Epidemiology
Biology
Anatomy
Diagnosis and Staging
Therapy
Combined Modality Therapy for Local or Regional Advanced Disease
Chemoprevention
Conclusions
References

Head and neck cancers are a diverse group of diseases, each with its own distinct epidemiologic, anatomic, and pathologic features, natural history, and treatment considerations. Despite improvements in diagnosis and local management, long-term survival rates for patients with this disease have not increased significantly over the past 30 years and are among the lowest for the major cancers. In this chapter we will review the major aspects of the epidemiology, biology, diagnosis, chemoprevention, and therapy for head and neck cancer. Because squamous-cell carcinoma accounts for almost 95% of the neoplasms in this region, the discussion will refer mainly to this histopathologic subtype.

Epidemiology

In the United States, head and neck cancers account for 3.2% (39,750) of all new cancers and 2.2% (12,460) of cancer deaths [1]. The disease is more common in many developing countries, with a worldwide annual incidence of more than 500,000. The incidence of head and neck cancer increases with age; most patients are older than age 50. Mainly because of the increasing use of tobacco among women, the male-to-female ratio has decreased from 4 to 5:1 to approximately 3:1 over the past 5 to 10 years. Furthermore, some studies have reported a higher risk for women at each successive pack-year stratum of smoking [2], a finding not shown for alcohol(Drug information on alcohol) [3]. From 1973 to 1989, the incidence of oral and pharyngeal carcinomas decreased in white men of all ages, whereas the African-American population experienced a significant increase [4].

The greatest risk factor is tobacco use. Although alcohol use produces a modest independent risk, it exponentially potentiates tobacco risk. A recent review showed heavy smokers to have a 5- to 25-fold higher risk of head and neck cancer than nonsmokers. Smoking unfiltered cigarettes, rather than filtered cigarettes, lowered the risk only minimally. Although the risk of developing head and neck carcinoma is lowered substantially by smoking cessation, the relative risk of former heavy smokers never reaches that of nonsmokers.

The use of smokeless tobacco, particularly prevalent among teenagers as they attempt to emulate sports personalities, is strongly associated with the formation of premalignant oral lesions (hyperkeratosis, epithelial dysplasia), at rates ranging from 16% to 60% [5,6]. Smoking of marijuana, cigars, and pipes is also associated with head and neck neoplasia, especially in the oral cavity [7,8].

Genetic susceptibility to environmental carcinogens may explain the fact that only a fraction of carcinogen (eg, tobacco)-exposed individuals develop these cancers. The best recent support for the importance of inherent susceptibility to carcinogens in the development of head and neck cancer comes from an assay of sensitivity of peripheral blood lymphocytes to mutagen-induced chromosomal damage. Mutagen sensitivity (eg, to bleomycin(Drug information on bleomycin) [Blenoxane]) has been shown to be a strong independent risk factor for the development of head and neck cancer and seems to have a multiplicative interaction with smoking. Mutagen-induced chromosome sensitivity also has been reported to correlate with the prospective development of second primary tumors [9,10].

Dietary factors also seem to play a prominent role in the risk of oral and pharyngeal cancers. Numerous epidemiologic studies have shown an increased risk of cancer in individuals whose diets lack sufficient quantities of nutrients. Other risk factors are listed in Table 1.

Certain viruses may contribute to the development of head and neck cancer [4,11]. The best-described agent is the Epstein-Barr virus (EBV), which is associated with nasopharyngeal carcinoma (NPC), an uncommon form of head and neck cancer in the United States but a common form in some north African and Asian countries. Most patients with NPC show evidence of an elevated serum titer of immunoglobulin (Ig)G and IgA antibodies against viral capsid antigen, and EBV viral genome has been found in NPC tissue [12,13]. The association of NPC and EBV (in antibody and DNA) is particularly strong in patients with endemic undifferentiated carcinoma [11]. In China, multiple other risk factors for NPC have been reported, including ingestion of salted fish and inhalation of smoke from cooking fires. In certain parts of China, the rate of NPC is as high as 54.7 in 100,000 persons, compared with fewer than 1 in 100,000 in the United States, where NPC has been linked to tobacco use [14]. Human papillomavirus, especially types 16 and 18, and herpes simplex virus type I have been detected in the sera and neoplastic tissues of patients with head and neck cancer, but no risk vs prevalence relationship has been found [15,16].

Biology

The mucosal surfaces of the upper aerodigestive tract, lungs, and esophagus are repeatedly exposed to the same carcinogens (eg, tobacco, alcohol). Therefore, multiple independent neoplastic lesions may arise and progress in the same patient, either simultaneously or sequentially, in a process known as “field cancerization.” [17] Metachronous second primary tumors develop at a constant rate of 4% to 7%, are generally of squamous pathology, are not treatment-related, and occur in the carcinogen-exposed regions of the aerodigestive tract [18,19].

Tumorigenesis in the aerodigestive tract is a multistep process believed to be driven by genetic damage caused by continuous exposure to carcinogens [20]. Specific genetic alterations linked with head and neck cancer include the activation of oncogenes (eg, cyclin D1), the inactivation or mutation of tumor-suppressor genes (eg, p53), and the amplification of growth factors and their receptors (eg, epidermal growth factor, epidermal growth factor receptor [EGFR], and transforming growth factor). These specific gene alterations may result in phenotypic changes in cell differentiation or proliferation or both (Table 2).

Multiple allelic abnormalities (eg, 3p, 9p, 11q, 13q, and 17p) have been documented, and in some studies they appear to have prognostic value [21,22]. The cyclin D1 gene, also known as PRAD-1, bcl-1, or CCND-1, has been the focus of several recent studies [23-25]. This oncogene, located on chromosome 11q13, is amplified in 30% to 50% of patients with head and neck cancers. Overexpression and amplification of cyclin D1 has been associated with more advanced disease, more rapid and frequent recurrence of disease, and shortened survival [24]. Alterations in the expression of other genes on 11q13 (eg, int-2, hst-1, and ems-1) are also under active study in the head and neck.

The tumor-suppressor gene p53, located on the short arm of chromosome 17, is the most thoroughly studied gene in head and neck neoplasia and many other cancers. Mutations and overexpression of p53 occur in 40% to 60% of cancer patients and have been associated with a poor prognosis [26]. Brachman et al [27] reported that tumors with p53 mutations recurred at a median time of 6 months, compared with a median time to recurrence of 17.4 months for tumors without mutations. The introduction of sensitive molecular techniques has made it possible to detect neoplastic changes in tissue “free of disease” by routine cytology. Brennan et al [28] recently reported that 13 of 25 patients who appeared to have complete tumor resection according to histopathologic assessment had p53 mutations in at least one tumor margin. Thirty-eight percent of the patients with p53-positive margins relapsed, compared with none of the 12 patients found by polymerase chain reaction (PCR) to have all tumor margins free of p53 mutations. This method may prove useful for identifying patients who are prone to local tumor recurrence and, therefore, in need of adjuvant therapy.

Mutations of p53 have been associated with tobacco and alcohol use. A recent study showed p53 mutations in 58% of tumors of cigarette smokers who also used alcohol. Among patients who smoked but did not drink alcohol, 33% had mutations, whereas only 17% of the patients who neither smoked nor drank alcohol showed mutations of the p53 gene. Furthermore, the number of types of p53 base-pair mutations was highest in patients who smoked and drank and lowest in patients who did neither. These findings provided molecular evidence that tobacco may play a role in the molecular progression of squamous-cell carcinoma of the head and neck [29].

EGFR is a cellular oncogene likely to play a role in head and neck tumorigenesis. Its gene amplification and overexpression have been demonstrated in preinvasive and invasive lesions, with high expression associated with an increased sensitivity to cytotoxic therapy [30]. The finding that anti-EGFR monoclonal antibodies upregulate EGFR may prove useful in enhancing chemotherapeutic efficacy. As a recent study demonstrated, in patients who smoked or drank alcohol or did both, increased expression of EGFR is found in histologically normal tissue surrounding cancers, compared with levels in controls never exposed to tobacco or alcohol [31]. These findings supported the theory of field cancerization in which epithelia exposed to carcinogenic insult may express early premalignant features.

Proliferating cell nuclear antigen (PCNA) is a nuclear protein whose expression is associated with DNA synthesis and cell proliferation. Shin et al [32] analyzed PCNA expression in squamous-cell carcinoma tissue samples and surrounding premalignant and normal tissues; PCNA expression sequentially increased 4- to 10-fold as tissue progressed from adjacent normal epithelium to squamous-cell carcinoma. Similarly high expression of TGF-alpha was also documented to be a strong mitogenic factor capable of inducing epithelial proliferation [32].

Anatomy

Cancers of the head and neck include a great variety of neoplasms, specifically those involving the upper aerodigestive tract. Tumors of the central nervous system, eye, skin, and thyroid, as well as tumors of lymphatic origin, are usually excluded.

Head and neck cancers originate in the area under the base of the skull to just below the larynx in a cephalocaudal orientation and by the anterior nasal cavity and the vermilion border of the lips anteriorly to the pharynx posteriorly (Figure 1). These boundaries enclose the nasal cavity, pharynx, larynx, and salivary glands. The oral cavity encompasses the oral mucosa, anterior two thirds of the tongue, and floor of the mouth. The oropharynx (which includes the posterior one third of the tongue), hypopharynx, and nasopharynx are subdivisions of the pharynx.

Diagnosis and Staging

Optimal therapy and survival depend on the proper identification of the primary tumor as well as the local, regional, and distant extent of disease. Patients with early-stage disease may present with vague symptoms and minimal physical findings, which is why a high index of suspicion is needed, especially if the patients use tobacco. Most patients will present with signs and symptoms of locally advanced disease that vary according to the subsite in the head and neck. Sinusitis, unilateral nasal airway obstruction, and epistaxis may be early symptoms of cancer of the nasal cavity and paranasal sinus. Persistent hoarseness demands visualization of the larynx. Otitis media that remains unresponsive to antibiotics may indicate a nasopharyngeal neoplasm. Chronic dysphagia or odynophagia (lasting 6 weeks or more) may be the presenting symptom of oropharyngeal or hypopharyngeal cancer. Supraglottic laryngeal neoplasms rarely present early symptoms; in some patients, a neck mass will be the presenting sign. Detailed examination of lymph nodes in the facial, cervical, or supraclavicular regions is important because the location of adenopathy provides clues to the specific subsite of a head and neck primary tumor, as shown in Figure 2. Subdigastric adenopathy, for example, suggests primary cancer of the oral tongue or oropharynx, and posterior cervical adenopathy is a frequent result of regional spread of a nasopharyngeal tumor.

Physical examination should include careful inspection of all mucosal surfaces, bimanual palpation of the floor of the mouth, and palpation of the neck. Leukoplakia (white mucosal patches that cannot be removed by scraping) and higher risk erythroplakia (red or mixed red-white patches) are the most common premalignant lesions in the head and neck. Up to 40% of cases of dysplastic oral leukoplakia transform into invasive carcinoma. Any suspicious surface in the oral mucosa (eg, erythroplakia) should undergo biopsy to rule out cancer [33].

Three-dimensional imaging with computerized tomography (CT), magnetic resonance imaging (MRI), and ultrasonography may be used to evaluate the extent of disease and to stage the neoplasm. These techniques are also helpful in evaluating a patient's response to therapy. Because the lungs are the most common site of distant metastasis, a chest x-ray should be performed as well [34]. Circulating tumor markers that would be useful for identifying squamous-cell carcinoma of the head and neck have not been discovered. Epstein-Barr virus DNA is found in nasopharyngeal carcinoma but not in other types of upper aerodigestive tract neoplasms. In patients with cervical lymphadenopathy and no obvious primary tumor, identification of EBV DNA in the lymph node may suggest a tumor of nasopharyngeal origin [35].

Patients who present with a suspicious neck mass and no obvious lesion in the oral cavity should undergo a flexible fiberoptic nasopharyngoscopy or indirect laryngoscopy to search for the primary tumor. A panendoscopy (direct laryngoscopy, bronchoscopy, esophagoscopy), performed while the patient is under anesthesia, is the definitive diagnostic and staging procedure, and it should be performed for all patients except those whose lesions originate in the oral cavity. Multiple biopsies of any visualized abnormality, or “blind biopsies” of random areas, are performed to define the extent of disease and to identify synchronous second primary tumors. If no obvious primary site is found, fine-needle aspiration of the lymph node is performed to establish the diagnosis. Open biopsy is a last resort, after fine-needle aspiration or panendoscopy fail to reveal the specific site and histopathology. In patients whose neck is surgically violated by open biopsy, subsequent curative therapy may be compromised.

Staging criteria for head and neck cancers are based on the tumor-nodes- metastasis (TNM) staging system, which classifies tumors according to anatomic site and extent of disease [36]. Head and neck primary (T) tumor staging is complex, varying with each primary subsite in the head and neck region. Classifications for lymph-node (N) and distant metastases (M) are uniform for all sites (Table 3).

TABLE 3: Staging System for Head and Neck Cancer

Tumor-node metastasis (NRM) system
TX	Primary tumor cannot be assessed
T0	No evidence of primary tumor
Tis	Carcinoma in situ
T1	Tumor < 2 cm in greatest dimension
T2	Tumor > 2 cm but < 4 cm
T3	Tumor > 4 cm
T4	Tumor invades adjacent structures (eg, cortical bone, deep muscle of tongue, maxillary sinus, skin)
NX	Regional lymph nodes cannot be assessed
N0	No regional node metastases
N1	Metastasis to a single ipsilateral lymph node 3 cm or smaller
N2	Metastasis to a single ipsilateral lymph node > 3 cm but not > 6 cm in greatest dimension; to multiple ipsilateral lymph nodes (none > 6 cm); or to bilateral or contralateral lymph nodes (none > 6 cm)
a	Metastasis to a single ipsilateral lymph node (> 3 cm but not > 6 cm)
b	Metastases to multiple ipsilateral lymph nodes (none > 6 cm)
c	Metastases in bilateral or contralateral lymph nodes (none > 6 cm)
N3	Metastasis to a lymph node > 6 cm
MX	Presence of a distant metasis cannot be assessed
M0	No evidence of distant metastasis
M1	Distant metastasis
Clinical stage
0	Tis N0 M0
I	T1 N0 M0
II	T2 N0 M0
III	T3 N0 M0 T1 N0 M0
IV	T4 N0-1 M0 Any T N2-3 M0 Any T Any N M1

Therapy

Surgery and radiation therapy have been the standard of care for most patients with head and neck neoplasms. Traditionally, chemotherapy has been used only for patients with recurrent or metastatic disease. More recently, chemotherapy has had an important role in the primary treatment of most patients with locally advanced resectable laryngeal cancer, in whom induction chemotherapy followed by radiotherapy may be used to avoid the need for total laryngectomy.

More than two thirds of patients with head and neck cancer will present with stage III or IV disease. For patients with early-stage disease (I or II), surgery or radiotherapy is used with curative intent, and more than 80% of patients with stage I disease and more than 60% with stage II disease achieve this goal. In patients with stage III disease and most patients with stage IV disease, surgery followed by radiation therapy is considered standard care unless the patient is unable to undergo surgery or the lesion is unresectable. In many cases, this curative treatment, especially for patients with disease at an advanced stage, is accompanied by severe cosmetic and psychological problems and long-term loss of organ function. Despite optimal local therapy, more than 50% of patients with stage III and IV disease will develop local or regional recurrence, and nearly 30% will develop distant metastases. The need to improve survival rates and decrease treatment-related morbidity has encouraged the investigation of new approaches. Chemotherapy is under intense study in locally advanced disease, with promising results. In this section we will review standard and new approaches to treating patients with head and neck cancer.

Nasopharynx: The three histopathologic subtypes of nasopharyngeal carcinoma are type 1, differentiated squamous-cell carcinoma; type 2, nonkeratinizing squamous-cell carcinoma; and type 3, undifferentiated or lymphoepithelioma [37]. About 50% to 75% of NPCs in the United States are type 1 or 2, whereas in Asian and African areas, type 3 NPC predominates. NPC is an aggressive neoplasm involving cervical lymphadenopathy in 60% to 90% of patients [38]. Because of the tumor's unique anatomic, biologic, and clinical characteristics, therapy for NPC is different from that of other head and neck cancers. Radiation is the therapeutic mainstay. Pilot studies of intracavitary radiotherapy for NPC have reported excellent palliation. The anatomic location of the nasopharynx limits the role of surgery to obtaining the initial biopsy and to resecting residual lymphadenopathy after radiotherapy. The lymphoepitheliomas (type 3) are the most chemoradiosensitive.

Five-year survival rates range from 30% to 65% depending on tumor stage, the radiation technique employed, and the percentage of patients with lymphoepitheliomas [38]. Patients with N0 disease have a 5-year survival rate of approximately 65%, whereas patients with cervical node involvement have a significantly worse prognosis [39].

Oral Cavity: The majority of oral-cavity neoplasms occur in the anterior two thirds of the tongue (oral tongue) and the floor of the mouth. Radiation or surgery, or both, are the main therapeutic methods. Treatment-related morbidity, especially with respect to speech and swallowing, are major issues in managing these patients; the treatment decision is also complicated by the high rate of regional metastases.

For most early-stage cancers (T1 and T2) that arise in the floor of the mouth, comparable results are obtained with either surgery or radiation therapy alone. Small, superficial lesions may be excised with little morbidity [40]. For T2 lesions that are deeply invasive, radiation therapy offers the advantage of simultaneously treating the bilateral neck lymph nodes, which have a 30% to 40% rate of involvement [18]. Interstitial radiotherapy is used in select cases and achieves higher control rates than external radiation alone. With either surgery or radiation, survival rates of patients with stage I and II tumors are 80% to 90% and 50% to 80%, respectively [41].

Oral-tongue cancers commonly arise on the lateral and ventral surfaces of the anterior part of the tongue. They are so aggressive that 40% of patients present with clinically evident lymph nodes and 35% with occult node metastasis. Bilateral node involvement is not uncommon. For T1 lesions, radiotherapy or surgery has a similar result, and both achieve good preservation of organ function. For infiltrative T2 lesions, surgery usually includes hemiglossectomy and selective nodal dissection, which achieve excellent control at the expense of compromising oral function. Brachytherapy combined with external radiotherapy may be as effective as surgery at this stage. More than 90% of patients with early recurrence after radiation may be salvaged by surgical intervention.

For patients with more advanced disease (T3, T4), surgery followed by radiation therapy is the most widely accepted approach. Neck dissection is usually performed when patients have cervical lymphadenopathy or are at high risk for occult metastasis. Radiation therapy alone has a failure rate of 80% to 90% in patients with advanced disease, compared with a failure rate of less than 30% for the combined modality of surgery and radiotherapy [42]. Survival rates for patients with advanced oral-cavity cancer are between 18% and 42% [18].

Oropharynx: The most common cancers of the oropharynx are of the base of the tongue and tonsils. Early tonsillar lesions (T1 and T2) may be treated with radiotherapy or surgery, with good results. The tendency is to irradiate the lesions, because local control rates of 90% for T1 lesions and 80% for T2 lesions have been obtained, and regional lymph nodes may be treated concurrently [43]. Intraoral wide local excision is possible for small, superficial lesions, with comparable results. For larger or deeper T1 or T2 lesions, radiation therapy is preferred to surgery because of the latter's higher morbidity (disfigurement and loss of oral function) [44]. For T3 tumors without clinical evidence of nodal involvement (N0), radiation therapy alone has resulted in a 5-year survival rate of 82%. Some studies have shown increased local control for combined-modality approaches [45]. For patients with T4 lesions, combined surgery and irradiation have achieved increased survival rates at the expense of a substantial decline in the quality of life.

Cancer at the base of the tongue presents a more difficult situation than tonsillar cancer because of its anatomic location, late diagnosis, and common metastasis. At presentation, 75% of patients have stage III or IV disease. For patients with T1 and T2 disease, surgery with or without neck dissection has shown a local control rate of 75% to 80%, whereas radiation therapy with or without brachytherapy has yielded local control rates of 70% to 100% [46]. Radiotherapy is preferred because of the preservation of oral function. The major prognostic determinant is lymph node involvement. For patients with N0 disease, the 5-year survival rate is higher than 60%, but for N1 to N3 patients it is lower than 30% [47].

Hypopharynx: With 75% of lesions occurring in the piriform sinus, carcinoma of the hypopharynx is relatively uncommon but highly lethal. At presentation, more than 80% of patients have advanced disease (T3 and T4), and more than 20% have distant metastases [48]. The rare patient with a T1 lesion is usually treated with radiation therapy; those with T2 to T4 lesions are mainly treated with total laryngectomy and partial pharyngectomy, radical neck dissection, and postoperative radiotherapy. The overall 5-year survival rate is lower than 25%. For early lesions CO2 laser surgery has been used with encouraging results and with preservation of more pharyngeal function than is possible with other surgical techniques [49]. A recent phase III trial showed that laryngeal preservation with sequential chemoradiotherapy is a feasible alternative to radical surgery [50].

Larynx: The most widely used treatment of T1 and T2 cancers of the larynx is radiotherapy, which has demonstrated 5-year survival rates of 96% to 98% for patients with T1 disease and 80% to 94% for those with T2 disease [51]. At early tumor stages, endoscopic laser surgery has shown local control rates of up to 90%, but the patients' voice quality was poor [52]. For early-stage carcinomas of the supraglottic larynx, supraglottic laryngectomy has resulted in excellent local control; an alternative is radiation therapy, for which rates of local control range from 50% to 90% in T1 or T2 disease. For fixed vocal cords (T3 disease) radiotherapy has produced local control rates of between 30% and 60%. In cases of localized recurrence, laryngectomy has achieved a cure rate of up to 80% [53]. For most patients with T3 lesions and virtually all patients with T4 lesions, total laryngectomy with postoperative radiation had been the standard treatment. Chemotherapy now has a more prominent role in the treatment of these cancers; most patients with locally advanced laryngeal carcinoma can now be offered the option of sequential chemoradiotherapy in an attempt to preserve the larynx (see "Chemotherapy for organ preservation").

Cervical Nodes: Management of the neck should be integrated with management of the primary tumor [40]. In patients with clinically negative nodes (N0) the decision to treat the neck electively depends mainly on the T stage and site of the primary tumor. Because of the relatively high incidence of nodal involvement in early stages, some authors have recommended treating the neck for N0 cancer in the nasopharynx, hypopharynx, oropharynx, and supraglottic larynx. For stage I or II glottic laryngeal carcinoma, however, the rate of nodal involvement is less than 5%, and neck treatment is not usually performed [52].

When surgery is the treatment chosen for a primary tumor with N0 disease, conservative nodal dissection could achieve good control in more than 90% of cases [54]. Similarly, when irradiation is the primary treatment, it may be applied to both sides of the neck.

When nodes are clinically present, combined-modality surgery and radiation may be used, depending on the primary treatment results. If radiation is primary, surgery is reserved for residual disease. For patients with advanced nodal disease (N2, N3) treated with surgery, irradiation is added (including to the contralateral nodes) to increase local control.

Salivary Glands: Tumors of the salivary glands (mucoepidermoid carcinoma, adenoid cystic carcinoma, adenocarcinoma, malignant mixed tumor, acinic-cell carcinoma, and epidermoid carcinoma) are considered different from other head and neck cancers because of their diverse histology (fewer than 3% are squamous carcinomas) and high percentage of benign lesions, compared with tumors at other sites. Benign lesions account for 80% of tumors arising in the parotid gland, 50% of tumors arising in the submandibular glands, and 25% of tumors arising in minor salivary glands. The most common malignant tumor of the major salivary glands is mucoepidermoid; its most common subsite is the parotid gland. Adenoid cystic carcinomas, which make up nearly 25% of malignant salivary gland tumors, are the most common type encountered in the minor salivary glands. This subtype is considered unique because of its prolonged natural history (10 to 15 years) and its resistance to radiation and chemotherapy. For this reason, chemotherapy and radiotherapy are not indicated unless the tumor is growing rapidly, is symptomatic, is causing local problems, or is associated with extrapulmonary metastatic visceral disease. Cranial nerve involvement is common and associated with poor local control and survival. Surgery and postoperative radiotherapy remain the major treatments for patients with cancer limited to the primary site and regional lymph nodes. Chemotherapy has been used mainly as palliation; trials of chemotherapy have been limited by the rarity of the disease. Doxorubicin(Drug information on doxorubicin) (Adriamycin, Rubex) and cisplatin(Drug information on cisplatin) (Platinol) seem to be the most active agents. Their combination with cyclophosphamide(Drug information on cyclophosphamide) (Cytoxan, Neosar) in CAP, the regimen studied most frequently, has shown response rates of from 22% to 100% and complete response rates of 0% to 40%. However, these responses lasted only 5 to 9 months [18,55].

Chemotherapy for Metastatic or Recurrent Disease

Patients with metastatic or recurrent disease who are undergoing chemotherapy, the standard treatment of diseases at these stages, have a median overall survival of 4 to 6 months. The role of chemotherapy is palliative; its goal is to improve quality of life while reducing disease symptoms. The single agents active in head and neck cancer—with response rates between 15% and 40%—include methotrexate(Drug information on methotrexate), cisplatin, carboplatin (Paraplatin), fluorouracil(Drug information on fluorouracil), ifosfamide(Drug information on ifosfamide) (Ifex), bleomycin, paclitaxel(Drug information on paclitaxel) (Taxol), and docetaxel(Drug information on docetaxel). There is no evidence that methotrexate or cisplatin has a dose-response effect; they are the only agents that have undergone phase III studies regarding this issue. Methotrexate, administered weekly at 40 to 60 mg/m², has been the standard chemotherapeutic approach because of low cost, low toxicity, and easy administration for outpatients. Attempts to improve methotrexate's therapeutic index, including the use of high-dose methotrexate with leucovorin rescue, sequential methotrexate–fluorouracil, and analog development, have been unsuccessful. A phase III, head-to-head comparison of methotrexate and the methotrexate analog edatrexate showed the analog to produce similar response and survival results but increased toxicity [56].

Paclitaxel and docetaxel are new agents active in head and neck neoplasms. In a multicenter Eastern Cooperative Oncology Group (ECOG) phase II trial, 28 patients had a 40% response rate to paclitaxel with a 9.2-month median survival, which exceeds the 15% to 30% response rate and 5- to 6-month survival observed with other agents used in this setting [57]. Paclitaxel was also found to have radiosensitizing activity and is under study with concomitant radiation [58]. Docetaxel, studied in a group of previously treated and untreated patients [59], showed a 32% response rate. Studies of these drugs in combination with other agents are currently under way. Two new promising agents being ivestigated in patients with head and neck cancer are topotecan(Drug information on topotecan) and gemcitabine(Drug information on gemcitabine) (Gemzar).

Randomized trials in patients with recurrent disease have failed to demonstrate the superiority of combination therapy over single-agent therapy in terms of overall survival, despite the combinations' higher overall and complete response rates. Cisplatin plus infusional fluorouracil, the combination most frequently studied, has produced overall response rates ranging from 11% to 79% and complete response rates ranging from 0% to 27% [60]. A randomized study in which this combination was compared with either cisplatin or infusional fluorouracil alone showed that with the combined drugs, the overall response rate, and time to disease progression were better. No significant difference in median survival was found, but of the patients treated with the combination, 40% were alive after 9 months of follow-up, compared with 24% of patients treated with cisplatin and 27% of patients treated with fluorouracil [61].

Another study of combined treatment with cisplatin and fluorouracil showed a significantly higher response rate but similar survival rate compared with results using methotrexate alone [62]. At present, treatment with methotrexate, cisplatin, carboplatin(Drug information on carboplatin), or infusional fluorouracil as single agents or with the combination of cisplatin and infusional fluorouracil may represent acceptable palliative therapy. In choosing the most appropriate therapeutic option for a patient with recurrent or metastatic disease, a careful risk and benefit assessment should be made to determine what level of treatment will improve the patient's quality of life and not affect his remaining time adversely.

Combined Modality Therapy for Local or Regional Advanced Disease

In patients with untreated locally advanced disease, combination chemotherapy achieves a higher response rate (70% to 90%) compared with the rate of less than 40% in previously treated patients [63]. By adding chemotherapy to the standard surgery and/or irradiation regimen, we pursue two major objectives: to decrease the morbidity associated with standard therapy, minimizing the need for radical resection (ie, organ preservation), and to increase overall survival. The three approaches to the use of primary chemotherapy are neoadjuvant chemotherapy, which is designed to reduce the patient's tumor burden before definitive local and regional treatment begins; adjuvant chemotherapy, which attempts to eradicate microscopic disease left after primary treatment; and concomitant chemoradiotherapy, which attempts to enhance the cytotoxic effect of radiation on resectable and unresectable tumor while increasing local control and decreasing systemic disease.

Neoadjuvant Chemotherapy

Neoadjuvant and adjuvant chemotherapy have been studied for 20 years. In the initial induction trials of patients with head and neck cancer, single-agent therapy consisted of methotrexate, cisplatin, or bleomycin. Patients receiving these drugs had response rates of between 10% and 40% and complete response rates of 0% to 5%. These early efforts evolved into trials of drug combinations, which yielded a significant increase in response rates. The most thoroughly studied combination active in locally advanced disease was cisplatin plus fluorouracil, with an overall response rate of approximately 80% and a complete response rate of 30% to 40%, but a 5-year survival rate of less than 25% [64].

The Head and Neck Contract Program [65] conducted the first major randomized trial of the neoadjuvant approach. It included three arms: induction chemotherapy followed by standard local therapy; induction chemotherapy followed by standard local therapy and then adjuvant chemotherapy; and standard local therapy alone. Only one cycle of induction chemotherapy (cisplatin-bleomycin) was administered. Overall survival and disease-free survival did not differ significantly among the three arms, but a significant decrease in distant metastases was seen in the group who received adjuvant chemotherapy.

In a recent randomized study of patients with operable and inoperable disease, the effects of four cycles of neoadjuvant chemotherapy followed by radiation therapy were compared with the effects of radiotherapy alone [66]. The only significant finding in the overall analysis was a decrease in the rate of distant metastases in patients in the chemotherapy-radiotherapy arm. However, a subset analysis of patients with inoperable disease revealed a significant improvement at 3 years in disease-free survival (34% vs 26%), time to distant metastasis (24% vs 42%) and overall survival (24% vs 10%) in the patients who had received neoadjuvant chemotherapy. A larger randomized trial involving only patients with inoperable disease is needed to support these results.

Chemotherapy for Organ Preservation: Although advances in surgical techniques have had some influence on morbidity, primary chemotherapy is being investigated as an alternative to the surgical component of current standard therapy, especially in the larynx, to preserve organ function. Several pilot trials using induction chemotherapy have suggested that surgery could be omitted without compromising survival in patients who respond to chemotherapy [67-69]. These organ preservation trials culminated in 1991 with the report of the Veteran's Administration Laryngeal Cancer Study Group [70,71]. In this trial, induction chemotherapy with cisplatin and infusional fluorouracil plus sequential definitive radiotherapy were compared with treatment with standard surgery (total laryngectomy) and postoperative radiotherapy. Patients who achieved at least a partial response after two cycles of cisplatin and fluorouracil entered a third cycle, which was followed by definitive radiotherapy, then by direct laryngoscopy and primary tumor-site biopsy to determine pathologic response. In these patients, surgery was reserved for later salvage of persis-tent or recurrent disease. Nonresponders to two cycles of induction chemotherapy underwent immediate surgery followed by radiotherapy.

After two cycles of induction chemotherapy, 85% of patients achieved a major response at the primary tumor site (31% complete response). After three cycles of chemotherapy, the overall response rate was 98% (49% complete), and 64% of patients had a pathologically complete response. At a median follow-up of 60 months, estimated 3-year survival rates were 56% (48% to 64%) for the surgery-radiotherapy group and 53% (45% to 61%) for the chemotherapy-radiotherapy group. Survival for the subset of patients with T4 disease was significantly worse (ie, shorter) in the chemotherapy arm than in the surgery arm.

The study also indicated that the short delay in definitive surgery in the nonresponders to chemotherapy was not detrimental (ie, survival in this subgroup was not significantly different from that of the standard arm). In 64% of all patients in the chemotherapy arm, the larynx was preserved. Given these rates of survival and larynx preservation, most patients with locally advanced laryngeal cancer now have the option of preserving their voice without risking shorter survival.

Early phase III results from two other large, multicenter chemotherapy-radiotherapy studies, including a large French study of all cancer sites and a European Organization for Research and Treatment of Cancer (EORTC) trial of patients with hypopharyngeal disease, seemed to confirm the Veterans Administration (VA) trial results and suggested a role for this approach for other locally advanced resectable head and neck cancer sites (Table 4). The French study design was similar to that of the VA trial, except that carboplatin was substituted for cisplatin [72]. The EORTC experimental arm design differed from that of the other two phase III studies in that only patients in complete remission after two cycles of cisplatin-fluorouracil were treated with sequential radiotherapy (70 Gy). Patients with a less than a complete response underwent the standard surgery and radiation [50].

TABLE 4: Phase III Trials of Neoadjuvant Chemotherapy for Organ Preservation in Head and Neck Neoplasms

			Responses
Author	Study arms	Number of patients	Overall	Complete	Overall survival	Organ preservation
VACSP [70] (larynx)	Cisplatin, fluorouracil, RT	166	81%	31%	53% (3 yr)	64%
	Surgery, RT	166	—	—	56%	12%
Martin et al [72] (all sites)	Carboplatin-fluorouracil, LRT	152	56%	31%	49% (4 yr)	48%
	LRT	154	—		38%	20%
EORTC [50] (hypopharynx)	Cisplatin, fluorouracil, RT	100	—		47% (3 yr)	28%
	Surgery, RT	97	—		44%	—
EORTC = European Organization for Research and Treatment of Cancer, LRT = local-regional treatment, RT = radiotherapy, VACSP = Veterans Affairs Cooperative Study Program Adapted, with permission, from Chemotherapy and chemoprevention, in Myers E, Suen JY (eds): Cancer of the Head and Neck, 3rd ed. Philadelphia, WB Saunders, in press.

Nasopharyngeal Carcinoma: Two randomized trials of neoadjuvant chemotherapy for patients with advanced NPC (N2, N3), WHO histopathology, have produced differing results. One was an international trial that included 399 patients and had a median follow-up of 24 months. A regimen of cisplatin, epirubicin(Drug information on epirubicin), and bleomycin, administered for three cycles followed by radiotherapy, was compared with standard radiotherapy alone. Patients so treated achieved increases in 3-year disease-free survival (47% vs 31%, P < .02) and overall survival (P = NS) [73]. In the other randomized trial, neoadjuvant cisplatin-fluorouracil (two cycles), followed by adjuvant therapy (four cycles) with the same agents, was compared to standard radiotherapy. This study included 77 patients and had a median follow-up of 29 months. Two-year overall survival (80% vs 81%) and disease-free survival (68% vs 72%) rates for the experimental arm and standard radiotherapy arm were similar, and there were no differences in patterns of treatment failure [74].

Adjuvant Chemotherapy

To date, seven phase III trials of adjuvant (maintenance) chemotherapy have been reported, but only one small trial (33 patients) has reported a significant increase in overall survival [75]. Another adjuvant trial of cisplatin plus concomitant radiotherapy with positive results is discussed in the section on concomitant chemotherapy [76]. From two recent, large phase III trials come reports of a decrease in the rate of distant metastases in patients in the chemotherapy arm [77,78].

Concomitant Chemoradiotherapy

Simultaneous chemoradiotherapy seems to be the most promising approach to prolong the survival of patients with locally advanced disease. The independent local activity of radiotherapy in combination with the systemic activity of chemotherapy has several advantages over the results achievable with sequential therapy, including the avoidance of a 2- to 3-month delay in initiating definitive local therapy [79].

Among the two different schedules used are the standard radiotherapy schedules to which low- or moderate-dose single-agent chemotherapy is added. Single agents with radiosensitizing effects as shown in preclinical studies are bleomycin, infusional fluorouracil, mitomycin(Drug information on mitomycin) (Mutamycin), hydroxyurea (Hydrea), methotrexate, cisplatin, carboplatin, and paclitaxel [80].

Of nine phase III trials of concomitant bleomycin and radiation, one demonstrated a significant improvement in disease-free survival, and another showed significant improvement in disease-free and overall survivals. Four phase III concomitant trials of fluorouracil [81-83] demonstrated significant increases in disease-free survival (three trials) and in disease-free and overall survival (two trials). Among two phase III trials of cisplatin and radiation, one was a study of primary therapy for locally advanced disease and reported negative results [84], the other was a trial of adjuvant therapy in patients with stage III to IV disease with extracapsular spread, and the findings were positive [76]. Results of the other phase III trials (of methotrexate, hydroxyurea, and mitomycin) were mostly negative. In most of these studies, increased toxic reactions (mucositis) were found in the concomitant-therapy group compared with the radiation alone group.

In the second approach, chemotherapy at full dosage is combined with radiotherapy. Because toxicity is substantially increased with this approach, patients require regularly scheduled interruptions of one or both treatment modalities to allow them to recover from normal tissue toxicity. A variation of the second approach is to alternate chemotherapy and radiation therapy. The detrimental effect of interrupted radiotherapy when used alone has been documented [80]. Despite this reservation, several studies have shown favorable long-term and overall survivals compared with those of standard radiotherapy. Five trials compared combination chemotherapy plus concomitant radiotherapy with induction chemotherapy followed by radiation therapy. All reported increased disease-free or overall survival in the concomitant-treatment group (Table 5)[85-89].

In four phase III trials, concomitant therapy was compared with radiotherapy alone, with improved survival in the concomitant arms found in three (statistically significant in two) trials (Table 6) [90-93]. From a recent study in which patients with advanced head and neck cancer were randomized to receive cisplatin and fluorouracil alternating with radiotherapy or radiotherapy alone, the investigators reported a rate of complete remissions in 53% of the concomitant-therapy group compared with a rate of 26% in the radiation-alone group. Overall survivals at 3 years in the two groups were 41% in the combined-therapy group and 23% in the radiotherapy group (P < .05) [92].

Chemoprevention

Chemoprevention is defined as a pharmacologic strategy to block or reverse carcinogenesis before the development of invasive cancer [94]. This approach has been studied in patients with premalignant lesions as well as patients in the adjuvant setting to prevent second primary tumors. In patients who are cured by radiation, surgery, or both, second primary tumors remain a major threat to long-term survival [95,96]. For this group, chemoprevention may be important in the long-term control of neoplasia of the upper aerodigestive airway.

Retinoids, including natural vitamin A and synthetic analogs, are the most thoroughly studied compounds for chemoprevention. In vitro, retinoids modulate tumor-cell differentiation, proliferation, and apoptosis and also affect other important functions, such as cell adhesion and invasion [97]. Retinoid activity in reversing premalignant lesions has been established in several clinical trials. Hong et al studied isotretinoin(Drug information on isotretinoin) (13-cis-retinoic acid [13cRA]) at high doses in 44 patients with oral leukoplakia. Major responses occurred in 67% of patients treated with isotretinoin compared with 10% of patients in the placebo group. Within 3 months of cessation of therapy, however, the lesions recurred in more than 50% of the patients, and they experienced substantial dose-related side effects; the major effects were cheilitis, conjunctivitis, and hypertriglyceridemia [98]. These results indicated the need for a less toxic dose and a longer period of treatment with isotretinoin.

In a subsequent randomized study of patients whose lesions responded or remained stable after high-dose isotretinoin, participants received maintenance therapy with either low-dose isotretinoin or beta-carotene. Only 8% of the lesions in the isotretinoin group progressed, compared with 55% in the beta-carotene group (P < .001). Low-dose isotretinoin was relatively well tolerated [99].

Another retinoid (fenretinide) has been studied for maintenance in an adjuvant setting, with results similar to those for isotretinoin [100]. Two other randomized trials of retinoid induction documented significant activity in reversing oral premalignant lesions [101]. Furthermore, beta-carotene and alpha-tocopherol (vitamin E) have been reported to have activity in patients in several nonrandomized trials [103,104].

In view of the suboptimal ability of adjuvant therapy to prevent primary disease recurrence and lower the risk of second cancer development, Hong et al designed an adjuvant chemoprevention trial of high-dose 13cRA [105]. In this phase III study, 103 patients with stage I to IV head and neck cancer were randomly assigned to receive either placebo or 13cRA for 1 year. Patients were eligible for the study after definitive local therapy with surgery and/or radiotherapy. At a median follow-up of 32 months, only two of the 51 patients in the 13cRA arm had developed a second primary tumor, compared with 12 of the 49 patients in the placebo group (P = .005). A recent update of this trial, at a median follow-up of 55 months, reported that 14% of patients who received 13cRA developed a second primary tumor, compared to 31% of the placebo group (P = .04) [106].

In a recent placebo-controlled trial in France, the synthetic retinoid etretinate was tested in the adjuvant setting. The study included patients with stage I to III squamous-cell cancer of the oral cavity and/or oropharynx after definitive local therapy. At a median follow-up of 24 months, the occurrence of second primary lesions and primary recurrence rates were similar in the two arms [107].

Conclusions

Although relatively uncommon in the United States, head and neck cancer continues to be a devastating disease with high mortality and morbidity. Management of this tumor requires a multidisciplinary approach, with surgery and radiotherapy being the mainstays. New combined-modality approaches with chemotherapy are starting to show promising results in overall survival and quality of life, especially in patients with locally advanced disease. Because organ preservation, particularly in the larynx, has shown encouraging results, other areas are being studied. Molecular biology has evolved, enabling a better understanding of tumorigenesis of the head and neck. Chemoprevention has been shown to reduce the incidence of second primary tumors and to reverse premalignant lesions; continuing trials will define its role in the management of head and neck neoplasia. Knowing that this disease is, for the most part, preventable, we must continue to support research efforts to decrease the rate of tobacco use.