Concurrent chemoradiotherapy for N2 or N3 squamous cell carcinoma of the head and neck from an occult primary

Transcription

1 Original article Annals of Oncology 14: , 2003 DOI: /annonc/mdg330 Concurrent chemoradiotherapy for N2 or N3 squamous cell carcinoma of the head and neck from an occult primary A. Argiris 1 *, S. M. Smith 2, K. Stenson 2, B. B. Mittal 1, H. J. Pelzer 1, M. S. Kies 1, D. J. Haraf 2 & E. E. Vokes 2 1 Northwestern University, The Feinberg School of Medicine and the Robert H. Lurie Comprehensive Cancer Center, Chicago, IL; 2 University of Chicago and University of Chicago Cancer Research Center, Chicago, IL, USA Received 17 January 2003; revised 27 March 2003; accepted 8 April 2003 Background: Our aim was to explore the use of concurrent chemoradiotherapy in the management of patients with squamous cell carcinoma of the head and neck from an occult primary (HNCOP). Patients and methods: From 1991 to 2000, 25 patients with T0N2M0 or T0N3M0 HNCOP were entered into five sequential phase II clinical trials. Chemoradiotherapy consisted of a split course of radiotherapy with concurrent 5-fluorouracil and hydroxyurea either alone or with cisplatin, or paclitaxel. Two of the five protocols incorporated induction chemotherapy. Results: Nodal stage was N2a in five patients (20%), N2b in 13 (52%), N2c in one (4%) and N3 in six (24%). Twenty-two patients (88%) underwent neck dissection; 14 of 22 patients underwent neck dissection before initiating protocol therapy. Total radiation doses of Gy (median 60 Gy) were delivered; radiation fields included the potential sites of mucosal primaries and the neck bilaterally. Selected patients received a radiation boost to the involved neck. With a median follow-up of 3.9 years, three patients have progressed (one local, two distant) and seven patients have died. Deaths were due to disease progression (three) or unrelated causes (four). No metachronous primaries developed. The 5-year progression-free and overall survival was 87% and 75%, respectively. Conclusion: Combined-modality treatment with intensive chemoradiotherapy results in excellent disease control and long-term survival for patients with N2 N3 HNCOP and compares favorably with traditional therapy. Key words: chemotherapy, head and neck cancer, occult primary, radiotherapy, squamous cell carcinoma Introduction Squamous cell carcinomas of the head and neck from an occult primary (HNCOPs) are uncommon, accounting for 1 2% of all head and neck cancers [1]. The majority of HNCOPs involve level II and III lymph nodes. The presence of an occult mucosal primary is assumed based on the well-studied patterns of spread for mucosal squamous cell malignancies of the upper aerodigestive track [2]. Therefore, patients with HNCOP are treated with locoregional therapies and are potentially curable [1, 3]. Traditional management of stage IV (N2 or N3) HNCOP involves neck lymph node dissection followed by postoperative radiation therapy [1, 3]. The radiation field often encompasses the neck lymph nodes as well as the putative mucosal primary sites, such as the nasopharynx, oropharynx and hypopharynx. However, the doses and field of radiation therapy remain controversial. Many series have shown that the addition of radiation therapy to surgery decreases the incidence of emergence of an overt mucosal primary and *Correspondence to: Dr A. Argiris, 676 N. St Clair, Suite 850, Division of Hematology-Oncology, Northwestern University, The Feinberg School of Medicine, Chicago, IL 60611, USA. Tel: ; Fax: ; a-argiris@northwestern.edu improves local control for advanced neck disease [3]. Although chemotherapy given concurrently with radiotherapy has now become an integral part of the standard non-surgical treatment for locally advanced head and neck cancer [4], the role of chemotherapy for HNCOP tumors has not yet been established. There is a dearth of literature in terms of the management of patients with HNCOP with combined-modality therapy including chemoradiotherapy, mainly because of the rarity of this presentation. In protocols conducted at the University of Chicago and Northwestern University and affiliated institutions, we have allowed the enrollment of patients with stage IV HNCOP. We now report a subset analysis of these consecutive studies that spans a decade and, to the best of our knowledge, represents the largest series of patients with HNCOP treated with concurrent chemoradiotherapy as a part of combined modality therapy. We also compare our results with other series that included patients with HNCOP tumors treated with traditional therapy without chemotherapy. Patients and methods Patients with HNCOP were identified through a search of a database that includes all patients with locoregionally advanced head and neck cancer treated on phase II chemoradiotherapy protocols at the University of Chicago and 2003 European Society for Medical Oncology

2 1307 Table 1. Multimodal treatment (n = 25) Regimen (protocol number). No. of patients (%) 1. PFL-INF induction then FHX (6126) 4 (16) 2. C-FHX (6950) 2 (8) 3. T(120-h)-FHX (7929) 7 (28) 4. T(1-h)-FHX (8626) a 3 (12) 5. CP induction then T-FHX (9502) b 9 (36) Neck dissection. No. of patients (%) Before chemotherapy/radiotherapy 14 (56) After induction chemotherapy 3 (12) After chemoradiotherapy 5 (20) None 3 (12) Radiation dose (Gy), median (range) 60 ( ) Fractionation. No. of patients (%) Once-daily (protocol 1) 4 (16) Twice-daily (protocols 2 5) 21 (84) PFL-IFN: Cisplatin 100 mg/m 2, leucovorin 300 mg/m 2 /day 5.5 days (132 h), interferon-α2b s.c IU/m 2 every 24 h for a total of six doses, 5-FU 640 mg/m 2 /day as a continuous infusion (CI) 5 days (120 h); cycle length 21 days. FHX: 5-FU 800 mg/m 2 as a CI 5 days (120 h), hydroxyurea 1000 mg orally b.i.d. 11 doses, radiation once-daily 200 cgy per fraction 5 days, 1 week on 1 week off to Gy. C-FHX: Cisplatin 100 mg/m 2 on cycles 1, 3 and 5, 5-FU 800 mg/m 2 /day 5 days, hydroxyurea 1000 mg orally b.i.d. 11 doses, radiation twice-daily 150 cgy per fraction (300 cgy daily) 1 week on 1 week off to Gy; cycles are repeated every 14 days until the completion of radiotherapy. T(120-h)-FHX: Paclitaxel 100 mg/m 2 as a continuous infusion for 5 days (120 h), 5-FU 600 mg/m 2 /day 5 days (120 h), hydroxyurea 500 mg orally b.i.d. 11 doses, radiation twice-daily 150 cgy per fraction (or 300 cgy daily) 1 week on 1 week off to Gy. T(1-h)-FHX: Paclitaxel 100 mg/m 2 in 1 h on day 1, 5-FU 600 mg/m 2 /day 5 days, (120 h), hydroxyurea 500 mg orally b.i.d. 11 doses, radiation twice-daily 150 cgy per fraction (300 cgy daily) 1 week on 1 week off to Gy. CP: Carboplatin area under the curve of 2 and paclitaxel 135 mg/m 2, over 3 h, weekly 6 followed by 2 weeks of rest 1 cycle. a Patients were randomized to receive or not erythropoietin IU s.c. once-weekly for 14 weeks starting with the first treatment week of cycle 1. b A subsequent cohort of patients on protocol 9502 was treated with a reduced total radiation dose. Three of the patients reported belong to this cohort. Northwestern University and affiliated institutions since Patients with stage IV (T0N2M0 or T0N3M0) HNCOP were eligible for participation. Results from these studies have been published previously [5 9]. All regimens included concomitant administration of 5-fluorouracil (5-FU) and hydroxyurea with either once-daily or twice-daily radiation therapy (FHX regimen) (Table 1). In two protocols (6126 and 9502) induction chemotherapy was given as well. The last three protocols (7929, 8626 and 9502) incorporated twice-daily radiation therapy with the addition of paclitaxel to the FHX regimen (T-FHX regimen), whereas the 6950 protocol incorporated cisplatin with FHX (C-FHX regimen). Informed consent has been obtained from all patients. The data collected included demographics, clinical presentation, performance status, smoking and alcohol habits, response to therapy, toxicities, disease-free survival and overall survival. All patients were staged according to the American Joint Committee on Cancer 1997 staging classification system. Actuarial curves for overall survival and progression-free survival were calculated using the Kaplan Meier method. Fisher s exact tests were used to analyze relapse rates. Statistical analysis was performed using a commercially available software package (SPSS for Windows; SPSS, Inc., Chicago, IL, USA). From August 1991 to August 2000, 25 patients with T0N2M0 or T0N3M0 squamous cell carcinomas were treated on the five consecutive chemoradiotherapy study protocols described above. We included in the analysis one patient who withdrew from protocol therapy after 1 week of the planned 4 weeks of concomitant chemoradiotherapy and was subsequently treated with radiation therapy alone, as well as another patient who had a history of tongue cancer that was treated surgically 13 years earlier. All patients underwent evaluation with panendoscopy with directed biopsies of potential mucosal primary sites, computed tomography (CT) scan of the chest, CT or magnetic resonance imaging of the neck, and bone scan, as previously reported [5 9], whereas two patients underwent tonsillectomy as a part of the evaluation for an occult primary. Results Patient characteristics are shown in Table 2. The median time from the onset of symptoms to diagnosis was 2 months. The most common presenting symptom was neck mass. Six patients reported neck pain and four patients had dysphagia. Eastern Oncology Cooperative Group (ECOG) performance status was 0 in 17 patients (68%), ECOG 1 in seven (28%) and ECOG 2 in one (4%). The nodal stage was N2a in five patients (20%), N2b in 13 (52%), N2c in one (4%) and N3 in six (24%). The left neck was involved in 16 patients, the right neck in seven patients and both sides in two patients. In 21 patients neck lymph node level information was available: 14 patients (67%) had level II involvement (eight with level II-only involvement), eight patients (38%) had level III involvement (two with level III-only involvement) and three patients (14%) had multiple lymph node involvement that included the supraclavicular region. Six patients were noted to have extracapsular tumor extension on neck dissection performed before initiating protocol therapy. Fifteen patients had grade 3 squamous cell carcinomas (60%), seven patients grade 2 (28%) and three patients grade 1 (12%). Twenty-two patients (88%) underwent modified radical neck lymph node dissection either before any other therapy (n = 14), after induction chemotherapy but before chemoradiotherapy (n = 3) or after chemoradiotherapy (n = 5). Three patients did not undergo neck dissection for the following reasons: one had previously undergone excisional biopsy of the involved nodes, one had severe co-morbidities and eventually died of an unrelated cause, and one had an attempted neck dissection before chemoradiation but was deemed unresectable (the latter patient was never diseasefree). Unilateral neck dissections were performed in 20 patients, and bilateral in two. Radiation doses delivered ranged from 55 to 75 Gy with a median of 60 Gy. The field of radiation included the potential mucosal primaries, including the nasopharynx, oropharynx and hypopharynx, and the neck bilaterally. Some patients received a radiation dose towards the higher end of the dose spectrum due to the delivery of a boost radiation dose to the involved

3 1308 Table 2. Patient characteristics (n = 25) Age, years (range) 58 (16 77) Gender Male 22 (88%) Female 3 (12%) Race Caucasian 23 (92%) African-American 2 (8%) Institution University of Chicago 13 (52%) Northwestern University 12 (48%) Smoking history Non-smoker 4 (16%) <20 pack years 4 (16%) pack years 8 (32%) >40 pack years 7 (28%) Pipe/cigar smoker 2 (8%) Alcohol history Non-drinker 6 (24%) Occasional use 9 (36%) Moderate 5 (20%) Heavy use 3 (12%) Non-smoker and non-drinker 2 (8%) Onset of symptoms, months median (range) 2 (0.3 24) Nodal stage N2a 5 (20%) N2b 13 (52%) N2c 1 (4%) N3 6 (24%) Tumor differentiation Well 3 (12%) Moderate 7 (28%) Poor 15 (60%) lymph nodes. Toxicities were expected (Table 3) and related to the protocol therapy prescribed. One patient died free of disease 6 months post chemoradiotherapy initiation and 1.5 months post neck dissection possibly from mucous plugging. This death was considered as possibly related to treatment. In no patient did the occult primary emerge or a second primary develop. Response and survival outcomes Pathological complete response was noted in two of three patients who underwent neck dissection after induction chemotherapy and in four of five patients who had neck dissection after chemoradiotherapy. The median follow-up of surviving patients was 3.9 years (range ). Of a total of 25 patients, three have progressed or relapsed; two patients relapsed at distant sites (in both cases the lung) and one patient never became disease-free and died of local Table 3. Worst grade toxicities of non-surgical therapy Toxicity No. of patients (%) Acute mucositis, grade 3 or 4 17 (68) Acute dermatitis, grade 3 or 4 10 (40) Diarrhea, grade 3 2 (8) Neutropenia, grade 3 or 4 7 (28) Infection 2 (8) Neuropathy, grade 3 1 (4) Gastrostomy tube placement 14 (56) Gastrostomy tube >1 year a 4 (16) Chronic xerostomia 11 (44) Death b 1 (4) a In one of the three patients a gastrostomy tube was placed before chemoradiotherapy due to dysphagia secondary to cranial nerve involvement by tumor. b Death occurred 1.5 months after neck dissection performed after chemoradiotherapy due to respiratory failure from possible mucous plugging. progression. The three patients who relapsed or progressed were treated on protocols 7929, 8626 and 9502, respectively (see Table 1 for protocol treatment details). The 5-year actuarial progression-free survival rate was 87% (Figure 1). Of the three patients who progressed or relapsed, two had N3 disease and one N2a disease; two of these three patients had supraclavicular lymph node involvement. Therefore, one of 19 patients with N2 and two of six patients with N3 disease relapsed/progressed (P = 0.13). The 5-year progression-free survival and overall survival for patients with N2 disease (n = 19) were 95% and 89%, respectively, whereas the 3-year progressionfree and overall survival for the six patients with N3 disease were 50% and 33%, respectively. Moreover, the presence of supraclavicular involvement was associated with poor outcome; one of 22 patients without supraclavicular lymphadenopathy, and two of three patients with supraclavicular lymphadenopathy relapsed/ progressed (P = 0.03). Finally, one of 15 patients with poorly differentiated squamous cell carcinomas and two of 10 patients with well or moderately differentiated squamous cell carcinomas relapsed (P = 0.54). Due to the small number of patients with disease progression, further prognostic factor analysis was not performed. The 5-year actuarial overall survival rate was 75% (Figure 2). Seven patients died; three due to disease progression and four due to other causes; one patient possibly due to treatment-related toxicity (see above), one patient due to pneumonia 9 years later, one patient by committing suicide within 2 months of treatment completion but after re-evaluation, and one patient of ruptured intraabdominal aneurysm more than 2 years after diagnosis. Discussion Due to the rarity of HNCOP, the role of chemotherapy has not been adequately investigated and no prospective studies have

4 1309 Figure 1. Kaplan Meier survival curve for progression-free survival. Figure 2. Kaplan Meier survival curve for overall survival. been conducted. We report an excellent progression-free survival and overall survival with the use of intensive chemoradiotherapy in five sequentially developed and similar therapeutic protocols. Our results indicate a potential survival benefit from comprehensive radiotherapy and concurrent chemotherapy in the management of HNCOP. However, this treatment is associated with considerable acute toxicity, and this needs to be considered in the therapeutic decision. Patients eligible for protocol therapy had good performance status and normal organ function, and therefore could be considered selected. On the other hand, the majority of our patients had multiple or >3 cm lymph nodes involved (52% had stage N2b) and may have been selected for this reason for protocol participation by their treating physicians. Moreover, a significant percentage of the patients reported had poorly differentiated tumors that may have contributed to increased sensitivity to chemoradiotherapy. In comparison with other patients with detectable primaries treated on the same protocols, the reported patients with HNCOP tumors appear to have performed better. This may be explained by the fact that T stage is a major prognostic factor. Accordingly, a retrospective analysis of our studies showed that T stage (T4 versus others) is an independent prognostic factor [10]. Nevertheless, differences in biological behavior between HNCOP and head and neck cancers with overt primaries may exist. To the best of our knowledge only one series of HNCOP patients treated with chemoradiotherapy has been reported in the English language literature [11]. De Braud et al. reported results on 16 patients who received various chemotherapy regimens before, during, or after radiation therapy [11]. Chemotherapy consisted of cisplatin used concurrently with radiation or cisplatin and 5-FU used as adjuvant therapy before or after radiation therapy with or without neck dissection. In this small series, patients treated with combined chemotherapy and radiation therapy had improved response rates (81% complete responses versus 60%) and median survival rates (37 versus 24 months) than a historical control of patients treated without chemotherapy despite the presence of a higher percentage of patients with N3 disease in the former group. Several retrospective studies have reported results on the management of HNCOP with surgery and/or radiotherapy (Table 4) [12 21]. We do not list reports that included non-squamous histologies or patients treated with a palliative intent. However, the nodal stage distribution as well as the treatment modalities employed in these series are heterogeneous and the reported survival rates vary considerably, especially for N3 disease. Colletier et al. [15], who reviewed 136 patients with HNCOP from M. D. Anderson Cancer Center, have reported the higher survival outcomes with surgery followed by postoperative radiation therapy to date. Unlike our series, the authors excluded patients who had macroscopic residual disease after surgery. The majority of patients had stage N2a disease (n = 49), whereas in our series most patients had more advanced, stage N2b, disease. They reported a disease-specific survival of 69% and 71% for patients with N2 and N3 disease, respectively [15]. Patients with multiple lymph node involvement (i.e. stage >N2a) performed worse in their series. The predominant pattern of failure in HNCOP is neck relapse or distant metastases [1], whereas mucosal primaries above the clavicle emerge in 10% of patients [3]. Some reports suggested an advantage for bilateral or comprehensive radiotherapy versus ipsilateral neck radiotherapy [14, 19]; however, there is no conclusive evidence to support the notion that the use of comprehensive radiotherapy is better than radiation to the ipsilateral neck [1, 3]. When surgery is the only treatment the rate of metachronous mucosal malignancies is high, up to 54% [14]. The most common metachronous primary sites are the oropharynx (base of tongue or tonsil), oral cavity and hypopharynx (mostly piriform sinus). A metachronous mucosal malignancy may either represent the occult primary that emerged or a second primary. In our series no patient developed a metachronous primary. This observation is intriguing and may suggest that systemic therapy can contribute to the eradication of occult primaries. In recent randomized studies, single-agent and multi-agent concurrent chemoradiotherapy have been shown to be superior to

5 1310 Table 4. Survival of patients with squamous cell carcinoma of the head and neck from an occult primary treated with surgery and/or radiotherapy (selected series); 3- to 5-year survival rates are reported Author/year/reference No. of patients (stage IV) Treatment modalities, no. of patients (%) a Older staging system used. b Disease-free or progression-free survival. c Disease-specific survival. CRT, chemoradiotherapy; NR, not reported; RT, radiotherapy; S, surgery. Survival (%) All patients N2 N3 Jesse et al [12] a 184 (121) S 104 (57%) 53 b RT 52 (28%) S, RT 28 (15%) Marcial-Vega et al [16] a 72 (60) RT 41 (57%) 45 b S, RT 31 (43%) Harper et al [20] 69 (51) RT 39 (57%) 66 c NR NR 12 S, RT 30 (43%) Maulard et al [13] 113 (83) S, RT 113 (100%) Reddy and Marks 1997 [19] 52 (43) RT 13 (25%) 51 b NR NR 19 S, RT 39 (75%) Medini et al [21] 24 (20) S, RT 24 (100%) 54 b Strojan and Anicin 1998 [17] 56 (50) S, RT 56 (100%) 66 c Colletier et al [15] 136 (105) S, RT 136 (100%) 74 c Grau et al [14] 277 (229) S 23 (8%) RT 224 (81%) S, RT 26 (9%) Friesland et al [25] 51 (48) RT 23 (45%) S, RT 28 (55%) Present series 25 (25) S, CRT 22 (88%) 87 b CRT 3 (12%) Metachronous mucosal primaries (%) radiation therapy alone for the non-surgical therapy of locoregionally advanced head and neck cancer [4]. The use of postoperative chemoradiotherapy appears promising but has not yet been established [22]. Given the rarity of HNCOP, prospective studies are difficult to conduct. Our data, which form the largest series reported, suggest that HNCOP should be treated in a similar manner to other head and neck squamous cell carcinomas with detectable primaries. Since all but three of our patients underwent neck dissection, we cannot discern the contribution of neck dissection in the outcome of HNCOP patients. Some reports have suggested that tumor control is superior with surgery and radiation therapy compared with radiation alone [19]. The role of neck dissection in the management of locoregionally advanced head and neck cancer is controversial [23, 24]. We recently reported that elective neck dissection may be safely omitted in N2 patients with head and neck cancer with a detectable primary who achieve a clinical complete response after induction chemotherapy or chemoradiotherapy [23]. This may apply to HNCOP tumors as well. Finally, induction chemotherapy may decrease the rate of distant metastasis, which is a common site of failure in patients with advanced N stage disease, and its potential use in advanced HNCOP warrants investigation. Patients with HNCOP with supraclavicular lymphadenopathy represent a different clinical entity possibly due to the association of supraclavicular lymphadenopathy with infraclavicular neoplasms, such as lung cancer. Their long-term survival is <15% at 5 years [12, 16]. In our series three patients had lymphadenopathy extending to the supraclavicular region. Their outcome was poor; one had distant failure and one local. However, none of these patients developed a lung primary. We conclude that the addition of systemic therapy to locoregional therapies results in high rates of locoregional and distant control and long-term survival for good performance status patients with stage IV (N2 or N3) HNCOP. Our results compare favorably with previous reports in which patients with HNCOP were treated with surgery and/or radiation therapy. Prospective multicenter studies will be required to establish further the role of concurrent chemoradiotherapy for HNCOP.

6 1311 Acknowledgements This study was supported in part by the University of Chicago/ Northwestern University Oral Cancer Center (P50 DE ), University of Chicago Cancer Research Center (P30 CA14599), the Francis Lederer Foundation, the Geraldi Norton Memorial Corporation, the Robert and Valda Svendsen Memorial, and Bristol- Myers Squibb, Princeton, NJ. References 1. Nieder C, Gregoire V, Ang KK. Cervical lymph node metastases from occult squamous cell carcinoma: cut down a tree to get an apple? Int J Radiat Oncol Biol Phys 2001; 50: Lindberg R. Distribution of cervical lymph node metastases from squamous cell carcinoma of the upper respiratory and digestive tracts. Cancer 1972; 29: Fu KK. Neck node metastases from unknown primary. Controversies in management. Front Radiat Ther Oncol 1994; 28: Argiris A. Update on chemoradiotherapy for head and neck cancer. Curr Opin Oncol 2002; 14: Kies MS, Haraf DJ, Athanasiadis I et al. Induction chemotherapy followed by concurrent chemoradiation for advanced head and neck cancer: improved disease control and survival. J Clin Oncol 1998; 16: Vokes EE, Kies MS, Haraf DJ et al. Concomitant chemoradiotherapy as primary therapy for locoregionally advanced head and neck cancer. J Clin Oncol 2000; 18: Kies MS, Haraf DJ, Rosen F et al. Concomitant infusional paclitaxel and fluorouracil, oral hydroxyurea, and hyperfractionated radiation for locally advanced squamous head and neck cancer. J Clin Oncol 2001; 19: Vokes EE, Stenson K, Rosen FR et al. Weekly carboplatin and paclitaxel followed by concomitant paclitaxel, fluorouracil, and hydroxyurea chemoradiotherapy: curative and organ-preserving therapy for advanced head and neck cancer. J Clin Oncol 2003; 21: Rosen FR, Haraf D, Brockstein B et al. Multicenter randomized phase II study of 1 h infusion paclitaxel, fluorouracil and hydroxyurea with concomitant hyperfractionated radiotherapy with or without erythropoietin for advanced head and neck cancer. Proc Am Soc Clin Oncol 2001; 20: A Brockstein BE, Haraf DJ, Kies MS et al. Survival, patterns of failure, and risk factors for recurrence after intensive concomitant chemoradiation (CRT) for squamous cell head and neck cancer (SCHNC). Proc Am Soc Clin Oncol 2002; 21: A de Braud F, Heilbrun LK, Ahmed K et al. Metastatic squamous cell carcinoma of an unknown primary localized to the neck. Advantages of an aggressive treatment. Cancer 1989; 64: Jesse RH, Perez CA, Fletcher GH. Cervical lymph node metastasis: unknown primary cancer. Cancer 1973; 31: Maulard C, Housset M, Brunel P et al. Postoperative radiation therapy for cervical lymph node metastases from an occult squamous cell carcinoma. Laryngoscope 1992; 102: Grau C, Johansen LV, Jakobsen J et al. Cervical lymph node metastases from unknown primary tumours. Results from a national survey by the Danish Society for Head and Neck Oncology. Radiother Oncol 2000; 55: Colletier PJ, Garden AS, Morrison WH et al. Postoperative radiation for squamous cell carcinoma metastatic to cervical lymph nodes from an unknown primary site: outcomes and patterns of failure. Head Neck 1998; 20: Marcial-Vega VA, Cardenes H, Perez CA et al. Cervical metastases from unknown primaries: radiotherapeutic management and appearance of subsequent primaries. Int J Radiat Oncol Biol Phys 1990; 19: Strojan P, Anicin A. Combined surgery and postoperative radiotherapy for cervical lymph node metastases from an unknown primary tumour. Radiother Oncol 1998; 49: McMahon J, Hruby G, O Brien CJ et al. Neck dissection and ipsilateral radiotherapy in the management of cervical metastatic carcinoma from an unknown primary. Aust N Z J Surg 2000; 70: Reddy SP, Marks JE. Metastatic carcinoma in the cervical lymph nodes from an unknown primary site: results of bilateral neck plus mucosal irradiation vs. ipsilateral neck irradiation. Int J Radiat Oncol Biol Phys 1997; 37: Harper CS, Mendenhall WM, Parsons JT et al. Cancer in neck nodes with unknown primary site: role of mucosal radiotherapy. Head Neck 1990; 12: Medini E, Medini AM, Lee CK et al. The management of metastatic squamous cell carcinoma in cervical lymph nodes from an unknown primary. Am J Clin Oncol 1998; 21: Cooper JS, Pajak TF, Forastiere AA et al. Patterns of failure for resected advanced head and neck cancer treated by concurrent chemotherapy and radiation therapy: an analysis of RTOG 9501/Intergroup phase III trial. Int J Radiat Oncol Biol Phys 2002; 54 (Suppl. 1): Argiris A, Stenson KM, Brockstein BE et al. Node dissection (ND) in the combined modality therapy of patients with locoregionally advanced head and neck cancer (HNC). Proc Am Soc Clin Oncol 2002; 21: A McHam SA, Adelstein DJ, Rybicki LA et al. Who merits a neck dissection (ND) after definitive chemoradiotherapy for N2-N3 squamous cell head and neck cancer (SCHNC)? Proc Am Soc Clin Oncol 2002; 21: A Friesland S, Lind MG, Lundgren J et al. Outcome of ipsilateral treatment for patients with metastases to neck nodes of unknown origin. Acta Oncol 2001; 40: