INTRODUCTION — Nasopharyngeal carcinoma (NPC) is an epithelial neoplasm arising from the lining of the nasopharynx, the narrow tubular passage behind the nasal cavity. Worldwide, there are 80,000 incident cases and 50,000 deaths annually, but there is a remarkable variation in racial and geographic distribution [1]. While rare in most parts of the world, NPC is endemic in southern China, southeast Asia, north Africa, and the arctic, where undifferentiated, nonkeratinizing squamous cell carcinoma is the predominant histology.

The treatment of locoregional nasopharyngeal cancer is presented here. The treatment of recurrent and metastatic NPC as well as the epidemiology, etiology, diagnosis, and staging of NPC is discussed separately. (See "Treatment of recurrent and metastatic nasopharyngeal carcinoma" and "Epidemiology, etiology, and diagnosis of nasopharyngeal carcinoma".)

STAGING AND CLASSIFICATION — NPC is clinically staged according to the International Union Against Cancer (UICC) and American Joint Committee on Cancer (AJCC) staging system (table 1) [2]. The World Health Organization (WHO) classifies NPC into three histopathologic types [3]:

• Keratinizing squamous cell carcinoma (SCC, WHO Type I)
• Nonkeratinizing carcinoma: differentiated (WHO Type II) and undifferentiated (WHO Type III)
• Basaloid squamous cell carcinoma

Staging and the histopathologic classification of NPC are discussed in more detail separately. (See "Epidemiology, etiology, and diagnosis of nasopharyngeal carcinoma".)

GENERAL TREATMENT PRINCIPLES — NPC differs from other head and neck squamous cell carcinomas (HNSCC) in its unique epidemiology, pathology, natural history, and response to treatment [4-6]. Radiation therapy (RT) is the mainstay of treatment for NPC but has evolved tremendously over the last decade owing to the integration of chemotherapy, improvement in tumor imaging and disease monitoring, and advances in high precision RT delivery [7]. Because of the deep anatomical location of the nasopharynx and its close proximity to critical neurovascular structures, surgery at the primary site is not typically used as first-line treatment. In selected cases, however, radical neck dissection may be indicated after RT for residual nodal disease or for an isolated neck recurrence, and nasopharyngectomy may be an option for a small local recurrence. (See "Treatment of recurrent and metastatic nasopharyngeal carcinoma".)

Prognostic risk groups — Based on the risk of treatment failure, both local and distant, three prognostic groups can be defined according to the UICC/AJCC staging system for NPC (table 1):

• Stage I and IIA - Early stage
• Stage IIB - Intermediate stage
• Stage III, IVA, IVB - Advanced stage

Whereas patients with early disease have good outcomes with RT alone, more intensive treatment strategies are required to manage intermediate and advanced stage disease.

Management of the neck — Given the propensity of NPC for early and bilateral spread to regional lymph nodes, all patients, even those with a clinically negative neck, are treated with bilateral neck irradiation [8,9]. (See "Management of the clinically negative neck in head and neck cancer".)

Endemic versus nonendemic NPC — Although there is some evidence that endemic NPC, ie, WHO type III, undifferentiated carcinoma, is particularly radiosensitive and may be associated with higher overall survival rates than WHO type I, keratinizing SCC, the histology more commonly encountered in the United States and Western Europe, there is otherwise insufficient evidence to suggest that treatment should differ based upon WHO histopathologic classification [10,11].

EARLY STAGE — NPC is a radiosensitive tumor, and because its anatomic location limits a surgical approach, it has been traditionally treated with RT. Early (stage I and IIA) cancers are treated with single modality RT with good locoregional control and survival. Five-year overall survival rates of 90 percent for stage I and 84 percent for stage II have been reported (table 2) [12].

Standard external beam radiation therapy (EBRT) is typically 70 to 72 Gray (Gy) to the primary tumor and 50 Gy to the uninvolved neck given in single daily fractions of 2.0 Gy, five days per week over six to seven weeks [5,13]. A similar dosing schedule is used for more advanced disease, with 66 to 70 Gy to involved lymph nodes. Conformal radiation therapy administration, particularly intensity modulated radiation therapy (IMRT), may improve local control but importantly, may diminish the risk of xerostomia in patients with NPC [14-16]. The use of three dimensional conformal RT or IMRT for delivery is strongly encouraged.

Evidence from retrospective studies suggest that outcomes for early stage NPC have improved as a result of advances in RT planning and delivery and improved staging with, as an example, magnetic resonance imaging (MRI) rather than computed tomography (CT) scanning [12,17]. However, it is unclear whether patients with early stage disease would benefit further from combined modality therapy, as these patients have not been included in any substantial number in trials of induction chemotherapy, adjuvant chemotherapy, or concurrent chemoradiation (CRT) [18,19]. The most established treatment for patients with early stage NPC is definitive RT.

INTERMEDIATE STAGE — Stage IIB disease, an intermediate risk group, is often considered for the combined modality treatment strategies, ie, concurrent CRT, that apply to stage III-IV disease because of elevated distant failure rates seen with stage IIB disease.

Retrospective data have suggested that some patients with stage IIB disease, in particular, those with T1N1 and T2N1 disease, have a worse outcome with RT alone than those with T2N0 disease [17,20]. Although patients with stage IIB disease have been included in prospective trials of induction chemotherapy or concurrent CRT, their numbers have been limited.

Evidence supporting a role for induction chemotherapy for intermediate stage disease is limited. Data for early stage patients (T1-T2N0-N1) was pooled from two phase III trials comparing induction chemotherapy followed by RT versus RT alone [18]. The addition of chemotherapy to these early stage patients appeared to significantly improve outcomes, particularly by reducing distant metastases. Five-year overall survival for induction chemotherapy versus RT alone was 79 versus 67 percent and the five-year distant metastases-free survival rate was 86 versus 71 percent.

Similarly, limited evidence supports concurrent CRT for intermediate stage patients because trials of concurrent CRT have predominantly enrolled locally advanced patients. A subgroup analysis from one trial found no benefit of concurrent CRT over RT alone for stage II patients [21]. In contrast, a single institution review suggested a marked improvement in three-year disease-free survival (DFS) for stage II patients treated with concurrent CRT such that their survival matched three-year DFS rates for stage I patients (97 percent for stage II patients versus 92 percent for stage I patients) [19].

Despite such limited supporting data, we and others suggest concurrent CRT for patients with stage IIB NPC to improve upon outcomes achieved with RT alone [22,23].

ADVANCED STAGE — Meta-analyses of randomized controlled trials have concluded that the addition of any adjunctive chemotherapy to definitive RT, whether concurrent CRT, induction chemotherapy, or adjuvant chemotherapy, reduces the risk of death by 18 percent and increases overall survival by 4 to 6 percent at five years [24,25]. A reduction in distant failure was what contributed most to the survival benefit, and concurrent CRT as opposed to induction chemotherapy or adjuvant chemotherapy demonstrated the most pronounced benefit (table 3). Concurrent chemoradiation, with or without adjuvant chemotherapy, has been established as the standard of care in advanced (stage III, IVA, and IVB) NPC.

Concurrent chemoradiotherapy — While a United States Intergroup trial first demonstrated the benefit of concurrent CRT for the management of locoregionally advanced NPC, several others have confirmed these results, importantly, in endemic regions. The rationale for combining chemotherapy concurrent with RT is discussed separately. (See "Methods to overcome radiation resistance in head and neck cancer: Concurrent chemoradiation".)

Intergroup Study 0099 randomized patients with stage III and IV NPC to RT alone (70 Gy in 35 to 39 fractions of 1.8 to 2.0 Gy daily) or chemotherapy with cisplatin (100 mg/m2 on days 1, 22, and 43) concurrent with RT followed by adjuvant cisplatin (80 mg/m2 on day 1) and 5-fluorouracil (1000 mg/m2 daily, days 1 through 4) repeated every four weeks for three cycles [26]. The study was terminated early at the time of the first planned interim analysis. Based upon analysis of 147 patients, concurrent CRT compared with RT alone significantly increased three-year progression-free survival (PFS, 69 versus 24 percent) and overall survival (78 versus 47 percent). This benefit remained with five years of follow-up, however, at the cost of increased toxicity [27].

Subsequently, many randomized controlled trials of concurrent CRT in locoregionally advanced NPC have been performed, particularly in endemic regions, with most confirming a survival advantage from concurrent CRT [21,28-33]. Furthermore, this benefit from concurrent CRT has been upheld in several meta-analyses, which have consistently demonstrated a survival benefit in favor of concurrent CRT (table 3) [24,25,34-36]. Consequently, concurrent CRT has become the standard of care for advanced NPC, in both nonendemic and endemic areas. Whether adjuvant chemotherapy following concurrent CRT adds additional benefit is discussed below.

Concurrent chemotherapy regimen — The regimen used in United States Intergroup Study 0099, cisplatin (100 mg/m2 on days 1, 22, and 43) concurrent with RT followed by adjuvant chemotherapy (cisplatin 80 mg/m2 on day 1 and 5-fluorouracil 1000 mg/m2 daily days 1 through 4, repeated every four weeks for three cycles) is considered by many to be the standard [26]. However, it also is associated with considerable severe acute and late toxicities. It can be difficult to administer the high-dose bolus cisplatin concurrent with CRT and to give patients adjuvant chemotherapy after concurrent CRT. Even in the clinical trial setting, compliance with the treatment protocol is suboptimal [26,28,32]. Therefore, alternative regimens have been investigated.

Weekly cisplatin (40 mg/m2) has demonstrated good efficacy, but has not been directly compared to high-dose bolus cisplatin [21,30,33]. Similarly, weekly oxaliplatin (70 mg/m2 over two hours) as concurrent CRT is superior to RT alone, but has not been compared to cisplatin [31].

The substitution of carboplatin, with a more favorable toxicity profile than cisplatin, has shown similar efficacy but improved tolerability when directly compared to cisplatin as a concurrent chemotherapy agent [37]. In this trial, patients were randomized to the Intergroup regimen, of high-dose bolus cisplatin (100 mg/m2 on days 1, 22, and 43) concurrent with RT followed by three cycles of adjuvant cisplatin and 5-fluorouracil, or carboplatin (100 mg/m2 on days 1, 8, 15, 22, 29, and 36) concurrent with RT and followed by three cycles of adjuvant chemotherapy consisting of carboplatin (area under the concentration x time curve [AUC] of 5 on day 1, (calculator 1) plus 5-fluorouracil (1000 mg/m2 over 96 hours) every four weeks, beginning four weeks after the end of RT [37]. Toxicity and survival rates were analyzed according to as-treated patient groups. More patients in the carboplatin arm versus the cisplatin arm completed concurrent CRT (73 versus 59 percent) and three cycles of adjuvant chemotherapy (70 versus 42 percent). Patients in the cisplatin arm had more renal toxicity (26 versus 0 percent), nausea and vomiting (59 versus 34 percent), anemia (47 versus 18 percent), weight loss, and need for enteral nutritional support, while patients in the carboplatin arm had more thrombocytopenia (4 versus 12 percent). At a median follow-up of 26 months, patients treated with either cisplatin or carboplatin had similar three-year DFS (63 versus 61 percent) and overall survival (78 versus 79 percent). In contrast, a retrospective study suggested that replacing cisplatin with carboplatin was associated with a poorer prognosis [38].

Confirmatory trials are needed before carboplatin-containing regimens can be recommended with the same confidence as cisplatin-containing regimens. However, for patients with borderline renal function, older age, or marginal performance status (less than ECOG 2, (table 4), substitution of carboplatin or lower dose, weekly cisplatin could be considered appropriate options.

Is adjuvant chemotherapy necessary? — It is unclear whether adjuvant chemotherapy adds benefit to concurrent CRT, since no trial has directly compared concurrent CRT plus adjuvant chemotherapy with concurrent CRT alone. Although several large concurrent CRT trials included three courses of postradiation adjuvant chemotherapy, results were similar in trials where concurrent CRT was used alone [21,26-30,32,33,39]. Individually and in meta-analysis, trials comparing the addition of adjuvant chemotherapy to RT alone showed no significant reduction in the risk of locoregional recurrence and/or distant metastases and no significant improvement in overall survival [24,25,40-42]. Furthermore, completion of adjuvant chemotherapy is difficult after concurrent CRT. In the Intergroup Study 0099, only 55 percent of patients completed three cycles of adjuvant chemotherapy as planned [26].

Whether adjuvant chemotherapy adds benefit to concurrent CRT remains debated. Due to the influence of the Intergroup Study 0099, some experts give three cycles of adjuvant chemotherapy as used in this trial, especially for patients who are fit and have good performance status. However, because of insufficient data in support of a benefit, we suggest not routinely giving adjuvant chemotherapy after concurrent CRT. A phase III trial of adjuvant chemotherapy in high risk patients (defined by residual EBV DNA following completion of RT or CRT) has been initiated (information available online at http://clinicaltrials.gov/ct2/show/NCT00370890?term=HKNPCSG+0502&rank=1).

Induction chemotherapy — Based upon the reduction in distant metastases seen with induction chemotherapy for squamous cell carcinomas in other head and neck sites, induction chemotherapy in combination with RT alone or concurrent CRT, has been investigated in NPC with the hope of improving survival above that achieved with concurrent CRT. While induction chemotherapy trials have shown disappointing results, sequential therapy holds more promise. (See "Induction chemotherapy for locoregionally advanced head and neck cancer".)

Trials of induction chemotherapy followed by RT alone have failed to show an improvement in overall survival or pattern of relapse compared to RT alone [43-46]. Since many of the studies were underpowered to show a significant difference in outcomes, a pooled analysis of two of the largest trials was performed [45]. Although the addition of induction chemotherapy to RT did result in a modest improvement in relapse-free and disease-specific survival, overall survival was not significantly improved.

Sequential therapy — Sequential therapy, the administration of induction chemotherapy followed by concurrent CRT, appears promising and may represent an alternative to concurrent CRT with adjuvant chemotherapy [47-53]. Furthermore, induction chemotherapy preceding concurrent CRT is more likely to be successfully administered than adjuvant chemotherapy following concurrent CRT. However, it is not known whether sequential therapy represents an improvement over concurrent CRT alone or concurrent CRT followed by adjuvant chemotherapy.

A randomized phase II trial that compared sequential therapy (induction chemotherapy with docetaxel and cisplatin followed by concurrent weekly cisplatin CRT) to concurrent CRT alone (with weekly cisplatin), demonstrated that although there were high rates of grade 3 and 4 neutropenia (97 percent) during induction chemotherapy, sequential therapy allowed for comparable dose intensities of concurrent cisplatin and resulted in comparable acute and late toxicities and quality of life scores [53]. Although the improvement in three-year progression free survival with sequential therapy was large, it was not significant (88.2 versus 59.5 percent); the increase in overall survival rate was significant, 94.1 versus 67.7 percent.

A multicenter phase III trial of sequential therapy in NPC with three-drug induction chemotherapy consisting of docetaxel, cisplatin, and 5-fluorouracil (TPF) followed by weekly cisplatin concurrent with RT compared to weekly cisplatin-based concurrent CRT alone has been initiated by Radiotherapy Oncology Group for Head and Neck Cancer (GORTEC).

At the present time, sequential therapy in NPC is considered experimental until phase III data are available to support its use. Some experts, however, would chose sequential therapy for large primary tumors (T4 lesions), advanced nodal disease (large burden or supraclavicular location), or when the tumor abuts the surrounding critical structures, such as the optic apparatus, the brainstem, or the temporal lobes, which limit the delivery of full dose RT.

Further treatment intensification — Although still investigational, molecularly targeted agents, such as cetuximab and bevacizumab, are being added to concurrent CRT and sequential therapy regimens. Modifications in RT delivery, including accelerated fractionation RT, stereotactic radiosurgery, and intracavitary brachytherapy boost, have been explored in an effort to improve upon results achieved with EBRT, but also remain investigational [54-60]. However, conformal radiation therapy administration, particularly intensity modulated radiation therapy (IMRT), may improve local control but importantly, may diminish the risk of xerostomia in patients with NPC [14,15]. (See "General aspects of radiotherapy for head and neck cancer".)

PROGNOSIS — The five-year overall survival for NPC according to disease stage in a contemporary case series was 90, 84, 75, and 58 percent for stage I through IV, respectively (table 2) [12].

In addition to prognostic factors of T and N stage and location of neck nodes (above or within the supraclavicular fossa), pre- and posttherapy Epstein-Barr Virus (EBV) DNA levels have prognostic significance, with higher levels conferring a worse prognosis, stage for stage [10,61-63]. Five-year survival rates according to UICC/AJCC stage grouping and pretherapy EBV DNA levels from one case series are as follows [63]:

• Stage I, II disease, low DNA (<4000 copies/mL) - 91 percent
• Stage I, II disease, high DNA (≥4000 copies/mL) - 64 percent
• Stage III, IV disease, low DNA - 66 percent
• Stage III, IV disease, high DNA - 54 percent

The presence of detectable posttreatment EBV DNA, which may reflect microscopic residual tumor, appears to provide an even greater estimate of the risk of disease recurrence and death than pretreatment EBV DNA [48,62,64-67].

POSTTREATMENT EVALUATION AND FOLLOW-UP — Documentation of complete remission in the nasopharynx and neck through clinical and endoscopic examination and imaging studies is important. Our preference is to obtain a posttreatment baseline MRI scan of the skull base and neck and a full body combined positron emission tomography (PET)/computed tomography (CT) scan at approximately three months after treatment completion. Sometimes, distinguishing between slowly regressing tumor, residual tumor, or posttherapy changes may be difficult. MRI and PET-CT scans may achieve higher accuracy when combined rather than individually in detecting residual disease [68,69]. Obtaining imaging studies too early, in particular combined PET/CT scans prior to 12 weeks following treatment, can lead to false positive results.

Postreatment surveillance is important for early detection of recurrent disease, whether local or metastatic, most commonly bone, lung, and liver, and for monitoring for toxicity. Follow-up for NPC patients includes periodic examination of the nasopharynx and neck, assessment of cranial nerve function, and evaluation of systemic complaints. In contrast to other head and neck cancers, periodic upper endoscopy is typically recommended. NPC is also different from other HNC sites in its propensity to recur, both locally and distantly, much later than other HNC sites. Consequently, we follow patients every three months for the first two years, every four to six months for years 3 to 5, and annually, thereafter. We also suggest an annual chest x-ray; other experts suggest reimaging only as indicated by signs and symptoms [67].

As noted above, posttherapy levels of EBV DNA are of prognostic significance, although there is no consensus about which level to use as a cutoff (both 0 and 500 copies/mL have been used) and the timing of posttreatment EBV DNA sampling [62,64]. At present prospective data in support of routine monitoring by plasma EBV DNA is lacking, although measuring plasma EBV DNA can be a useful diagnostic aid in suspected recurrence. Posttreatment monitoring of patients by assay of a nasopharyngeal swab for detection of the EBNA-1 and latent membrane protein-1 genes by RT-PCR has also been suggested as a means of early detection of recurrent disease, but this methodology is expensive, not widely available, and has not been prospectively validated [66].

Neck irradiation places patients at risk for developing hypothyroidism, thus monitoring of serum thyroid stimulating hormone (TSH) levels is routine. Surveillance after treatment for head and neck cancer is discussed in detail elsewhere. (See "Posttreatment surveillance of head and neck cancer".)

TREATMENT-RELATED COMPLICATIONS — Mucositis is the predominant acute toxicity associated with RT alone, while the addition of chemotherapy into NPC treatment introduces the risk of emesis and hematologic toxicity, mainly neutropenia [26,33,53]. Late treatment-related complications after RT or concurrent CRT include injury to the temporal lobes, sensorineural hearing loss, hypothalamic-pituitary dysfunction, second cancers, cranial nerve palsies from injury to the brainstem, and xerostomia [55,70-81].

• Temporal lobe necrosis, characterized by memory loss, complex partial seizures, and hypodense areas in one or both temporal lobes on CT scanning occurs in two to three percent of patients and is significantly increased with higher doses of RT, altered fractionation techniques, and shorter overall treatment times [55,70,79]. A preliminary report suggests that high doses of vitamin E (eg, 2000 international units daily for one year) may ameliorate this [80].
• Skull base osteoradionecrosis with bleeding from the internal carotid artery is an uncommon but potentially fatal complication of irradiation for NPC [81].
• Delayed bulbar palsy, developing 1 to 18 years post radiation, is reported in up to 20 percent of cases and can result in moderate to severe functional disability [76]. Deficits may include any combination of deafness, dysarthria, dysphagia, tongue and palatal weakness, and motor weakness of the sternocleidomastoid, trapezius, supraspinatus, infraspinatus, and rarely, the deltoid muscle.
• Xerostomia can be a long lasting or even permanent problem, although it often improves with time.
• Radiation-induced second cancers are frequently EBV-negative, squamous cell carcinomas and occur in the tongue and temporal bone [73,74].

(See "Complications of cranial irradiation" and "Complications of radiotherapy for head and neck cancer".)

SUMMARY AND RECOMMENDATIONS — Because of the deep anatomical location of the nasopharynx and its close proximity to critical neurovascular structures, radiation therapy (RT), rather than surgery, is the mainstay of first-line treatment for early stage nasopharyngeal carcinoma (NPC). For more advanced disease, concurrent chemoradiation (CRT) reduces the rate of distant metastasis and improves local control and overall survival compared to RT alone.

• Treatment approaches are based upon the stage of the disease. Despite differences in prognosis, there is insufficient evidence to suggest that treatment should differ based upon WHO histopathologic classification. (See 'Prognostic risk groups' above.)

• For patients with early (stage I and IIA) disease, we recommend RT alone (Grade 1B). (See 'Early stage' above.)

• For patients with intermediate (stage IIB) disease, we suggest concurrent chemoradiation (CRT) rather than RT alone (Grade 2C). This approach may improve survival, but also increases toxicity. (See 'Intermediate stage' above.)
• For patients with advanced (stage III, IVA, and IVB) disease, we recommend concurrent CRT (Grade 1A). We suggest not using adjuvant chemotherapy for most patients (Grade 2C). While adjuvant chemotherapy has been a standard part of concurrent CRT regimens, its benefit is uncertain and toxicity is substantial. Adjuvant chemotherapy may be a reasonable option for patients with high-risk disease and a good performance status. (See 'Advanced stage' above.)

• For eligible patients being treated with concurrent CRT, cisplatin (100 mg/m2 on days 1, 22, and 43) concurrent with RT is a standard option for patients with good performance status. (See 'Concurrent chemotherapy regimen' above.)

• Low dose cisplatin (40 mg/m2 weekly) concurrent with RT or the substitution of carboplatin for cisplatin are options for patients with a poor performance status or comorbidities. (See 'Concurrent chemotherapy regimen' above.)

• Until more data in support of sequential therapy are available, we suggest not using sequential therapy for most patients with advanced NPC (Grade 2C). Some experts would choose sequential therapy for large or extensive primary tumors or advanced nodal disease. (See 'Sequential therapy' above.)

• Given the propensity of NPC for early and bilateral spread to regional lymph nodes, we recommend bilateral neck irradiation for all patients, even those with a clinically negative neck (Grade 1B). (See 'Management of the neck' above.)

• We obtain a posttreatment MRI scan of the skull base and neck and a full body combined positron emission tomography (PET)/computed tomography (CT) scan at approximately three months after treatment completion to evaluate treatment response and to serve as a baseline. (See 'Posttreatment evaluation and follow-up' above.)

• NPC has a propensity to recur, both locally and distantly, much later than other HNC sites; consequently, we follow patients every three months for the first two years, every four to six months for years 3 to 5, and annually, thereafter. (See 'Posttreatment evaluation and follow-up' above.)

ACKNOWLEDGMENT — The editors at UpToDate, Inc., would like to acknowledge Drs. Gary S Gordon and Bruce E Brockstein, who contributed to earlier versions of this topic review.