification, soft-tissue swelling, and occasionally bone destruction (170,213,233). Octreotide is a somatostatin analog that, when coupled to a radioisotope, produces a scintigraphic image of neuroendocrine tumors (NET) expressing somatostatin type 2 (SSR 2) receptors. Octreotide scintigraphy (OS) may be useful in confirming the preoperative diagnosis of certain head and neck NET, such as paragangliomas, Merkel cell carcinomas, medullary thyroid carcinomas, and esthesioneuroblastomas. Bustillo et al. (34) carried out a retrospective study that compared the results of OS with the histopathologic diagnosis in 74 patients with head and neck NET. Of the 60 patients undergoing evaluation for suspected paraganglioma, OS was correctly positive in 36/37 patients with PG and correctly negative in 19/23 patients who did not exhibit PG (sensitivity of 97% and a specificity of 82%). There were 14 patients in the nonparaganglioma group. OS detected or diagnosed locoregional recurrences in two with esthesioneuroblastoma.

Table 45.1 outlines the suggested diagnostic work-up (214,253). MRI, especially with gadolinium contrast, may be used as a supplement or alternative to CT scanning (213). Kairemo et al. (135) described imaging findings in 17 olfactory neuroblastomas; CT provided the best information about the tumor and its local invasion into surrounding bone structures. MRI allowed an estimate of tumor spread into surrounding soft-tissue areas, such as the anterior cranial fossa and the retromaxillary space. Bone scintigraphy scan detected distant metastases.

In a review of 22 patients with a histologically proven olfactory neuroblastoma the tumors displayed a variety of imaging characteristics and aggressiveness (71). The expansile tendency of olfactory neuroblastoma is characterized by bowing of the sinus walls. The destructive aspect is manifested as tumor replacing the turbinates, septum, and sinus walls with extension into contiguous areas (Fig. 45.9 and 45.10). The density/signal and enhancement characteristics are nonspecific of olfactory neuroblastoma.

Figure 45.10.A,B: The sagittal and coronal views of a preoperative magnetic resonance imaging of a 56-year-old patient who was initially seen with a Kadish stage C tumor involving left nasal cavity and extending intracranially (arrows).

Although dopamine β-hydroxylase and catecholamines are produced by these tumors, their measurements or vanillylmandelic acid excretion levels have not proven clinically useful (71).

A staging system has been proposed by Kadish et al. (134) (Table 45.11).

Table 45.11. Kadish System for Staging of Esthesioneuroblastoma

Pathologic Features and Prognostic Factors

ENBs are polypoid, frequently reddish, soft, and vascular tumors with neuroblasts and neurocytes. Gerard-Marchant and Micheau (97) classified three histologic types. ENB contains epithelial components serving as a supporting stroma and have a nerve component that corresponds to the olfactory cells. Rosettes are the main feature, consisting of several rows of cells arranged around the central area (11,252). ENBs may be confused with lymphoma or anaplastic carcinoma and have diffuse, regular distribution. ENBs contain many fibrils, which fill the central space of the rosette (called a pseudorosette). It has been suggested that the presence of chromaffin granules indicates a derivative from primitive neural crest cells. ENB must be distinguished from other poorly differentiated neoplasms including sinonasal undifferentiated carcinoma, which is derived from the Schneiderian epithelium. Sinonasal undifferentiated carcinoma lacks rosettes and intercellular fibrils (252).

Extension of the primary tumor based on the Kadish staging system (134) has been identified as the most important determinant of treatment outcome although this was not confirmed by Chao et al. (40). High-grade tumors had worse outcome in the reports from the Mayo Clinic and UCLA (85,91).

Argiris et al. (5) found in 16 patients with ENB, 11 of whom had Kadish stage 8 (50%) had brain involvement at presentation. Craniofacial resection was performed in 13 patients (81%); 14 received either preoperative or postoperative therapy (radiation therapy in 11 and chemotherapy in 4). The actuarial 5-year survival was 60%, diseasefree survival 33%, with a median follow-up of 4.3 years. The first site of failure was locoregional alone in 10/12 patients who progressed, and in six patients involved the brain or the meninges. Two patients were successfully salvaged.

Hyams (126) proposed a histologic grading system for ENB in which grade I tumors have an excellent prognosis and grade IV tumors are uniformly fatal. The Hyams grading system predated advanced craniofacial techniques, extensive use of immunohistochemistry, and the recognition of sinonasal undifferentiated carcinoma (SNUC) as a distinct entity. Miyamoto et al. (186) in a retrospective review of 12 patients with esthesioneuroblastoma and 14 with SNUC used the Kadish clinical stage and Hyams histopathologic system. Kadish staging was available for 26 patients (two patients with stage A tumors; seven with stage B, and 17 with stage C). Of the eight evaluable patients with Kadish stage A or B tumors, six remained disease free for more than 2 years compared with only 5/7 Kadish stage C tumors. Slides were available for Hyams grading in 21 patients (two patients with grade I tumors, four with grade II, four with grade III, and 11 with grade IV). Of the six patients with Hyams grade I or II tumors, four remained disease free for more than 2 years compared with only 4/15 patients with Hyams grade III or IV tumors. Three patients with Kadish stage C tumors (two with esthesioneuroblastoma, one with SNUC) and two patients with Hyams grade IV tumors (one with esthesioneuroblastoma and one with SNUC) survived for more than 5 years. They concluded that both the Hyams grading and the Kadish staging system can be used as independent predictors of outcome; patients with either advanced clinical stage or pathologic grade of ENB or SNUC have poor prognosis, but long-term survival is possible in these patients if aggressive treatment is used.

Papadaki et al. (204) analyzed 18 formalin-fixed paraffinembedded olfactory neuroblastoma specimens (12 primary tumors and six recurrences or metastases) from 14 patients and concluded that p53 point mutation does not play an important role in the initial development of olfactory neuroblastoma; however, p53 wild-type hyperexpression may occur in subsets, show local aggressive behavior, and have a tendency for recurrence.

General Management

Surgery alone appears to be adequate treatment for small, low-grade tumors confined to the ethmoids in which negative surgical margins can be obtained (25). An ethmoidomaxillary resection with or without orbital sparing is usually necessary. This procedure is combined with preoperative or postoperative irradiation (85,226). A complete resection with preservation of vital structures is achievable by using a craniofacial approach. The experience from the University of Virginia, however, has yielded no firm conclusions regarding whether craniofacial resection or more conservative surgery could be performed in early-stage disease (91).

Dias et al. (62) reported on 35 patients with ENB treated with gross tumor resection through a transfacial approach with postoperative RT in 11 patients, craniofacial resection (CFR) and postoperative RT in seven, exclusive RT in 14, CFR alone in one, and a combination of chemotherapy and RT in two. Radiation therapy median dose was 48 Gy. Analysis of survival showed that the Kadish classification best predicted diseasefree survival (p = 0.046). The presence of regional and distant metastases adversely affected prognosis (p <0.0001 and p = 0.01, respectively). Craniofacial resection plus postoperative RT provided a better 5-year diseasefree survival rate (86%) compared with the other therapeutic options used (p = 0.05). The 5-year disease-specific survival rate was 64% and 43% for the low- and high-grade tumors, respectively (p = 0.20). Diseasefree survival was 46% and 24% at 5 and 10 years, respectively. Overall survival was 55% and 46% at 5 and 10 years of follow-up, respectively. Aggressive multimodality therapeutic strategies, particularly CFR and adjuvant RT, yielded the best treatment outcome.

Early lesions involving the ethmoids with little or no bony destruction or nerve invasion can be treated adequately by high-energy (photon or electron) radiation therapy with good cosmetic and functional results (20,85,143). Those with more extensive local disease benefit from surgery and adjuvant irradiation (20,40,214), although some have spoken against combined surgery and radiation therapy because of complications (13,256,302). Patients with locally advanced disease or high-grade tumors should receive aggressive treatment with combined modalities, such as surgery, radiation therapy, and chemotherapy.

Monroe et al. (187) described treatment results in 22 patients who received RT for ENB (equal numbers of males and females, median age of 54 years). The modified Kadish stage was A in one patient, B in four patients, C in 15 patients, and D in two patients. Treatment modalities included primary RT in six patients, preoperative RT in one patient, postoperative RT after craniofacial resection in 12 patients, and salvage RT in three patients treated for recurrence after surgery. Elective neck RT was performed in 11/20 patients (two patients had cervical metastases at presentation for RT). Rates of local tumor control, cause-specific survival, and absolute survival at 5 years were 59%, 54%, and 48%, respectively. The cause-specific survival rate at 5 years was lower after primary RT (17%) than after craniofacial resection and postoperative RT (56%). Cervical metastases occurred in 6/22 patients (27%). No neck recurrences occurred in 11 patients treated with elective neck RT compared with four neck recurrences in nine patients (44%) not receiving elective neck RT (p = 0.02). Their data and review of the current literature suggest a higher cervical failure rate than previously recognized; elective neck RT seems to correlate with improved nodal tumor control and should be considered in the treatment of esthesioneuroblastoma.

Rosenthal et al. (226) treated 72 adults with nonmetastatic, primary sinonasal neuroendocrine tumors (31 with ENB, 16 with SNUC, 18 with neuroendocrine carcinoma [NEC], and seven with small cell carcinoma [SmCC]). Patients with ENB usually were treated with surgery and/or radiotherapy; only 3/31 patients (9.7%) received radiation to regional lymphatics, and only 5/31 received chemotherapy. In contrast, patients with non-ENB histologies usually received chemotherapy (10/16 patients with SNUC, 12/18 patients with NEC, and 5/7 patients with SmCC). With a median follow-up for surviving patients of 81.5 months, overall survival at 5 years was 93.1% for patients with ENB, 62.5% for SNUC, 64.2% for NEC, and 28.6% for SmCC (p = 0.0029). The local control tumor rate at 5 years also was superior for patients who had ENB (96.2%) compared with patients who had SNUC (78.6%), NEC (72.6%), or SmCC (66.7%) (p = 0.04). The regional failure (RF) rate at 5 years was 8.7% for patients with ENB, 15.6% for patients with SNUC, 12.9% for patients with NEC, and 44.4% for patients with SmCC. Additional late events increased the RF rate for patients with ENB to 31.9% at 10 years. The distant metastasis rate at 5 years was 0.0% for patients with ENB, 25.4% for patients with SNUC, 14.1% for patients with NEC, and 75.0% for patients with SmCC.

For advanced lesions, in which disseminated disease is likely, chemotherapy may improve tumor control and decrease the incidence of distant metastases. A combination of thiotepa, cyclophosphamide, doxorubicin, vincristine, nitrogen mustard, and actinomycin-D has been used (178). Wieden et al. (285) reported complete tumor regression and 2.7-year survival n a patient with extensive olfactory esthesioneuroblastoma treated with a combination of wide local excision, chemotherapy with cisplatin and 5-fluorouracil (5-FU), and irradiation (55.8 Gy).

A retrospective review of 10 patients with recurrent esthesioneuroblastoma treated with chemotherapy at the Mayo Clinic suggested that cisplatin-based chemotherapy is active in advanced, high-grade tumors (227). Survival from initial chemotherapy treatment was 44.5 months (range, 3 to 130 months) in patients with low-grade tumors and 26.5 months (range, 2 to 67 months) in patients with high-grade tumors.

Treatment, which could be classified in 898 reported cases, consisted of surgery alone in 24% (226 cases), radiation therapy alone in 18.4% (165 cases), combined surgery and radiation therapy in 43.2% (388 cases), chemotherapy in 13.2% (119 cases), and in 11 cases (1.2%) bone marrow transplant. In the reported cases follow-up could be evaluated in 477 cases, while in only 234 cases a 5-year follow-up was done; on these 20.5% had surgery only, 11.1% radiation therapy, and 68.4% combined surgery and radiation therapy. The best survival rates were obtained by combined therapy, 72.5% versus 62.5% with surgery alone and 53.8% with radiation therapy (31).

Elective Neck Treatment

Esthesioneuroblastoma has been shown to metastasize to the neck and remote sites. Although the sites of metastases are widely variable and often atypical, Olsen and DeSanto (196) reported cervical lymph nodes to be the most common site, developing in 10/21 patients (48%) in their series. Beitler et al. (16) found cervical lymph node metastases to be as common as local recurrence. In a literature review of 110 patients by Bailey and Barton (7), 24 patients (22%) had metastatic disease, with cervical lymph nodes being the most common site. Davis and Weissler (57) compiled a retrospective review of patients and found that the cumulative cervical metastasis rate reached 27% (55/207 patients). In general, because of the low incidence of cervical lymph node metastasis (≤10%) in early-stage disease, elective irradiation of the neck or a dissection is not indicated (75). However, in patients with Kadish stage C disease, the cervical metastatic rate climbed to 44% (25/57 patients). As noted previously, Monroe et al. (187) observed cervical node metastasis in 6/22 patients (27%), incidence similar to that reported by other authors. In 11 patients they treated with elective neck RT no recurrences were noted, in contrast to 4/9 (44%) in patients not receiving elective neck RT. Thus, with advanced-stage disease, cervical nodes should be initially managed by irradiation, radical neck dissection, or a combination of both (57,75).

Radiation Therapy Techniques

A combination of photons and electrons with anterior fields provides good coverage for limited ethmoidal disease when the tumor is confined anteriorly. Beam arrangement can be modified for disease extending into the orbit or maxillary sinus. Obturator or bolus may be needed postoperatively to compensate for tissue deficit. When intracranial or posterior extension is present or tumor has spread into the maxillary sinus, a pair of perpendicular (anteroposterior and lateral) portals with wedges or two lateral wedge fields in conjunction with an open anterior photon field will give good coverage of the treatment volume with the dose inhomogeneity around 10% to 20%. Incorporation of a vertex field eliminates the high inhomogeneous dose along the junction line of the conventional three-field technique. Treatment techniques are similar to those described for treatment of paranasal sinuses (see Chapter 39). The orbits can be spared or treated as the degree of extension dictates. Occasionally, an anterior electron beam field may be needed to supplement lowdose areas. When the electron beam is used over air cavities, some dosimetry problems result. Eye blocks must be positioned precisely to avoid undesirable side effects.

When combined therapy is used, preoperative doses of 45 Gy and postoperative doses of 50 to 60 Gy are indicated, depending on the status of the surgical margins. Doses of 65 to 70 Gy are delivered with irradiation alone in patients with inoperable tumors (137). Contrast-enhanced CT or MRI scans before initiation of treatment are crucial to demarcate extension of the tumor. Treatment planning with CT for determination of tumor extension is extremely important (239). Because of the proximity of esthesioneuroblastoma to the optic nerves, optic chasm, and the brainstem, the precision of treatment setup, target volume definition, and dose homogeneity dictate tumor control and the sequelae of treatment. Treatment techniques similar to those for paranasal sinuses may create “hot spots” along the optic tracks. High doses per fraction (exceeding 2 Gy) increase the possibility of late sequelae such as blindness and bone and brain necrosis (10,91).

Three-dimension CRT or IMRT provide alternatives to the conventional three-field technique frequently used to treat these tumors (Fig. 45.11). Special attention should be directed to reduce unnecessary irradiation to ocular structures. When occasionally a patient presents with cervical node metastasis IMRT is very helpful to optimally treat the primary tumor and the cervical lymphatics (Fig. 45.12)

Figure 45.11. Esthesioneuroblastoma in a 35-year-old female, initially treated with a craniofacial surgical resection.Patient received postoperative intensity-modulated radiation therapy (2-Gy fractions). A: Cross-section illustrating coverage of ethmoidnasal and left maxillary antrum volume. B: Cross-section showing dose distribution in target volume with excellent sparing of ocular structures. C: Dose-volume histogram:

Structure         Dose Range (cGy)      Mean Dose (Gy)

Planning target volume 1       30-70   65

Planning target volume 2 (+ radiotherapy nasal??)   40-70   58

Optic chasm and nerves         13-42   24

Figure 45.12. Patient with stage C esthesioneuroblastoma of ethmoid cells and nasal cavity who presented with a large left upper cervical lymph nodemetastasis.Intensity-modulated radiation therapy plans to deliver 70 Gy to primary tumor and cervical lymphadenopathy. A: Coronal, (B) sagittal, (C) cross-section dose distributions illustrate excellent coverage of all target volumes.

Results of Therapy

Surgery and Irradiation

Radiation therapy is an important component in the management of esthesioneuroblastoma, but the optimal sequence when integrated with surgery is unknown. Dulguerov and Calcaterra (70) reported that 83% of patients who were treated with combined irradiation and surgery did not have recurrences. Eden et al. (72) reported no significant difference in survival whether preoperative or postoperative irradiation was given, but suggested improved local tumor control with preoperative irradiation. Technical factors may have contributed to a higher incidence of postoperative radiation therapy failures because three of five postoperative cases received <50 Gy; all three patients were treated with a single anterior field, which gives less homogeneous dose distribution throughout the treatment volume.

At M.D. Anderson Cancer Center, 11 patients were treated with combined surgery and irradiation in most instances. Four patients were alive and disease free at 2.5, 5, 8, and 12 years, respectively; two died of disease 4 and 5 years after treatment. One patient was lost to follow-up at 9 months with extensive disease and is presumed dead. Two patients died of complications but were free of disease.

Spaulding et al. (246) compared 30 patients treated in two time periods: 1969 to 1975 and 1976 to 1985. With the introduction of craniofacial resection, complex field megavoltage irradiation, and chemotherapy in stage C disease in 1976, the overall 2-year survival rate increased from 70% to 87%. For stage C disease, the survival rate increased from 50% to 88%.

Dzalitian et al. (71) described 19 cases of esthesioneuroblastoma, 18 of which were followed. Four patients with advanced disease received radiation therapy exclusively; three soon died of disease, and one patient died 5 years later, presumably of coronary occlusion. Combination radiation therapy and rhinotomy failed in five other patients, but three were salvaged surgically. Three of nine patients treated with surgery only were free of disease at 3 to 14 years. Five of six failures were salvaged with irradiation and surgery.

Foote et al. (85) updated the experience of the Mayo Clinic. Seventeen patients had disease confined to nasal cavity or paranasal sinuses (Kadish stage A and B), and 32 patients had more advanced disease. Treatment included gross total resection alone or combined with radiation therapy. The 5-year actuarial survival, diseasefree survival, and local tumor control rates were 69.1%, 54.8%, and 65.3%, respectively. Local tumor control was improved in patients who received postoperative irradiation (55.5 Gy) even after complete tumor resection.

Levine et al. (160) conducted a retrospective review of 35 patients; 6% of them presented with cervical metastasis, but ultimately 25.7% developed cervical metastases. Fourteen percent of the patients developed a local recurrence an average of 6 years after diagnosis, and in 37% ultimately at least one episode of metastatic disease occurred. The diseasefree survival was 80.4% at 8 years. Central nervous system complications occurred in 25.7% of patients, orbital complications in 22.9%, systemic posttreatment problems in 20%, and had chemotoxic sequelae in 18%. Eriksen et al. (78) carried out a retrospective review of 13 patients with esthesioneuroblastoma; according to the Kadish classification one patient had stage A disease, 5 patients stage B, and 7 patients stage C. The 5-year diseasefree survival was 51%. Forty-six percent of the patients experienced relapse, and despite intensive salvage therapy, median survival after recurrence was only 12 months.

Chao et al. (40) reported on 25 patients with esthesioneuroblastoma, ages ranging from 16 to 73 years (median, 37 years). The tumors were Kadish stage A in 3, stage B in 13, C in 8, and modified D in 1 (cervical nodal metastasis). Seventeen patients were treated with surgery and radiation therapy, six with irradiation alone, and two with surgery only. Eight patients received neoadjuvant chemotherapy. Median follow-up was 8 years (range, 2 to 24 years). The 5-year actuarial overall survival, diseasefree survival, and local tumor control rates were 66.3%, 56.3%, and 73%, respectively. Kadish stage was not a significant prognosticator for local control or diseasefree survival. Five-year local tumor control was 87.4% for the combination of surgery and radiation therapy and 51.2% for irradiation alone. Two patients with Kadish stage A and B disease underwent surgical resection alone; both failed locally. In contrast, only 3/9 with Kadish stage A or B disease who received adjuvant radiation therapy had a local recurrence. With adjuvant radiation therapy, the surgical margin status did not influence local tumor control. Among the eight patients who received neoadjuvant chemotherapy, six patients showed no response, one had partial response, and one a complete response.

Simon et al. (241) reported on 13 patients with esthesioneuroblastoma or olfactory neuroblastoma; none of the patients had Kadish stage A disease, five had stage B, and eight had stage C. The majority of the patients were treated with a craniofacial resection or tumor removal through a rhinotomy approach. Two patients received neoadjuvant chemotherapy before surgical resection because of locally advanced tumor (cisplatin, ifosfamide, and etoposide). Twelve of the 13 patients received radiation therapy either initially or for salvage. Median dose of postoperative irradiation was 59.4 Gy; 1.8 Gy per day to fields encompassing the involved anatomy with a margin. The overall actuarial 5-year survival was 61% and 10 year 24%, and diseasefree survival rates were 56% and 42%, respectively.

Elkon et al. (75) compiled results for 97 patients treated with different modalities reported in the literature. Survival and tumor control in 78 patients, staged according to the Kadish system, are summarized in Table 45.12.

Table 45.12. Results of Treatment Correlated with Modality and Stage for Esthesioneuroblastoma

Chemotherapy and Irradiation

Wade et al. (279) reported that 8/13 patients (62%) had an objective response to cytotoxic agents (cyclophosphamide and vincristine). Eden et al. (72) described results in 16 patients with stage A or B disease and 24 with stage C disease treated with irradiation (median dose 50 Gy) and surgery for stage A and B disease, with the addition of chemotherapy (cyclophosphamide and vincristine) for stage C disease. Actuarial survival rates at 5 and 10 years were 78% and 71%, respectively. Locoregional failure developed in 15/40 patients; 68% of the failures were locoregional (including brain, neck, facial bone, and sinus). They had no recurrences at the primary tumor bed; all recurrences were either outside the irradiation field or at distant sites.

Preoperative neoadjuvant therapy may provide a valuable complement to radical craniofacial resection. Polin et al. (214) reviewed 34 patients with biopsy-proven esthesioneuroblastoma. In multivariate regression analysis, advanced age was predictive of decreased diseasefree survival (p = 0.008), whereas advanced Kadish stage was associated with a borderline higher rate of disease-related mortality (p = 0.056). Two thirds of the patients showed a significant reduction in tumor burden with adjuvant therapy. Patients with response to neoadjuvant chemotherapy demonstrated a significantly lower rate of disease-related mortality (p = 0.50). The overall 5- and 10-year survival rates were 81% and 54.5%, respectively.

Forty patients were treated for esthesioneuroblastoma at Institut Gustave Roussy, France (97). Three had stage T1, seven T2, 15 T3, and 15 T4 lesions. At presentation the cervical metastatic rate was 18% and distant metastases were detected by bone marrow biopsy and bone scan in three patients. Treatment modalities included surgery alone in eight patients, radiation therapy alone in three patients, surgery plus radiation therapy in 11 patients, chemotherapy alone in two patients, chemotherapy plus radiation therapy in 10 patients, and hemotherapy plus surgery and radiation therapy in six patients. The 5-year survival rate was 51%. Multimodality treatment offered better survival (63% at 5 years) and diseasefree interval (54 months). Overall local, regional, and distant failure rates were 58%, 15%, and 40%, respectively. Distant metastases commonly occurred in bone (82%). Cervical metastasis was an unfavorable prognostic indicator (0% survival at 2 years).

Bhattacharyya et al. (24) reported on nine patients with esthesioneuroblastoma or neuroendocrine carcinoma of the paranasal sinuses treated with two cycles of cisplatin and etoposide followed by photon and stereotaxic proton radiation therapy totaling approximately 68 Gy to the primary site. Poor responders were treated with surgical resection followed by postoperative irradiation; in two cases, this was combined with two additional cycles of cisplatin and etoposide chemotherapy. Nine patients with a median Dilguerov T3 stage (range, T2 to T4) completed the treatment protocol, with mean follow-up after diagnosis of 20.5 months. Eight of nine patients exhibited a dramatic response to therapy, and resection was not required. One patient failed to respond to induction chemotherapy and received surgical therapy followed by postoperative radiation therapy. There have been no recurrences (mean diseasefree interval of 14 months). Complications were limited and generally transient.

In the University of Virginia series of five patients salvaged with high-dose chemotherapy and bone marrow transplantation, three were alive with no evidence of disease, whereas only 4/17 patients (24%) salvaged with surgery and chemotherapy or irradiation were alive with no evidence of disease. Although the indications for high-dose chemotherapy and bone marrow transplantation must be better defined, it is a promising alternative for patients with large tumors or those with recurrent tumor to whom no further local therapy (e.g., surgery or irradiation) can be safely given.

Sequelae of Treatment

In a few patients, depending on the dose of irradiation, longterm sequelae include bone necrosis, blindness, or painful eye reactions requiring enucleation (10,196,206).

Simon et al. (241), in 13 patients with olfactory esthesioneuroblastoma treated with surgery and radiation therapy, noted that visual impairment was the most common complication. One patients lost vision as a result of glaucoma and radiation retinopathy after receiving 67.34 Gy in 34 fractions. One patient treated with 61.76 Gy in 34 fractions developed a visual field defect and optic atrophy; she also had a nasal cutaneous fistula. One patient sustained intraoperative rupture of the ocular globe and subconjunctival hemorrhage.

Extramedullary Plasmacytomas

Solitary plasmacytomas are rare tumors of plasma cell origin making up 4% of all plasma cell tumors. Multiple myeloma occurs about 40 times more frequently than solitary plasmacytoma (142,200,288). Monoclonal extramedullary plasmacytoma (EMP) is a rare, low-grade lymphoma found predominantly in the head and neck region. Only since the introduction of immunophenotyping techniques two decades ago has it been possible to differentiate EMP from benign polyclonal plasma cell proliferation. Hotz et al. (122) reviewed the records of 24 patients with morphologically diagnosed EMP treated at their institution; only 14 patients had true monoclonal plasmacytoma. No EMP-related deaths occurred. Two patients had local recurrence, and two patients developed multiple myeloma. Diagnostic procedures exclude a benign polyclonal plasmacytoma, multiple myeloma, and solitary bone plasmacytoma. The slow natural progression of the disease and the rarity of secondary multiple myeloma favor nonmutilating local surgery whenever possible to avoid the long-term sequelae of radiation.

Epidemiology

The annual incidence of EMP is 0.04 cases per 100,000 population (178). They constitute only 0.5% of all upper respiratory tract malignancies. Male patients exceed female patients by a ratio of 4:1, and 75% of patients are 40 to 60 years of age (16,200,292).

In a detailed literature search of more than 400 publications between 1905 and 1997, EMP mainly occurred between the fourth and seventh decades of life (3). Seven hundred fourteen cases (82.2%) were found in the upper aerodigestive tract.

The most common sites in the head and neck are the nasopharynx, nasal cavity, paranasal sinuses, and tonsils.

Clinical Presentation and Diagnostic Work-Up

EMP of the head and neck area should be considered a separate entity because of its clinical behavior. The most common symptoms are nasal obstruction, local pain and swelling, and epistaxis.

Grossly, plasmacytomas tend to be sessile in the nasal cavity and paranasal sinuses and pedunculated in the nasopharynx and larynx. The masses are soft, pliable, and pale gray. The lesion may remain localized or may infiltrate and destroy the surrounding soft tissue and bone. The usual criteria for solitary plasmacytomas, either medullary or extramedullary, include a biopsy-proven plasma cell tumor with one or, at the most, two solitary foci, absence of Bence-Jones protein in the urine, bone marrow taken some distance from the primary site not involved by tumor (<10% of plasma cells), hemoglobin of 13 g/mL or more, and a normal serum protein level or serum electrophoresis at the time of the diagnosis. Basically, the diagnosis of solitary plasmacytoma is made by exclusion, that is, by eliminating the possibility of multiple myeloma (51). Diagnosis is based on histology along with special immunoperoxidase staining for immunoglobulin lambda and kappa light chains (281).

Dimopoulos et al. (65) noted that strict staging criteria, including normal MRI studies of the axial skeleton and the long bones and absence of monoclonal plasma cells detected by flow cytometry or PCR, are required for diagnosis of solitary plasmacytoma. Many patients enjoy prolonged diseasefree survival, but the incidence of systemic relapse is high. Careful microscopic and immunohistochemical studies are required for the correct diagnosis, because this disease can be confused with other malignancies, particularly lymphomas.

Six patients with primary EMP in the head and neck were examined with MRI (276); five lesions were oval and sharply demarcated without signs of infiltration, while the other lesion filled the parapharyngeal space bilaterally. On T2-weighted sequence, the lesions had moderate signal intensity. On plain T1-weighted sequences, the tumors were isointense or slightly hyperintense with respect to surrounding muscles; after administration of contrast medium, four lesions showed notable enhancement, with distinct central inhomogeneity.

Bone destruction is not a particularly bad prognostic sign, although some investigators report that it adversely affects prognosis (276). Bony invasion is common in the more malignant types (11).

Cervical lymph node metastasis from EMP varies with the site of the primary lesion and follows the same pattern of spread as squamous-cell carcinoma arising in a similar site. The reported incidence of lymph node metastasis ranges from 12% to 26% (142). The diagnostic work-up for EMP arising in the head and neck region is shown in Table 45.1.

The exact relationship between EMP and multiple myeloma is unclear; however, approximately 20% to 30% of EMP cases will convert to multiple myeloma (51,121,287).

General Management

Pedunculated EMP lesions may be treated by surgical excision because the chance of local recurrence is low. The treatment of choice for all other lesions is radiation therapy alone or combined with other modalities (109). In a review of 714 cases in the literature the following therapeutic strategies were used to treat patients with EMP of the upper aerodigestive tract: radiation therapy alone in 44.3%, combined therapy (surgery and irradiation) in 26.9%, and surgery alone in 21.9%. The median overall survival or recurrencefree survival was longer than 300 months for patients who underwent combined intervention (surgery and irradiation), for surgical intervention alone (median survival time, 156 months), and for radiation therapy alone (median survival time, 114 months). Overall, after treatment for EMP in the upper aerodigestive tract, 61.1% of all patients had no recurrence or conversion to systemic involvement (i.e., multiple myeloma); however, 22% had recurrence of EMP, and 16.1% had conversion to multiple myeloma.

Radiation Therapy Techniques

Irradiation techniques vary with the location of the primary tumor. The techniques are similar to those used for primary tumors in comparable locations (i.e., nasopharynx, tonsil, paranasal sinuses). Solitary plasmacytomas respond well to doses of 50 to 60 Gy in 2-Gy fractions. The local tumor control rate with radiation therapy alone is about 85%. Harwood et al. (109) summarized the literature but could not draw a doseresponse curve from the data because of a lack of cases receiving low-dose radiation therapy. Nevertheless, there is a high risk of local recurrence with tumor doses below 30 Gy and a negligible risk for those treated at or above 40 Gy.

Wax et al. (281) reported on seven patients. Treatment consisted of radiation therapy in three cases, with doses ranging from 31.75 to 60 Gy. All patients have maintained local tumor control and had been followed for a minimum of 1.5 years, with an average of 3 years. One patient, treated with surgical excision, experienced a relapse at a distant site 6 years later.

The response to therapy of 32 patients with localized plasmacytoma were described by Shih et al. (238); 22 patients had solitary plasmacytoma of bone, and 10 had EMP. Median age for EMP was 63 years. Most EMPs occurred in the oronasopharynx (six cases) and paranasal sinuses (two cases). Seven patients with EMP received radiation therapy (47 to 65 Gy), and all achieved initial local tumor control. There was one local recurrence and multiple myeloma conversion in the EMP group. Local recurrence or dissemination was associated with the appearance of or an increase in M protein.

Holland et al. (121) reported on 14 cases of EMP, eight of which were in the head and neck. With doses of 46 to 62 Gy, the complete tumor response was 72%. No dose-response effect was observed.

Liebross et al. (162) described results in 22 patients with solitary EMP, in the head or neck in 19, usually in the nasal cavity or maxillary sinuses, bone destruction was found in 10/11 patients. Among all patients, serum myeloma protein was present in three patients (14%) and Bence Jones protein alone in two (9%). Radiation therapy was the sole treatment in 18/22 patients (median dose 50 Gy; range, 40 to 60 Gy); 5/7 patients with an EMP of oral cavity, oropharynx, nasopharynx, parotid, or larynx also received elective neck irradiation. Local tumor control was achieved in 21/22 patients (95%), and disease never recurred in regional nodes. Disappearance of myeloma protein occurred in 3/5 patients with an evaluable abnormality. Multiple myeloma developed in seven patients (32%), all within 5 years. The 5-year rate of freedom from progression to multiple myeloma was 56% and the median survival was 9.5 years. Chao et al. (40) reported on 16 patients with EMP and a median follow-up of 66 months. The head and neck region accounted for the majority of presentations (88%). A serum monoclonal paraprotein was found in three patients and bone erosion was identified in seven patients. All patients received local RT, although two patients also received elective nodal irradiation. The median RT dose was 45 Gy (range 40 to 50.4 Gy). Local tumor control was achieved in all patients (100%), however, regional recurrence outside the RT fields occurred in 2/16. Multiple myeloma (MM) developed in five patients (31%) all within 5 years. The 10-year myelomafree survival is 75% and 10-year overall survival is 54%.

Galieni et al. (92) reviewed 46 cases of EMP most frequently localized in the upper airways (37/46, 80%), with the mass being limited to a single site in all but seven patients in whom two contiguous sites were involved. The most frequent form of treatment was local radiation therapy. Thirty-nine patients (85%) achieved complete remission, five (11%) a partial remission, and two (4%) did not respond to therapy. Local recurrence or recurrence at other sites occurred in 7.5% and 10%, respectively. Seven patients (15%) developed multiple myeloma. The 15-year survival rate was 78%.

Michalski et al. (182) described 10 patients with EMP treated with radiotherapy. One patient treated at relapse underwent surgical resection followed by postoperative RT. The disease was most frequently localized in the paranasal sinuses (50%). All nine patients who received definitive RT (40 to 50 Gy) achieved a complete response. Median follow-up period was 29 months. Four patients (40%) relapsed, three have died of their disease. Two patients with paranasal sinus disease subsequently relapsed with multiple myeloma at 10 months and 24 months, respectively. The relapse rate in neck nodes of 10% does not justify elective irradiation of the uninvolved neck.

Miller et al. (183) reported that tumor arose in the sinonasal/nasopharyngeal region in 11/20 patients (55%). The primary modality of treatment was radiation therapy (45 to 60 Gy). The mean follow-up was 60.2 months. In 15/20 cases, immunohistochemistry staining for immunoglobulin light chain production was conducted. One of the two cases (50%) classified as medullary plasmacytoma demonstrated conversion to multiple myeloma, whereas only 2/18 cases of EMP (11%) converted to multiple myeloma.

Ozsakin et al. (199) published a compilation of solitary plasmacytoma (42) in the head and neck. There were 258 patients with bone (n = 206) or extramedullary (n = 52) plasmacytomas without evidence of MM. Most (n = 214) of the patients received RT alone; 34 received chemotherapy and RT; and eight had surgery alone. The median radiation dose was 40 Gy. Median follow-up was 56 months (range 7 to 245 months). The median time for MM development was 21 moths (range 2 to 135 months), with a 5-year survival probability of 45% (Fig. 45.13A). The 5-year overall survival, diseasefree survival, and local control rates were 74%, 50%, and 86% respectively (Fig. 45.13B). On multivariate analyses, favorable factors were younger age and tumor size <4 cm for survival; age, extramedullary localization, and RT for diseasefree survival; and small tumor and RT for local control. Bone localization was the only predictor of MM development. No dose-response relationship was found for doses >30 Gy, even for larger tumors.

Figure 45.13.A: Probability of progression to multiple myeloma according to bone (dotted line) or extramedullary (solid line) solitary plasmacytoma (p = 0.0009). B: Overall survival correlated with bone (dotted line) or extramedullary (solid line) solitary plasmacytoma (p = 0.04).

Tournier-Rangeard et al. (265), in a review of 17 patients with solitary EMP in the head and neck, noted a local tumor control of 100% for patients who received 45 Gy or greater dose to the CTV versus 50% with doses below 45 Gy (p = 0.034). Prognostic factor for 5-year disease-specific survival (81.6%) was local tumor control (p = 0.058). Prognostic factors for diseasefree survival (64.1%) were monoclonal immunoglobulin secretion (p = 0.008) and CTV dose >45 Gy (p = 0.056).

Table 45.13 summarizes the doses of irradiation and probability of tumor control reported by various investigators. Our limited experience confirms the efficacy of tumor doses of 45 to 50 Gy for local tumor control. In patients who had extensive disease, a higher dose (50 to 60 Gy) was used, as recommended by several investigators (51).

Table 45.13. Extramedullary Plasmacytoma of Head and Neck Treated by Radiation Therapy