Controlling advanced neck disease: efficacy of neck dissection and radiotherapy.

OBJECTIVE: Neck dissection remains the standard method of treating cervical metastasis from head and neck squamous cell carcinoma. In light of recent trends to modify the classic radical neck dissection (RND) for early neck disease, we reviewed our experience with radical and modified RND (MRND) plus radiotherapy as treatment for N2/N3 neck disease in head and neck squamous cell carcinoma. METHODS: We retrospectively reviewed our clinical records from July 1989 to June 1996 to identify 43 neck dissections in 39 patients who were found to have pathologically N2 or N3 neck disease treated primarily by neck dissection and postoperative radiotherapy. All patients had head and neck squamous cell carcinoma with a minimum follow-up of 24 months. RESULTS: Nine percent (4/43) of the dissected necks were pathologically N2a, 72% (31/43) were N2b, 7% (3/43) were N2c, and 12% (5/43) were N3. Of these, 28% (12/43) underwent a RND and 72% (31/43) underwent a MRND. The most common modification of RND was preservation of the spinal accessory nerve. All patients underwent postoperative radiotherapy with a mean dose of 55 Gy. Only 4 of 43 dissected necks had isolated treatment failure, for a regional control rate of 91%. CONCLUSIONS: The combination of RND or MRND and radiotherapy is highly effective in controlling neck disease in the absence of persistent or recurrent local disease. Also, in our experience, MRND appears to be as effective as RND in controlling even advanced neck disease, which supports preservation of the spinal accessory nerve whenever oncologically feasible.

Sequential chemotherapy and radiotherapy for organ preservation in advanced resectable nonlaryngeal head and neck cancer.

Although there is no definite survival advantage to the use of sequential induction chemotherapy (CT) followed by radiotherapy (RT) in advanced resectable laryngeal cancer, this approach does succeed in preserving the larynx in many of these patients. The authors performed this study to analyze their results using a similar approach for patients with advanced resectable cancer located outside the larynx who would have required a total laryngectomy for oncologic or functional reasons. A retrospective study was performed at a single institution that included all patients with advanced resectable nonlaryngeal head and neck cancer treated with induction CT between January 1990 and August 1995. A total of 19 patients were included, with primary cancers located in the oropharynx in 14 patients, the hypopharynx in four, and the oral cavity and oropharynx in one. Eight patients had clinical stage III disease, and 11 patients had stage IV disease. Our treatment protocol consisted of two cycles of induction CT with cisplatin and 5-fluorouracil, followed by a third cycle of CT and subsequent RT in patients who achieved at least a clinical partial response (PR) after two courses of induction CT. Eighteen of 19 patients were evaluable for response. Overall, 13 patients (72%) had a major response (PR or CR) to induction CT at the primary site, and eight patients (57%) had a major response to chemotherapy in the neck. With a mean follow-up of 53 months (range, 24-71 months), the disease-specific survival was 57% for those patients with cancer of the oropharynx and oral cavity. In the subset of patients with hypopharynx cancer, 3 of 4 patients died of cancer despite achieving major response to induction CT. Organ preservation using sequential CT and RT for advanced resectable nonlaryngeal head and neck cancer is feasible, and the results in our experience with cancer of the oropharynx were similar to those reported for primary laryngeal cancer. Our limited experience using this protocol for cancer of the hypopharynx has been disappointing.

Role of angiogenic factors: coexpression of interleukin-8 and vascular endothelial growth factor in patients with head and neck squamous carcinoma.

OBJECTIVE/HYPOTHESIS: Angiogenesis has been used as a prognostic indicator in a variety of cancers and is believed to be controlled by angiogenic factors, including the cytokines interleukin-8 (IL-8) and vascular endothelial growth factor (VEGF). We hypothesized that the in vivo coexpression of both IL-8 and VEGF in head and neck tumors contributes to perpetuating tumor growth and metastasis by enhancing angiogenesis. METHODS: Immunohistochemical analysis for IL-8 and VEGF was performed using specimens from 33 cancer patients and 6 control patients. We quantitatively evaluated levels of IL-8 and VEGF in tumor tissue homogenates from those same patients using enzyme-linked immunosorbent assay and radioimmunoassay. Comprehensive histories of each patient were taken and later analyzed for clinical correlations with IL-8 or VEGF levels. RESULTS: IL-8 and VEGF were found to be colocalized within the head and neck squamous cell carcinoma (HNSCCA) tumor cells. In the head and neck tumor specimens, IL-8 levels ([38,152+/-1.8]x10(5) pg/mg total protein [TP]) were 22-fold greater than controls (1,721+/-2,122 pg/mg TP). The tumor levels of VEGF (1,304+/-6,037 pg/mg TP) were nearly fourfold higher than the controls (317+/-400 pg/mg TP. Interleukin-8 and VEGF levels were found to have a positive correlation (P< or = .0001). Patients exhibiting high levels in picograms per milligram of TP and/or number of moles of IL-8 and VEGF were found to clinically have more aggressive disease manifested by higher TNM stage, more recurrences, and shorter disease-free intervals (P< or =.03) CONCLUSIONS: Marked increase in HNSCCA of IL-8 and VEGF underscores the importance of these angiogenic factors in this disease. Understanding the roles and interplay of angiogenic factors such as IL-8 and VEGF may have value in the treatment of HNSCCA.

Coexpression of interleukin-8 receptors in head and neck squamous cell carcinoma.

BACKGROUND: Interleukin 8 (IL-8) is an important cytokine involved in tumor growth and angiogenesis in a variety of malignancies. We hypothesize that IL-8 plays an important role in the cellular proliferation and angiogenesis seen in head and neck squamous cell carcinoma (HNSCC) and set out to identify its receptors, IL-8RA and IL-8RB. METHODS: Immunohistochemical analysis was performed on specimens from 38 HNSCC patients with stage I to IV disease and control tissues. RESULTS: All of cancer specimens demonstrated positive staining for IL-8RA. The IL-8RA staining of microvessel endothelial cells was seen in 51%. The IL-8RB pattern was similar to the IL-8RA pattern in that 97% of cancer sections demonstrated positive cancer cell staining, and 74% of the specimens demonstrated positive staining for microvessel endothelial cells. CONCLUSION: Our studies demonstrate that IL-8 receptors are expressed by cancer cells and microvessel endothelial cells in HNSCC, suggesting that IL-8 may act in an autocrine/paracrine fashion to stimulate cellular proliferation and angiogenesis.

Vascular endothelial growth factor expression in head and neck squamous cell carcinoma.

BACKGROUND: Angiogenesis is an essential process required for growth and metastasis in cancer. In breast, gastric, and prostate cancer, vascular endothelial growth factor (VEGF) has been implicated in angiogenesis; however, little is known about VEGF in HNSCC. In this study, we hypothesize that VEGF is present in elevated levels in HNSCC and may therefore play a role in promoting angiogenesis. METHODS: We obtained tumor tissue from 63 HNSCC patients undergoing primary resection. All tissue samples were analyzed by immunohistochemistry (IHC) techniques for the presence and localization of VEGF; however, only 36 had sufficient amounts of tissue for quantitative analysis of VEGF by ELISA. Nine control specimens taken from patients undergoing uvulopalatopharyngoplasty were also analyzed. RESULTS: In all 63 of our patient samples we found VEGF to be present and localized to the cancer cells and endothelial cells. The poorly differentiated cancer cells stained more intensely in comparison with the well-differentiated ones. There was a 20-fold increase in the patient levels when compared with controls levels (P > or =0.05). Analysis by enzyme-linked immunosorbent assay revealed elevated mean levels of VEGF (241 +/- 326 pg/mg total protein [TP]) with a range of 2 to 1484 pg/mg TP. The control specimens had mean levels of 13 +/- 11 pg/mg TP and a range of 1 to 78 pg/mg TP. Patients who exhibited higher levels of VEGF tended to have a higher rate of disease recurrence (P < or =0.048) and shorter disease-free interval (P < or =0.05). CONCLUSIONS: The expression of VEGF in elevated levels in the HNSCC tumor microenvironment appears to be associated with more aggressive disease. Based on our results, VEGF may be an important angiogenic factor associated with cancer cells and endothelial cells in HNSCC. Further studies are needed to better define the role of VEGF in HNSCC and its role as a potential target for therapeutic intervention.

Decreased expression of transforming growth factor beta receptors on head and neck squamous cell carcinoma tumor cells.

BACKGROUND: Transforming growth factor beta (TGF-beta) has antiproliferative effects on normal and neoplastic cells that express specific TGF-beta receptors. We hypothesize that diminished expression of TGF-beta and/or its receptors may contribute to the uncontrolled proliferation of head and neck squamous cell carcinoma (HNSCCA) cancer cells. METHODS: Using immunohistochemical techniques, we characterized the expression of TGF-beta isoforms and TGF-beta receptors, TGF-beta(RI) and TGF-beta(RII), in HNSCCA. Tumor production of TGF-beta was evaluated in culture supernatants from a cytokine-stimulated HNSCCA tumor line (HTB-43). RESULTS: All control specimens displayed strong cell-associated staining of TGF-beta as well as both receptors. Forty-seven of 47 cancer specimens exhibited positive staining for TGF-beta in the tumor matrix. Forty of the 47 cancer specimens demonstrated no expression of TGF-beta(RI), and 43 of the 47 expressed no TGF-beta(RII). Only interleukin 1 alpha (IL-1 alpha) and IL-1 beta induced significant TGF-beta expression from the HTB-43 cells. CONCLUSIONS: Decreased expression of TGF-beta receptors may play a significant role in the pathogenesis of HNSCCA by allowing uncontrolled cell proliferation.

Interleukin-1 receptor antagonist in head and neck squamous cell carcinoma.

BACKGROUND: We hypothesized that in head and neck squamous cell carcinoma, the overexpression of protumorigenic interleukin-1 (IL-1) activity within the tumor tissue is a result of decreased expression of the specific antagonist or inhibitor (ie, IL-1 receptor antagonist) by the tumor cells. Ultimately, this local overexpression of IL-1 activity increases tumor growth and metastasis. DESIGN: To test our hypotheses, immunologic analysis for IL-1 alpha, IL-1 beta, and IL-1 receptor antagonist was performed on histologic sections and tumor homogenates of human head and neck squamous cell carcinomas. SETTING: University teaching hospital. PATIENTS OR OTHER PARTICIPANTS: Normal and tumor specimens were obtained from patients undergoing surgical resections of the head and neck for benign and malignant disease. RESULTS: Immunohistochemical analysis demonstrated the presence of IL-1 alpha, IL-1 beta, and IL-1 receptor antagonist within tumor cells and inflammatory cells in the tumor stroma in 19 of 19 tumor specimens. Quantitatively, IL-1 alpha was present in 19 of 19 tumor specimens (1.97 +/- 0.46 ng/mg of total protein [mean +/- SD]) and 5 of 9 normal specimens (0.23 +/- 0.12 ng/mg of total protein). All specimens contained IL-1 beta in detectable quantities (1.60 +/- 0.29 ng/mg of total protein in tumor specimens and 0.189 +/- 0.04 ng/mg of total protein in normal specimens). All specimens contained IL-1 receptor antagonist (368.87 +/- 57.63 ng/mg of total protein in tumor specimens and 585.10 +/- 166.03 ng/mg of total protein in normal specimens). The mean total IL-1/IL-1 receptor antagonist ratio was 13.26 +/- 2.31 in patients with cancer compared with 0.997 +/- 0.26 in normal patients. CONCLUSIONS: The increased IL-1 index in the cancer state compared with the normal state reflects an imbalance of IL-1 and IL-1 receptor antagonist, which may contribute to unrestricted growth and metastasis of head and neck squamous cell carcinoma.

Enhanced tumor cell expression of tumor necrosis factor receptors in head and neck squamous cell carcinoma.

BACKGROUND: To determine if tumor necrosis factor (TNF) receptors are upregulated in tumor cells, we measured the distribution and levels of TNF-alpha and TNF-beta, and TNF receptors RI and RII, in head and neck squamous cell carcinoma (HNSCC) tumor specimens and normal control specimens. METHODS: HNSCC and control tissue specimens were analyzed qualitatively using immunohistochemistry and quantitatively using immunoassays. RESULTS: Immunohistochemical analysis revealed that TNF-alpha, TNF-beta, TNF RI, and TNF RII antigens were associated predominately with tumor cells in the tissue. Quantitative analysis of TNF factors and receptors in tissue homogenates (mean levels +/- standard error of the mean, in pg/mg of total protein) indicated that: (1) TNF-alpha levels in cancer patients were not statistically different from levels in normal tissues (7.27 +/- 0.91 versus 4.62 +/- 1.33, respectively, P < 0.11); (2) TNF-beta levels in cancer patients were one third of those in normal tissue (5.07 +/- 1.83 versus 16.06 +/- 3.26, respectively, P < 0.01); and (3) both TNF RI and TNF RII levels were consistently elevated two- to four-fold in the cancer tissue when compared to normal tissue levels (1,228.72 +/- 125.67 versus 650.33 +/- 187.70, P < 0.01; and 823.39 +/- 95.90 versus 230.03 +/- 153.01, P < 0.002, respectively). CONCLUSIONS: In HNSCC, enhanced expression of TNF receptors on the cancer cells occurs and is likely to contribute to the regulation of TNF and its activation of tumor cells within the tumor microenvironment; targeting these receptors in cancer cells may provide a new approach to controlling tumor growth and metastasis.

Cytokine regulation of gelatinase production by head and neck squamous cell carcinoma: the role of tumor necrosis factor-alpha.

Gelatinases (GLs) belong to a family of enzymes known as matrix metalloproteinases (MMPs), which are produced by both normal and neoplastic cells. These enzymes have been implicated in tumor invasion and metastasis, although the mechanism of regulation of tumor MMP production is unknown. Since our previous studies have shown that numerous cytokines are present in the tumor microenvironment, our goal was to establish the effect of selected cytokines on GL production by both established tumor cell lines and primary cultures of head and neck squamous cell carcinoma (HNSCC). Supernatants of HNSCC cell lines SCC-25 and FADU stimulated with interleukin (IL)-1 alpha and IL-1 beta demonstrated modest induction of 92 kd GL production by zymogram analysis when compared with controls; IL-2, IL-6, and interferon-gamma had no consistent effect on MMP production. Stimulation of cell lines with tumor necrosis factor (TNF)-alpha (10(4) to 10 U/mL), however, dramatically enhanced production of 92 kd GL by both cell lines in a dose-dependent fashion, although tissue inhibitor of metalloproteinase (TIMP) expression was unaffected. Northern blot analysis showed that this enhancement of 92 kd GL occurred at the messenger RNA level. Stimulation of short-term primary tumor cultures with TNF-alpha resulted in significant enhancement of 92 kd GL expression in one of four cultures and enhancement of 72 kd GL expression in all cultures. The observed increase in GL expression by TNF-alpha suggests a role for this cytokine in the regulation of GL expression by tumor cells during invasion and metastasis.

Nonsquamous cell malignant neoplasms of the nasal cavities and paranasal sinuses.

BACKGROUND: Nonsquamous cell malignant tumors of the nasal cavity and paranasal sinuses are an uncommon and diverse group of neoplasms. Our goal was to define the relative frequency and distribution of these neoplasms, as well as treatment and outcome. METHODS: This retrospective study analyzed 106 patients with nonsquamous cell malignant neoplasms of the nasal cavity and paranasal sinuses treated at our institution between 1966 and 1982. RESULTS: Pathology included: salivary type carcinoma (33 patients); sarcoma (25 patients); melanoma (18 patients); esthesioneuroblastoma (11 patients); lymphoma (11 patients); and anaplastic cancer (9 patients). Most tumors arose in the nasal cavity (50%), followed by the antrum (39%), ethmoid sinus (9%), and frontal sinus (2%). Fifty-four percent of all patients were treated with surgery alone. Determinate 5-year and 10-year cure rates were: esthesioneuroblastoma, 70% and 50%; lymphoma, 45% and 30%; anaplastic, 33% and 25%; salivary, 31% and 18%; sarcoma, 25% and 21%; and melanoma, 19% and 0%. CONCLUSIONS: Overall survival for patients with nonsquamous cell malignant neoplasms of the nasal cavity and paranasal sinuses is poor, particularly with follow-up extended to 10 years. Local recurrence is the major source of treatment failure. We were unable to demonstrate any survival advantage in the group of patients who received surgery and radiotherapy in combination.

Interleukin-8 expression by head and neck squamous cell carcinoma.

OBJECTIVE: To test the hypothesis that interleukin-8 (IL-8) is produced by human head and neck squamous cell carcinomas (HNSCCAs) and may therefore be a possible mediator for lymphocyte recruitment and neovascularization by these tumors. METHODS: Nine fresh samples of HNSCCA were analyzed for expression of IL-8 antigen using radioimmunoassay and immunohistochemical staining techniques. Also, four short-term primary cultures of HNSCCA and two continuous HNSCCA cell lines were then analyzed for production of IL-8 expression under both baseline conditions and following stimulation with other cytokines. RESULTS: The IL-8 antigen was detected in all fresh tumor homogenates by radioimmunoassay (5.58 to 331.69 ng of IL-8 per gram of tissue), and immunohistochemical results localized staining predominantly within the tumor cells. Primary cultures of HNSCCA and continuous HNSCCA cell lines produced only low levels of IL-8 (0.04 to 4.49 ng of IL-8 per 10(6) cells) under baseline (unstimulated) conditions. Stimulation of both primary cultures and cell lines with interleukin-1 and tumor necrosis factor induced significant increases in IL-8 antigen, while other cytokines failed to induce a significant increase. CONCLUSIONS: This study demonstrates that IL-8 antigen is expressed by HNSCCA in vivo, and that cultured HNSCCA in vitro can be stimulated to express IL-8 antigen by both interleukin-1 and tumor necrosis factor. Local production of IL-8 by HNSCCA cells, and its regulation by other cytokines, may be important in both the lymphocyte recruitment and tumor neovascularization seen in HNSCCA, and may thus ultimately affect the natural history of the disease.

The pectoralis major myocutaneous flap in reconstructive head and neck surgery revisited: a recent experience.

Although the pectoralis major myocutaneous flap (PMMF) has proved to be a reliable method of reconstruction for the head and neck surgeon, the current emphasis in tissue replacement has shifted to microvascular free-tissue transfer. This has prompted us to review retrospectively all PMMFs performed for reconstruction following head and neck cancer resection at the University of Connecticut Health Center over a recent four-year period. A total of 24 flaps were utilized consecutively in 23 patients, following a variety of ablative procedures. Postoperative complications occurred in 14 of the 24 flaps (59%), and 11 of the 14 patients who experienced postoperative complications had received previous radiation therapy. Seven of the complications were felt to be major because they prolonged hospital stay, or they required secondary reconstructive procedures (four of seven cases). There were no instances of total flap failure. In our experience, the PMMF has proved to be reliable and useful in a variety of head and neck reconstructive procedures. We conclude that the PMMF remains an excellent option for repair of defects resulting from head and neck cancer resection.

Phospholipid metabolite expression by head and neck squamous cell carcinoma.

OBJECTIVE: To characterize the presence and production of various phospholipid metabolites by head and neck squamous cell carcinoma (HNSCC) and squamous cell carcinoma cell lines in vitro and in vivo. DESIGN: The HNSCC tumor homogenates and supernatants of HNSCC tumor cultures and established squamous cell carcinoma cell lines were assayed for prostaglandin E2 (PGE2), leukotriene B4 (LTB4), and platelet activating factor (PAF). In vitro experiments were carried out under baseline conditions or with exposure to several known immunomodulators (epidermal growth factor, bacterial lipopolysaccharide, and interleukin 1). PATIENTS: The HNSCC tumor tissue was obtained from primary tumor or cervical lymph node metastasis of surgical resections. MAIN OUTCOME MEASURES: Prostaglandin E2, LTB4, and PAF were measured in tumor homogenates and cell culture supernatants using standardized radioimmunoassay kits. RESULTS: All tumor homogenates (eight of eight) contained detectable levels of PGE2 (range, 324 to 2258 pg/g of tumor tissue) and LTB4 (range, 790 to 41,900 pg/g of tumor tissue); PAF was detected in six of eight homogenates (range, 7362 to 40,788 pg/g of tumor tissue). All of the short-term primary HNSCC tumor cultures and squamous carcinoma lines produced PGE2 (range, 90 to 1160 pg/10(6) cells), and half of the cultures produced LTB4 (range, 100 to 1700 pg/10(6) cells); none of the cultures or cell lines produced detectable levels of PAF. Interleukin 1 significantly enhanced production of PGE2 by tumor cultures (P < .02). Characterization of tumor cultures with a fibroblast antibody marker, BR2, revealed that 26% to 64% of tumor culture cells were fibroblasts. CONCLUSIONS: Prostaglandin E2, LTB4, and PAF are present in the tumor microenvironment, where they may be involved in the local immunosuppression phenomenon seen in HNSCC. Both PGE2 and LTB4 were produced in vitro by tumor cultures and squamous cell carcinoma cell lines; PAF was not produced by tumor cultures in vitro and therefore may be a product of local immune cells in HNSCC in vivo. Interleukin 1 and PGE2 may interact in immunoregulation in the HNSCC tumor microenvironment.

Cytokine expression by head and neck squamous cell carcinomas.

Cytokines are known to play an important role in host defense by regulating the function, growth, and differentiation of the cells of the immune system. We hypothesize that, in the tumor microenvironment, tumor cells and resident tissue cells (e.g., fibroblasts) also produce cytokines that may regulate the local immune response to tumors. Initially, homogenates of eight head and neck squamous cell carcinomas (HNSCC) were assayed for the presence of interleukin-1 (IL-1), interleukin-4 (IL-4), interleukin-6 (IL-6), and granulocyte-macrophage colony-stimulating factor (GM-CSF) to establish the presence of these cytokines in the tumors in vivo. We detected IL-1 in all tumor homogenates and IL-4, IL-6, and GM-CSF in some homogenates. To assess the ability of HNSCC to produce these cytokines, supernatants of short-term primary cultures of HNSCC were assayed for the same cytokines. No IL-1 was detected, although baseline levels of IL-4, IL-6, and GM-CSF were present. However, the stimulation of primary tumor cultures with exogenous IL-1 induced or significantly enhanced production of IL-4 (p < 0.01), IL-6 (p < 0.001), and GM-CSF (p < 0.02). These results support our hypothesis that HNSCC secrete cytokines that may influence the response of local immune cells. Our data also suggest that IL-1 may have a central role in regulating the local immune response through the enhancement or induction of cytokine production by tumor and/or resident tissue cells.

Olfactory loss and allergic rhinitis.

Olfactory loss is of importance for allergists to investigate in their patients, because if it is due to either allergic rhinitis or nonallergic rhinitis, it is potentially reversible. One should be sure to consider nasal polyposis and inflammation from chronic sinusitis, especially of the ethmoidal sinuses. Simple screening in the office can be achieved with an odor identification test of widely available substances as described above. Should there be no response to treatment or if the patient has a history of chronic sinusitis, recalcitrant nasal polyposis, or previous otolaryngologic procedures, further evaluation including rhinoscopy may be required. Recent olfactory loss in the absence of nasal symptoms and in the absence of abnormalities in the nasal cavity should suggest further investigation to look for a more central process. Morphologic investigation with electron microscopy of the olfactory epithelium and the superior nasal cavity is just beginning. The impact of inflammation in this area awaits investigation.

The management of chyle fistula.

Over a recent 4-year period, 823 neck dissections that included the lower jugular lymph nodes were performed. Of the 823, 14 (1.9%) patients developed chyle fistulas. Two other patients developed fistulas, one after undergoing a gastric transposition, and the other after a scalene node biopsy. All 16 patients were initially managed conservatively with closed-wound drainage and low-fat nutritional support; this was successful in only 4 patients, 3 of whom had peak 24-hour chyle drainage of less than 600 cc. The remaining 10 patients required open-wound management, which included operative ligation in 4 instances. Continued conservative treatment with an open neck wound resulted in significant additional hospitalization. Our experience indicates that closed-wound management of a chyle fistula is likely to fail when peak 24-hour fistula output exceeds 600 cc. Considering the cost and morbidity of conservative treatment, early reoperation may be appropriate in those patients with high fistula output.

Squamous carcinoma of the nasal cavity and paranasal sinuses.

This study retrospectively analyzed 105 patients with squamous carcinoma of the nasal cavity and paranasal sinuses. The primary tumor was located in the maxillary sinus in 65 patients (62 percent), the nasal cavity in 27 (26 percent), the ethmoid sinus in 11 (10 percent), and the sphenoid sinus in 2 (2 percent). Over half of the patients with antral cancer were treated with surgery and radiotherapy, whereas one-third of the remaining patients received combination therapy. Most procedures were radical, including sacrifice of the orbital contents in half of the surgically treated patients. The 5-year determinate cure rate was 45 percent for patients with nasal cavity tumors, 38 percent for those with maxillary sinus lesions, and 13 percent for those with ethmoid tumors. Local recurrence remains a major problem despite aggressive surgery and increased use of adjunctive radiotherapy.