In the United States, nasopharyngeal carcinoma (NPC) accounts for 2% of all HNSCCs. Its unusual epidemiologic and natural history features include a remarkable tendency toward early regional and distant dissemination. NPC also is extremely sensitive to radiotherapy and cytotoxic chemotherapy.
In the adult, the nasopharynx is a chamber that is approximately cuboidal in shape and 4 cm on an edge. It is bounded anteriorly by the choana of the nasal cavity, superiorly by the clivus, and inferiorly by the soft palate. Its posterior wall is the mucosa that overlies the superior constrictor muscles of the pharynx and the C1 and C2 vertebral bodies. The lateral walls contain the Eustachian tube orifices.
The epithelium of the superior lateral walls contains pseudostratified columnar cells and occasional goblet cells while the inferior lateral and posterior walls are stratified squamous in nature. The region is richly endowed with lymphatics that drain to the retropharyngeal and deep cervical nodes.
Malignant neoplasms of the nasopharynx are primarily epithelial, with the presence of keratin associated with a poorer prognosis. The World Health Organization (WHO) recognizes three histopathologic types of NPC: type 1, differentiated SCC (of varying degrees); type 2, nonkeratinizing carcinoma; and type 3, undifferentiated or lymphoepithelial carcinoma. Mixed patterns are common. About 75% of nasopharyngeal carcinomas are type 1 or 2 (or predominantly one or both of these types). The term “lymphoepithelial carcinoma” (type 3) is used when numerous infiltrating lymphocytes are seen.
About one-third of patients present with a neck mass without other complaints, and about 70% to 75% of patients have enlarged neck nodes at presentation. Other common complaints are epistaxis, nasal stuffiness, headache, or hearing loss (generally unilateral). The tumor can spread laterally and superiorly to cause bony destruction of the base of the skull. Frequently, there are cranial nerve findings, with the sixth nerve being most commonly involved. There are two principal cranial nerve syndromes associated with NPC: (1) the retroparotidian syndrome, involving cranial nerves IX, X, XI, and XII, and (2) the petrosphenoidal syndrome, involving cranial nerves III, IV, V, and VI (and occasionally cranial nerve II via extension through the foramen lacerum into the middle cranial fossa). Evaluation of the nasopharynx should consist of direct visualization with a fiberoptic scope. A CT and/or MRI scan is important in evaluating base-of-skull involvement and the possible presence of occult involved lymph nodes.
The most recent revision of the AJCC/Union Internationale Contre le Cancer (UICC) staging system recognizes the uniqueness of nasopharyngeal carcinoma among other head and neck tumors. Both the criteria for T and N staging have been revised, as has the stage grouping. These are summarized in
Standard treatment for NPC is radiotherapy for early and locally advanced disease. Surgical resection even for early-stage disease is technically difficult because of the anatomic location of the primary tumor and the frequent bilateral cervical and retropharyngeal node involvement. The role of the surgeon is limited to obtaining tissue for diagnosis and occasionally to resecting residual adenopathy after definitive radiotherapy. Fortunately, these tumors tend to be fairly radiosensitive, and even large lymph nodes often respond to moderate doses of radiotherapy. Prior to initiating therapy, a dental consultation is advised since it is necessary to irradiate the parotid glands bilaterally and the resulting xerostomia predisposes to serious oral problems.
The initial radiation fields encompass the adjacent base of the skull as well as the nasopharynx itself. The fields are bilaterally directed and include the retropharyngeal drainage and the anterior and posterior cervical chains. A dose of 4,500 cGy is given, using megavoltage photons, and then the fields are reduced to spare the spinal cord and an additional 500 cGy are given. Megavoltage electrons are used to bring the posterior cervical nodes to this same dose. The fields are then reduced in size and an additional 2,000 to 2,200 cGy are given to the nasopharyngeal primary. Regions of positive cervical adenopathy are also boosted with megavoltage photons and/or electrons to total doses of 6,500 to 7,500 cGy, depending on the original size of the node and its response to the first phase of therapy. In selected patients, the boost dose to the nasopharynx itself can be given with an intracavitary implant. Critical normal structures in the treatment region include the cervical cord, the brainstem, the optic nerves, and orbital contents. Proper shielding and limiting the delivered dose to these structures are necessary to avoid untoward complications. An anterior supraclavicular field is generally matched to the initial large lateral fields, and approximately 5,000 cGy are given to treat submicroscopic disease in this area.
Treatment results are related to both stage and histopathology, but many series do not adequately document outcome as a function of these variables. Huang combines the above-listed T1, T2, and T3 stages into his T1/T2 categories. For a clinically negative neck, he reports a 5-year survival of 65%. For groups corresponding to T4N0-N2 and T4N3, he found respective 5-year survivals of 41.3% and 23%. Vikram and colleagues noted a 5-year locoregional control rate for early T-stage N0 patients of 65%.358 Scanlon and colleagues found a clear worsening of prognosis with increasing cervical adenopathy with 5-year survivals of 67%, 24%, and 14% when the patient had no clinical adenopathy, unilateral adenopathy, or bilateral adenopathy. It is important to note that these series were treated prior to routine CT/MRI scanning, which would have the tendency to increase the clinical stage of the neck disease.
A clear correlation exists between the degree of cervical adenopathy and the subsequent development of distant metastases, with patients with bilateral adenopathy having a 5-year actuarial risk of approximately 80% of developing distant metastases. Common sites of distant metastases are the lung, bone, and liver. In selected cases, a failure at the primary site alone can be salvaged using a combination of external beam radiotherapy and an intracavitary implant. However, the morbidity associated with this may be substantial. Long-term study has shown that brachytherapy with permanent radioactive gold grain interstitial implantation is an effective salvage treatment in persistent and recurrent NPC patients who have nasopharynx-confined disease. Five-year local control rates were the best for patients with persistent disease (87.2%) versus those with a first recurrence (62.7%) or second recurrence (23.4%, p = .0004). Overall 5-year survival rates for these three patient groups were 79.1%, 53.6%, and 42.9%, respectively. Lesion size was not an independent prognostic factor for local control. Complications of therapy (including headache, palatal fistula, and mucosal radiation necrosis at site of gold grain implantation) of this brachytherapy approach occurred most often in patients with persistent disease (28.3% versus 18.9% and 16% in patients with first and second recurrence, respectively).
Although effective in early stages, standard radiotherapy (despite achieving high complete response rates) produces 5-year survival rates in stage III disease of only 10% to 45% and in stage IV disease of 0% to 30%. Despite major differences between NPC and other HNSCCs, many chemotherapy studies have included NPC patients, which confounds study results. Chemotherapeutic strategies for NPC now treat this disease as a distinct entity. With the exception of parts of China, the problem of small patient numbers is obviously even greater in NPC trials than in many other HNSCC trials.
Early US reports of chemotherapy for NPC were single-institution retrospective surveys of recurrent or metastatic NPC patients treated over a period of many years with a variety of agents. Active single agents include cisplatin, bleomycin, methotrexate, 5-fluorouracil, doxorubicin, and Vinca alkaloids. Interferon has very limited activity, with response rates of less than 10%. Retrospective surveys from the early 1980s reported 40% to 70% response rates (less than 20% complete responses) with a variety of cisplatin- and noncisplatin-based combination regimens in recurrent disease. More recent series with intensive cisplatin-based regimens in recurrent disease have reported higher and more durable complete response rates
The use of combined-modality treatments (sequential chemotherapy and radiotherapy) is under active study in advanced NPC. Neoadjuvant series have reported 70% to 90% response rates (20% to 40% complete response) with cisplatin- and noncisplatin-based regimens. Results of a large series from Taiwan treating 1,206 patients with a variety of chemotherapeutic agents given with split-course radiotherapy suggested that the combined-modality approach was more effective than radiotherapy alone in historic controls.
Several single-arm sequential-therapy studies have been reported, all but one with cisplatin-based regimens. Tannock and colleagues reported the largest series of 49 consecutive patients treated with methotrexate, cisplatin, and 3-day continuous-infusion bleomycin followed by radiotherapy. The overall response rate was high, but complete response was low - 22% (8 of 36) in patients with measurable nodal disease. After radiotherapy, the complete response rate jumped to 82%. This group compared its results with 140 stage-matched historic controls and reported no differences in disease-free or overall survival. Furthermore, there was no apparent reduction in distant metastases in the chemotherapy group. This study was not confirmed by three similar series. Although complete response rates were equivalent in the combined-modality and radiotherapy groups, the disease-free and median survivals of the combined-modality group were higher.
Two phase III trials using neoadjuvant chemotherapy have reported conflicting results. One international trial tested epirubicin, cisplatin, and bleomycin for three cycles in 399 patients in addition to standard radiotherapy. This showed an increase in the 3-year disease-free survival rate from 4% to 31% (p = .02) with the neoadjuvant treatment. Advanced-stage patients with N2 or N3 disease were in this trial. Another trial, from Hong Kong, tested three cycles of neoadjuvant cisplatin and 5-fluorouracil followed by radiation followed by four more cycles of cisplatin and 5-fluorouracil chemotherapy. With a median follow=up time of 29 months, the experimental arm and the radiotherapy-alone arm showed respective 2-year overall survivals of 80% and 81% and respective disease-free survivals of 68% and 72% (differences not statistically significant). More recent prospective randomized studies also have shown that neoadjuvant chemotherapy (5-fluorouracil and carboplatin, cisplatin and epirubicin, or mitomycin, epirubicin, cisplatin, 5-fluorouracil and leucovorin) does not improve overall or disease-free survival over that with radiotherapy alone.
The use of concomitant chemotherapy and radiotherapy is more promising. A phase III trial from Taiwan compared neoadjuvant versus concomitant chemotherapy along with radiotherapy. Although only 68 patients were entered, the acuarial disease-free survival at 5 years favored the concomitant arm (65% vs 41%), with the difference being primarily due to a reduced rate of distant metastases. In the United States, an intergroup cooperative study (IG0099) tested radiotherapy alone versus an experimental arm that used concomitant cisplatin given every week during radiotherapy followed by three courses of adjuvant cisplatin and 5-fluorouracil chemotherapy. A total of 147 evaluable patients with stages III and IV tumors were entered onto this study. At 3 years there was improved progression-free survival (69% vs 24%, p = .001), improved overall survival (76% vs 46%, p = .005) and reduced distant metastases (13% vs 35%, p = .002) on the experimental arm (the p value for distant metastases was calculated by us, using the Fisher exact test). The results of this trial changed the standard of care in the United States to the experimental-arm regimen. The results of this study also were corroborated by another randomized phase III trial, which compared concurrent cisplatin (40 mg/m2 weekly) and radiotherapy with radiotherapy alone in 350 patients with local and regionally advanced (Ho’s stage N2 or N3, or N1 with nodal disease ≥ 4 cm) NPC. This study found that combined chemotherapy plus radiotherapy was able to prolong progression-free survival (hazard ratio, 1.367; 95% CI, 0.93-2.0). The treatment effect had a significant covariate interaction with tumor stage, and subgroup analysis showed a significant difference in patients with Ho’s stage T3 disease in favor of the concomitant-therapy arm (p = .0075; hazard ratio, 2.328; 95% CI, 1.26-4.28). The time to first distant failure also was statistically prolonged in patients with T3 tumors of the concomitant-treatment arm versus T3-tumor patients of the radiotherapy-alone arm (p = .016). Treatment in the chemotherapy-plus-radiation arm was well tolerated.
Revision date: July 7, 2011
Last revised: by Andrew G. Epstein, M.D.