ICAM-5 (telencephalin) gene expression in head and neck squamous carcinoma tumorigenesis and perineural invasion!

ICAM-5 (telencephalin) is an intercellular adhesion molecule reported to be expressed only in the somatodendritic membrane of telencephalic neurons. We recently identified high ICAM-5 expression in a cDNA array study of head and neck neoplasms with a propensity for perineural invasion. To determine the association of this gene in tumorigenesis and perineural invasion, we analyzed the expression and functional status of ICAM-5 mRNA transcripts in 30 different human cancer cell lines and 25 head and neck squamous carcinoma specimens by reverse-transcriptase polymerase chain reaction (cell lines and specimens) and in vitro functional assays (cell lines). ICAM-5 transcripts were detected in 28 (93%) of 30 cell lines derived from primary head and neck, colon, thyroid, cervical, pancreatic, skin, and adenoid cystic carcinomas. In cell lines, small interfering RNA blocked ICAM-5 expression and inhibited cell proliferation. Treatment with the phosphatidylinositol 3'-kinase (PBK) inhibitor LY294002 resulted in ICAM-5 down-regulation. In tissue specimens, none of the 25 histologically normal oral mucosal specimens had detectable ICAM-5 level, whereas 16 (64%) of the 25 matched primary squamous carcinomas showed expression. Carcinoma specimens high ICAM-5 expression had a high incidence of perineural invasion. Our study indicates that ICAM-5 may play a role in tumorigenesis and perineural invasion, most likely through the P13K/Akt-signaling pathway.

Advanced oropharyngeal carcinoma treated with surgery and radiotherapy: oncologic outcome and functional assessment.

INTRODUCTION/PURPOSE: Surgery and postoperative radiotherapy (XRT) is a standard therapy for locally advanced resectable oropharyngeal carcinoma. This maximizes local-regional control, but does not address the potential for occult distant metastases. Additionally, some patients may suffer poor functional outcome after this intensive local therapy. This report reviews our institutional experience with modern radical surgery and XRT for this disease. METHODS: A retrospective chart review was performed on 51 consecutive patients treated from 1991 to 1997 at the University of Pennsylvania with radical surgery and postoperative XRT. This study included patients with locally advanced, stage III/IV (exclusive of T1-2N1) squamous carcinoma of the oropharynx. All patients had a good performance status (ECOG 0-1). Patients who received adjuvant chemotherapy were excluded. No patient had gross residual disease after surgery; the median XRT dose was 63.7 Gy. Survival, local-regional control (LRC), and freedom from distant metastases (DM) were calculated actuarially. In patients who remained free of disease, functional status was determined using the List Performance Status Scale (PSS). RESULTS: With a median follow-up in surviving patients of 34 months, the 3-year actuarial overall survival was 51%. The 3-year LRC was 73%, and the freedom from DM was 69%. The most significant factor predicting for failure was the number of pathologically positive nodes (P <.001 for survival and DM; P =.003 for LRC). In 29 patients who were evaluable for the List PSS, the mean normalcy-of-diet score was 48; the mean eating-in-public score was 53; and the mean understandability-of-speech score was 75. There was a trend toward better PSS scores in patients with T1-2 tumors versus T3-4 tumors, although this did not reach statistical significance. CONCLUSIONS: Surgery and postoperative XRT offer relatively good LRC and moderate overall survival rates. Results, however, remain suboptimal, particularly with respect to the risk of DM and the functional outcome. These data provide a baseline for comparison with maturing results from multimodality trials in which radical surgery is not used in all patients with locally advanced oropharyngeal carcinoma.

Feasibility and outcome of endoscopic staple-assisted esophagodiverticulostomy for Zenker's diverticulum.

OBJECTIVES/HYPOTHESIS: Endoscopic staple-assisted esophagodiverticulostomy (ESED) is a newly described method of surgically correcting Zenker's diverticulum. Initial reports on the ease and success of the surgery have been quite enthusiastic, making it seem the procedure of choice. We initiated the procedure in an algorithm of treatment of Zenker's diverticulum, to further explore the feasibility and outcome of this new technique. STUDY DESIGN: This is a case series of 23 patients with Zenker's diverticulum who have undergone surgical repair. For each patient, an attempt at ESED was made. If unsuccessful, an open approach was then taken. RESULTS: Seven of 23 patients (30%) were unable to be treated with ESED because of inability to expose the diverticulum or unfavorable anatomy of the diverticulum itself. Of the remaining 16 patients, ESED was successful in resolving the symptoms of diverticulum in 14 (87%). Two patients (13%) were somewhat improved but had persistent dysphagia. No significant complications occurred. All patients resumed oral diet within the first 24 hours after surgery. CONCLUSION: Esophagodiverticulostomy is an excellent method of surgically correcting Zenker's diverticulum in many patients, but anatomical considerations may prevent its use, making open approaches of continued importance in a surgeon's armamentarium.

Internal jugular vein versus external jugular vein anastamosis: implications for successful free tissue transfer.

BACKGROUND: Microvascular free flaps are becoming the reconstructive option of choice for many head and neck defects. Many previous studies have examined factors predicting free flap survival. No study has compared differences in free flap survival when anastomosed to the internal or external jugular systems. METHODS: Retrospective review of all free flaps performed at an academic medical center by a single head and neck microvascular surgeon during the period July 1995 to December 1999. Flaps were closely monitored postoperatively and taken back to the operating room urgently for arterial insufficiency or venous congestion. RESULTS: On hundred fifty-six free flaps were performed during this time period. Sixty-five free flaps were anastomosed to the external jugular (EJ) vein and 86 to the IJ system (62 to the proximal common facial vein, 17 end-side on the IJ, and 7 to other branches). Five had either two venous anastomoses or were anastomosed to other veins and were excluded from statistical analysis. Six (4%) vascular thromboses occurred; 5 were venous and 1 arterial. Success by group was 99% for IJ anastomosis (1 arterial thrombosis) and 92% for EJ anastomosis (5 venous thromboses, p =.03). Urgent anastomotic revision and reperfusion salvaged 5 of the 6 flaps (overall success 99%). CONCLUSIONS: Although the overall success rate (96% success with 99% success with salvage) is comparable to other large series, microvascular free flaps anastomosed to the external jugular vein failed at a significantly higher rate than those anastomosed to the IJ system. This suggests that the IJ system should be used as a recipient vessel when feasible.

Phase I study of paclitaxel given by seven-week continuous infusion concurrent with radiation therapy for locally advanced squamous cell carcinoma of the head and neck.

PURPOSE: Paclitaxel is one of the most active agents for squamous cell carcinoma of the head and neck (SCCHN) and an in vitro radiosensitizer. The dose-response relationship for paclitaxel may depend more on exposure duration than on peak concentration. This National Cancer Institute-sponsored phase I trial was designed to determine the feasibility of combining continuous-infusion (CI) paclitaxel with concurrent radiation therapy (RT). PATIENTS AND METHODS: Patients with previously untreated stage IVA/B SCCHN were eligible. Primary end points were determination of the maximum-tolerated dose, dose-limiting toxicity, and pharmacokinetics for paclitaxel given by CI (24 hours a day, 7 days a week for 7 weeks) during RT (70 Gy/7 weeks). RESULTS: Twenty-seven patients were enrolled and assessable for toxicity. Nineteen of the patients who completed > or = 70 Gy were assessable for response. Grade 3 skin and mucosal acute reactions occurred at 10.5 mg/m(2)/d, but uninterrupted treatment was possible in five of six patients. At 17 mg/m(2)/d, skin toxicity required a 2-week treatment break for all three patients. The mean paclitaxel serum concentration at dose levels > or = 6.5 mg/m(2)/d exceeded that reported to achieve in vitro radiosensitization. Initial locoregional control was achieved in 14 (58%) of 24 of patients treated to 70 Gy, and control persisted in nine (38%). CONCLUSION: CI paclitaxel with concurrent RT is a feasible and tolerable regimen for patients with advanced SCCHN and good performance status. Preliminary response and survival data are encouraging and suggest that further study is indicated. The recommended phase II dose of paclitaxel by CI is 10.5 mg/m(2)/d with RT for SCCHN.

Tumor deposition of laminin-5 and the relationship with perineural invasion.

OBJECTIVES/HYPOTHESIS: Perineural invasion (PNI) is increasingly being recognized as an important indicator of aggressiveness in head and neck squamous cell carcinoma. The mechanisms of PNI are poorly understood. Laminin-5, an important basement membrane constituent, has been shown to be essential in head and neck squamous cell carcinoma invasion and motility. We hypothesized that tumors exhibiting increased expression of laminin-5 are more likely to be neurotropic. STUDY DESIGN: Analysis of archived surgical specimens with and without PNI for presence and intensity of laminin-5 tumor staining. METHODS: Immunohistochemistry of archived head and neck squamous cell carcinoma specimens with known PNI was performed with anti-laminin-5 antibodies and appropriate positive and negative control specimens. The staining patterns were characterized as follows: A, few to no tumor cells positive; B, some peripheral cells positive; C, all peripheral cells positive; and D, almost all tumor cells positive. Statistical analysis was by chi2 analysis. RESULTS: Forty-six PNI-positive and 18 PNI-negative specimens were analyzed. The staining distribution for the PNI-positive specimens was as follows: 2% for A, 41% for B, 46% for C, and 11% for D. For tumors without PNI, the distribution was 28% for A, 50% for B, 22% for C, and 0% for D (P = .005). In PNI-positive tumors, no significant difference in staining was seen between areas with and without PNI. CONCLUSIONS: We found a significant correlation between laminin-5 staining and the presence of PNI in head and neck squamous cell carcinoma. Expression of laminin-5 by tumors is, possibly, an important step in the process of PNI. These preliminary findings support the concept that deposition of basement membrane constituents are required in the multistep process of nerve invasion.

Evidence for diet effects on the composition of silk proteins produced by spiders.

Silks are highly expressed, secreted proteins that represent a substantial metabolic cost to the insects and spiders that produce them. Female spiders in the superfamily Araneoidea (the orb-spinning spiders and their close relatives) spin six different kinds of silk (three fibroins and three fibrous protein glues) that differ in amino acid content and protein structure. In addition to this diversity in silks produced by different glands, we found that individual spiders of the same species can spin dragline silks (drawn from the spider's ampullate gland) that vary in content as well. Freely foraging ARGIOPE: argentata (Araneae: Araneoidea), collected from 13 Caribbean islands, produced dragline silk that showed an inverse relationship between the amount of serine and glycine they contained. X-ray microdiffraction of the silks localized these differences to the amorphous regions of the protein that are thought to lend silks their elasticity. The crystalline regions of the proteins, which lend silks their strength, were unaffected. Laboratory experiments with ARGIOPE: keyserlingi suggested that variation in silk composition reflects the type of prey the spiders were fed but not the total amount of prey they received. Hence, it may be that the amino acid content (and perhaps the mechanical properties) of dragline silk spun by ARGIOPE: directly reflect the spiders' diet. The ability to vary silk composition and, possibly, function is particularly important for organisms that disperse broadly, such as Argiope, and that occupy diverse habitats with diverse populations of prey.

Synchronous appearance of germ cell tumor and papillary carcinoma of the thyroid.

Synchronous appearance of 2 different malignancies in one patient is a rare phenomenon. We describe our experience of 2 patients with synchronous malignancies of the testis and thyroid gland, and of a third patient who developed a thyroid neoplasm unrelated to recent treatment for a germ cell tumor. The medical records of 3 male patients treated for both a germ cell tumor and a thyroid cancer between 1989 and 1994 were reviewed. Two patients with nonseminomatous germ cell tumor received postoperative chemotherapy after orchiectomy and developed a papillary carcinoma of the thyroid during treatment. A third patient, who received radiation therapy for a clinical stage 1 seminoma, recurred with biopsy proven seminoma in the neck in association with a thyroid nodule 2 years later. All 3 patients had their thyroid cancer treated by surgical resection, and one received adjuvant radioactive iodine. Two of the patients are currently alive and disease-free. One patient died of pulmonary complications that stemmed from bleomycin toxicity. Synchronous appearance of germ cell tumor and papillary carcinoma of the thyroid has not been previously described. Genetic predisposition may play a role in the development of such simultaneous neoplasms.

Morbidity after midline mandibulotomy and radiation therapy.

PURPOSE: To assess the morbidity of mandibulotomy in patients treated for neoplasms of the oropharynx and oral cavity, and to determine if postoperative radiation therapy to the mandibulotomy site carries an increased risk of complications. PATIENTS AND METHODS: The medical charts of 30 patients treated between 1992 and 1996 undergoing midline mandibulotomy for tumors of the oral cavity (7 patients) and oropharynx (23 patients) were retrospectively reviewed. Three patients presented with recurrent disease, 1 of whom was previously irradiated. Twenty-five patients received postoperative radiation after mandibulotomy to a median dose of 60 Gy to the primary tumor bed, whereas 5 patients were treated with surgery alone. The patients were separated into those whose mandibulotomy site was within the radiation treatment field (n = 9), and those whose site was shielded (n = 10). Median follow-up was 27.8 months (range 5-81 months). End points included significant pain involving the mandibulotomy site, trismus, malocclusion, wound infection, osteoradionecrosis, and time to oral intake. RESULTS: There were no postoperative deaths. Minor wound infection or breakdown occurred in 4/30 patients (13%). All of these resolved with local care and parenteral antibiotics. More serious complications involving the mandibulotomy occurred in 2 patients (7%). One patient had chronic wound drainage at the mandibular osteotomy site, which healed after plate removal. Another patient developed osteoradionecrosis. No patient developed trismus or malocclusion. With a median follow-up of 27.8 months, 4 patients have recurred locally. The complication rate was 11% for patients whose mandibulotomy site was irradiated, and 30% for those whose site was shielded. CONCLUSION: Mandibulotomy can be safely performed in patients who are likely to require postoperative external radiation.

Pilot study of organ preservation multimodality therapy for locally advanced resectable oropharyngeal carcinoma.

The purpose of this study was to determine the early efficacy and toxicity of a new multimodality organ-preservation regimen for locally advanced, resectable oropharyngeal squamous cell carcinoma (SCC). Patients with T3-4N0-3M0 or T2N2-3M0 oropharyngeal SCC were eligible for this Phase II study. Patients needed the physiologic reserve for surgery and technically resectable tumors. Induction carboplatin (area under the curve = 6) and paclitaxel (200 mg/m2) x 2 cycles (q21 days) were given. Objective responders received definitive radiotherapy (XRT), 70 Gy/7 weeks with concurrent weekly paclitaxel. Initially, the dose of paclitaxel was 50 mg/m2/week; because of mucosal toxicity it was reduced to 30 mg/m2/week. Patients with N2-3 disease received post-XRT neck dissection and 2 more cycles of "adjuvant" chemotherapy. In the first 22 patients, the neutropenic fever rate was 27%. Although there has been no grade IV-V toxicity from induction therapy, grade II-III toxicity resulted in an unacceptable delay in starting XRT in 14% of patients. The response rate to induction chemotherapy was 91%. Grade III mucositis occurred in all patients during concurrent chemoradiotherapy. One patient died of pneumonia during concurrent chemoradiotherapy after receiving 26 Gy and 3 doses of paclitaxel 50 mg/m2. No dose-limiting toxicity occurred in 15 patients treated with concurrent paclitaxel 30 mg/m2/week. Actuarial overall survival at 18 months is 82%; local-regional control is 86%. To date, distant metastases have not developed in any patients. This regimen has intense but acceptable acute toxicity. The maximum tolerated dosage of weekly paclitaxel during standard continuous-course XRT is confirmed to be 30 mg/m2/week. The treatment efficacy of this regimen (response rate and short-term local-regional and distant control) is encouraging. Accrual continues to obtain long-term toxicity, efficacy, and quality-of-life data.

Allograft dermal implant (AlloDerm) in a previously irradiated field.

OBJECTIVE: To evaluate the integration of AlloDerm (LifeCell Corp., The Woodlands, TX) in a field exposed to external-beam radiation (EBR) by analyzing graft thickness, fibroblast recellularization, and neovascularization. STUDY DESIGN: Randomized control. METHODS: Thirty-six male Sprague-Dawley rats (n = 36) were randomly assigned to four groups. One hind leg of each rat was exposed to 20 Gy of EBR; the other limb served as the nonirradiated control. Two weeks after irradiation, AlloDerm was implanted into both hind legs. Grafts were harvested at 3, 4, 6, and 14 weeks after implantation and underwent histological analyses. RESULTS: There was no statistically significant difference in graft thickness, fibroblast count, or neovascularization between the grafts placed in the irradiated bed and the controls (n = 33, P = .332, P = .336, and P = .057, respectively). However, at week 3, fibroblast counts in the graft placed in the field exposed to EBR were significantly lower than those of controls (P = .019), although at week 14 the counts in the experimental limb were higher than those of the controls (P = .002). Graft thickness (P = .001) and fibroblast count (P < .004) were lower at week 14 than at earlier time periods for both the experimental and control grafts. CONCLUSIONS: In the rat model, graft thickness and neovascularization of the AlloDerm dermal implant do not appear to be adversely affected by a field that has received EBR. Fibroblast ingrowth may be hindered in the early postimplantation period but appears to normalize in the long term. Furthermore, overall graft thickness and fibroblast counts decrease over time, regardless of irradiation status.

Thyroid sclerosing mucoepidermoid carcinoma with eosinophilia: mimic of Hodgkin disease in nodal metastases.

We present the clinical and pathologic findings of a case of sclerosing mucoepidermoid carcinoma with eosinophilia of the thyroid in a 39-year-old woman. This particular case is notable because it initially presented as a cervical lymph node metastasis, and the initial clinical and histologic impression was Hodgkin disease, nodular sclerosis type. Sclerosing mucoepidermoid carcinoma with eosinophilia is a differentiated malignant neoplasm of the thyroid that can be confused with anaplastic carcinoma, medullary carcinoma, squamous cell carcinoma, or, as in this case, Hodgkin disease. A correct diagnosis of sclerosing mucoepidermoid carcinoma with eosinophilia involves awareness of this entity and appropriate immunohistochemical analysis. In this article, we briefly review the literature and stress the histologic and cytologic findings characteristic of sclerosing mucoepidermoid with eosinophilia of the thyroid.

Effects of external beam radiation on the allograft dermal implant.

OBJECTIVE: The purpose of this study was to define the effects of external beam radiation (EBR) on AlloDerm (LifeCell Corp) through the analysis of graft thickness, fibroblast recellularization, and neovascularization as a function of time. METHODS AND MATERIAL: Thirty-six male Sprague-Dawley rats (n = 36) were randomly assigned to 1 of 4 groups (A, B, C, and D). AlloDerm was implanted subcutaneously into the hind legs of each rat, and 20 Gy of EBR was administered to one side. Grafts harvested 1, 2, 4, and 12 weeks after radiation were subjected to blinded histologic analysis. RESULTS: In groups A, B, and C, the irradiated grafts showed a significant decrease in recellularization versus nonirradiated (P < 0.001). At 12 weeks (group D), recellularization equalized, but neovascularization was significantly less (P = 0.048) in the irradiated group. Graft thickness was unaffected. CONCLUSIONS: In the rat model, EBR of the implanted AlloDerm graft hinders recellularization in the early posttreatment period. However, EBR did not adversely affect graft thickness, recellularization or ultimate graft survival.

Clinical care pathways: decreasing resource utilization in head and neck surgical patients.

In this era of decreasing reimbursement, health systems have been forced to become more efficient and decrease resource utilization to remain financially viable. One of the methods of internal cost control has been the use of clinical pathways. Given the complexity of treatment of head and neck cancer patients, clinical pathways can help to standardize decision making and introduce uniformity in resource utilization. The objective of this study is to compare resource utilization and outcomes before and after implementation of a clinical pathway for head and neck surgical patients. We observed significant decreases in hospital costs as well as shorter lengths of stay after pathway implementation. It is our belief that a uniform management tool is beneficial in this complex disease.

Detection of recurrent head and neck squamous cell carcinomas after radiation therapy with 2-18F-fluoro-2-deoxy-D-glucose positron emission tomography.

OBJECTIVES/HYPOTHESIS: Fluorodeoxyglucose positron emission tomography (FDG-PET) has been proposed as a sensitive method to diagnose and stage various malignancies. We assessed the efficacy of FDG-PET imaging in distinguishing tumor persistence/recurrence from posttreatment changes following radiation therapy for squamous carcinomas of the head and neck STUDY DESIGN: Retrospective analysis of FDG-PET results compared with biopsy results or outcome, or both. METHODS: Twenty-eight patients who had undergone radiation therapy with or without surgery for treatment of squamous cell carcinoma were studied with FDG-PET imaging. There was clinical suspicion for recurrence in each patient, but no obvious mass or lesion to biopsy was found on physical examination or anatomic imaging. The results of FDG-PET imaging were compared with those of biopsy or clinical follow-up of at least 6 months, or both. RESULTS: FDG-PET imaging was positive in 13 patients, and the presence of active disease was confirmed in 12. Two thirds of the 12 received further cancer treatment. There were 15 negative FDG-PET images. Thirteen of these were confirmed true-negative images, but two studies were false-negative images. The sensitivity and specificity of FDG-PET were 86% and 93%, respectively, with positive and negative predictive values of 92% and 87%, respectively. The overall accuracy was 89%. CONCLUSION: FDG-PET imaging is a useful modality to distinguish tumor persistence/recurrence from radiation-induced tissue changes in the neck following treatment for head and neck cancer. FDG-PET can identify patients who may benefit from further treatment, and may lead to improved outcome for individual patients.

Nerve cell adhesion molecule expression in squamous cell carcinoma of the head and neck: a predictor of propensity toward perineural spread.

OBJECTIVE: To evaluate head and neck squamous cell carcinomas (SCCAs) for the expression of nerve cell adhesion molecule (N-CAM). We propose that expression of N-CAM by tumor cells may be associated with perineural invasion in SCCA of the head and neck. METHODS: Seventy-six archived specimens of histologically proven SCCA were analyzed by immunohistochemistry for the expression of N-CAM. Positive and negative controls were used to assess staining. Two sections of each specimen were reviewed for the presence of perineural invasion. A retrospective chart review was performed for each patient that corresponded to the above specimens. RESULTS: Perineural invasion was present in 28 (37%) of the 76 patients evaluated for the expression of N-CAM. N-CAM expression was demonstrated in 38 (50%) of the 76 specimens. The incidence of N-CAM expression was significantly associated with perineural invasion (P = .002). There was no significant association between the presence of staining or the presence of perineural invasion and the incidence of locoregional recurrence, distant metastasis, or survival status; however, the mean follow-up was only 13.6 months (range, 1-49 mo). CONCLUSION: There is a positive correlation between the presence of N-CAM expression and perineural invasion in SCCA of the head and neck. The expression of this adhesion molecule by tumor cells may facilitate both homophilic cell-to-cell and heterophilic cell-to-substrate adhesion, thereby enabling the tumor cells to use the perineural tissues or neural cells, or both as a conduit for perineural spread.

Seven-week continuous-infusion paclitaxel concurrent with radiation therapy for locally advanced non-small cell lung and head and neck cancers.

The goal of these National Cancer Institute-sponsored phase I trials is to determine the feasibility, toxicity, and pharmacokinetics of continuous-infusion (24 hr/d, 7 d/wk, 7 weeks total) intravenous paclitaxel combined with standard, curative-intent thoracic radiation therapy (XRT) for previously untreated, locally advanced non-small cell lung cancer and squamous cell cancer of the head and neck (HNSCC). Eligible patients have locally advanced (T4NXM0 or TXN2-3M0) non-small cell lung cancer ineligible for potentially curative surgical resection or locally advanced HNSCC with an expected 5-year survival rate of less than 25%, as well as a good performance status, adequate hematologic, hepatic, and renal function, and no distant metastases. Non-small cell lung cancer patients receive a total tumor dose of 64.8 Gy megavoltage XRT in 7 weeks at 1.8 Gy once daily, 5 d/wk. Patients with HNSCC receive 70 Gy megavoltage XRT in 7 weeks at 2 Gy once daily, 5 d/wk. Paclitaxel is delivered by continuous intravenous infusion starting 48 hours before XRT and continuing for its duration. The dose of paclitaxel is escalated in cohorts of three patients in a standard phase I design. To date, 49 patients have been entered on both studies and 43 are evaluable for toxicity. Paclitaxel dose is currently at the 17 mg/m2/d dose level, with no dose-limiting toxicity thus far. Clinical outcomes suggest significant activity for this combination. This therapy is feasible and has been well-tolerated through current dose levels. Dose escalation is ongoing.

Malignant tumors of the nose and paranasal sinuses: hospital of the University of Pennsylvania experience 1990-1997.

We reviewed our experience with sinonasal malignancies, which comprise less than 1% of all cancers, in order to determine the spectrum of disease and outcome after treatment. The medical records of 48 patients with sinonasal malignancies treated between 1990-1997 were reviewed for epidemiologic characteristics, tumor location and histology, treatment modalities, and tumor recurrence. Mean age was 58.5 years and 46% were male. Multiple sites of origin were common, including maxillary sinus (83%), ethmoid sinus (35%), and nasal cavity (40%). The histologic spectrum included squamous cell carcinoma (46%), adenoid cystic carcinoma (6%), and miscellaneous others (48%). Treatment included surgery and adjuvant radiotherapy (XRT) (58%), surgery alone (27%), XRT and chemotherapy (6%), surgery and chemotherapy (4%), and XRT alone (4%). Mean follow-up was 15 months (range 2-58). Recurrence was evident in nine patients (19%), 3 (33%) of whom had prior treatment before presenting to HUP. Of the six who recurred after initial treatment at HUP, five (83.3%) were treated with surgery and XRT and one (16.7%) was treated with surgery alone. Of the three that recurred after undergoing attempts at salvage (prior treatment and then treatment at HUP), one had received surgery alone followed by surgery and XRT, one had surgery and XRT followed by surgery and one had XRT followed by surgery alone. Our experience reveals surgery and XRT to be the modality of choice, particularly for advanced tumors, whereas surgery alone may be sufficient for small, well localized tumors. Neoadjuvant chemotherapy may offer improved local control; the future role of endoscopic surgery warrants further investigation.