Figure 1. Survival determined by Kaplan-Meier methods. A, Overall survival; B, disease-free survival; and C, cause-specific survival for all patients. Filled circles indicate censored events.
Figure 2. Overall survival grouped by number of involved lymph nodes. Filled circles indicate censored events.
Figure 3. Overall survival grouped by p16 positivity determined by immunohistochemical analysis. Filled circles indicate censored events.
Figure 4. Overall survival grouped by comprehensive radiation treatment (COMP) vs ipsilateral (IPSI). Filled circles indicate censored events.
Perkins SM, Spencer CR, Chernock RD, Haughey BH, Nussenbaum B, Adkins DR, Kuperman DI, Thorstad WL. Radiotherapeutic Management of Cervical Lymph Node Metastases From an Unknown Primary Site. Arch Otolaryngol Head Neck Surg. 2012;138(7):656–661. doi:10.1001/archoto.2012.1110
Author Affiliations: Departments of Radiation Oncology and Mallinkrodt Institute of Radiology (Drs Perkins, Spencer, and Thorstad), Pathology and Immunology (Dr Chernock), and Otolaryngology (Drs Haughey and Nussenbaum), Division of Medical Oncology (Drs Adkins and Kuperman) and Washington University School of Medicine, St Louis, Missouri.
Objective To determine whether ipsilateral radiotherapy affects overall survival, cause-specific survival, or local control in patients with a cancer from an unknown primary of the head and neck compared with comprehensive radiotherapy.
Design Retrospective medical record review.
Setting Academic tertiary care hospital.
Patients The study population comprised 46 patients with cervical metastases from an unknown primary cancer treated with radiotherapy from 1989 through 2008. Median follow-up was 4.6 years.
Interventions All patients were treated with radiation therapy. Radiotherapy target volumes were categorized as either ipsilateral neck only (IPSI) or comprehensive (COMP), including both the potential mucosal surfaces and ipsilateral or bilateral neck. Human papillomavirus (HPV) status, as determined by p16 immunohistochemical analysis, was evaluated for 36 patients (74%).
Main Outcome Measures Overall survival, cause-specific survival, locoregional control, and rate of distant metastases were analyzed.
Results Overall survival at 2 years and 5 years was 87% and 77%, respectively. Cause-specific survival at 2 years and 5 years was 89% and 81%, respectively. There were no ipsilateral neck failures. There was no difference in overall survival between patients treated with IPSI or COMP radiation therapy. The contralateral neck was controlled in all patients receiving bilateral neck irradiation and in 95% receiving ipsilateral neck irradiation. Of the 34 patients evaluated with p16 immunohistochemical analysis, results for 16 (47%) were positive. There was a nonsignificant trend toward improved overall survival in p16-positive patients (P = .06).
Conclusion IPSI radiation therapy demonstrated excellent locoregional control with no adverse effect on disease-free survival or overall survival.
Cervical metastasis from an unknown primary tumor is an uncommon diagnosis, comprising 2% to 4% of head and neck malignant diseases.1 If, after thorough evaluation, no primary tumor is identified, deciding on the best treatment approach can be difficult. Although definitive radiotherapy or chemoradiotherapy is an option, these tumors are most often treated with lymph node dissection followed by radiation therapy. The optimal radiotherapeutic treatment remains controversial.
At issue is the morbidity of comprehensive irradiation balanced with the risk of an emerging primary tumor or failure in the contralateral neck. Many institutions irradiate the bilateral necks along with potential mucosal surfaces (ie, the oropharynx, nasopharynx, or oral cavity).1,2 However, previous data from our institution demonstrated that for patients with disease on 1 side of the neck, there was no difference in the appearance of a primary tumor in patients treated with or without prophylactic mucosal irradiation. In addition, there was a low risk of failure in the untreated contralateral neck.3 This led to a shift in treatment approach at our institution with selected patients receiving ipsilateral neck radiation with no prophylactic mucosal irradiation.
To date, there are no randomized data to guide treatment decisions. A study proposed by the European Organization for Research on Treatment of Cancer and the Radiation Therapy Oncology Group to randomize patients to ipsilateral neck irradiation vs mucosal and bilateral neck irradiation closed early owing to poor accrual. Herein, we retrospectively reviewed disease control and clinical outcome in patients treated with comprehensive (COMP) or ipsilateral neck only (IPSI) irradiation.
This retrospective study was undertaken after approval from the Human Research Protection Office. From 1989 through 2008, 46 patients with level I to V cervical metastases from an unknown primary site were treated with radiation therapy at Washington University in St Louis, Missouri. Pathologic diagnosis was obtained for all patients either by fine-needle aspiration, excisional biopsy, or neck dissection.
Patients were determined to have a cancer of unknown primary site after thorough evaluation failed to detect a primary tumor (Table 1). Evaluation at the time of diagnosis included medical history, physical examination, nasopharyngoscopy, computed tomographic (CT) scan or magnetic resonance imaging (MRI), and directed or random biopsies for all patients. Twenty-three patients (50%) were evaluated with a fluorodeoxyglucose–positron emission tomography (FDG-PET) scan. Four patients (9%) had transoral laser microsurgery (TLM) directed examination under anesthesia (EUA), a new approach to detect unknown primary cancer recently implemented at our institution.4
Immunoperoxidase staining was done on formalin-fixed paraffin-embedded, 4-μm tissue sections using an antibody to p16 (MTM Laboratories Inc; mouse monoclonal; 1:1 dilution). Immunostaining was performed on a Ventana Benchmark automated immunostainer (Ventana Medical Systems Inc) according to standard protocols with appropriate positive controls. Antigen retrieval, standard on the machine, used the Ventana CC1, EDTA-Tris, pH 8.0 solution. For p16, cases were classified in a binary manner as positive when 50% or more of cells showed nuclear and cytoplasmic staining or negative when less than 50% of the cells stained for statistical analysis.
All patients were treated with curative intent. Forty (87%) were treated with surgery followed by postoperative radiation therapy. Six patients (13%) could not undergo surgery owing to either comorbidities or extent of disease, and they were treated with radiation therapy alone or in combination with chemotherapy. For the surgical patients, the extent of surgery ranged from local excision in 3 patients (7.5%) to bilateral neck dissection for 3 (7.5%). Ipsilateral radical or modified radical neck dissections were performed in 32 patients (80%). The remaining 2 patients underwent an ipsilateral selective neck dissection. Only 1 patient with disease confined to 1 side of the neck underwent a bilateral neck dissection. The other 2 bilateral neck dissections were performed in patients with N2C disease.
Fourteen patients (30%) received chemotherapy or a targeted agent. Patients receiving chemotherapy were equally balanced between IPSI and COMP irradiation (Table 2). Chemotherapy was administered neoadjuvantly in 1 patient (2%), concurrently with radiation in 10 patients (22%), and neoadjuvantly and concurrently in 3 patients (7%). Neoadjuvant chemotherapy consisted of docetaxel, cisplatin (Platinol), and fluorouracil. All concurrent chemotherapy consisted of cisplatin. One patient received concurrent cetuximab (Erbitux). Regarding the selection of patients to receive chemotherapy, most patients (64%) received chemotherapy in or after 2004, the year data were published on the benefits of postoperative chemoradiotherapy.5,6 Of the 5 patients receiving chemotherapy prior to 2004, 1 was treated with definitive chemoradiotherapy; 4 had extracapsular extension, and 1 had bilateral disease.
All patients received conventionally fractionated radiation therapy. COMP treatment volumes consisted of radiation therapy to the potential mucosal sites and bilateral neck in 21 patients (46%) or radiation to potential mucosal sites and the ipsilateral neck in 4 patients (8%). IPSI treatment consisted of ipsilateral neck radiation only with no planned radiation therapy to the mucosa in 21 patients (46%). Two-dimensional (2D) or 3-dimensional (3D) conformal techniques were used for 18 patients (39%). Intensity-modulated radiation therapy (IMRT) was used to treat the remaining 28 patients (61%). A detailed multimodal treatment approach for all patients is summarized in Table 2.
Twenty-five patients (54%) were treated with COMP radiation. Fourteen of these patients (56%) were treated with IMRT, and the remaining 11 (44%) were treated with 2D- or 3D-conformal techniques. Median treatment dose was 66 Gy (range, 56-71.8 Gy). Twenty-one patients (46%) were treated with IPSI radiation. Fourteen (67%) were treated with IMRT. The median treatment dose was 66 Gy (range, 54-70.2 Gy). For the IPSI patients treated with IMRT, the dose was prescribed to cover the planning target volume, which included the involved lymph node levels as well as contiguous lymph node levels above and below the involved levels. Median time to complete the course of radiation therapy was 49 days (range, 39-77 days). Analysis of the dose to the nasopharynx, base of tongue, and tonsils was performed in 14 patients treated with IMRT in whom detailed dosimetric data could be recovered. This included 9 patients treated with ipsilateral neck radiation only and 5 patients treated comprehensively with bilateral neck and mucosal irradiation. The mean doses to the ipsilateral/contralateral nasopharynx, ipsilateral/contralateral base of tongue, and ipsilateral/contralateral tonsils were evaluated and compared between the IPSI- and COMP-treated patients. Comparison of dose received to the contralateral neck structure were performed using the Wilcoxon 2-sample test.
Overall, cause-specific, and disease-free survivals were determined using the Kaplan-Meier method. Tumor characteristics, patient characteristics, and treatment methods were evaluated for association with survival. P values were generated using the log-rank method.
Patient and disease characteristics are shown in Table 3. The median duration of follow-up for this study was 4.6 years (range, 7 months–18 years). The median follow-up for living patients was 5.1 years, with a minimum of 1.5 years (range, 1.5-16.0 years). Overall survival at 2 years and 5 years was 87% and 77%, respectively (Figure 1A). Using univariate analysis, there was no difference in overall survival based on sex (P = .16) or race/ethnicity (P = .98). Patients undergoing definitive vs postoperative radiation therapy had similar overall survival (P = .44), as did patients treated with or without chemotherapy (P = .52). Disease-free survival at 2 years and 5 years was 78% and 69%, respectively (Figure 1B). Cause-specific survival at 2 years and 5 years was 89% and 81%, respectively (Figure 1C).
The mean radiation doses to the tonsils, base of tongue, and nasopharynx were evaluated for 14 patients treated with IMRT (9 IPSI patients and 5 COMP patients). The mean dose to the ipsilateral nasopharynx, ipsilateral tonsil, and ipsilateral base of tongue did not significantly differ between the 2 treatment groups owing to the proximity of these structures to the treated lymphatic regions. However, the mean dose to the contralateral nasopharynx, contralateral tonsil, and contralateral base of tongue was significantly less for patients treated to the ipsilateral neck only (Table 4).
Forty patients (87%) underwent surgical resection of their neck disease. In these patients, overall survival was significantly worse in patients with more than 3 positive nodes (P = .03) (Figure 2). Survival at 2 years was 91% in patients with 3 or fewer nodes involved compared with 71% for those with more than 3 nodes involved. Similarly, cause-specific survival was improved in patients with 3 or fewer involved nodes (P = .02). There was a nonsignificant trend for worse overall survival for patients with tumors exhibiting extracapsular extension (P = .06). Material was available for p16 immunohistochemical analysis in 34 patients (74%). Of these patients, 16 (47%) were p16 positive. There were no differences in age, race/ethnicity, or sex between patients who were p16 positive and those who were p16 negative. There was no difference in p16 positivity between the IPSI and COMP groups because 47% of patients were p16 positive in each group. There was a nonsignificant trend toward improved overall survival in p16-positive patients (P = .06) (Figure 3).
There was no difference in overall survival (Figure 4), cause-specific survival, or disease-free survival between the IPSI and COMP radiation groups. There were no ipsilateral neck failures in either group. In the IPSI treatment group there was 1 contralateral neck failure. After undergoing a neck dissection, this patient has remained disease-free to date (4 years after surgery).
Three patients (6%) developed recurrent disease in a mucosal site. The failures occurred in 2 patients from the IPSI group and 1 patient from the COMP group. The COMP-treated patient failed in-field at the base of tongue 22 months after treatment. This patient underwent surgical resection and remains disease-free to date (3.5 years later). In the IPSI group, 2 mucosal primary lesions emerged. One patient failed both in the larynx and with widespread distant metastases 1 year after completion of therapy. The other patient developed an epiglottis primary tumor 59 months after initial diagnosis. This patient underwent surgical resection and remained disease-free until developing distant metastases 4.5 years later. Ten patients (22%), equally distributed between the IPSI and COMP groups, developed distant metastases. Appearance of distant metastases occurred within 3 years of initial diagnosis in all but 2 patients.
With no available randomized data, the best radiation technique for head and neck unknown primary lesions remains controversial. Our radiation approach has shifted over time to a more limited treatment volume based on our previous institutional data and other studies suggesting that ipsilateral neck radiation can lead to acceptable control of disease.3,7,8 Four of 16 patients (25%) treated prior to 1998 received IPSI radiation, whereas 17 of 30 (57%) received IPSI radiation from 1999 through 2008. After 1998, the decision regarding IPSI vs COMP treatment was based on the extent of neck disease and treatment modality, although physician preference cannot be eliminated as a factor in this retrospective analysis. Comprehensive radiation was delivered to patients with bilateral neck disease, those treated with definitive nonsurgical therapy, and in 7 of 9 patients with N3 disease. Therefore, the data presented herein do not apply to all patients with occult head and neck primary tumor. Limitations of the study include the retrospective format and small patient numbers, in common with other reports of unknown primary head and neck cancer.
Workup was thorough in this group of patients in that 100% had directed biopsy aimed at identifying the primary tumor, and 50% had FDG-PET scans at the time of diagnosis. This is similar to the data reported by Madani et al,9 in which all patients underwent endoscopy, biopsy of suspicious lesions, CT, and/or MRI, and FDG-PET scan was performed on 13 of 23 patients. The TLM-directed EUA approach has become a standard at our institution for workup of patients with unknown primary cancer of the head and neck, and this approach was used in 9% of the patients in this study. In an institutional pilot study, the primary site detection rate with traditional EUA was 25% (3 of 12) and with TLM-directed EUA was 94% (17 of 18).4 For patients with human papillomavirus (HPV) related tumors, the TLM-directed EUA approach may be extended to include a palatine and lingual tonsillectomy with careful sectioning to search for the primary tumor.
The 5-year overall survival in our previous experience published in 1990 was 28%.3 Overall survival in the present study at 2 and 5 years was 87% and 77%, respectively. Similar data from a contemporary series of 21 patients treated at Memorial Sloan-Kettering reported a 2-year overall survival of 85%.10 The rates in these studies compare favorably with other reported survival rates. A recent publication from Beldì et al11 and a large study from the Netherlands1 both reported a 5-year overall survival of approximately 50%; however, they also reported fewer patients undergoing directed biopsy or FDG-PET imaging. The improvement in outcome compared with historical patients is likely reflective of several factors, including emergence of HPV-related tumors as well as modern imaging techniques including FDG-PET, CT, and MRI.
We performed p16 immunohistochemical analysis, a surrogate marker for HPV status, in 74% of patients in this study.12 Human papillomavirus has been identified as an independent cause of oropharyngeal cancers and has been reported in up to 70% of oropharyngeal primaries.13 The rate of p16 positivity in unknown primary tumors of the head and neck has not been extensively studied but has been previously reported as positive in 11 of 25 (44%) by Compton et al.14 Similarly, we found that 47% of the patients (16) tested had positive results for p16, and there was a trend toward improved overall survival for these patients.
In patients who completed the prescribed course of treatment, ipsilateral neck control was 100%. One patient, treated definitively for clinical N3 disease, completed 59 Gy of a planned course of 70 Gy. The tumor progressed shortly after therapy, and the patient died 6 months later. Two other patients with N3 disease were treated definitively and were disease free with 4 and 6 years of follow-up. Contralateral neck control rates were also very high, with 100% control in the COMP group. Previous data for patients treated at our institution from 1964 to 1986 demonstrated no contralateral neck failures in 19 patients treated to the ipsilateral neck only.3 Therefore, in 40 patients treated with ipsilateral neck radiation from 1964 to 2008 there was a contralateral failure rate of 2.5%. The median neck relapse rates from the literature were notably higher than this at 19% and 51.5% for patients treated comprehensively and unilaterally, respectively.15 A closer look at these data shows that many of these failures were ipsilateral in-field neck failures in patients treated with radiation alone. Grau et al1 reported a 5-year neck control rate of only 51% in a study in which 91% of patients were treated with radiation alone. However, the contralateral neck failure rates in patients treated with bilateral or ipsilateral radiation were low at 2% and 4%, respectively. Weir et al8 evaluated 144 patients treated with radiotherapy alone. Again, local control of nodal disease was 51% with 94%, of these failures occurring in the field of radiation. In their series of 52 patients, investigators at Loyola University reported a contralateral failure rate of 23%.16 Patients treated only to the ipsilateral neck were more likely to fail in the opposite side of the neck. However, 44% of the patients treated to the ipsilateral neck failed in a primary mucosal site and contralateral failure occurred simultaneously with mucosal site failure in 77%.
The issue of primary mucosal site failure is of great importance when deciding to irradiate the ipsilateral neck alone. However, it is important to realize that patients with head and neck cancer are at risk to develop a second primary malignant tumor in the upper aerodigestive tract. A risk of 3% per year has consistently been reported.17,18 In this series, we observed 3 failures (6% of patients) at a mucosal site. One occurred 4.9 years after the initial diagnosis, and a difference in pathologic grade from the original neck node indicates this is likely a new primary tumor. Our mucosal failure sites included the base of tongue, epiglottis, and larynx. As reported in the “Results” section, 2 of these patients underwent salvage surgery and remain disease-free. It is also important to realize the dose of radiation received to the ipsilateral sides of the nasopharynx and oropharynx. As shown in Table 4, the mean radiation doses to these ipsilateral structures were the same whether the patient was treated with COMP or IPSI radiotherapy. Even with IMRT, the proximity of these structures to the treated neck lymph nodes leads to a high dose of radiation therapy. It is highly likely that the incidental dose to these structures aided in our low mucosal failure rates.
Distant metastatic disease was the predominant pattern of failure in this study and occurred more than twice as frequently as all types of locoregional failure combined. Overall, 10 patients (22%) developed metastases, which is similar to the rate of 31% reported for patients treated in our department prior to 1989. The reported rate of distant disease in other studies ranges from 10% to 38%.11,19
In summary, neck metastasis from unknown primary cancer is a rare diagnosis but has the potential for favorable outcome similar to that of known primary lesions of the head and neck. Our data suggest that overall and disease-free survival is the same for selected patients treated to the ipsilateral neck alone rather than with comprehensive radiotherapy. Exceptions may include patients with bilateral neck disease, N3 disease, or those receiving definitive nonsurgical treatment. Approximately 50% of the patients tested had HPV-related tumors based on p16 positivity, and there was a nonsignificant trend for improved survival for this group. The use of FDG-PET/CT and the incorporation of TLM-directed EUA in the workup may have limited the number of patients with small primary cancers included in this cohort of patients with unknown primary tumors. The predominant pattern of failure for this group of patients was distant metastatic disease.
Correspondence: Stephanie M. Perkins, MD, Department of Radiation Oncology and Mallinkrodt Institute of Radiology, Washington University School of Medicine, 4921 Parkview Pl, PO Box 8224, St Louis, MO 63110 (Stephanie.firstname.lastname@example.org).
Submitted for Publication: January 26, 2012; final revision received April 5, 2012; accepted May 3, 2012.
Author Contributions: Drs Perkins, Spencer, Haughey, Nussenbaum, Adkins, Kuperman, and Thorstad had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Perkins and Thorstad. Acquisition of data: Perkins, Spencer, Chernock, Haughey, and Thorstad. Analysis and interpretation of data: Perkins, Spencer, Chernock, Haughey, Nussenbaum, Adkins, and Kuperman. Drafting of the manuscript: Perkins and Adkins. Critical revision of the manuscript for important intellectual content: Perkins, Spencer, Chernock, Haughey, Nussenbaum, Adkins, Kuperman, and Thorstad. Statistical analysis: Perkins and Spencer. Administrative, technical, and material support: Spencer, Haughey, Adkins, and Thorstad. Study supervision: Nussenbaum and Thorstad.
Financial Disclosure: None reported.
Additional Contributions: Jianping Li, BS, provided expert technical assistance with the immunohistochemical experiments.