Harry Quon
SUMMARY OF KEY POINTS
|
Incidence
Epidemiology and Risk Factors
Pathology and Tumor Biology
Clinical Findings
Differential Diagnosis
Staging
Primary Therapy, Salvage Therapy
Complications
Prognosis
Future Directions
|
INTRODUCTION
In the United States for the year 2003, cancers of the head and neck accounted for 2.8% of all new cancer cases and 2% of all cancer deaths.[1] Because these tumors are relatively uncommon, misdiagnosis along with patient neglect not uncommonly contributes to an advanced stage at presentation and limited survival.[2] Despite their relatively low numbers in clinical practice, research and management of head and neck cancers continue to receive significant emphasis owing to the rich anatomic and functional complexity of this body site, which is critical to issues of self-esteem, communication, and social integration. Although squamous cell carcinomas constitute a majority of adult histopathologic patterns (95%)[3] that may be seen in head and neck cancers, the variation in histopathologic types and the differing features of the possible anatomic subsites of involvement result in tremendous discordant variability in the natural course of the diseases for this small anatomic region.
The site begins at the base of the skull and extends only to the clavicles and includes the base of the skull, temporal bone (external auditory canal, middle and inner ear), paranasal sinuses, nasopharynx, oropharynx (soft palate, tonsil structures, base of tongue, oropharyngeal wall), larynx, hypopharynx (pyriform sinuses, postcricoid, posterior pharyngeal wall), oral cavity (lips, buccal mucosa, alveolar ridges, floor of mouth, oral tongue, retromolar trigone), major and minor salivary glands, skin, and the neck.
Accordingly, management of head and neck cancers has evolved to require a multidisciplinary approach with effective integration of various skills and treatment modalities to achieve the desired goals of cure and functional organ preservation. Despite improvements in the diagnosis and local-regional management of head and neck cancer, few significant increases in long-term survival rates over the past 30 years have been realized. The desire for more effective yet organ-preserving therapies with acceptable toxicity is currently being addressed with recent technological advancements in surgical and radiotherapy techniques and the development of novel biologic agents. These agents are particularly attractive because they have been demonstrated in cell model systems to target various aberrant molecular proteins that appear to have dominant roles in mediating biologic aggressiveness or therapeutic resistance. These advances, coupled with tremendous clinical research activity, have resulted in a number of therapeutic options for patients with head and neck cancer.
EPIDEMIOLOGY
An estimated 37,000 new cases of head and neck squamous cell carcinoma (HNSCC) were diagnosed in the United States in 2003, with 68% being diagnosed in males, accounting for approximately 11,000 cancer deaths, with 71% in males.[1] Additionally, cancers of the oral cavity and oropharynx represent 3% of all malignancies in men and 2% in women in the United States.[4] Worldwide, 15% of male cancers, 600,000 cases annually, are HNSCCs. Head and neck cancers affect both sexes and all races, but a preponderance continues to be seen in males, African Americans, and Asians.[4] Oral and pharyngeal cancers have decreased significantly in white males over the past 20 years and in white females younger than 65 years of age.[5] Conversely, the incidence and mortality rates have significantly increased in African-American males but have decreased in black women. The incidence of laryngeal cancer appears to mimic that of lung cancer; a small decline in the incidence in white men younger than 65 has been noted, but the incidence continues to increase in white and black males older than 65 and in women in all age groups.[6]
Limited knowledge exists concerning the epidemiology of in situ head and neck cancers (lip, oral cavity, larynx, and pharynx). Data extrapolated from population-based cancer registries in the National Cancer Institute's (NCI) Surveillance, Epidemiology, and End Results (SEER) program revealed that annual age-adjusted incidence rates for in situ HNSCC rose from 6.33 per 1 million person-years in 1976 to 8.04 per 1 million in 1995 (a 35% change).[7] The predominant anatomic sites of change were the larynx and lip, with this significant rise thought to be attributable to improved surveillance.
Of significant concern are the rising trends of head and neck malignancies in the pediatric population and younger adults. Again, according to data from SEER, the average annual rate in children younger than 15 years of age rose 35% from 1.10 to 1.49 between 1973 to 1975 and 1994 to 1996, respectively.[8] This incidence reflects a greater increase than for childhood cancer in general, which increased 25% during that same period. In a selected population at the M.D. Anderson Cancer Center, the percentage of adults younger than 40 years with oral tongue squamous cell carcinomas rose from 4% to 18% between 1971 and 1993.[9] This finding has been substantiated by SEER data indicating that the incidence of tongue cancer (oral cavity and oropharynx) increased by 60% in adults younger than 40.[10]The prognosis tended to be worse in black than in white males in the same age category. Trends in laryngeal cancers, documented most accurately in the Veterans Affairs (VA) system increased as well between 1983 and 1993 but more notably among those older than 65 years.[11]
Successful therapy for HNSCC may be limited by the recognized risk of second primary malignancies in the aerodigestive tract. Using data from a population-based cancer registry, investigators from the United Kingdom observed 5.5% of males and 3.6% of females to have developed a second primary cancer after an initial diagnosis of head and neck cancer.[12] They also noted a significantly increased risk for a second cancer in most of the upper aerodigestive tract sites commonly associated with tobacco exposure, with a standardized incidence ratio for subsequent oral cancer of 5.56 in men and 15.31 in women. Excluding these tobacco-associated sites resulted in a nonsignificant risk for a second malignancy. Patients with a first detected pharyngeal cancer experienced the highest incidence of a second malignancy. The relative risk for multiple primary cancers was higher in younger patients and among patients who received radiotherapy for their first primary malignancy. These investigators estimated that within 20 years of diagnosis of the first primary HNSCC, subsequent primary malignancies would develop in 30% of males and 20% of females.
ETIOLOGY AND PATHOGENESIS
Tobacco and alcohol continue to remain the two major risk factors for HNSCC in developed countries, with their carcinogenic risks summarized in several working group reports by the International Agency for Research on Cancer (IARC). [13] [14] It is estimated that 75% to 90% of all head and neck carcinomas are attributable to tobacco consumption, particularly cigarette smoking. [15] [16] The cumulative evidence easily fulfills the criteria for causality between cigarette smoking and the development of HNSCC. [17] [18] Supporting a causal relationship has been the demonstration that the risk of developing HNSCC rises with increasing numbers of cigarettes smoked per day and increasing years’ duration of the habit. [3] [19] Current smokers have an approximate 20-fold higher risk of oropharyngeal and laryngeal cancers than that in lifelong nonsmokers. [19] [20] The relative risk of developing head and neck cancer in the heaviest smokers is quoted as 20 to 40 times that of nonsmokers. [19] [20] [21] Even light or occasional cigarette smoking leads to an increased risk of cancer. [3] [19] [21] In addition to the duration of smoking and number of cigarettes, the type of cigarettes and the age, sex, and race of the smoker also influence the relative risk. By contrast, it has been estimated that the relative risk of developing head and neck cancer for heavy consumers of alcohol is two- to sixfold.[22]
Approxiamtely 300 known carcinogens are present in tobacco, with tobacco-specific N-nitrosamines (TSNAs) being the most harmful.[23] These substances are metabolites of nicotine, the major alkaloid responsible for addiction to tobacco.[17] TSNAs are known to bind to DNA and to cause mutations that can activate proto-oncogenes or inactivate tumor suppressor genes.[24] Other harmful mutagenic compounds found in tobacco include polycyclic aromatic hydrocarbons (PAHs), carbon monoxide, and hydrogen cyanide.[23]
Despite limited data, an association between cigar and pipe smoking and an increased risk for HNSCC also is recognized. Cigar consumption in the United States has increased substantially since 1993. Cigars are known to contain even higher concentrations of TSNAs than in cigarettes.[25] With regard to environmental smoke, compared with cigarettes, cigars emit 20 times the carbon monoxide and twice the PAHs, owing in large part to their greater size.[25] The increased risk of HNSCC associated with cigar smoking is between 1.9 and 10.3 times that in nonsmokers. [15] [25] [26] [27] [28] Smoking more than 4 cigars per day increases the risk more than 20-fold.[29] A recent study analyzed prospectively the rates of cancer deaths among cigar-smoking men. Men who never smoked cigarettes or pipes were excluded. Risk of death from oral cavity or pharynx and larynx cancers was 4.0 (95% confidence interval [CI], 1.5 to 10.3) and 10.3 (95% CI, 2.6 to 41.0), respectively.[26] Although fewer dedicated studies have been performed, pipe smoking has been similarly implicated. [15] [30] [31]
Active cigarette smoking is not the only form of tobacco exposure that poses a risk. Exposure to secondhand, or environmental, smoke also has been implicated in several publications as a risk factor for the development of head and neck cancer. [32] [33] [34] A recent study of more than 300 patients and control subjects found that secondhand smoke increased the risk of HNSCC, with a confirmed dose-response pattern. Persons who were exposed to the highest levels of environmental cigarette smoke were up to four times more susceptible to the development of HNSCC.[34]
Tobacco is consumed in a smokeless form in many cultures around the world. In the West, the most common form of smokeless tobacco is termed snuff, whereas in parts of Asia, it often takes the form of betel quid. Betel quids consist of a betel pepper leaf wrapped around a mixture of areca nut, slaked lime, and tobacco; the slaked lime releases an alkaloid that causes a sense of euphoria in the user. Approximately 200 million persons throughout the western Pacific basin and south Asia regularly chew betel quid.[35] Habitual use of betel quid can lead to a progressive scarlike formation known as oral submucous fibrosis. The data regarding an increased risk of oral cancer in persons who chew betel are incontrover-tible, with an estimated odds ratio of 17 for the development of HNSCC. [35] [36] [37] [38] In Europe and the Americas, a large variation has been found in the amounts of TSNA among the various brands and forms of smokeless tobacco available.[39] Therefore, a discrepancy exists between estimated cancer risks associated with smokeless tobacco consumption in Europe and in the Americas. [18] [39] Recent studies of Swedish moist snuff users showed no increase in cancer incidence over that in control subjects, [18] [40] which may be due to the lower concentrations of TSNA in Swedish snuff.[39] The link between other forms of smokeless tobacco and oral cancer is well established, however.[41] [42] [43] A recent review of the known literature found relative risks ranging between 0.6 and 13, depending on the type of smokeless tobacco studied.[42]
Alcohol consumption also is a known independent risk factor for HNSCC. [19] [44] [45] [46] [47] For those individuals that drink heavily but do not smoke, the increased relative risk reportedly ranges from 5.0 to 11.6, increasing in significance with higher numbers of drinks consumed. [45] [46] When alcohol and tobacco are consumed together, the risk increases multiplicatively, rather than additively. [15] [19] [48]Extremely high odds ratios have recently been quoted for heavy consumers of both tobacco and alcohol. [19] [44] Franceschi and colleagues reported an odds ratio of 228 for oral cancer in consumers of more than 25 cigarettes per day and more than 11 drinks per day.[44] One other report noted an odds ratio of 177 for laryngeal cancer in heavy tobacco and alcohol users.[19]
The mechanism for alcohol-induced carcinogenesis is not fully understood. Pure ethanol has been shown not to be carcinogenic.[49] Alcohol, however, often is regarded more as a cocarcinogen, facilitating carcinogenesis rather than initiating it.[50] Several mechanisms for this effect have been proposed. It has been suggested that alcohol may increase the penetration of carcinogens across the oral mucosa,[51]as well as causing mucosal atrophy, which may result in an enhanced susceptibility to chemical carcinogens.[52] It also has been suggested that alcohol may have an effect on DNA repair mechanisms.[53]Other suggested mechanisms include nutritional deficiencies associated with heavy drinking, the effects of contaminants and congeners in alcoholic beverages, and the induction of microsomal enzymes that enhance the metabolic activation of tobacco or other carcinogens.[3] These mechanisms also may explain the observation of an increased risk of cancer with heavy alcohol consumption in nonsmokers. Several reports suggest that women may be more susceptible than men to alcohol-induced carcinogenesis in the head and neck. [46] [54] [55]
Several studies have implicated the lack of vitamin and fresh fruit intake as a risk factor for head and neck cancer. [56] [57] [58] [59] [60] [61] It often is difficult, however, to separate malnutrition from other confounding variables, including alcohol consumption and cigarette smoking. Related to malnutrition is the problem of poor oral hygiene, which also has been found to be associated with HNSCC. [62] [63] [64]
It has been proposed that patients with HNSCC have increased chromosomal sensitivity to carcinogen exposure that predisposes them to developing cancer. [65] [66] [67] [68] In vitro laboratory studies have shown that cells from these patients suffer more chromosome breaks on exposure to a mutagen than are seen in normal control cells. [67] [69] In addition, mucosa from young patients with HNSCC who were nonsmokers was particularly mutagen sensitive.[70]
Immunosuppression may predispose individuals to an increased risk of HNSCC. The risk of carcinomas of the lip is particularly increased in renal transplant recipients. [71] [72] [73] Cutaneous malignancies including lip cancers appear to be increased in cardiothoracic transplant recipients.[74] Other sites of malignancies may include the oral cavity, although the risk does not appear to be as significantly increased as it is with cutaneous malignancies including the lip in transplant recipients. An increased risk of oral cavity carcinomas has been reported in human-immunodeficiency virus (HIV)-infected patients.[75] These tumors generally are more aggressive, with decreased patient survival [74] [76] and a demonstrated association between the degree of medical immunosuppression, most notably involving prednisone, and advanced-stage presentation and an adverse survival.[76]
A potential causal relationship is believed to exist between viral infections and carcinomas of the head and neck. Human papillomaviruses (HPVs) have been associated with a risk for oral cavity [77] [78] and oropharyngeal carcinomas. [78] [79] Of interest, these carcinomas may carry a better prognosis [80] [81] and may respond better to therapy such as radiotherapy.[79] A nested case-control study suggested that the risk may be with the HPV-16 serotype, with 50% and 14% of oropharyngeal and oral tongue carcinomas, respectively, containing HPV-16 DNA. In a large retrospective study, 90% of HPV-positive tumors were of the HPV-16 subtype. Two oncoproteins encoded by HPVs, E6 and E7, are known to inactivate p53 and the tumor suppressor protein for retinoblastoma, presenting potential mechanisms of action. [82] [83] The EBV is a human herpesvirus that has been implicated in a number of human malignancies, including nasopharyngeal carcinoma (NPC). A consistent association between EBV and less differentiated types of NPC has been reported. [84] [85] [86] Although the extent to which EBV infection may contribute to NPC carcinogenesis has not been completely elucidated, an early role is supported by evidence of clonal EBV infection in preinvasive lesions (carcinoma in situ).[87] The expression of various viral gene products, EBV nuclear antigen (EBNA), and latent membrane proteins (LMP1, LMP2) has been demonstrated to have the capacity to induce transformation in vitro, consistent with a carcinogenic role for EBV infection.
PATHOLOGY
Squamous Cell Carcinoma
Oral Cavity
Squamous cell carcinoma accounts for 95% of all malignant tumors in the oral cavity. Other malignancies involving the oral cavity include malignant salivary gland lesions, mucosal melanoma, lymphoma, and sarcoma. Although squamous cell carcinoma can occur anywhere in the oral cavity, the most common locations include the floor of the mouth, tongue, and hard palate. In the earliest recognizable stage, squamous cell carcinoma appears as firm, pearly plaques or as irregular, roughened, or verrucous areas of mucosal thickening, which can be mistaken for leukoplakia, a premalignant lesion. Larger lesions seldom are mistaken for leukoplakia and form either exophytic masses ( Fig. 72-1A ) or endophytic lesions, often with associated ulceration and heaped-up edges (see Fig. 72-1B ).
|
Figure 72-1 A, Squamous cell carcinoma growing as an exophytic lesion in the floor of the mouth. B, Squamous cell carcinoma growing as an ulcerated endophytic lesion along the alveolar ridge of the mandible. C, Invasive squamous cell carcinoma and adjacent precursor lesion of the tongue. D, Squamous cell carcinoma, moderately differentiated, of the tongue. |
Histologic examination frequently reveals an association with in situ lesions, sometimes with surrounding areas of epithelial dysplasia of varying degrees (see Fig. 72-1C ). Histologically, these tumors range from well-differentiated keratinizing neoplasms to poorly differentiated squamous cell carcinoma without keratinization to anaplastic and sometimes sarcomatoid growth patterns (see Fig. 72-1D ). Less common patterns include verrucous carcinoma.
Larynx
Epithelial changes in the larynx include a spectrum of pathologic changes beginning with hyperplasia and progressing through dysplasia, carcinoma in situ, and invasive carcinoma.[88] As in the oral cavity, these early pathologic lesions appear as white (leukoplakia) or reddened (erythroplakia) thickenings and cannot be reliably distinguished from early invasive carcinoma. The earliest lesions (hyperplasia) have little or no malignant transformation potential, whereas mild dysplasia progresses to carcinoma in 1% to 2% of cases over a 10-year period and rises to 5% to 10% with high-grade dysplasia over a similar time span. In general, the more severe the dysplasia, the greater the risk of progression to carcinoma.
Approximately 95% of laryngeal carcinomas are the typical squamous cell carcinoma. Squamous cell carcinoma may develop on the vocal cords, but it also may develop in a supraglottic or subglottic location. Similar to oral squamous cell carcinomas, these lesions begin as in situ carcinoma and, left untreated, grow into infiltrating, ulcer ated, and fungating lesions ( Fig. 72-2A–C ). The degree of differentiation of squamous carcinoma is highly variable and similar to that in tumors of the oral cavity. Rare cases show sarcomatoid differentiation.
|
Figure 72-2 A, Squamous cell carcinoma of the larynx, supraglottic opened posteriorly. B, Squamous cell carcinoma of the larynx invading between thyroid and cricoid cartilages. C, Coronal section of larynx showing a supraglottic squamous cell carcinoma. |
Nasal Sinus and Nasopharynx
Epithelial carcinomas in the nasal sinus are grouped under the entity of NPC. Within this clustered entity are three distinct histopathologically recognizable tumor patterns: (1) keratinizing squamous cell carcinoma; (2) nonkeratinizing squamous cell carcinoma; and (3) undifferentiated carcinoma, also referred to as the lymphoepithelial type of NPC. The nonkeratinizing and undifferentiated carcinoma is most closely associated with EBV infection.
Keratinizing and nonkeratinizing squamous cell carcinoma in the nasopharynx is morphologically similar to lesions found in the larynx and oral cavity. The nonkeratinizing forms also may take the appearance of transitional epithelium—hence the designation of some nasal carcinomas as transitional type. The undifferentiated form of NPC, by contrast, shows a unique morphology characterized by the growth of large tumor cells with round to oval nuclei, prominent nucleoli, fine vesicular chromatin, and indistinct cell borders (which produces a syncytial appearance). An abundant lymphoid response is seen within this tumor, sometimes obscuring the malignant epithelial cells. EBV genome frequently is found within the epithelial tumor cells ( Fig. 72-3A–C ).
|
Figure 72-3 A, Nasopharyngeal carcinoma, undifferentiated type, lymphoepithelial type (hematoxylin-eosin stain). B, Nasopharyngeal carcinoma, undifferentiated type, lymphoepithelial type (cytokeratin stain). C, Nasopharyngeal carcinoma, undifferentiated type, lymphoepithelial type (Epstein-Barr virus in situ hybridization). D, Basaloid squamous cell carcinoma of nasopharynx showing in situ squamous cell carcinoma. |
A rare and highly aggressive variant of squamous cell carcinoma, the so-called basaloid squamous cell carcinoma, can be found in the upper aerodigestive tract and nasopharynx and less commonly in other sites within the head and neck. [89] [90] This tumor is characterized by cells infiltrating in small to large nests with prominent central comedo-type necrosis. At the edges of the nests, the tumor cells form an organized palisade. In addition, variable areas of the tumor show malignant squamous cell morphology with keratinization. Cytologically, the tumor cells show marked nuclear pleomorphism with single cell necrosis and high mitotic rates (see Fig. 72-3D ).
A rare and highly aggressive carcinoma found within the nasal sinus is the sinonasal undifferentiated carcinoma (SNUC) believed to originate from the schneiderian mucosa.[91] These lesions may manifest with or without neuroendocrine differentiation but without evidence of squamous or glandular differentiation. SNUCs are distinguished from undifferentiated nasopharyngeal carcinoma by the absence of a surrounding lymphoid reaction. SNUC cells are medium in size and grow in nests. Cytologically, the tumor cells show a fine chromatin pattern, hyperchromasia, and variably sized nucleoli. Mitotic rates are high, and necrosis is common.
Malignant Salivary Gland Neoplasms
Salivary gland tumors constitute a rare and interesting heterogeneous group of tumors. The vast majority of salivary gland tumors are benign and develop in the parotid gland. A subset of salivary gland tumors is malignant, however. Malignant salivary gland tumors are more common in the minor salivary glands (50% to 60%) and sublingual salivary glands (80% to 90%) in contrast to the parotid (20% to 30%) and submandibular (30% to 40%). A brief discussion of the major types of malignant salivary gland tumors follows.
Mucoepidermoid Carcinoma
Mucoepidermoid carcinomas are the most frequent type of malignant salivary gland tumor. Although occurring most often in the parotid gland, they also can be seen with high frequency in the minor salivary glands. These tumors, as their name implies, are composed of a mixture of squamous cells (epidermoid component) and mucus-secreting cells (mucoid component). A third cell type found within these lesions is the intermediate cell.
Mucoepidermoid carcinomas vary in size and lack a well-defined capsule. Microscopically, the tumor shows a pushing or infiltrative border. They vary in appearance from white to gray and frequently contain small to microscopic mucinous cysts ( Fig. 72-4A ). Microscopically, the tumor is composed of cells arranged in cords, nests, and sheets with varying amounts of squamous, mucous, and intermediate cells. These tumors can vary from well differentiated to highly aggressive, poorly differentiated histopathology. Various grading schemes to account for these differences cluster tumors into low, intermediate, and high-grade categories based on the amount of mucinous cells, solid squamous nests, mitotic rate, necrosis, and pleomorphism. [92] [93] These tumor grades also correlate with patient outcome. Low-grade tumors (see Fig. 72-4B ) rarely metastasize, often are locally infiltrative, and recur in 10% to 15% of cases, and therefore have an excellent 5-year survival rate of greater than 90%. By contrast, high-grade tumors (see Fig. 72-4C ) are highly infiltrative, recur in 30% to 40% of cases, and have metastasized in 30% to 40% of cases at presentation, producing 5-year survival rate of only 50%.[94] [95] [96] Grading appears to have less prognostic importance when the submandibular glands are involved. Involvement of this site appears to be associated with an increased risk of distant relapse regardless of grade. [96] [97]
|
Figure 72-4 A, Mucoepidermoid carcinoma of the minor salivary gland eroding mandibular bone. B, Mucoepidermoid carcinoma, low grade of the palate showing microcystic architecture. C, Mucoepidermoid carcinoma, high grade showing solid growth and high mitotic rate (other areas of the tumor showed focal mucinous pattern). D, Minor salivary gland with adenoid cystic carcinoma growing in a typical cribriform pattern, with many lumina showing eosinophilic basement membrane material. E, Parotid gland with adenoid cystic carcinoma growing in cribriform plates and cords with perineural invasion. F, Parotid gland with carcinoma ex pleomorphic adenoma. G, Parotid gland with carcinoma ex pleomorphic adenoma showing adenocarcinoma not otherwise specified, with comedo-type necrosis. |
Adenocarcinoma, Not Otherwise Specified
These tumors are now restricted to lesions that do not show histopathologic patterns of the other salivary gland tumor types. These tumors can occur in the parotid gland as well as minor salivary glands and typically present in the sixth to eighth decade of life. At presentation, varying histologic patterns including tubular, papillary (with varying degrees of differentiation from well to poorly differentiated), and adenocarcinoma may be observed. Important in the workup for these patients is to exclude an adenocarcinoma from another body site because salivary gland adenocarcinomas can mimic adenocarcinomas from other organ sites.
Adenoid Cystic Carcinoma
Adenoid cystic carcinomas account for some 20% of malignant salivary gland tumors. They are frequently located in the minor salivary glands (40% to 50% of cases) and less often in the parotid gland (20% to 30%). The most common histologic pattern is the classic cribriform type or “Swiss cheese” pattern, characterized by neoplastic cells forming oval or circular spaces or nests. Within these nests is dark, eosinophilic, hyaline-like basement membrane material. The amount of the hyaline material can distort the cell nests and produce a small acinar- or cordlike appearance to the tumor nests (see Fig. 72-4D ). Perineural invasion is almost invariably observed in these tumors (see Fig. 72-4E ) and a diagnosis of adenoid cystic carcinoma without finding perineural invasion should be carefully reconsidered. Adenoid cystic carcinomas are graded into low, intermediate and high-grade tumors based on the amount of solid growth, mitoses, necrosis, and pleomorphism[98]; as with mucoepidermoid carcinoma, the degree of differentiation is correlated with long-term survival.
Acinic Cell Carcinoma
More than 90% of acinic cell carcinomas arise in the parotid gland. These tumors occur between the ages of 50 to 70, with men affected more than women (2:1 male-to-femael ratio). The gross appearance of these tumors varies, but they often are whitetan in color and present as a well-circumscribed but unencapsulated mass. Microscopically, the most common histologic pattern consists of solid sheets of cells with low-grade cytology and granular basophilic cytoplasm similar to the serous acinar cells of a normal salivary gland. Less common variants include a cystic and microcystic pattern, in which the tumor cells are arranged at the edges of the cyst, and cells with granular basophilic cytoplasm are arranged among cells, with a tombstone appearance and capitation-type secretion, as well as cells with a bubbly vacuolated appearance. Rare cases of acinic cell carcinoma have been reported to undergo dedifferentiation into a high-grade aggressive malignancy[99]; most acinic cell carcinomas, however, behave in a more indolent fashion. [100] [101]
Malignant Mixed Tumor (Carcinoma ex Pleomorphic Adenoma)
Malignant mixed tumors of the salivary gland are predominantly tumors that started as pleomorphic adenomas in which the epithelial tumor component subsequently underwent malignant transformation. This transformation may be confined to the adenoma, resulting in the noninvasive carcinoma ex pleomorphic adenoma.[102] A majority of carcinomas that arise out of pleomorphic adenomas, however, are infiltrative and aggressive lesions (see Fig. 72-4F and G ). The transformation to a malignant phenotype is heralded by a change in the behavior of the patient's underlying pleomorphic adenoma characterized by sudden enlargement of an otherwise stable adenoma, or by the onset of pain in a previously asymptomatic adenoma. Morphologically, the malignant transformation can take on any malignant epithelial salivary tumor phenotype, the most common being adenocarcinoma not otherwise specified.[103] Occasionally, multiple recognizable histologic patterns may be present. These tumors are highly aggressive, with 5-year survival rates as low as 25%. Another form of the malignant mixed tumor is the true carcinosarcoma, which again often arises from a pleomorphic adenoma background. These tumors have both malignant epithelial and malignant stromal elements.
Polymorphous Low-Grade Carcinoma
Polymorphous low-grade adenocarcinoma (PLGA) is a rare low-grade carcinoma of minor salivary glands. These tumors grow as infiltrative masses and are composed of uniformly low-grade tumor cells arranged in a variety of histologic patterns, including tubular, solid, papillary, microcystic, cribriform (with true lumina), pseudoadenoid cystic (without true lumina), fascicular, single file, and cordlike.[104]Perineural invasion is common. In contrast with adenoid cystic carcinoma, PLGA stains positively for EMA, which may be relied on to distinguish between these two pathologic lesions. Rare cases of PLGA have metastasized, and these cases have been associated with more than focal areas of papillary growth. Rare cases have dedifferentiated into high-grade aggressive carcinomas.[105]
Salivary Duct Carcinoma
Salivary duct carcinomas are rare tumors with a male predilection. These are rapidly growing tumors that can produce pain and nerve palsies. Morphologically, these tumors resemble comedo-type duct carcinoma of the breast and also may grow in a cribriform or papillary pattern. Cytologically, these tumors tend to be high grade, with nuclear pleomorphism, single-cell necrosis, and high mitotic rate. Clinically, these tumors behave in an aggressive fashion, with poor long-term survival. [106] [107]
Lymphoid Tumors of Head and Neck
Extranodal Lymphoma
Extranodal lymphomas of the head and neck can take on the same morphologic spectrum as that seen anywhere in the body. Large cell lymphoma predominates, but any morphologic pattern can be seen. Although many of these lymphomas arise in association with lymphoid tissue within Waldeyer's ring, they are not limited to this location.
Angiocentric Lymphoma
Formerly called by a variety of terms (polymorphic reticulosis, lethal midline granulomatosis), this lymphoid malignancy manifests as an aggressive destructive lesion characterized by ulceration and tissue necrosis. The tumor cells are visible, growing in an angiocentric location. Immunophenotypically, these tumors often are either NK or T cells, with rare B cell phenotype reported. [108] [109] [110] It is important to distinguish this entity from nonmalignant sinonasal processes, such as Wegener's granulomatosis, that also can produce an ulcerating and necrotizing lesion.
Plasmacytoma
Extramedullary plasmacytomas occur in the area of Waldeyer's ring of lymphoid tissue. These tumors represent clonal plasma cell tumors. Some 30% of patients with plasmacytomas will eventually develop multiple myeloma if followed for 20 years. Cytologically, the plasma cells can vary from well-differentiated plasma cells to very atypical-appearing plasmablasts. Occasionally, plasmacytomas with increased plasmablasts are misdiagnosed as poorly differentiated carcinoma, unless the pathologist considers plasmacytoma in the differential diagnosis and orders the appropriate immunohistochemical studies.
Olfactory Neuroblastoma
Olfactory neuroblastomas or esthesioneuroblastomas are rare tumors that arise from the olfactory neuroepithelium. This epithelium can be found along the roof of the nasal cavity to the midportion of the nasal septum and onto the superior turbinate. These tumors appear to be morphologically similar to neuroblastomas that arise in the adrenal gland. They are characterized by proliferation of small round blue cells arranged in nests and surrounded by a vascular network in a loose connective tissue stroma. Rosette formations can be seen as well as neurofibrillary material in the center of the rosette and within the connective tissue matrix. These tumors stain for neuroendocrine markers including S100, NSE, chromogranin, and Leu-7, which aids in distinguishing them from poorly differentiated carcinoma. [111] [112]
Melanoma
In addition to cutaneous melanomas that can occur anywhere on the head and neck, melanomas may also involve the mucosal surface of the oral cavity or nasal sinus. These tumors often manifest in the fifth or sixth decade of life. Mucosal melanomas often appear histologically similar to their cutaneous counterparts and may display a similar range of architectural and cytologic variability to that exhibited by cutaneous melanomas. One interesting exception is that some mucosal melanomas display a small cell phenotype that on biopsy can easily be mistaken for a nonkeratinizing squamous cell carcinoma (Fig. 72-5A and B ).
|
Figure 72-5 A, Sinus with primary mucosal melanoma showing infiltrative destructive growth by large epithelioid cells beneath respiratory mucosa. B, Sinus with primary mucosal melanoma showing infiltrative destructive growth by large epithelioid cells beneath respiratory mucosa. Cells exhibit a blastic appearance, with prominent single-cell necrosis and high mitotic rate. |
Sarcomas
Angiosarcoma
Angiosarcomas may occur in the head and neck and most commonly manifest on the face and scalp, usually in the sixth or seventh decade of life. Angiosarcomas of the oral or nasal cavity are exceedingly rare. Morphologically, these tumors show anastomosing vascular channels, with varying degrees of nuclear atypia in the malignant endothelial cells. Those with better differentiation are associated with a better prognosis.
Chondrosarcoma
Chondrosarcomas may arise in any bone within the head and neck as well as within the larynx. They are typically seen in the sixth or later decade of life. As with chondrosarcomas occurring in other areas, malignancy is defined by infiltrative, permeative growth as nuclear crowing and atypia within cartilage lacunae.
Osteosarcoma
Osteosarcomas may occur in any bone of the head and neck. This lesion often is seen in younger patients, with a peak in the third decade of life. In addition, osteosarcomas have been seen after radiation therapy or in association with Paget's disease. [113] [114] Morphologically, these tumors are characterized by infiltrative and destructive growth of surrounding bone and soft tissue.
Chordoma
Chordomas are low-grade tumors derived from notochord remnants. The two most common locations for the tumors are at the base of the skull (spheno-occiput) and sacrum. In the head and neck area, these tumors may involve the clivus, sphenoid, upper nasopharynx, occipital bone, maxilla, ethmoid, pterygoids, or cervical vertebrae.[115] These tumors grow as lobulated, infiltrative lesions that destroy adjacent normal structure. Microscopically, a myxoid matrix predominates with epithelioid cells growing singly or in small nests. These epithelioid cells are granular or vacuolated (physalipherous cells) and show staining for both keratin and S100 on immunohistochemistry studies, helping to differentiate them from low-grade chondrosarcomas, which are keratin negative.
TUMOR BIOLOGY
In the simplest terms, cancer formation can be regarded as the escape of cells from the normal regulatory mechanisms of the cell cycle. Consequently, the cells proliferate without regard to their surroundings. Their progeny often become less differentiated, and they become able to invade surrounding tissues and spread to distant sites.
The genetic alterations associated with these changes are myriad. With regard to head and neck squamous cell carcinoma, much of the seminal work has elucidated the potential role of the tumor suppressor gene in the pathogenesis of cancer. These genes are expressed at appropriate levels in normal tissue and have a negative effect on cellular growth. If one or more of these genes is inactivated through mutation, chromosomal deletion, or DNA methylation, however, the loss of negative regulatory control may lead to unregulated growth and subsequently to cancer. The tumor suppressor genes encoding p53, p16, and p19ARF are the major tumor suppressor genes involved in HNSCC pathogenesis. The p53 gene lies on the short arm of chromosome 17 (17p), and p53 is a negative regulator of the cell cycle by means of the cyclin/cyclin-dependent kinase complex. The p53 gene frequently is mutated in HNSCC, with 40% to 70% of tumors harboring mutations. [116] [117] The proteins p16 and p19ARF also are cell cycle inhibitors, and these genes also frequently are inactivated in HNSCC.[118] The mechanism of inactivation, however, includes not only gene mutation but also chromosomal deletion and promoter methylation.[119] Approximately 70% to 80% of cancers have allelic loss of the chromosomal region 9p, where p16 resides. [120] [121] The loss of these cell cycle inhibitors is thought to contribute to progression from normal mucosa to preneoplastic lesions to invasive cancer.[122]
Certain genes are overexpressed in HNSCC as well. These oncogenes include those encoding cyclin D1, C-myc, and epidermal growth factor (EGF), and their increased expression also can contribute to the pathogenesis of HNSCC. [123] [124] [125] More recently, attention has turned to cell adhesion receptors such as integrins and E-cadherin with regard to tumor invasion and recurrence. [126] [127] E-cadherin has been shown to be highly underexpressed in oral tongue carcinomas, and weak expression of the molecule was significantly correlated with a higher recurrence rate and worse 5-year survival.[127] Integrin expression is closely tied to the invasiveness of HNSCC cell lines.[126]
Theoretically, many of these genetic alterations should result in the expression of altered proteins that the host immune system would recognize as foreign. The cells would then be identified and targeted for immune-mediated destruction. It is clear, however, that tumor cells are able to escape from the host immune system through a variety of mechanisms.[128] A significant proportion of tumor cells have decreased expression of human leukocyte antigen (HLA) class I, a molecule necessary for the presentation of foreign peptides to the immune system.[129] Furthermore, T lymphocytes that infiltrate thetumor are not as competent at killing tumor cells, presumably due to factors within the tumor microenvironment.[128]
New technologies will soon enable researchers to analyze further the genetic and immunological changes that occur in HNSCC. Complementary DNA microarrays allow for large-scale analysis of gene expression. Researchers have already identified hundreds of genes that are either overexpressed or underexpressed in HNSCC. [130] [131] Future studies will elucidate the relationship of these genes to the pathogenesis of cancer, and novel tumor markers will be identified against which new chemotherapeutic and immunotherapeutic modalities can be targeted.
CLINICAL PRESENTATION AND PATIENT EVALUATION
Patients with head and neck cancers present with certain symptoms according to the location of the tumor. Some signs and symptoms and patient characteristics are common to many of these patients regardless of the site of tumor origin, however. Patients with head and neck cancer are mostly male and in their sixth decade of life or older. Most have a history of tobacco and alcohol use. They often present with cachexia due to dysphagia, tumor burden, or malnutrition. Rapid weight loss is sometimes reported. Hemoptysis also is frequently found. Pain, either primary or referred to the ipsilateral ear, also is often encountered. Many tumors manifest with cervical metastasis as the first sign of disease, and a thorough examination must be performed to identify the primary site.
More specifically, signs and symptoms associated with oral cavity tumors include bleeding, dysphagia, dysarthria, and halitosis. The patient may complain of ill-fitting dentures that previously fit properly. Oropharyngeal tumors cause similar signs and symptoms. In addition, pain often is experienced at the site within the oropharynx or is referred to the ear. A neck mass is often the presenting sign of an oropharyngeal tumor. Supraglottic tumors can manifest with dyspnea, dysphagia, and voice change. Glottic and subglottic tumors are likely to cause hoarseness and dyspnea. Dysphagia and ear pain often is experienced by patients with hypopharyngeal cancer.
Initial Head and Neck Examination
An initial examination of the head and neck should be performed with the patient sitting upright in a chair. A standard and complete head and neck examination should then be conducted. All 12 cranial nerves are assessed. Otoscopy and anterior rhinoscopy also are indicated. Examination of the oral cavity is preferably done using a head light to allow bimanual examination of the lips, buccal mucosa, gingiva, floor of the mouth, and oral tongue. Palpation of the tongue and the base of the tongue often is forgotten but is crucial for a thorough evaluation. The tongue base can be further examined using a mirror and a flexible fiberoptic laryngoscope. The larynx also is visualized using these instruments. While viewing the larynx fiberoptically, the physician should ask the patient to perform several maneuvers: vocalization to allow assessment of vocal cord motion; tongue protrusion to aid in viewing the vallecula; and filling the cheeks with air to help visualize the pyriform sinuses. After the subglottis is evaluated, the laryngoscope can be removed and the examination completed with palpation of the neck for masses. The size, mobility, and consistency of any mass should be carefully noted.
Staging Investigations
Radiography is a necessary component of the evaluation. It is now standard practice to order a CT scan of the head and neck for any patient with suspected cancer. Not only does this examination help to evaluate the size and location of the primary, but also any possible metastasis. An MRI scan can provide useful additional information in certain cases, such as in previously treated patients or in those lesions in which skull base involvement must be ruled out. Routine preoperative laboratory values should be obtained, as well as appropriate medical consultations as indicated.
If accessible in the clinic, biopsy of any suspicious mucosal lesion should be undertaken after local administration of anesthetic. Generally, oral cavity lesions and selected oropharyngeal lesions are easily biopsied in the clinic. For most lesions of the tongue base, larynx, and hypopharynx, direct laryngoscopy with biopsy under anesthesia is required. During this procedure, flexible esophagoscopy should be undertaken to rule out tumor spread into the cervical esophagus. In addition, a tracheostomy or gastrostomy tube should be placed at this time if indicated. If no primary tumor is identified, but a firm node 1 cm or greater is present, the node should be sampled by means of fine-needle aspiration (FNA). An open biopsy of a neck node in an adult suspected of head and neck cancer should be undertaken only after FNA diagnosis is inconclusive and a primary site cannot be identified.
Based on this extensive clinical and radiologic evaluation, the cancer is assigned a stage. Generally, T1, T2, and T3 represent increasing tumor size, whereas T4 is defined by invasion of a surrounding structure (skin, nerve, vessel, cartilage). The node, or N, stage is identical for all head and neck cancer sites and is defined in Table 72-1 . The absence or presence of distant metastases is defined as M0 or M1. The T, N, and M stages are combined into overall groupings that are presented in Table 72-1 . Because the natural history of head and neck cancer varies somewhat according to specific anatomic location of the primary disease and since stages III and IV include a large number of different T and N stages, it is customary to refer to specific head and neck cancers by their individual T, N, and M stage and the primary site.
Table 72-1 -- Overall Group Staging of Head and Neck Cancer: Tumor-Node-Metastasis (TNM) Classification
|
Stage |
Grouping |
|
|
|
0 |
Tis |
N0 |
M0 |
|
I |
T1 |
N0 |
M0 |
|
II |
T2 |
N0 |
M0 |
|
III |
T3 |
N0 |
M0 |
|
T1 |
N1 |
M0 |
|
|
T2 |
N1 |
M0 |
|
|
T3 |
N1 |
M0 |
|
|
IVA |
T4a |
N0 |
M0 |
|
T4a |
N1 |
M0 |
|
|
T1 |
N2 |
M0 |
|
|
T2 |
N2 |
M0 |
|
|
T3 |
N2 |
M0 |
|
|
T4a |
N2 |
M0 |
|
|
IVB |
T4b |
Any N |
M0 |
|
Any T |
N3 |
M0 |
|
|
IVC |
Any T |
Any N |
M1 |
From American Joint Committee on Cancer: AJCC Cancer Staging Manual, 6th ed. New York, Springer, 2002, p 65.
Follow-up Program
After completion of initial treatment, patients are seen every 4 to 6 weeks for 2 years in the office; during this period a complete set of laboratory data and chest radiographs are obtained every 6 months to look for metastatic disease or second malignancies. It is important to monitor levels of thyroid-stimulating hormone, because many patients who have received therapeutic irradiation to the neck will become hypothyroid and will require thyroid hormone supplementation. A baseline post-treatment CT scan (or MRI after skull base procedures) is obtained approximately 2 months after the end of postoperative radiation therapy as a reference point for comparison in the event that disease recurs.
During the third year, the patient is seen every 3 months, and every 4 months during the fourth year. After 5 years, the patient should be examined once per year.
PROGNOSIS
Prognostic and treatment-predictive factors play a central role in the treatment decision-making process, because the current therapeutic approach uses a risk stratification paradigm. This approach recognizes that currently accepted treatment modalities are associated with significant risk and a spectrum of toxicities. Identification of these factors facilitates this risk assessment and helps to guide decisions on whether to initiate therapy and also may influence what type of therapy to administer. An ideal prognostic factor would provide information about the biologic behavior of a tumor, permitting the prediction of the outcome and response to therapy. The use of such factors has become confusing, however, owing to the spectrum of factors reported, often with conflicting results. This dilemma has led to efforts to systematically classify head and neck prognostic factors based on the level of significance and reliability, including a recent meta-analysis.[132]
This heterogeneity results from several sources, including statistically underpowered studies, absence of statistical modeling for independent prognostic effects, variability in the nature of the therapy applied, the composition of prognostic factors as represented in the study group, and the inherent biologic heterogeneity of cancer. The heterogeneity is compounded further by the heterogeneity associated with the tests and instruments that often are used to measure aspects of the tumor biology. Although it may be said that clinical prognostication often is imprecise, several clinical factors have an established role. In the head and neck, tumor site is a strong prognostic factor for survival, regardless of the treatment modality, and reflects the precision with which it can be identified by the clinician, although exceptions may exist, particularly for large tumors that overlap anatomic subsites. [133] [134] [135] Within each subsite, the current American Joint Committee on Cancer (AJCC) tumor-node-metastasis (TNM > staging criteria have been demonstrated to have prognostic significance and continue to evolve through efforts to stratify further the staging classification based on prognosis.[136] Most problematic are patients who constitute the heterogeneous locally advanced AJCC stage IV group.
The sixth edition of the TNM classification introduces a distinction of the heterogeneous T4 disease group based on the probability of disease control, with T4a disease representing a reasonable probability and T4b disease representing such extensive disease that an adverse outcome is certain. In essence, the watershed between T4a and T4b reflects for most subsites the transition to unresectable disease that has been a recognized prognostic factor. Criteria for T4 subclassification by each anatomic subsite have been developed.[137] Consideration of this subclassification scheme in the context of N and M staging facilitates recognition of three distinct prognostic groups: advanced lower-risk stage IVA (potentially curable); advanced high-risk stage IVB (of dubious curability); and stage IVC with distant metastatic disease (undoubtedly incurable).[137] The critical distinction is the definition of IVB as N3 disease or T4b with any N category (excluding the nasopharynx subsite), for which aggressive local-regional therapy may be undertaken but with a potentially low expectation of success—a philosophical view that has been termed “aggressive palliation.” In this subgroup, other patient-based prognostic factors, such as performance status, comorbid conditions, and the patient's ability to tolerate and comply with local-regional aggressive therapy, may have significant effect on the anticipated outcome and hence treatment decisions. Although the committee acknowledges that comparative outcome data do not exist for these new stage IV groupings, these groupings are likely to be surrogate groupings for tumor volume and resectability that have established prognostic effects and form the basis for the current therapeutic paradigm.
Within each subsite, various tumor features, noteworthy for their prognostic impact and not necessarily reflected in the new TNM staging criteria, warrant additional comments. In the oral cavity, tumor thickness or depth of invasion has been well recognized to influence the risk of nodal metastases [138] [139] and to reflect biologic aggressive disease with an adverse survival. [138] [139] Po and colleagues recently demonstrated that tumor thickness in oral tongue carcinomas was the only significant factor that had significant predictive value for subclinical nodal metastasis, local recurrence, and survival in multivariate analysis. With the use of 3-mm and 9-mm division, tumor of up to 3 in mm thickness is associated with an 8% risk of subclinical nodal metastasis, zero local recurrence, and 100% 5-year actuarial disease-free survival; tumor thickness of more than 3 mm and up to 9 mm carried a risk of 44% subclinical nodal metastasis, 7% local recurrence, and 76% 5-year actuarial disease-free survival; tumor of more than 9 mm carried a 53% risk of subclinical nodal metastasis, 24% local recurrence, and 66% 5-year actuarial disease-free survival.[139] With disease of the nasopharynx, the staging classification reflects the prognostic importance of soft tissue extension beyond the nasopharynx, particularly the presence and the degree of extension into the parapharyngeal space.[140] In this regard, MRI is superior to CT imaging for such prognostication. For well-differentiated thyroid carcinomas, young age (younger than 45 years) is an important favorable prognostic factor such that patients with other favorable prognostic factors including small tumors (1 to 1.5 cm) that are well encapsulated and confined to the thyroid, may be monitored by observation, with no adjuvant therapy recommended.[141]Anaplastic thyroid carcinomas carry an adverse prognosis irrespective of the local disease extension at the time of presentation.
For patients with HNSCC treated with radiotherapy, CT-based tumor volume assessment [142] [143] [144] [145] [146] and the presence of anemia [144] [147] [148] [149] [150] [151] appear to be prognostic for local-regional control and survival. Of interest, in a series of 258 patients with T1 to T4 glottic carcinoma treated with surgery only, anemia and positive surgical margins were independently associated with an adverse 5-year local-regional control rate.[151] Protracted overall treatment duration for a course of radiotherapy is well established as an adverse prognostic factor for local-regional control. [152] [153] [154] [155] [156] Limited data suggest that this adverse prognostic factor also may be important in patients receiving concurrent chemoradiotherapy, when the risk of toxicity and treatment interruptions is increased.[157] For early-stage glottic larynx carcinomas treated with radiotherapy, a dose per fraction less than 2 Gy appears to be associated with poor local control rates in various univariate analyses, but this has not been conclusively demonstrated to be independent of other time-dose parameters. [156] [158] [159] Nevertheless, current consensus recommendations are for dose per fraction of 2 to 2.25 Gy.[160]Tumor response to radiotherapy has been demonstrated in a large prospective series of 228 patients with HNSCC treated uniformly with conventionally fractionated radiotherapy to be associated with local control. Multivariate analysis showed that the probability of local relapse was significantly and independently increased for minor regression at 5 weeks (less than 75%) (relative risk, 2.3), for nonlaryngeal tumors (relative risk, 2.4) and for T3-T4 disease (relative risk, 2.4).[134]
In patients requiring postoperative radiotherapy who demonstrate high-risk adverse prognostic features, the time interval between surgery and radiotherapy and the overall treatment time from the time of surgery to completion of the radiotherapy appear to be important.[161] Investigators from the M.D. Anderson Cancer Center have previously demonstrated that the most adverse prognostic postoperative risk factor is the presence of extracapsular extension.[162] These investigators have proposed and prospectively validated a risk stratification criterion that identifies patients as high-risk if extracapsular extension is present or two or more pathologic risk factors are present. These include oral cavity primary, mucosal margins close or positive, nerve invasion, more than one positive lymph node, more than one positive nodal group, the largest node greater than 3 cm, and treatment delay longer than 6 weeks. [161] [162] Cooper and colleagues noted similar risk factors in a review of the Radiation Therapy Oncology Group (RTOG) database.[163] In the high-risk group of patients, progressively protracted treatment time was associated with worsening local-regional control rates (P = 0.005) and overall survival (P = 0.027), with the most favorable outcome noted when the overall treatment time was less than 11 weeks.[161] Other retrospective studies have reported similar observations. [164] [165]
The heterogeneity and imprecision of clinical prognostic factors have encouraged recent efforts to evaluate the potential prognostic significance of various molecular factors[166] and various quantitative measures describing functional tumor imaging such as magnetic resonance spectroscopy. [167] [168] Grandis and colleagues reported a strong independent prognostic value for the epidermal growth factor receptor (EGFR) and its ligand, transforming growth factor-α (TGF-α), in a mature but heterogeneous group of 91 patients with a mixture of stages and sites treated with surgery and adjuvant therapy. Of the 91 patients, 56 (62%) received postoperative external beam radiation therapy (EBRT) and 16 (18%) received adjuvant chemotherapy. Protein expression was quantitated by computer immunohistochemistry image analysis on paraffin-embedded specimens. Increasing levels of overexpression of either EGFR or TGF-α were associated with an increasing adverse disease-free and cause-specific survival. Other investigators using various quantitative assays of this surface receptor also have demonstrated an independent prognostic effect of EGFR overexpression. [169] [170] [171] [172] These series are noteworthy for the strong prognostic significance observed in multivariate analysis, possibly owing to the reduced test heterogeneity with such quantitative assays and the biologic significance of EGFR overexpression.
PRIMARY TREATMENT AND TREATMENT COMPLICATIONS
General Principles
Effective management of head and neck cancers requires comprehensive consideration of several often-competing treatment goals. This routinely requires the efficient integration of various treatment modalities and supportive services for appropriate patient care. Accordingly, representation from disciplines including head and neck surgery, reconstructive surgery, radiation oncology, medical oncology, pathology, neuroradiology, dentistry, oral and maxillofacial surgery, nutrition, nursing, rehabilitation medicine, social services, and psycho-oncology is routinely required.
Before the start of any therapy, it is important not only to evaluate issues of histologic diagnosis and the anatomic extent of disease but also to review issues that may have an impact on treatment compliance. This assessment should include a review of the level of social support a patient has, with appropriate referrals to support services as indicated. Similarly, effective attention to nutritional support and pain management can significantly improve patient compliance to any subsequent therapeutic plan. Depending on the treatment modalities that are required, pretreatment evaluation by oral and maxillofacial services may be indicated. If radiotherapy is indicated, prophylactic enteral tube placement for nutritional support may be appropriate, especially if concurrent chemoradiotherapy will be given. Whenever possible, review of a patient's case within the context of a multidisciplinary tumor conference is strongly advocated.
On establishment of a histologic diagnosis of cancer in the head and neck, subsequent treatment decisions follow from a hierarchy of considerations. Because treatment for head and neck cancer is associated with significant risk and spectrum of toxicities, it becomes important to identify patients with a poor prognosis, for whom treatment may be appropriately tailored. Determining a poor prognosis can be complex and imprecise, influenced not only by the anticipated clinical outcome but also by various patient factors including treatment tolerance and toxicity. In large part, the imprecision of prognostication results from the clinical heterogeneity of established prognostic and treatment predictive factors, as already discussed. When a poor prognosis is uncertain, it may be appropriate to adopt a curative intent, but the potentially low expectations of treatment success should be discussed with the patient.
In tailoring treatment to the prognosis, it is the hope that patients with a poor prognosis will only be subjected to treatment toxicities that are appropriate for the goals of any palliative treatment. Hence, palliative (or curative) intent should be clearly distinguished from palliative (or curative) treatment goals. Unfortunately, palliative treatment goals for advanced head and neck carcinomas (HNSCCs) often require achieving some degree of local-regional disease control but with constraints to minimize toxicity. Although various palliative radiotherapy schedules have been reported, in general, it may be more appropriate to consider the role of systemic chemotherapy to minimize the adverse quality of life that results from radiotherapy-induced xerostomia and taste alteration. In appropriate circumstances, surgical intervention including the use of high-dose rate brachytherapy implants may achieve treatment goals with the most favorable therapeutic ratio. When and how best to achieve such palliative treatment goals become issues of judgment that follow from engaging a thoughtful dialogue between various members of the multidisciplinary team and the patient and family members.
In treating patients thought to be appropriate candidates for potentially curative therapy, a higher threshold for unacceptable toxicity on the part of the treating physician and patient is implicitly accepted. As a result, an evidence-based approach in deciding between various treatment options is preferred to justify the increased tolerance of toxicity. This has favored a systematic categorization of the quality of clinical reports reflecting the study methodology employed and the confidence that the results reported are free of biases and random variability. In general, randomized trials with sufficient statistical power provide the greatest unbiased level of evidence and confidence that the results are reproducible. Not uncommonly, treatment effects often are smaller than anticipated, leading to variable results among similar randomized trials, which limits consensus treatment recommendations. For head and neck carcinomas, competing comorbid conditions and the risk of second malignancies contribute to underpowered trials in demonstrating improvements in patient survival. As a result, prospective and retrospective comparative analyses of different treatment options often are relied on to provide significant treatment guidance. These studies are, however, subject to potential biases from patient selection and the generalizability of the reported results may be limited by the nature of the patients selected. Attempts often are made to match groups by stage to ensure their comparability, leaving only the treatment to vary. This strategy may be limited, however, by the effects of stage migration and should be carefully reviewed for this potential influence. For HNSCC, this is particularly important in light of the significance of anatomic tumor extent on treatment prognosis[137] and the recent evolution of various sensitive imaging modalities. Finally, the usefulness of data from institutional series, case series, and reports also may be limited by the size of the sample study population and the absence of comparative analysis of different therapeutic strategies. Such data often provide insight into the management of rare histologic diagnoses, however. When insufficient evidence-based literature exists, consensus treatment principles have evolved to further guide treatment recommendations. Alternative systematic quantitative tools such as decision tree analysis techniques and cost-effectiveness analysis have not been well studied and remain limited in their impact on the management of head and neck cancers.
For HNSCC, therapeutic principles traditionally have considered competing treatment goals of local-regional control and the level of risk and spectrum of toxicity in defining the concept of a therapeutic ratio. Both quantitative and qualitative issues in toxicity must be considered, because a severe but low-probability toxicity may be important in defining what is an acceptable therapeutic ratio for the patient. This concept of a therapeutic ratio is important, because a dose-response relationship exists for both disease control and toxicity in several treatment modalities, including chemotherapy and radiotherapy, and may even be extended conceptually to surgery. Although treatment toxicity is an important consideration in evaluating treatment options, the emphasis remains on local-regional control as a principal measure of treatment success. This focus underscores the natural history of the disease and the cosmetic and functional impact of cancer in this body site. Accordingly, definitive therapeutic decisions often involve evaluation of surgery- or radiotherapy-based strategies. Although rigorous randomized comparisons of these two treatment approaches are extremely limited, several generalizations may be noted.
When local-regional control rates appear comparable between competing treatment modalities, such as surgery and radiotherapy, the use of a single modality associated with the most favorable therapeutic ratio is preferred. This principle is most applicable to the use of surgery or radiotherapy for early-stage disease (typically T1-2N0-1). For early-stage disease, local-regional control rates reported from surgical and from radiotherapy institutional series appear to be comparable, with the possible notable exception of the oral tongue subsite, for which anecdotal experience suggests that local control rates with EBRT are inferior. (Rather, the incorporation of a brachytherapy implant appears to provide comparable results to surgery.) Surgery is often preferred for disease involving sites that are readily accessible, such as the oral cavity site, to minimize the risk of complications. By contrast, early-stage NPC typically is treated with radiotherapy owing to the risks associated with surgery in this location.
Radiotherapy has traditionally been considered to be more attractive for early-stage tumors involving sites at which surgical resection may compromise organ function, such as involvement of the oropharynx, larynx, and hypopharynx. Again, this consideration is appropriate when local control rates appear comparable. Evolving experience and sophistication with various organ-preserving surgical techniques, however, offer the potential for good function preservation with no apparent compromise in oncologic results. Examples include various innovative larynx preservation techniques, including the supracricoid laryngectomy. [173] [174] These techniques are influenced by patient and tumor selection, with some techniques being very operator-dependent and thus not necessarily generalizable. When available, these techniques provide a potential single-treatment modality for patients in the early stage of disease that involves critical sites of organ function, such as the larynx, without the issues of radiotherapy-induced late complications. It also allows radiotherapy to be reserved for the management of relapses and second head and neck malignancies. Currently, the relative efficacy with regard to oncologic results and organ function between these surgical options and radiotherapy remains unexplored.
This single-modality principle for treatment of early-stage disease attempts to minimize toxicity and emphasizes the importance of patient and tumor selection. When postoperative radiotherapy appears likely, definitive radiotherapy may be more appropriate, particularly for small volume disease, as it is unclear that combined-modality therapy is superior to radiotherapy alone. This consideration is particularly relevant when initial definitive en bloc resection may, in combination with postoperative radiotherapy, further complicate treatment toxicity and organ function. Not uncommonly, these issues arise with more extensive early-stage HNSCC or in cases in which the anatomic location limits achieving adequate surgical margins. For these and more advanced lesions, the role of debulking surgery also has been proposed.[175] This has been favored as a strategy for efficient reduction of the number of tumor clonogens without the surgical morbidity of more extensive tissue resection. Altough attractive, the value of this approach remains controversial, because it has not been subjected to critical scientific scrutiny and remains ill defined in the literature.
For more advanced primary lesions (typically defined as T3 and T4 disease), increased treatment-related toxicities generally have been accepted owing to inferior local control rates attained. Accordingly, a combined-modality approach typically has been employed. Traditionally, this approach has included surgery and radiotherapy with either preoperative or postoperative conventionally fractionated radiotherapy. A postoperative protocol typically has been favored because the radiotherapy can be delayed until more accurate delineation of the tumor extent of disease and histopathologic stratification of patients requiring postoperative radiotherapy (PORT).[176] When evaluated in a randomized trial in 277 patients with supraglottic larynx and hypopharynx carcinomas (RTOG 73–03), PORT demonstrated superior mature local-regional control rates although the results were confounded by use of a higher dose delivered in the postoperative setting.[177] Recent randomized trials have permitted optimization of not only selection of patients requiring PORT[161] but the dose required,[162] as well as the identification of prognostic factors adversely influencing the outcome particularly in high-risk patients.[161] Of greater significance, it is now clear that the time from surgery to the start of PORT and the overall treatment time of PORT are important determinants of local-regional control in patients with high-risk features. Because the issue of patients with more extensive disease undergoing higher risk surgical resection may confound these observations, the magnitude of benefit from manipulating these time variables remains to be determined. Nevertheless, it remains prudent to regard these observations, emphasizing the importance for effective coordinated management, which is best achieved in the context of a multidisciplinary team of health care providers. Strategies to improve the results of surgery and PORT have included adjuvant chemotherapy, concurrent postoperative chemoradiotherapy, and altered fractionated radiotherapy schedules such as various accelerated radiotherapy schedules. These strategies remain subjects of ongoing study to further define their role and efficacy in patients with resectable locally advanced HNSCC.
For unresectable advanced disease, conventionally fractionated radiotherapy alone has been traditionally recommended. Results of suboptimal local-regional disease control in the absence of surgical options, however, have favored an acceptance of intensifying treatment with increased treatment toxicities. Of note, however, the increased toxicity associated with this strategy warrants careful patient selection, because the therapeutic ratio is clearly reduced. This has included the use of concurrent chemotherapy with daily-fractionated radiotherapy,[178] altered fractionated radiotherapy alone,[179] and recently developed combinations of concurrent chemotherapy with altered fractionated radiotherapy schedules.[180] In general, modest success has been realized with both daily-fractionated chemoradiotherapy and altered fractionated radiotherapy, and they are regarded as acceptable therapeutic options. For concurrent chemoradiotherapy, determining the optimal agents and schedule remains largely undefined and is the subject of ongoing evaluation. Although a recent updated patient-based meta-analysis has demonstrated a modest survival benefit with concurrent chemoradiotherapy,[181] the generalizability of these results to current popular taxane-based chemoradiotherapy regimens is unclear. [157] [182] Randomized data currently also support the use of either a dose-escalated hyperfractionated schedule or an accelerated schedule reducing the overall treatment time by 1 week with twice daily treatments in the final 2.4 weeks for this group of patients.[179] The relative efficacy and indications for these two treatment strategies also remain largely undefined. Despite the improvements in local-regional control rates observed with the various altered fractionated schedules, the predominant relapse pattern continued to be local-regional (approaching 50%), with distant relapse rates of less than 20%, leading current opinion to favor the use of concurrent chemoradiotherapy for large volume disease manifesting at the primary site or in the neck.[183] The integration of chemotherapy has been particularly favored for bulky neck disease, because the risk of distant relapses is increased with some chemoradiotherapy regimens, demonstrating an impact on this risk.[135]
In recent years, treatment decisions have been progressively influenced by obligations to achieve functional organ preservation for advanced but resectable HNSCC. This treatment goal has been buoyed by the fact that the superiority of altered fractionated radiotherapy schedules[179] and concurrent chemoradiotherapy[178] demonstrated in randomized trials has, in general, been observed in study populations that have been unselected based on the resectability of the disease.[184] This treatment goal has been particularly paramount for disease in the larynx [185] [186] and oropharynx subsites[187] and continues to evolve; the optimal definitions, methodology, and instruments for studying functional organ preservation remain to be defined. Because surgical resection for more advanced disease often results in greater normal tissue extirpation and loss of function, radiotherapy-based management strategies typically have been favored. It is noteworthy, however, that the oncologic efficacy of surgical resection, often followed by PORT, relative to these radiotherapy-based strategies has not been rigorously evaluated. In fact, studies have demonstrated that patients will prioritize length of life if compromises in speech or swallowing function are required.[188] Accordingly, surgical resection remains an appropriate treatment consideration, particularly when organ function already has been compromised by tumor. Invasion of bone or cartilaginous structures by advanced HNSCC has been also considered an indication for surgery on the basis of historic experience of poor responsiveness to radiation therapy, although this is debated.[189] Recent results with treatment intensive intra-arterial chemoradiotherapy strategies have challenged this indication.[190]
In general, two therapeutic paradigms involving radiotherapy have been adopted for organ preservation. A nonselective strategy may be adopted whereby the tumor response serves as an in vivo test for radiosensitivity reserving surgery for salvage.[191] This strategy poses increased risks of difficulties and complications with salvage surgery but increases the number of patients achieving organ preservation. It also is limited by the recognition that not all patients who fail at the primary site after radiotherapy are amenable to surgical salvage. Alternatively, selecting tumors and patients who may have a more favorable probability of local control with radiotherapy minimizes the proportion of patients subjected to increased surgical complication rates, but may subject patients with radiosensitive lesions to unnecessary surgery. A hybrid strategy that continues to have proponents is based on treatment response to either chemotherapy [187] [192] [193] or an intermediate radiotherapy dose, typically between 50 and 55 Gy. [134] [135] Responses to chemotherapy have been argued to predict for biologically favorable and responsive disease to subsequent radiotherapy. It also has been argued that significant radiotherapy responses possess sufficient predictive power, allowing for the selection of patients who are destined to require salvage surgery, in turn limiting the subsequent toxicity from cumulative therapy with higher radiotherapy doses. A large prospective radiotherapy series has demonstrated this to have independent prognostic value for local control.[134]
Early experiences of organ preservation with conventionally fractionated radiotherapy were limited to the larynx, and the results were disappointing. [143] [194] [195] [196] [197] Salvage surgery was possible in 50% to 80% of patients with early-stage disease treated with radiotherapy. [194] [195] [196] Inferior local-regional control rates were reported for bulky disease such that a majority of patients required salvage surgery, with a success rate of 50%, and also were associated with increased surgical complication rates. [191] [197] This gave rise to the use of various neoadjuvant chemoradiotherapy regimens in attempts to achieve organ preservation. The value of the neoadjuvant chemotherapy in contributing to local disease control was questioned, however. In a recently completed randomized trial, it was subsequently demonstrated to have minimal value and the trial indicated that intensive local-regional therapy was a more appropriate strategy.[198] Recent studies have now demonstrated that the use of an altered fractionation schedule (hyperfractionated and accelerated schedules)[179] or the use of concurrent chemotherapy [135] [198] in combination with daily-fractionated radiotherapy offers superior local-regional control rates. This gain in local-regional control rate, however, may still come at the increased risk of surgical complications with salvage surgery. The risk of pharyngocutaneous fistula increased from 15% to 30% in patients randomized to receive radiotherapy alone or concurrent chemoradiotherapy with cisplatin for larynx preservation, respectively.[185] It remains to be demonstrated whether this risk is observed with other treatment-intensive regimens (and for other head and neck disease sites) that aim to improve on local-regional disease control.
Accordingly, patients deciding between initial surgical therapy and organ preservation therapy must balance the issues of immediate organ loss with lower complication rates and a reduced probability of delayed organ extirpation but at an increased risk of surgical complications. To guide these decisions, attention has focused on the probability of organ preservation with radiotherapy-based strategies and the degree of function that will be achieved. To date, radiotherapy predictive factors remain limited to considerations of clinical tumor features such as tumor volume. [142] [143] [146] Although predictive factors of organ function remain in evolution, it is clear that when nonexisting pretreatment function is observed, successful radiotherapy is unlikely to restore function.
In the evolution of organ preservation as a curative treatment goal, several additional concerns have been raised. Initial concerns focused on the potential risk of an adverse outcome due to the potential risk of ongoing tumor metastasis. Early randomized trials comparing surgery and PORT with neoadjuvant chemoradiotherapy have demonstrated that organ preservation can be achieved without immediate surgical extirpation and with no compromise in overall survival. [192] [193] The second trial has centered on treatment toxicity, especially with recent evidence favoring the concurrent integration of chemoradiotherapy for organ preservation. [185] [186] [198]
With concurrent chemoradiotherapy, increased swallowing dysfunction and secondary risks of aspiration, events contrary to the goal of functional organ preservation have been observed. [187] [199] The mechanisms contributing to these events remain to be fully understood but in general may relate to the severity and location of the treatment-induced mucosal edema with subsequent fibrosis and damage to peripheral nerves, all compounded by significant radiotherapy-induced xerostomia. Despite these concerns, preliminary data suggest that patient quality of life may still be improved as a consequence of organ preservation.[200]
Traditionally, the management of the primary tumor site dictated the treatment modality for the neck as a strategy to facilitate efficient management of HNSCC. Treatment decision-making for the neck has adopted a similar therapeutic paradigm of risk stratification to optimize the therapeutic ratio, but the decision-making becomes more complex as a result of the primary treatment considerations. Conceptually, the issue of functional organ preservation also may be extended to the neck, because neck dissection may result in cosmetic changes and compromised neck and shoulder function. Similarly, radiotherapy-induced edema and fibrosis may compromise the goal of functional neck preservation and may be exacerbated by a neck dissection.
Typically, single-modality treatment is favored for early-stage neck disease and combined-modality strategies for advanced neck disease. In the clinically negative neck, when surgical resection has been elected for the primary site management, elective neck dissection may be omitted if preoperative evaluation determines a high risk of requiring PORT and the risk of occult nodal metastasis is sufficiently high to warrant elective management. When radiotherapy has been selected for management of the primary site, neck dissection in the clinically negative neck is not indicated. In fact, preradiotherapy neck dissection may alter the lymphatic flow of the neck, necessitating larger volumes of the neck to be irradiated and requiring surgical wounds to be irradiated to higher doses. It also may contribute to delays in the delivery of radiotherapy, which have been reported to contribute to an adverse overall survival when compared with postradiotherapy neck dissection in a retrospective analysis.[201]
In the clinically positive neck with adverse risk features, combined-modality therapy (surgery and radiotherapy) also has been favored owing to the increased recurrence risk and morbidity associated with regional relapses and the limited salvage options. Postoperative radiotherapy after a neck dissection is indicated in accordance with the presence of adverse pathologic nodal risk factors such as extracapsular extension. Again, delays in the start of PORT may be detrimental, particularly when high-risk nodal factors are present.
When conventionally fractionated radiotherapy alone has been employed for locally advanced HNSCC, suboptimal regional control rates coupled with the morbidity and the limited success of subsequent salvage neck dissection have prompted the incorporation and general acceptance of a planned neck dissection. This approach, generally accepted in the presence of residual adenopathy after radiotherapy, also has been selectively applied to patients with adverse risk factors such as large nodal size (typically 3 cm or greater). This risk stratification follows from radiotherapy series demonstrating an inverse relationship between nodal size and control rate. [202] [203] In a study of radiotherapy alone for treatment of HNSCC in 1251 patients, Dubray and colleagues noted 3-year neck control rates, by maximum nodal size, of 0.5 cm, 77%; 2 cm, 67%; 4 cm, 60%; 6 cm, 52%; 8 cm, 37%; and 10 cm, 7%. Multivariate analysis revealed that regional relapses independently increased with increased nodal size (P = 0.0001), decreasing radiation dose (P = 0.0001), T4 primary disease (P = 0.0001), node fixation (P = 0.02), bilateral neck disease (P = 0.03), and geographic miss (P = 0.0001).[203]
Controversy continues, however, regarding the benefit of a planned neck dissection in the setting of a complete clinical response in the neck, particularly in large pretreatment lymph nodes, after completion of radiotherapy.[204] Treatment with conventionally fractionated radiotherapy has demonstrated that the prognosis of large neck nodes with a complete response is associated with a prognosis comparable with that of smaller nodal metastases, and that the risk of relapse is low. [202] [205] This suggests that perhaps the subgroup of patients with a complete response in the neck may have more radiosensitive disease and may not require a neck dissection. A neck dissection continues to be favored, particularly for advanced neck disease, however, because the likelihood of achieving a complete response with radiotherapy alone is limited. The efficacy of this approach is further supported by several retrospective series demonstrating improved regional control rates, [205] [206] [207] with a possible improvement in survival[208] despite the absence of a randomized trial. Attempts to definitively treat advanced neck disease also increase the risk of subsequent wound complications with any subsequent salvage neck dissection,[209] if that is even possible given that salvage options are almost always limited owing to disease encasing critical structures such as the carotid artery. [210] [211]
This issue has been further compounded by the use of concurrent chemoradiotherapy strategies that have reported improved local-regional control rates. One trial of concurrent hyperfractionated chemoradiotherapy versus hyperfractionated radiotherapy suggested a lower frequency of residual disease in the dissected neck specimen in favor of concurrent chemoradiotherapy (21% for chemoradiotherapy versus 37.5% for radiotherapy).[212] No neck relapses occurred in either arm in cases for which a planned neck dissection was performed. McHam and colleagues retrospectively reported on 109 patients who received concurrent chemoradiotherapy with or without a neck dissection for indications of residual neck disease or as a planned procedure.[211] Residual neck disease was observed in 33% of the neck specimens (25% and 39% of neck specimens with complete clinical responses or partial clinical responses, respectively). Neck relapses were observed in 5 out of 76 (6.6%) and 4 out of 33 (12%) patients receiving and not receiving a neck dissection, respectively. The only factor significantly correlating with neck relapses was the presence of residual disease (5 out of 25 versus 0 out of 51). These findings suggest that both chemoradiotherapy and neck dissection are contributing to high neck control rates, with clinical evaluation of the neck not possessing sufficient predictive power to select for a subgroup of patients for whom a planned neck dissection may be held. The role of PET imaging, which carries a risk of false-negative results, remains to be defined. [213] [214] Although there is no appreciable increased risk of surgical complications with the addition of concurrent chemotherapy, [215] [216] [217] increased neck and mucosal edema, particularly with bilateral neck dissections, may contribute significantly to treatment morbidity.
Management of patients with HNSCC is further complicated by the risk for development of second primary carcinomas and relapses within an aerodigestive tract that may have been extensively exposed to prior treatment. This challenging scenario further emphasizes the single-modality treatment principle when appropriate and favors surgery alone when possible. An irradiated aerodigestive tract not only limits reirradiation but also may preclude an effective surgical salvage, because concerns of residual microscopic disease often exist in cases that would otherwise warrant consideration of postoperative radiotherapy. In general, repeat surgery is associated with fewer normal tissue toxicity constraints but greater immediate functional consequences than are seen with reirradiation. Accordingly, when a second primary or relapse occurs within a previously irradiated field, surgical resection should be the primary treatment option.[218] Experiences with various repeat EBRT strategies, including the integration of chemotherapy, have unfortunately demonstrated limited success, often at the risk of significant toxicities.[218] The exception appears to be NPCs that are more radiosensitive. Even then, significant late complications may arise but may be more accepted in view of the limited surgical options.
Institutional series have demonstrated that higher repeat radiation doses are more likely to be successful; this has favored the incorporation of conformal radiotherapy techniques when possible. Such techniques may include stereotactic radiosurgery or radiotherapy, three-dimensional conformal radiotherapy, IMRT, and brachytherapy techniques. The first option has been used as a preferred strategy with nasopharyngeal cancer in which infiltrative disease extends beyond the limits of an intracavitary brachytherapy implant. The last has been particularly favored when the disease is well defined and for its ability to deliver a biologically effective dose to a limited volume. In fact, limited institutional series have reported on the efficacy of a brachytherapy implant for the base of tongue and tonsil for selected well-defined lesions as an alternative to surgery for lesions that also are suitable for surgical resection. In general, a coordinated brachytherapy implant at the time of complete gross surgical resection of either the second primary relapse or in the dissected neck may facilitate successful reirradiation. When surgical reconstruction and wound closure incorporate unirradiated tissue, the risk of complications from brachytherapy reirradiation may be reduced. This approach is limited by the ability to define the extent of microscopic disease in the head and neck when normal anatomic barriers and lymphatic drainage patterns have been altered from prior therapies. It thus becomes important to apply a brachytherapy implant appropriately, because these results are not necessarily generalizable.
Treatment Modality Considerations
Surgery
GENERAL.
The decision to treat HNSCC with surgical therapy must be undertaken carefully. To assess whether a patient is a surgical candidate, the input of a multidisciplinary team is invaluable. Of the many issues to be considered, among the most important is the medical condition of the patient. Severe cardiac or pulmonary disease, profound malnutrition, and generalized debilitation are relative contraindications to immediate surgical intervention. Therefore, a detailed assessment by an experienced internist or cardiologist is mandatory when a major resection is planned.
The next important consideration is whether the lesion can be removed safely with adequate margins. The resectability of a tumor is assessed via physical examination and radiographic studies. In general, tumor involvement with certain anatomic landmarks, including the base of skull and the prevertebral fascia, renders the lesion unresectable. This is due to the inability to achieve an adequate normal tissue margin in these areas. In addition to tumor location, massive tumor size also can preclude total extirpation. In such cases, tumor-free margins may be impossible to obtain. In addition, adequate reconstruction may be exceedingly difficult, and other forms of therapy should be considered.
Postoperative function also is a primary concern in considering surgical therapy. The ability to speak clearly and swallow effectively is greatly affected by surgery in the head and neck region. Although complete resection of the malignancy is of paramount importance, reconstruction of the surgical defect must be designed so that the patient has an opportunity to regain as much speech and swallowing function as possible. Site-specific considerations are discussed in the following sections.
The cognitive ability of the patient to participate in postoperative rehabilitation should be reviewed as well. Patients who are neurologically or emotionally unable to participate should be identified early in the decision-making process. They may be better served with different forms of therapy. The social situation of the patient also should be considered. Patients who live alone may require admission to a nursing home or skilled nursing facility to ensure adequate postoperative care.
In general, primary surgical therapy is reserved for patients with tumors of the oral cavity, selected early staged larynx cancer, skin cancer, salivary gland tumors, paranasal sinus tumors, and thyroid neoplasms. Patients with very advanced tumors invading bone, destroying cartilage, or extending into the soft tissues of the neck are considered for primary surgical therapy. When patients have severe organ dysfunction as a consequence of cancer infiltration, surgery also should be considered since functional restoration with nonsurgical therapy is unlikely.
NECK DISSECTION.
It is important to understand the different types of neck dissections that can be performed ( Fig. 72-6 ). The radical neck dissection is a procedure wherein the lymph nodes from all five levels of the neck, the sternocleidomastoid muscle, the internal jugular vein, and the spinal accessory nerve are all removed. The specimen is removed en bloc, theoretically so that there is no spillage of tumor and so that there is a complete resection of any metastases. This procedure is indicated only in the setting of massive neck metasases involving most of the levels of the neck as well as the nonlymphatic structures.
|
Figure 72-6 Types of neck dissection. A, Radical. B, Modified radical: One or more of the nonlymphatic structures are preserved. C, Supraomohyoid. D, Lateral. E, Posterolateral. F, Anterior compartment. |
A modified radical neck dissection removes all of the lymph nodes in the neck but spares one or more of the nonlymphatic structures. Three types of modified radical neck dissection are in general use:Type I spares the spinal accessory nerve, type II spares the nerve and the internal jugular, and type III spares the nerve, vein, and sternocleidomastoid muscle. A type III dissection also is termed a functional or Bocca neck dissection, named after the Italian surgeon who pioneered the surgery.[219]
Selective neck dissections do not involve the resection of all five levels of lymph nodes but usually involve three or more according to the site of the primary cancer. These neck dissections are usually perfomed in the setting of the N0 neck. There is some controversy regarding the use of a selective neck dissection versus a modified radical or radical neck dissection when known metastases are present, although a consensus is forming that a selective neck dissection is appropriate for patients with N1 and selected N2 disease. [220] [221] [222]
Radiotherapy
GENERAL.
Historical experiences with EBRT have demonstrated that acute treatment-limiting radiotherapy-induced dermatitis may be limited by fractionation and the use of higher energy radiotherapy. which results in less surface dose. Accordingly, current standard radiotherapy practices have evolved to utilize a fractionated radiotherapy prescription using modern linear acceleration (linac)-model radiotherapy machines that can produce a spectrum of beam energies. With fractionation, issues of patient immobilization and treatment setup reproducibility become important considerations. For the head and neck, several critical normal tissue structures such as the spinal cord and optic chiasm often are in close proximity to the irradiated target. For these reasons, a prerequisite treatment simulation whereby patients are immobilized with various devices including a custom-made face mask and frame, with the setup referenced to a laser light coordinate system in the treatment rooms, is required before treatment can be initiated (Figs. 72-7 and 72-8 [7] [8]).
|
Figure 72-7 Linear accelerator with reference laser lights shown for reproducible patient setup. |
|
Figure 72-8 Patient in mask in the treatment position. A, Supraglottic carcinoma. B, Glottic cancer. (From Liebel S, Phillips TL (eds): Clinical Radiation Oncology, p 503.) |
Various immobilization devices exist, achieving different degrees of immobilization. Immobilization addresses the issue of the precision of the treatment delivery as a strategy to optimize the therapeutic ratio. The most examples of this are recently described stereotactic radiotherapy and radiosurgery techniques that involve immobilization of the patient in a rigid head frame system that may be removed daily for fractionated treatments in the former case or bolted to the cranium for a single large dose of radiation in the latter case.
More sophisticated treatment planning may involve obtaining axial images, typically with a dedicated CT scanner, of the immobilized referenced patient to facilitate non-coplanar three-dimensional beam arrangements or the use of IMRT techniques. The formerutilizes geometric shielding to effect shaping of the radiotherapy beam; the latter utilizes various approaches that result in modulation of the radiotherapy beam fluence to achieve additional degrees of radiotherapy beam conformality. Increasing the conformality provides an additional strategy to optimize the therapeutic ratio, separate from the issue of radiotherapy beam precision. Radiotherapy beam conformality, however, often is dependent on achieving beam precision for it to be successful. This again places emphasis on achieving sufficient reproducible patient immobilization.
Although these new conformal techniques often are advantageous in the sparing of dose to critical head and neck normal structures, it is important to recognize that this benefit comes at the expense of increasing the volume of normal tissue that is exposed to low doses of the entry and exit radiotherapy beams. This issue is particularly emphasized with the IMRT technique, which often may dose many small segments of a radiotherapy beam or exploit the automated delivery process to deliver many more radiotherapy beams to achieve the degree of conformality sought. Concerns have been raised with regard to the potential long-term consequences of exposing larger volumes of normal tissue to low doses of radiation. In particular, a genotoxic and possibly carcinogenic effect has been speculated. These concerns remain to be fully evaluated, but they emphasize the importance of judiciously and appropriately applying these techniques in counseling young patients, particularly those with an anticipated good prognosis.
The success of these techniques is also dependent on the ability to accurately identify anatomic sites that may harbor subclinical disease. Currently, the basis for this determination is derived from surgical and clinical documentation of disease extension that is often unique to each head and neck subsite. Recent functional imaging techniques, such as PET-based studies, remain promising active areas of investigation. As a result, the transition to the use of axial images for radiation treatment planning places a significant emphasis on the knowledge and experience of the radiation oncologist with regard to the natural history of the disease as it pertains to axial anatomy. In particular, the identification of nodal groups that typically are at risk for harboring subclinical nodal metastases may be more problematic. To aid in this regard, several reports have been published delineating axial anatomic structures that may be used to delineate the various nodal groups ( Table 72-2 ).[223] The incidence of nodal metastases has also been summarized and can significantly aid the radiation oncologist ( Table 72-3 ). Of note, however, when prior treatment to the neck has occurred, altered flow of lymphatics is a significant concern. As with the head and neck surgeon, skill and judgment must be exercised with these precise treatment techniques. The alternative may be particularly detrimental.
Table 72-2 -- Recommendation for the Radiologic Boundaries of the Neck Node Levels
|
|
ANATOMIC BOUNDARY |
|||||
|
Level |
Cranial |
Caudal |
Anterior |
Posterior |
Lateral |
Medial |
|
Ia |
Geniohyoid m. |
Platysma m. |
Symphysis menti; platysma m. |
Body of hyoid bone |
Medial edge of anterior belly of digastric m. |
n.a.[*] |
|
Ib |
Mylohyoid m., cranial edge of submandibular gland or caudal edge of medial pterygoid m. |
Platysma m. |
Symphysis menti |
Body of hyoid bone; posterior edge of submandibular gland |
Basilar edge of mandible; platysma m. |
Lateral edge of anterior belly of digastric m. |
|
II |
Bottom edge of the body of CI |
Bottom edge of the body of hyoid bone |
Posterior edge of submandibular gland; posterior edge of posterior belly of digastric m. |
Posterior border of sternocleidomastoid m. |
Medial edge of sternocleidomastoid m. |
Internal edge of internal carotid artery, paraspinal (levator scapulae) m. |
|
III |
Bottom edge of the body of hyoid bone |
Bottom edge of cricoid cartilage |
Posterolateral edge of sternohyoid m. |
Posterior edge of sternocleidomastoid m. |
Medial edge of sternocleidomastoid m. |
Internal edge of carotid artery, paraspinal (scalenius) m. |
|
IV |
Bottom edge of cricoid cartilage |
Cranial border of clavicle |
Posterolateral edge of sternohyoid m. |
Posterior edge of sternocleidomastoid m. |
Medial edge of sternocleidomastoid m. |
Internal edge of internal carotid artery, paraspinal (scalenius) m. |
|
V |
Skull base |
Cranial border of clavicle |
Posterior edge of sternocleidomastoid m. |
Anterior border of trapezius m; scalenius m. |
Platysma m; skin |
Paraspinal (levator scapulae, splenius capitis) m. |
|
VI |
Bottom edge of the body of hyoid bone |
Sternal manubrium |
Skin; platysma m. |
Posterolateral edge of sternohyoid m. |
Medial edge of common carotid artery, skin and anterior-medial edge of sternocleidomastoid m. |
n.a. |
|
Retropharyngeal |
Base of skull |
Cranial edge of the body of hyoid bone |
Levator veli palatini m. |
Prevertebral m. (longus colli, longus capitis) |
Medial edge of internal carotid artery |
Midline |
From Grégoire V, Coche E, Cosnard G, et al: Selection and delineation of lymph node target volumes in head and neck conformal radiotherapy. Proposal for standardizing terminology and procedure based on the surgical experience. Radiother Oncol 2000;56:135–150.
|
* |
Midline structure lying between the medial borders of the anterior belly of the digastric muscle. |
Table 72-3 -- Distribution of Clinical Metastatic Neck Nodes from Head and Neck Squamous Cell Carcinomas [3] [28] [49]
|
|
DISTRIBUTION OF METASTATIC LYMPH NODES PER LEVEL (% OF NODE-POSITIVE PATIENTS) |
||||||
|
Tumor Size |
Patients with N+ (%) |
I |
II |
III |
IV |
V |
Other[*] |
|
Oral cavity (N = 787) |
36 |
42/3.5[†] |
79/8 |
18/3 |
5/1 |
1/0 |
1.4/0.3 |
|
Oropharynx (N = 1479) |
64 |
13/2 |
81/24 |
23/5 |
9/2.5 |
13/3 |
2/1 |
|
Hypopharynx (N = 847) |
70 |
2/0 |
80/13 |
51/4 |
20/3 |
24/2 |
3/1 |
|
Supraglottic larynx (N = 428) |
55 |
2/0 |
71/21 |
48/10 |
18/7 |
15/4 |
2/0 |
|
Nasopharynx (N = 440) |
80 |
9/5 |
71/56 |
36/32 |
22/15 |
32/26 |
15/10 |
From Grégoire V, Coche E, Cosnard G, et al: Selection and delineation of lymph node target volumes in head and neck conformal radiotherapy. Proposal for standardizing terminology and procedure based on the surgical experience. Radiother Oncol 2000;56:135–150.
|
* |
Parotid buccal nodes. |
|
† |
Ipsilateral/contralateral nodes. |
EXTERNAL BEAM RADIOTHERAPY TIME-DOSE-FRACTIONATION CONSIDERATIONS.
Conventional or daily radiotherapy fractionation (typically, daily 1.8- to 2-Gy fractions to a total dose of 70 Gy) has permitted the delivery of higher radiotherapy doses that currently are limited by normal tissue tolerances, such as in the mandible. Conceptually, radiotherapy failures may result from insufficient doses of radiotherapy relative to the number of tumor clonogens or may be due to cellular mechanisms of radioresistance. The former has proved to be more amenable to therapeutic manipulation with studies of various altered fractionation schedules, which may be generalized into two groups: hyperfractionation and accelerated fractionation ( Fig. 72-9 ). Various randomized trials have been conducted and have been recently summarized.[183]
|
Figure 72-9 Schematic illustration of the fractionation regimens investigated by the Radiation Therapy Oncology Group. Each bar represents one radiation fraction. Bars above the lines represent large-field irradiation and those below the lines stand for coned-down boost irradiation. CF, conventional fractionation; 70 Gy in 35 fractions over 7 weeks. HFX, hyperfractionation; 81.6 Gy in 68 fractions over 6.8 weeks. AFX-S, accelerated split-course fractionation; 67.2 Gy in 42 fractions over 6 weeks. AFX-CB, accelerated fractionation with concomitant boost strategy; 72 Gy in 42 fractions over 6 weeks. (From Nguyen LN, Ang KK: Radiotherapy for cancer of the head and neck: altered fractionation regimens. Lancet Oncol 2002;3:693–701.) |
A hyperfractionation schedule, or the use of lower doses per treatment fraction, has been hypothesized to reduce the risk of late radiotherapy-induced complications associated with an increase in the total radiotherapy dose. Typically, the dose per fraction is reduced to 1.15 to 1.2 Gy and exploits a differential radiosensitivity between normal late-responding tissues and most cancers, including HNSCC. To ensure that the overall treatment time is not adversely protracted, fractions often are delivered twice a day, with an interfraction period of 6 hours.
Several randomized trials support the concept of hyperfractionation as a strategy to increase the biologically effective dose delivered without a significant increase in the risk of late radiotherapy-induced complications [179] [224] [225] [226] ( Table 72-4 ). In general, improved local-regional control rates were observed, with improved survival rates, in two trials. [225] [226] The largest trial (RTOG 90–03) of radiotherapy fractionation schedules randomized over 1000 patients with locally advanced AJCC stage III and IV HNSCC (greater than 60% stage IV) for all subsites except nasopharynx (stage II disease permitted for base of tongue and hypopharynx) to one of four fractionation schedules including hyperfractionation to a dose of 81.6 Gy in 1.2 Gy per fraction twice daily in 68 fractions over 7 weeks, treated every Monday to Friday.[179] The standard control regimen consisted of 70 Gy in 2 Gy per day in 35 fractions over 7 weeks. With a median follow-up period of 23 months, the 2-year local-regional control rate, 2-year disease-free survival rate, and 2-year overall survival rate were 54.4% (versus standard arm of 46%; P = 0.045; Fig. 72-10 ), 37.6% (versus 31.7%; P = 0.067) and 54.5% (versus 46.1%; not significant), respectively. The modest but superior local-regional control rates in the hyperfractionated schedule reflected a 2-year local relapse and 2-year regional relapse rates of 37.8% (versus 43.7% in the control arm) and 26.6% (versus 32.1%), respectively. A planned neck dissection was permitted for residual neck abnormalities and for N2 and N3 neck disease regardless of the response. The 2-year distant relapse rate was 16.8% (versus 17.8%).
Table 72-4 -- Phase III Trials Addressing Hyperfractionation in Patients with Head and Neck Cancer
|
Study |
Tumor Site and Stage |
No. of Patients |
Dose per Fraction (Gy) |
Fractions per Day |
Total Dose (Gy) |
Overall Treatment Time (weeks) |
Tumor Response |
Complications |
|
Fu et al, 2000 |
Various sites, stage III–IV, stage II of tongue base, hypopharynx |
1073 |
1.2 |
2 |
81.6 |
6.0 |
LRC, higher with HF and CB (P = 0.045 and 0.05); DFS, trend in favor of HF and CB (P = 0.067 and 0.054); no difference in OS |
More acute mucositis with all altered fractionations; no difference in late complication rate |
|
1.8[*] |
1–2 |
72.0 |
7.0 |
|||||
|
1.6 |
2 |
67.2 |
6.0 |
|||||
|
2.0 |
1 |
70.0 |
7.0 |
|||||
|
Horiot et al, 1992 |
Oropharynx, T2–3 N0-1 |
356 |
1.15 |
2 |
80.5 |
7.0 |
5-year LRC, 59% vs 40% (P = 0.02); improved local control of T3 tumors |
More acute mucositis with HF: no difference in late complication rate |
|
2.0 |
1 |
70.0 |
7.0 |
|||||
|
Pinto et al, 1991 |
Oropharynx, stage III–IV |
98 |
1.1 |
2 |
70.4 |
6.5 |
Tumor response, 84% vs 64% (P = 0.02) 3–5 year OS, 27% vs 8% (P = 0.03) |
Earlier onset of acute reactions with HF, late complications, no details |
|
2.0 |
1 |
66.0 |
6.5 |
|||||
|
Cummings et al, 2000 |
Various sites, T3–4, N0 or any TN |
331 |
1.45 |
2 |
58.0 |
4.0 |
5-year LRC, 45% vs 37% (P = 0.01); 5-year OS, 40% vs 30% (P = 0.01) |
More acute mucositis with HF; 5-year grade 3–4 late toxic effects, 8% vs 14% (P = 0.31) |
|
2.55 |
1 |
51.0 |
4.0 |
From Nguyen LN, Ang KK: Radiotherapy for cancer of the head and neck: altered fractionation regimens. Lancet Oncol 2002;3:693–701.
|
CB, constant boost; DFS, disease-free survival; HF, hyperfractionation; LRC, locoregional control; OS, overall survival. |
|
* |
Boost dose given in 1.5 Gy fractions. |
|
Figure 72-10 Local-regional control rates for concomitant boost accelerated and hyperfractionation regimens relative to that for conventional fractionation. (From Nguyen LN, Ang KK: Radiotherapy for cancer of the head and neck: altered fractionation regimens. Lancet Oncol 2002;3:695.) |
Increased acute mucositis was the predominant acute toxicity in patients receiving the hyperfractionated protocol in the RTOG 90-03 trial. No increased risks of late toxicities were observed through the follow-up at the time of report, although this report would be regarded as immature for this endpoint. Of note, however, a prior prospective dose-escalation hyperfractionation study (also using 1.2 Gy per fraction twice daily) conducted by the RTOG (83-13) demonstrated that the risk of late toxicities was significantly increased with an interfraction time of less than 4.5 hours.[227] A large retrospective review also observed the importance of interfraction time.[228] Because this strategy can be logistically demanding for the patient, it becomes important to recognize that the current recommended interfraction time of 6 hours should be maintained.
Alternatively, an accelerated fractionation radiotherapy schedule attempts to deliver the prescribed total dose over a shorter treatment duration. This strategy is founded on observations of adverse local-regional control rates with protracted treatment durations (with conventional fractionated schedules) such that higher total doses are required to maintain the same probability of tumor control. These results have been interpreted to be consistent with a model whereby tumor clonogens surviving each daily radiotherapy fraction undergo an accelerated rate of repopulation. As a consequence, a larger tumor burden would be expected with increasing duration of treatment interruptions. It has been rationalized that by reducing the overall treatment time, the opportunity and impact of accelerated tumor repopulation would be minimized. As the severity of acute toxicities is increased, some accelerated schedules studied have attempted to modify the risk of unacceptable acute toxicities by modifying either the dose per fraction or the total dose as a strategy to achieve an acceptable therapeutic ratio. Accelerated radiotherapy schedules may therefore be categorized into two groups: schedules that do not modify the dose per fraction or the total dose (pure accelerated schedules) [229] [230] [231] [232] ( Table 72-5 ) and those that do (hybrid accelerated schedules) [179] [233] [234] [235] [236] ( Table 72-6 ). Examples of the former are two fractions delivered per day on some or all treatment weekdays and daily treatments for 6 to 7 days per week. A variety of hybrid schedules reflecting a spectrum of dose modifications has been studied. Conceptually, the success of these hybrid schedules is dependent on the dose equivalent of the reduction in the overall treatment time being greater than the biologic equivalent dose reduction in the fractionation schedule.
Table 72-5 -- Phase III Trials of Pure Accelerated Fractionation in Patients with Head and Neck Cancer
|
Study |
Tumor Site and Stage |
No. of Patients |
Dose per Fraction (Gy) |
Fractions per Day |
Total Dose (Gy) |
Overall Treatment Time (weeks) |
Tumor Response |
Complications |
|
Jackson et al, 1997 |
Various sites, stage III–IV |
82 |
2.0 |
2 (at least 6 h apart) |
66.0 |
3.4 |
CR, 35% vs 29% |
Grade 3–4 reactions |
|
2.0 |
1 |
66.0 |
6.8 |
No difference in 3-year relapse-free survival |
Grade 4 late toxicity (P = 0.10) |
|||
|
Skladowski et al, 2000 |
Various sites, T2–4, N0-1 |
100 |
1.8–2.0 |
1 |
- 70.0 |
5.0 |
3-year LC, 82% vs 37% (P < 0.0001); 3-year OS, 78% vs 32% (P < 0.0001) |
Severe mucositis, 62% vs 26%; late complications, 10% vs zero |
|
1.8–2.0 |
1 |
- 70.0 |
7.0 |
|||||
|
Overgaard et al, 2003 |
Various sites, all stages |
1485 |
2.0 |
1 |
- 66.0 |
6.0 |
5-year LRC, 66% vs 57% (P = 0.01); 5-year DFS, 72% vs 66% (P = 0.04); no difference in OS |
More acute mucositis with AF; no difference in late complication rate |
|
2.0 |
1 |
- 66.0 |
7.0 |
|||||
|
Hliniak et al, 2002 |
Laryngeal carcinomas, T1–3, N0 |
396 |
2.0 |
1–2 (at least 6 h apart) |
66.0 |
5.5 |
||
|
2.0 |
1 |
66.0 |
6.5 |
LRC, higher with AF (P = 0.03) |
More acute reactions with AF; no difference in late complications except for telangiectasia |
From Nguyen LN, Ang KK: Radiotherapy for cancer of the head and neck: altered fractionation regimens. Lancet Oncol 2002;3:693–701.
|
AF, accelerated fractionation; CR, complete response; DFS, disease-free survival; LC, local control; LRC, locoregional control; OS, overall survival. |
Table 72-6 -- Phase III Trials of Hybrid Accelerated Fractionation in Patients with Head and Neck Cancer
|
Study |
Tumor Site and Stage |
No. of Patients |
Dose per Fraction (Gy) |
Fractions per Day |
Total Dose (Gy) |
Overall Treatment Time (weeks) |
Tumor Response |
Complications |
|
ACCELERATED FRACTIONATION WITH TOTAL DOSE REDUCTION |
||||||||
|
Dische et al, 1997 |
Various sites, mainly stage II–IV |
918 |
1.5 |
3 (every 6 h) |
54.0 |
2.0 |
No difference in LRC, disease-free interval, or ulceration |
More acute mucositis, less epidermal telangiectasia, mucosal ulceration, and edema with AF |
|
2.0 |
1 |
66.0 |
6.5 |
|||||
|
Poulsen et al, 2001 |
Various sites, stage III–IV |
350 |
1.8 |
2 (at least 6 h apart) |
59.4 |
3.5 |
5-year LRC, 52% vs 47% (P = 0.30); 5-year DFS, 41% vs 35% (P = 0.32); 5-year DSS, 46% vs 40% (P= 0.40) |
More severe acute mucositis (P = 0.00008) but lower frequency of grade ≥2 late soft-tissue effects (P < 0.05) with AF (except for mucosal late effect) |
|
2 |
70.0 |
7.0 |
||||||
|
Bourhis et al, 2000 |
All sites, 75%; T4, 70% |
269 |
2.0 |
2 |
- 63.0 |
3.3 |
2-year LRC, 68% vs 34% (P < 0.01) |
Grade 3–4 mucositis, 83% vs 28% (P < 0.01); similar late toxic effects |
|
2.0 |
1 |
70.0 |
7.0 |
No difference in OS |
||||
|
ACCELERATED FRACTIONATION WITH SPLIT-COURSE (TYPE B) OR CONCOMITANT BOOST (TYPE C) |
||||||||
|
Horiot et al, 1997 |
Various sites, T2–4 N0-1 |
500 |
1.6 |
3 |
72.0 |
5.0 |
5-year LRC, 59% vs 46% (P = 0.02); trend for higher 5-year DFS (P = 0.08); no difference in OS (P = 0.95) |
More severe acute mucositis and higher frequency of severe late morbidity (P < 0.001) with AF |
|
2.0 |
1 |
70.0 |
7.0 |
|||||
|
Fu et al, 2000 |
Various sites, stage III–IV; stage II of tongue base, hypopharynx |
1073 |
1.8[*] |
1–2 |
72.0 |
6.0 |
LRC, higher with CB and HF (P = 0.06 and 0.045); DFS, strong trend in favor of CB and HF (P = 0.054 and 0.067); no difference in OS |
More acute mucositis with all altered fractionations: no difference in late complication rate |
|
1.2 |
2 |
81.6 |
7.0 |
|||||
|
1.6 |
2 |
67.2 |
6.0 |
|||||
|
2.0 |
1 |
70.0 |
7.0 |
|||||
From Nguyen LN, Ang KK: Radiotherapy for cancer of the head and neck: altered fractionation regimens. Lancet Oncol 2002;3:693–701.
|
AF, accelerated fractionation; CB, conconstant boost; DFS, disease-free survival; DSS, disease-specific survival; HF, hyperfractionation; LRC, locoregional control; OS, overall survival. |
|
* |
Boost dose given in 1.5-Gy fractions. |
A review of the 4 randomized trials of pure accelerated fractionation demonstrates that the overall treatment time may be reduced by 1 week without unacceptable acute and late toxicities, achieving modest improvements in local-regional control with no evidence of consistent survival gains. [229] [230] [231] [237] The most aggressive of these schedules was conducted at the BC Cancer Agency, where the experimental treatment protocol consisted of 66 Gy in 33 fractions with twice-daily fractions of 2 Gy, with an interfraction time of 6 hours, resulting in an overall treatment time of 3.4 weeks (versus 6.6 weeks). Patients receiving the accelerated schedule were more likely to experience RTOG grade 3–4 acute toxicity (seen in 27 of 41 versus 8 of 41; P = 0.00005). Increased grade 4 late toxicity (seen in 8 of 41 versus 2 of 41) resulted in premature termination of the study, which precluded definitive conclusions regarding the therapeutic efficacy of this schedule. Of the eight cases of late toxicity, four occurred after salvage surgery, two were soft tissue necroses, and two followed from persistent acute toxicity. Investigators at the Sklodowska-Curie Institute observed similar observations of consequential late effects involving the mucosa in their experimental arm of 70 Gy in 35 fractions with daily fractions of 2 Gy 7 days per week with an overall treatment time of 5 weeks. These investigators observed no additional late toxicities when the dose per fraction was reduced to 1.8 Gy per day. Although a significant gain in local-regional control and overall survival was observed, concerns regarding the validity of these observations have been raised owing to the unexpected poor outcome of the control arm, especially when the study did not include patients with N2 or N3 disease.
In contrast, investigators from both the Danish[232] and Polish[238] Cooperative Groups conducted randomized trials accelerating treatment with six treatments per week, either as an additional treatment on the weekend [232] [238] or as a single second daily fraction, with no evidence of any increased late toxicities.[232] Acute toxicities were increased, as would be expected. Significant improvements in local-regional control rates were observed, but with no survival gains. Hence, the evidence to date suggests that late toxicities, possibly resulting from severe acute pathologic processes that exceed normal tissue repair capacities, limit the increased biologic effect to only an approximately 1-week reduction in overall treatment time. It may be possible to realize additional therapeutic gains with a modest reduction in the dose per fraction to 1.8 Gy, but this needs to be further validated.
Various hybrid accelerated radiotherapy schedules have been studied as strategies to further achieve increased biologic effects with greater reductions in the overall treatment time (see Table 72-6 ). [179] [233] [234] [235] [236] These schedules may be further differentiated depending on whether or not the total dose was reduced (type A versus types B and C). Typically, the total dose has been reduced where more than a 2-week treatment time reduction (from a conventional 7-week course) has been attempted. The most aggressive schedule was the CHART schedule, which reduced the treatment time by 4.5 weeks by delivering 1.5 Gy three times a day with an interfraction time of 6 hours over a total duration of 2 weeks to a total dose of 54 Gy (representing an 18% dose reduction).[233] In a randomized trial of 918 patients with stage II to IV HNSCC involving various sites (a majority had laryngeal carcinomas), no improvement in local-regional control or overall survival was observed. Although this trial did demonstrate increased acute toxicities that occurred earlier, it is noteworthy that fewer late radiotherapy toxicities were observed in the accelerated arm. Poulsen and colleagues performed a randomized trial of 350 patients with stage III or IV HNSCC through the Trans-Tasman Radiation Oncology Group (TROG) studying a schedule with a 3.5-week reduction in the treatment time and a total dose reduction of 10.6 Gy (15%).[234] The experimental arm consisted of twice-daily1.8 Gy (6-hour interfraction time) to a total dose of 59.4 Gy. No improvement in local-regional control rates, disease-free survival, or survival was noted. Again, acute toxicities were more severe and occurred earlier, but no increased late toxicities were observed. Lastly, a French Cooperative Group Study (GORTEC 94–02) reported the preliminary results of a randomized trial of 268 patients with the majority having advanced T4 oropharyngeal carcinomas with the experimental arm receiving a 4-week reduction in treatment time and a modest 10% (7 Gy) reduction in the total dose.[235] The schedule used consisted of 2 Gy twice daily over 3 weeks to 62 Gy versus 70 Gy in 2 Gy per day. With only a median follow-up period of 28 months, a significant improvement in 2-year actuarial local-regional control rate was observed (58% versus 34%; P < 0.01) with no difference in overall survival. Increased acute mucosal toxicities were reported with no increased late toxicities noted with the limited follow-up. Hence, the data to date suggest that with only a modest total dose reduction of 10%, reducing the overall treatment time by more than 3 weeks may achieve improvements in local-regional control. With only a modest total dose reduction, however, it remains to be seen if late toxicities are increased.
Two other randomized trials have attempted to accelerate the overall treatment but without reductions in the total dose. The European Organization for Treatment of Cancer (EORTC) RadiotherapyCooperative Group randomized 512 patients with T2 to T4 HNSCC of all sites excluding the hypopharynx to conventional arm of 70 Gy in 35 fractions daily over 7 weeks or 72 Gy in 45 fractions over 5 weeks (EORTC 22851).[236] The experimental arm introduced a split in the treatment duration with the first half delivering 28.8 Gy in 18 fractions over 8 days with 1.6 Gy per fraction three times a day. This was followed by a 12- to 14-day treatment interruption, followed by a second course of 43.2 Gy in 27 fractions over 17 days again with 1.6 Gy per fraction three times a day. Although the 5-year local-regional control rate improved by 13% (59% versus 46%; 95% CI 3% to 23%), with a 24% reduction in local failure rate, this treatment regimen was associated with unacceptable toxicities including twice as many grade 3 or 4 acute morbidities, with grade 5 toxicity reported. Significantly more grade 3 fibrosis (P < 0.001) and severe neurologic complications, including permanent peripheral neuropathy, occurred in the accelerated arm. In contrast, the RTOG conducted a four-arm randomized trial in 1073 patients with stage III or IV HNSCC (stage II base of tongue and hypopharynx permitted), with one of two accelerated schedules also using a treatment interruption and a 4% total dose reduction (RTOG 90-03).[179] This treatment protocol consisted of 1.6 Gy twice daily (6-hour interfraction time) to a dose of 67.2 Gy in 42 fractions over 6 weeks. No improvement in local-regional control or overall survival was noted, suggesting that the treatment interruption employed also contributed to the absence of treatment benefit that was not sufficiently compensated in the dose intensity of fractionation schedule.
In the RTOG 90-03, a second accelerated schedule consisted of twice-daily fractions in the final 2.4 weeks, with the second fraction delivered with a 6-hour interfraction interval and limited to only the boost volume.[179] A total dose of 72 Gy was delivered, with the morning fraction 1.8 Gy and the afternoon boost fraction 1.6 Gy in the final 2.4 weeks of a 6-week treatment schedule. This limitation in volume was developed as a strategy to minimize the toxicity with the twice-daily fractionation. The timing of the concomitant boost followed from prior work by Ang and colleagues, who demonstrated slightly better local control rates than if the concomitant boost was delivered at the beginning of the radiotherapy schedule.[239] The results of the RTOG 90–03 demonstrated that the local-regional control rate was significantly improved (2-year survival rate of 54.2% versus 46.1%), with a trend to improved disease-free survival (see Fig. 72-10 ). The modest but superior local-regional control rates with this delayed concomitant boost accelerated schedule reflected a 2-year local relapse rate and 2-year regional relapse rate of 36.9% (versus 43.7% in the control arm) and 33.3% (versus 32.1%), respectively. A planned neck dissection was permitted for residual neck abnormalities and for N2 and N3 neck disease regardless of the response. The 2-year distant relapse rate was 16.6% (versus 17.8%). No improvement in overall survival was observed. Comparable increased acute toxicities were observed as with the other altered fractionation schedules including the hyperfractionated arm. No significantly increased late toxicities were observed though the incidence of late toxicities was higher in the accelerated schedule. Because the benefits from this fractionation schedule appear to be comparable to the hyperfractionated schedule reported in RTOG 90-03, investigators have concluded that an accelerated schedule with a delayed concomitant boost may be preferred due to the more favorable logistical treatment delivery issues.[183]
Although the use of an altered fractionated radiotherapy schedule offers an improved local-regional control rate of approximately 15%, this comes at the price of increased acute toxicities. Most notable is the increased mucositis, which can occur earlier and be more severe depending on how the various dose and time parameters are manipulated. The results from the Conventional Accelerated Irradiation (CAIR) trial highlight the potential for excessive mucosal toxicities to exceed their normal repair capacity, leading to consequential late effects. [238] [240] A similar potential for increased mucosal toxicities with consequential late effects also appears to have emerged with various concurrent chemoradiotherapy schedules including both conventionally fractionated [187] [199] and altered fractionated schedules.[241] These results highlight that a limit to local-regional intensive therapy does in fact exist and have led to interest in the development of normal tissue protectants.
TREATMENT TOXICITIES AND NORMAL TISSUE PROTECTANTS.
The most developed of these normal tissue protectants is amifostine. This thiol-containing compound and its metabolite, WR-2721, are believed to function as free radical scavengers and appear to have preferential normal tissue uptake, with the highest concentration found in salivary glands and kidneys. [242] [243] Amifostine has been shown to reduce cisplatin-induced nephrotoxicity.[244] It has been the subject of tremendous clinical interest, initially for the protection of salivary glands from radiotherapy-induced xerostomia, and most recently as a mucosal normal tissue protectant. To date, clinical trials have reported improvements in the rate and severity of radiation-induced xerostomia [245] [246] [247] [248]; reduced hematologic effects, particularly with concurrent chemoradiotherapy[245]; and reduced radiotherapy-induced mucositis with altered fractionated radiotherapy and from concurrent chemoradiotherapy. [246] [249] No compromise in treatment outcome has been reported, either by a small randomized trial using definitive chemoradiotherapy[246] or by a larger randomized trial containing a mixture of patients receiving postoperative radiotherapy and definitive therapy.[247] The latter trial led to the current approved indication for amifostine limited to the postoperative setting, owing to concerns of potential tumor protection that may not be adequately detected.
This trial randomized 315 patients undergoing conventionally fractionated radiotherapy to receive an intravenous 3-minute infusion of amifostine (200 mg/m2) 15 to 30 minutes before each daily fraction.[247] The main toxic effects were nausea (any grade, 44% versus 16%; P < 0.001), vomiting (any grade 37% versus 7%; P < 0.001), hypotension (any grade, 15% versus 2%; P < 0.001), and a hypersensitivity reaction (any grade, 5% versus 0%; P = 0.003) with 21% discontinuing the amifostine before completing the scheduled treatment. The mean quantity of unstimulated saliva 1 year after treatment was significantly higher, correlating with a lower frequency of late grade 2 or higher xerostomia in the group receiving amifostine. In this trial, mucositis was not reduced (grade 3 or higher, 35% with amifostine versus 39% no amifostine; P = 0.48), in contrast with to the results of other randomized trials. [246] [249] It has been postulated that this discrepancy may relate to dose, becuase Buntzel administered 500 mg as a flat dose and only on the days during which daily concurrent carboplatin was administered (days 1 to 5 and days 21 to 26).[246] Ongoing trials are under way to verify these results. It is clear, however, that the toxicities related to amifostine are dose-related, particularly the emetogenic side effects. These concerns along with logistical issues with coordinated daily administration have prompted studies with a subcutaneous schedule of administration.
A French Cooperative Group study (GORTEC) compared intravenous amifostine as a 3-minute infusion of amifostine (200 mg/m2) 15 to 30 minutes before each daily fraction with a subcutaneous schedule delivering 500 mg 20 to 60 minutes before each daily fraction.[248] Preliminary results have been promising. Early reporting describes reduced incidence of hypotension (6% versus zero, in favor of the subcutaneous route), with nausea and vomiting remaining dominant side effects in both treatment groups. The rates of acute xerostomia appeared to be similar in both treatment groups, offering a potential alternative schedule for administration.
INTENSITY-MODULATED RADIOTHERAPY.
In recent years, significant technological advances have enabled the ability to vary the fluence of the radiotherapy beam, permitting an additional degree of dose conformality and some exciting potential therapeutic applications. These may include an improved therapeutic ratio with irradiation near the base of skull, parotid sparing to minimize the risk of xerostomia, and manipulation of the effective radiotherapy dose per fraction that is delivered to the tumor or surgical bed. Coupled with promising advances in functional imaging, there exists the potential to manipulate the dose to critical areas within the tumor that may harbor radioresistant cells. In particular, interest has focused on identifying areas of tumor hypoxia that may be amenable to in vivo hypoxia imaging with various promising compounds including Cu-ATSM[250] and EF-5.[251] Although this technique remains promising, it is important to recognize its evolving nature and the potential for geographic tumor “misses,” particularly areas of subclinical tumor extension, which also may be complicated by a lower dose per fraction delivered. In addition, successful sparing of normal tissues requires knowledge of the dose and volume constraints that are associated with acceptable risks for toxicities. This knowledge base remains in evolution. The generalizability of not only the technique but also target delineation is the subject of several ongoing trials through the RTOG. Nevertheless, early reports are promising.
Several prospective reports demonstrated that with IMRT, dose and volume constraints to the parotid glands may be successful in reducing the xerostomia associated with radiotherapy to the head and neck.[252] [253] [254] [255] Quality of life instruments have been used and suggest that there may be additional benefits resulting from reduced xerostomia.[256] These investigators noted that the probability of a geographic tumor miss was low, [255] [257] with the majority of relapses within field, emphasizing the importance to identify potential radioresistant subvolumes. [254] [257] Lee and colleagues reported a promising 4-year local-regional progression free rate of 98% without any increased acute toxicities for 67 patients with stage I to IV NPC, with stage III or IV disease in 70%.[255] Although toxicities to important critical structures may be manipulated, this may come at the expense of increased skin toxicities,[258] as a result of increased radiation dose resulting from multiple complex beam arrangements that expose more normal tissue that would otherwise have been excluded with conventional techniques. It is this observation that has led to concerns regarding potential long-term adverse effects and a call for prudence in the application of IMRT.[259]
POSTOPERATIVE RADIOTHERAPY.
The indications for postoperative radiotherapy (PORT) may follow a risk-stratification paradigm that identifies patients as low, intermediate, or high risk for local-regional relapse, based on the absence, presence of one, or presence of two or more risk factors, respectively ( Fig. 72-11 ). These risk factors include: oral cavity primary, mucosal margins close or positive, nerve invasion, more than one positive lymph node, more than one positive nodal group, largest node greater than 3 cm, and treatment delay longer than 6 weeks. [161] [162] In addition, the presence of nodal extracapsular extension by itself placed patients in the high-risk group. [162] [163] Although no randomized trial exists to demonstrate the efficacy of postoperative radiotherapy, Ang and colleagues demonstrated in a prospective trial in patients with an intermediate risk of relapse a local-regional control rate (greater than 90%) comparable to that in patients with no adverse risk factors, deemed to be low risk, and not subjected to PORT (see Fig. 72-11 ).[161] The overall survival for the intermediate group appeared to be inferior to that for the low-risk group; nevertheless, PORT continues to be recommended owing to the importance of local-regional control.
|
Figure 72-11 Actuarial local-regional control rates (A) and overall survival (B) by postoperative risk stratification. (Ang KK, Trotti A, Brown BW, et al: Randomized trial addressing risk features and time factors of surgery plus radiotherapy in advanced head-and-neck cancer. Int J Radiat Oncol Biol Phys 2001;51:571–578.) |
Peters and colleagues previously reported on a dose-finding randomized trial demonstrating that a minimum tumor dose of 57.6 Gy to the whole operative bed should be delivered, with a boost of 63 Gy to sites of increased risk, especially regions of the neck in which extracapsular nodal disease is present.[162] These investigators did not find any benefit with dose escalation above 63 Gy at 1.8 Gy per day and postulated that this might be offset by tumor repopulation. Accordingly, Ang and colleagues recently reported the results of a multi-institution prospectively registered trial of 288 patients with HNSCC deemed to require PORT.[161] Of these 288 patients, 151 patients were stratified as high-risk and subsequently randomized to receive radiation therapy in either a conventionally fractionated schedule or an accelerated schedule using the delayed concomitant boost technique, with a total dose of 63 Gy delivered in each arm. These investigators report in this mature trial a nonsignificant trend for higher local-regional control (P = 0.11) and survival (P = 0.08) in favor of the accelerated schedule that appeared to result from an underpowered sample size ( Fig. 72-12 ). Acute confluent mucositis (62% versus 36%) was significantly greater in the experimental arm as would be expected. The actuarial probability of a patient sustaining one or more late complications between the two fractionation schedules (P = 0.94) did not significantly differ between the two arms. These investigators therefore concluded that the benefits of an accelerated PORT schedule had not been definitively established, but comment that, in practice, an accelerated PORT schedule may be used to keep the overall treatment time to less than 11 weeks in unavoidable situations where there has been protracted time before PORT. The latter follows from the demonstration of a significant adverse impact on both local-regional control and survival rates when analyzed by the overall treatment time ( Fig. 72-13 ). These results must be interpreted with caution, however, as the study design did not stratify by the time interval before starting PORT.
|
Figure 72-12 Actuarial local-regional control rates (A) and overall survival (B) for high-risk patients according to the postoperative radiotherapy fractionation schedule. (Ang KK, Trotti A, Brown BW, et al: Randomized trial addressing risk features and time factors of surgery plus radiotherapy in advanced head-and-neck cancer. Int J Radiat Oncol Biol Phys 2001;51:571–578.) |
|
Figure 72-13 Actuarial local-regional control rates (A) and overall survival (B) for high-risk patients according to the overall treatment time (from the time of surgery to the completion of radiotherapy). (Ang KK, Trotti A, Brown BW, et al: Randomized trial addressing risk features and time factors of surgery plus radiotherapy in advanced head-and-neck cancer. Int J Radiat Oncol Biol Phys 2001;51:571–578.) |
Care should be exercised in the use of an accelerated schedule as concern has also been raised of a possible increased risk of late toxicities in the postoperative setting.[260] Rather, patients anticipated to require postoperative radiotherapy should be appropriately identified with the appropriate arrangements made for the patient so as to prevent unnecessary interruptions to starting PORT. In a prospective comparative trial with a median follow-up period of 6 years Trotti and colleagues demonstrated that patients initiating PORT within 4 weeks had a significantly lower rate of crude in-field relapses (seen in 0 of 10 versus 10 of 32).[260]
It has also been proposed that the integration of chemotherapy with postoperative radiotherapy be favored in light of the increased risk of distant relapses also observed in the high-risk group. Ang and colleagues noted a 5-year actuarial distant relapse rate of 33% (versus 3% in the low-risk group). Although attractive, this remains an active area of investigation with no established chemoradiotherapy regimen. Two randomized trials of postoperative chemoradiotherapy with concurrent cisplatin and daily fractionated radiotherapy have been reported. [261] [262] Bachaud and colleagues reported the results of a randomized trial prematurely closed due to poor patient accrual, demonstrating improved local-regional control rate, disease-free survival, and overall survival with weekly concurrent cisplatin. Only 88 patients with stage III or IV disease with the presence of extracapsular extension were were accrued, however, so the authors suggest that these results require validation in a larger study. Recently, the RTOG presented the preliminary results of a randomized trial of 459 patients with high-risk postoperative features including extracapsular extension receiving radiotherapy or radiotherapy and bolus cisplatin on weeks 1, 4, and 7 (RTOG 9501).[262] Modest improvement in local-regional disease control and disease-free survival (54% versus 43%; P = 0.049) in favor of the chemoradiotherapy arm, with no difference in overall survival, was reported.
Several institutional retrospective reports have suggested that the addition of a brachytherapy implant to postoperative radiotherapy when surgical margins are positive or close may improve the local control rates for a mixture of tumor sites. [263] [264] [265] [266] [267] This strategy may be particularly attractive for management of early oral cavity lesions, for which the risk of nodal metastasis is low, thereby avoiding the morbidity associated with EBRT. Site-specific indications for the floor of the mouth[263] and the oral tongue[264] have been reported. Pernot and colleagues reported a 5-year local control rate of 89% for 97 patients who received either postoperative EBRT followed by an implant or brachytherapy with an implant alone for oral cavity tumors.[265] Self-resolving grade 1 and 2 complications occurred in 19% and 12%, respectively, with only 6% of complications requiring surgical intervention (grade 3). These results are particularly promising in light of the fact that one third of the patients who received this treatment had T3 and T4 lesions. Accordingly, this therapeutic may be considered when the surgical margin of concern can be located and when experience exists for the safe administration of the implant.
BRACHYTHERAPY.
Brachytherapy has a significant role in the management of head and neck squamous cell carcinomas [268] [269] with various techniques described.[270] Owing to the unique physical properties of this region, treatment morbidity is minimized as a result of reduced irradiation in the surrounding normal tissues. An implant may be used in the definitive setting for several tumor sites including the tonsil and soft palate, [271] [272] [273] [274] [275] [276] oral tongue, [277] [278] [279] [280] base of tongue, [281] [282] [283] [284] [285] [286] [287] [288] and lip. [289] [290] The use of an implant in the base of tongue has the advantage of being a functional organ-preserving treatment strategy validated with quality of life instruments, [291] [292] with the results suggested to be superior to those achievable with EBRT alone and comparable with those for surgery with PORT.[284] It may be used as an alternative to surgery for selected cancers of the floor of mouth in specific circumstances. In early-stage lesions, for which the risk of nodal metastases is low, brachytherapy may be employed definitively or in an adjuvant fashion after surgery. In more advanced lesions, it often is combined with external irradiation of the head and neck. The ability to provide specific high local irradiation also permits the selective use of brachytherapy in the setting of recurrent [293] [294] or second HNSCC occurring within a previously irradiated region.[295]
Appropriate application of a brachytherapy implant begins with patient selection. This process requires assessment of the patient's understanding and ability to comply with the inherent radiation precautions associated with brachytherapy implants, especially for continuous low dose rate (LDR) implants. Patients should be selected for their ability to provide for their baseline self-care needs, in addition to the treatment-related needs such as the care of a tracheostomy, nasogastric feeding tube, and patient-controlled analgesic pump as indicated. Patients subject to periods of confusion and disorientation may not be suitable for this mode of therapy.
Several considerations influence the decision of a permanent or a temporary implant. Permanent implants, emitting radiation over the lifetime of its radioactivity, use sources that provide LDR irradiation. Suboptimal placement of a permanent implant and the potential adverse dosimetric effects of organ swelling and movement pose potential risks for an unfavorable therapeutic ratio. A permanent implant affords the delivery of a very high total dose delivered, however, and may be advantageous when implanting complex and irregular surfaces not amenable to placement of temporary catheter-based implants where chinking of the catheters is a significant risk. The judicious use of permanent sources with low-energy photons, such as iodine 125 (125I), may be advantageous when critical normal structures, such as the spinal cord, are adjacent to the implant.
Temporary implants more commonly are applied in the head and neck because they permit a more deliberate and accurate placement of the implant applicator system without the radiation exposure concerns that occur with a permanent implant. Typically, nylon catheters are placed, approximating the desired position of the radioactive sources, which then may be subsequently afterloaded with LDR radioactive seeds embedded at defined positions within a nylon strand. This technique affords optimization of the implant dosimetry after placement of the implant applicator system. Commonly, this has involved obtaining orthogonal plain x-ray films of the implant with dummy seeds placed within the selected applicator system, with digitization of the relative seed positions into a treatment planning software. Variations in the activity, number of radioactive sources, and loading duration and, for high dose rate (HDR) computer- guided remote afterloading systems, variations in the dwell time and position allow for dosimetric optimization. Optimization cannot obviate the adverse dosimetry associated with poor implant geometry, however.
Temporary LDR implants also offer several radiobiologic advantages, including a reduced treatment time, the capability to irradiate a potentially less hypoxic tumor bed early in the postoperative period, a reduced adverse influence of hypoxia itself, and exploitation of cell cycle-specific radiosensitization. These implants further exploit the differential repair capacities between tumor and normal tissues, reducing the risk of normal late complications. The risk of radiation exposure to personnel, however, necessitates good source-handling skills and strict radiation precautions. Alternatively, HDR sources with computer-guided remote afterloading significantly reduce the exposure risks and required precautions. Fractionated radiotherapy is delivered with a single iridium 192 (192Ir) source fixed to the end of a cable wire that may be variably stepped along the length of each catheter. HDR implants confer greater flexibility in conforming the implant dosimetry to the target volume, yield a relatively more homogeneous dose distribution to those of LDR implants, and, because the delivery of the radiation occurs over a shorter time period, are less subject to the effects of organ movement. These advantages may yield a lower complication rate as a result of this precise geometric sparing. Concerns remain, however, with regard to the risk of increased late complications from the higher dose rate of radiation.[296]This has prompted ongoing studies to define the optimal fractionation schedules to reduce this risk. Several other promising but investigational techniques include the use of pulsed dose rate (PDR) radiation, which has been studied as a technique to exploit the logistical advantages and reduced radiation exposure of remote afterloading and the LDR biologic advantages that may be expected with this technique. [268] [297] [298] HDR intraoperative radiation therapy (HDR-IORT) remains a promising investigational technique that has the advantage of accurately delivering radiation to the areas at risk for tumor recurrence potentially at a time when the tumor burden is the lowest.[299]
Chemotherapy
GENERAL.
The role of chemotherapy in the primary treatment of locally advanced disease became more prominent in the 1970s due to the poor outcome of stage III and IV disease treated with surgery or radiation, or both. Definitive treatments with chemotherapy emerged in nasopharyngeal cancer and in organ preservation protocols for the larynx,[193] the hypopharynx,[192] and, most recently, the oropharynx.[300]Systemic chemotherapy continues to have a role for palliation in patients with locally advanced-stage disease, locally recurrent disease beyond salvage techniques such as surgery, and metastatic disease.
PROGNOSTIC FACTORS.
The decision to treat with chemotherapy remains dependent on various factors that can contribute to response. Those factors include the patient's performance status, nutritional status, the tumor burden and extent, disease stage, degree of tumor differentiation, and primary cancer site. [192] [301] [302] [303]
NEOADJUVANT AND INDUCTION CHEMOTHERAPY.
Over the last 10 to 15 years, approaches to the treatment of patients with stage III and IV disease began to include the use of chemotherapy as induction therapy before planned surgical resection and, more recently, before radiotherapy. The concept of induction chemotherapy arose from several principles. It has been postulated that chemotherapy may promote regression of tumor, enhancing local-regional therapy through sensitization, and also may identify patients who may be candidates for a more conservative surgical approach as the need for improved quality of life through functional preservation has arisen. Thus, organ preservation, rather than extensive, potentially morbid surgical procedures, came into vogue as a philosophical consideration in the management of advanced-stage disease. An additional attractive feature with this approach was the conceptual ability to treat micrometastatic disease in hopes of reducing distant failure rates, which can be 40% or greater with conventional local-regional surgical/radiation approaches. Finally, it was felt that the use of chemotherapy before the tumor and vascular bed are altered by surgery or irradiation may improve the ability to identify responding tumors for which adjuvant chemotherapy may be beneficial.
Nonrandomized phase II trials in the 1970s used single-agent chemotherapy based on strategies used in the recurrent and metastatic setting. These single-agent trials reported 30% to 40% response rates.[304] Induction strategies subsequently involved multiple chemotherapy regimens. The first reported trials by Wittes and colleagues[304] showed a 71% response rate with complete responses noted in 21% of patients using cisplatin and continuous infusion bleomycin in 21 patients. Other studies followed, using cisplatin/bleomycin with other drugs such as hydrea and revealing increased toxicity with no improvement in response rates or survival.[305] Investigators from Wayne State University reported the first trial using neoadjuvant cisplatin with infusional 5-fluorouracil, with an overall response rate of 88% and a complete response rate of 54%.[306] The investigators reported that 120-hour infusional 5-fluorouracil showed improvement over 96-hour infusions and that complete response rates after three total cycles were double those achieved after two. [305] [306] [307] More recent studies appear to confirm that complete response rates will increase after three to five cycles; other studies have confirmed the activity of this combination but at varied response rates (38% to 100%) and complete response rates (13% to 54%). [308] [309] A 38% complete response rate also was achieved by the RTOG.[310] Numerous other combinations reported have included high-dose cisplatin with fixed-dose 5-fluorouracil, high-dose 5-fluorouracil with fixed-dose cisplatin, intra-arterial cisplatin, and additional drugs, such as bleomycin, cyclophosphamide, mitoguazone, taxanes, and methotrexate, have been given with the cisplatin plus 5-fluorouracil regimen, with significant toxicity and no overall difference in response or survival. [307] [310] [311] [312] Trials using carboplatin with 5-FU have shown similar response rates of 70% to 80% and complete response rates of 30% to 40%, which are similar to those with cisplatin plus 5-fluorouracil.[313]
Although these experiences with chemotherapy plus irradiation included several randomized trials with neoadjuvant chemotherapy, many were characterized by methodological problems. In 1985, the increasing interest in laryngeal preservation coupled with disappointing experiences with upfront radiotherapy for advanced disease, laid the foundation for the Veterans Affairs Cooperative Studies Program (VACSP) to initiate a multi-institutional randomized trial of neoadjuvant chemotherapy as an organ preservation strategy.[193] Patients with previously untreated, locally advanced but potentially resectable stage III (T2-3N1 or T3N0) or stage IV (T1-3N2-3 or T4N0-1) disease of the supraglottic or glottic larynx were randomized either to receive neoadjuvant chemotherapy or to undergo surgical resection (Fig. 72-14 ). The chemotherapy regimen was cisplatin (CDDP) with continuous-infusion 5-fluorouracil for 5 days with response after 2 cycles used to stratify patients to either continue with an additional cycle of chemotherapy followed by radiotherapy or, for nonresponders, salvage surgery followed by postoperative radiotherapy. In total, 332 patients were enrolled: 216 patients with T3, 85 with T4, and 240 patients with N0-1 disease. Laryngeal preservation was achieved in 64% of patients enrolled in the chemotherapy arm. The local failure rate was significantly higher in the chemotherapy plus irradiation arm, but the distant failure rate was significantly lower in this arm. The long-term overall survival rate was approximately 30% in each arm. On subset analysis, sequential chemotherapy plus irradiation was less effective in the T4 tumors or those with N2 or greater disease; 50% of these patients required salvage laryngectomies. Although this trial demonstrated that laryngeal conservation was achievable in 64% of patients with advanced laryngeal carcinoma, the incremental role neoadjuvant chemotherapy contributed to this is unclear. This laryngeal preservation rate appears to be comparable to the results achieved with radiotherapy alone followed by salvage surgery, which ranged from 50% to 73%. [191] [197] The value of neoadjuvant chemotherapy has also been questioned as only a subset of patients with advanced laryngeal cancers had chemoresponsive tumors.
|
Figure 72-14 Study schema of Veterans Affairs larynx preservation trial. CR, complete response; NR, no response; PR, partial response. (From Department of Veterans Affairs Laryngeal Cancer Study Group: N Engl J Med 1991;324:1685–1690.) |
These results subsequently were approximated in a similar randomized trial conducted by the EORTC.[192] The EORTC randomized a smaller population of 194 patients with locally advanced hypopharyngeal cancer (6% with stage II, 57% with stage III, and 37% with stage IV) to receive either cisplatin plus 5-fluorouracil or total laryngectomy, partial pharyngectomy or radical neck dissection with adjuvant radiation therapy. Only patients who achieved a complete response received radiation alone; 54% had a complete response at the primary site, 51% at the nodal site, and 43% achieving a complete response both at the primary and nodal site. Treatment failures at the local, regional, and second primary sites occurred at approximately the same frequencies in the immediate surgery arm (12%, 19%, and 16%, respectively) and in the induction chemotherapy arm (17%, 23%, and 13%, respectively). By contrast, there were fewer failures at distant sites in the induction chemotherapy arm than in the immediate surgery arm (25% versus 36%, respectively; P = 0.041). The median duration of survival was 25 months in the immediate surgery arm and was 44 months in the induction chemotherapy arm, which the investigators concluded were equivalent. The 3- and 5-year estimates of retaining a functional larynx in patients in the induction chemotherapy group were 42% (95% CI, 31% to 53%) and 35% (95% CI, 22% to 48%), respectively. It would appear that with a more stringent response to chemotherapy, fewer local relapses may be anticipated, in contrast with the results of the VA study.[193]
Recently, a French Cooperative trial (GETTEC) reported the randomized results using a neoadjuvant chemotherapy approach for organ preservation for the oropharynx site.[300] Patients with a squamous cell carcinoma of the oropharynx for whom curative radiotherapy or surgery was considered feasible were randomized to undergo either three cycles of neoadjuvant chemotherapy, followed by local-regional treatment determined by the treating physician, or the same local-regional treatment without chemotherapy. The local-regional treatment consisted either of surgery plus radiotherapy or of radiotherapy alone. The chemotherapy regimen consisted of cisplatin (100 mg/m2) on day 1, followed by a 24-hour intravenous infusion of fluorouracil (1,000 mg/m2 per day) for 5 days delivered every 21 days. A total of 318 patients were enrolled in the study between 1986 and 1992; the study was prematurely closed as a result of a loss of clinical equipoise, because the treating physicians believed that neoadjuvant chemotherapy was efficacious. The investigators note, however, that this decision was independent of any knowledge of the trial results, minimizing the impact of any bias. Overall survival was significantly better (P = 0.03) in the neoadjuvant chemotherapy group than in the control group, with a median survival of 5.1 years versus 3.3 years in the no chemotherapy group. The effect of neoadjuvant chemotherapy on event-free survival was less and of borderline significance (P = 0.11). In summary, it is clear that systemic chemotherapy can have an impact on the risk of distant relapses, with possible improvements in overall survival.
The preceding studies led to the more recent intergroup trial, RTOG 9111, conducted in patients with stage III or IV resectable disease of the larynx.[198] Patients were randomized to three treatment groups: chemotherapy (cisplatin plus 5-fluorouracil) followed by radiation therapy; concurrent chemoradiation therapy with high-dose cisplatin as the radiosensitizer; or standard fractionated EBRT daily. Patients with T4 lesions were not included in this trial. If a patient initially had N2 or N3 neck disease, a modified neck dissection was performed independent of response. Two-year laryngectomy-free survival was superior in the group of patients receiving concurrent chemoradiotherapy (P = 0.018), reducing the number of laryngectomies performed by approximately 50%, with the number of laryngectomies performed identical in the other two treatment arms (43 versus 21 versus 49, respectively). Two-year local-regional control rates were also superior in the concurrent chemoradiotherapy arm (61%, 78%, and 56%, respectively), with the overall survival (approximately 75%) not differing among the treatment arms.
Efforts continue to improve on the clinical efficacy of neoadjuvant chemotherapy in hopes of achieving significant activity to yield consistent survival benefits.[314] Recent strategies have focused on the use of neoadjuvant chemotherapy followed by concurrent chemoradiotherapy. Various phase II studies [187] [301] [315] [316] [317] [318] have been reported. These also include variations of the “gold standard” chemotherapy regimen, cisplatin and 5-fluorouracil, incorporating leucovorin rescue (Platinol-fluorouracil-leucovorin [PFL] regimen), with or without interferon-a, efforts pioneered by Vokes and colleagues in Chicago. [301] [317] A recently completed Eastern Cooperative Oncology Group (ECOG) phase II trial of neoadjuvant carboplatin and paclitaxel followed by concurrent weekly paclitaxel and daily-fractionated radiotherapy for oropharynx carcinomas included functional swallowing assessments as a measure of functional preservation. This trial follows from promising preliminary data reported by Machtay and colleagues, who noted a major clinical response rate of 89% after induction chemotherapy, with a 90% complete response rate after concomitant chemotherapy.[187] The 3-year survival and 3-year progression-free survival rates were 68% and 60%, respectively. Local-regional control was 82%, and the 3-year distant failure rate was reported at 18%. Organ preservation was achieved in 77% of all patients. In general, increased toxicities have been observed in these trials with promising activity that will require study in the context of a randomized trial. These strategies offer the promise of not only improved local-regional control rates with organ preservation but also a reduced risk of late distant relapses and more consistent improvements in overall survival. Recent results from a large patient-based meta-analysis, however, have drawn attention to the potential negative effects of neoadjuvant chemotherapy when used as a larynx preservation strategy, because a nonsignificant hazard ratio of death (1.19; range, 0.97 to 1.46) was noted.[181]
CONCURRENT AND CONCOMITANT CHEMORADIOTHERAPY.
A significant body of literature exists, including numerous randomized trials of concurrent chemotherapy that have been systematically summarized by several investigators. [181] [319] [320] [321] These independent reviews have consistently favored the concurrent integration of chemotherapy. Pignon and colleagues reported the largest and recently updated of these meta-analyses.[181] This patient-based meta-analysis of more than 10,000 patients derived from 63 randomized trials confirmed an absolute survival benefit of 4% at 2 and 5 years, with the greatest benefit of 8% observed in the group receiving concurrent chemotherapy. The group that contributed to this survival benefit was found to be the group of studies that used radiotherapy as the local-regional treatment. The survival benefit was found to be significantly greater with multiagent versus single-agent concurrent chemotherapy. A nonsignificant increase in the risk of death was noted with multiagent chemotherapy regimens containing a platinum agent ( Fig. 72-15 ). When results were analyzed with covariants, a significant decreasing benefit on survival for concurrent chemotherapy was noted with increasing age, which may be partly explained by lower compliance and higher toxicities ( Fig. 72-16 ). These results are consistent with anecdotal clinical experiences and highlight the importance of patient selection for concurrent chemoradiotherapy in light of the small incremental benefit. No survival benefit was observed with either neoadjuvant or adjuvant chemotherapy, leading the study investigators to recommend its use only within the context of a clinical trial.
|
Figure 72-15 Hazard ratio of death with local-regional treatment plus chemotherapy (CT) with local-regional treatment by chemotherapeutic regimen: Platin (cisplatin or carboplatin) + fluorouracil (FU), combination CT with platin (polyCT + P), combination CT without platin (polyCT w/o P), single-agent CT (monoCT) including platin. Test for heterogeneity between types of chemotherapy (P = 0.02). (From Pignon JP, Bourhis J, Domenge C, Designé L: Chemotherapy added to locoregional treatment for head and neck squamous-cell carcinoma: three meta-analyses of updated individual data. MACH-NC Collaborative Group. Lancet 2000;355:949–955.) |
|
Figure 72-16 Hazard ratio of death with local-regional treatment with or without chemotherapy (CT) by age, sex, performance status, stage, or tumor site. Test for trend for age was significant (P = 0.05). (From Pignon JP, Bourhis J, Domenge C, Designé L: Chemotherapy added to locoregional treatment for head and neck squamous-cell carcinoma: three meta-analyses of updated individual data. MACH-NC Collaborative Group. Lancet 2000;355:949–955.) |
Although concomitant chemoradiotherapy is probably the most promising and feasible approach in locally advanced patients, there remain several issues including the generalizability of these survival benefits to specific subsites, specific chemoradiotherapy regimens, including recent popular taxane-based regimens, and whether or not, differences exist between resectable and unresectable disease. The latter distinction becomes important, because concurrent chemoradiotherapy has emerged as a popular strategy to achieve functional organ preservation.[198] Because the addition of concurrent chemotherapy increases mucosal toxicity and poses the potential of adversely affecting swallowing function, [187] [199] if the survival benefits, particularly with multiagent regimens, are generalizable to those patients with resectable disease, compromises in organ function may result. Even if survival benefits are not observed, multiagent regimens may still be preferred, as recently noted by Adelstein and associates.[135] In their mature report of 100 patients with resectable HNSCC at various tumor sites, improved local-regional control rates without surgical salvage (77% versus 45%; P < 0.001), although with no survival benefit, were observed with concurrent chemotherapy (both 5-fluorouracil [1000 mg/m2 per day] and cisplatin [20 mg/m2 per day] given as a continuous intravenous infusion over 4 days beginning on day 1 and again on day 22). Successful primary salvage surgery was possible in 73% of cases, which contributed to the absence of any demonstrable survival benefit.
Although concurrent chemoradiotherapy has been studied in patients with stage III and IV HNSCC, often without selection for resectability, it becomes important to recognize the narrowed therapeutic ratio in its application.[184] In this regard, the relative efficacy of altered fractionated radiotherapy schedules is an important consideration, because the randomized trials establishing their efficacy have used comparable selection criteria. To date, limited randomized data exist comparing conventionally fractionated radiotherapy with concurrent chemotherapy with altered fractionated radiotherapy. Olmi and colleagues recently reported the results of a three-arm multi-institutional randomized trial of concurrent chemoradiotherapy versus altered fractionated radiotherapy versus conventionally fractionated radiotherapy alone.[322] These investigators randomized 192 patients with previously untreated, stage III and IV oropharyngeal carcinoma (excluding T1N1 and T2N1) to undergo one of three therapy regimens described as arms A, B, and C: for arm A, conventionally fractionated radiotherapy to a dose of 66 to 70 Gy in 33 to 35 fractions, 5 days a week over 6.5 to 7 weeks; for arm B, altered fractionated radiotherapy to a dose of 64 to 67.2 Gy, in two fractions of 1.6 Gy every day, with an interfraction interval of at least 4 hours and preferably 6 hours, 5 days a week with a 2-week split at 38.4 Gy, with radiotherapy resumed at the same fractionation after the split (arm B); or for arm C, chemotherapy in a regimen consisting of carboplatin and 5-fluorouracil (CBDCA 75 mg/m2, days 1 to 4, plus 5-fluorouracil 1000 mg/m2 given by intravenous infusion over 96 hours, days 1 to 4, every 28 days [at weeks 1, 5, and 9]), plus radiation therapy using the same daily fractionation schedule as described for the standard arm (arm A). No significant differences were detected in overall survival (P = 0.129): 40% of patients in arm A, 37% of those in arm B, and 51% of those in arm C were alive at 24 months. The 2-year disease-free survival rates, however, were significantly different among the three arms (P = 0.022), with the highest rate in the chemoradiotherapy arm. At 24 months, the proportion of patients without relapse was 42% for arm C, 23% for arm A, and 20% for arm B. Increased grade 3 skin and mucosal toxicities were noted in the concurrent chemoradiotherapy arm and the altered fractionated schedule. A suggestion of increased late skin and mucosal toxicities was noted in the chemoradiotherapy arm. Nguyen and Ang also have favored concurrent chemoradiotherapy in patients who are able to tolerate the increased toxicity, particularly in the setting of advanced T3 or T4 primary disease or in patients with advanced N2 or N3 neck disease, stratifying patients with T2 or exophytic T3N1 disease to altered fractionated radiotherapy.[183]
Multiagent regimens continue to be favored over single-agent regimens owing to concerns of late distant failures that have become more evident with improved local-regional management. A favorable effect for many multiagent regimens on the risk of distant relapses, however, remains to be clearly established. [178] [323] A reduced risk of distant relapse has been suggested in several reports [135] [324] and may indicate that the local-regional therapy is effective in addressing micrometastases, but control of the local-regional disease is necessary for this benefit to manifest. Accordingly, significant interest has been focused on the incorporation of altered fractionated radiotherapy schedules with concurrent chemotherapy ( Table 72-7 ). In total, six randomized trials have been reported and may be characterized by the type of altered fractionation used. [212] [241] [323] [324] [325] [326] In the trials combining an accelerated schedule, acute mucosal toxicities were significantly increased, with two trials demonstrating unacceptable toxicity, [241] [326] one of which included a high rate of chronic swallowing dysfunction.[326] Jeremic and colleagues reported an improved 5-year local-regional control rate (50% versus 36% at 5 years; P = 0.041), 5-year overall survival (46% versus 25% at 5 years; P = 0.0075), and distant metastasis-free survival (86% vs. 57% at 5 years; P = 0.0013) with the concurrent administration of daily cisplatin (6 mg/m2) with a hyperfractionated schedule delivering 77 Gy as 1.1 Gy twice daily in 70 fractions over 7 weeks compared with this same hyperfractionated schedule alone.[324] No significantly increased acute or late toxicities were reported. This remains a promising aggressive regimen that requires further validation of the concept of daily radiosensitization with an altered fractionation schedule. Finally, both 5-fluorouracil and cisplatin with or without leucovorin have been administered concurrently with a split-course schedule. Wendt and colleagues demonstrated improved local-regional control rates with an accelerated split-course schedule with concurrent bolus schedule of cisplatin, 5-fluorouracil, and leucovorin, but the results were overall disappointing (36% versus 17%; P = 0.004).[323] A significant reduction in the total dose was used. Brizel and colleagues administered concurrent cisplatin and 5-fluorouracil with a split-course hyperfractionated schedule of 1.25 Gy twice daily over 47 days with a 7- to 10-day treatment interruption after 40 Gy.[212] Improved local-regional control rates with no difference in overall survival were noted, with increased risk of sepsis and enteral feeding noted in the experimental arm. To date, the generalizability of any of these regimens is limited not only by the increased toxicity, but also by the intensive resources required on the part of the treating team and the patients.
Table 72-7 -- Phase III Trials of Concurrent Chemotherapy and Altered Fractionation in Patients with Head and Neck Cancer
|
Study |
Tumor Site and Stage |
No. of Patients |
Therapy Regimens |
Tumor Response |
Complications |
|
ACCELERATED FRACTIONATION PLUS CHEMOTHERAPY |
|||||
|
Dobrowsky and Naude, 2000 |
Various sites, T1-4 N0-3 |
188 |
55.3 Gy over 17 days (2.5 Gy on day 1, then 1.65 Gy, twice a day) ± mitomycin; conventional fractionation, 60 Gy over 7 weeks |
Combination treatment yielded higher LRC (P < 0.05) and survival (P < 0.03) |
More mucositis than in the combination group but not intensified by mitomycin; late toxic effects not reported |
|
Staar et al, 2001 |
Various sites, stage III–IV |
240 |
69.9 Gy over 5.5 weeks plus carboplatin (70 mg/m2 per day) and fluorouracil (600 mg/m2 per day) for 2 cycles of 5 days; 69.9 Gy over 5.5 weeks (1.8 Gy once daily for 3.5 weeks, then, individual fractions of 1.8 Gy and 1.5 Gy, daily for 2 weeks) |
2-year OS, 48% vs 39% (P = 0.11); 2-year LC, 51% vs 45% (P = 0.14); patients receiving radiochemotherapy had worse LRC |
Grade 3–4 mucositis, 68% vs 52% (P = 0.01); grade 3–4 vomiting, 8.2% vs 1.6% (P = 0.02); late swallowing problems and feeding tube dependency, 51% vs 25% (P = 0.02) |
|
Bourhis et al, 2001 |
Various sites, advanced-inoperable |
109 |
62.64 Gy over 5 weeks plus cisplatin (100 mg/m2 on days 1, 16, and 32) and fluorouracil (1 g/m2 on days 1–5 and 31–35); 62–64 Gy over 3 weeks |
Not yet reported |
Early stopping due to higher number of treatment-related deaths in the combined-treatment group |
|
HYPERFRACTIONATION PLUS CHEMOTHERAPY |
|||||
|
Jeremic et al, 2000 |
Various sites, stage III–IV |
130 |
77 Gy over 7 weeks plus cisplatin (6 mg/m2 per day); 77 Gy over 7 weeks (1.1 Gy, twice daily) |
5-year LRPFS, 50% vs 35% (P = 0.04); 5-year PFS, 46% vs 25% (P = 0.007); 5-year DMFS, 86% vs 57% (P = 0.001); 5-year OS, 46% vs 25% (P = 0.008) |
No significant difference in acute morbidity (except for leucopenia, P = 0.006) or late toxic effects |
|
SPLIT-COURSE ALTERED FRACTIONATION PLUS CHEMOTHERAPY |
|||||
|
Wendt et al, 1998 |
Various sites, stage III–IV |
270 |
70.2 Gy over 51 days plus cisplatin, fluorouracil, and leucovorin; 70.2 Gy over 51 days (23.4 Gy in 1.8-Gy fractions, twice daily × 3 cycles with a 10-day break) |
3-year LRC, 36% vs 17% (P < 0.004); 3-year OS, 48% vs 24% (P < 0.0003) |
Grade 3–4 acute mucositis, 38% vs 16% (P < 0.001); serious late side effects, 10% vs 6.4% (NS) |
|
Brizel et al, 1998 |
Various sites, T2–4 N0–3 |
122 |
70 Gy over 47 days as 1.25 Gy, twice daily (7- to 10-day break after 40 Gy) plus cisplatin and fluorouracil in weeks 1 and 6; 75 Gy over 42 days as 1.25 Gy, twice daily |
3-year LRC, 70% vs 44% (P = 0.01); 3-year RFS, 61% vs 41% (P = 0.07); 3-year OS, 55% vs 34% (P = 0.07) |
Similar mucositis; increased enteral feeding and sepsis with combination therapy; similar late complications |
From Nguyen LN, Ang KK: Radiotherapy for cancer of the head and neck: altered fractionation regimens. Lancet Oncol 2002;3:698.
|
DMFS, distant metastasis-free survival; LC, local control; LRC, locoregional control; LRPFS, locoregional progression-free survival; NS, not significant; OS, overall survival; PFS, progression-free survival. |
POSTOPERATIVE AND ADJUVANT CHEMOTHERAPY.
The use of postoperative adjuvant chemotherapy in patients at high risk for local and regional recurrence from HNSCC remains under evaluation.[181] Patients with two or more positive regional nodes, extracapsular extension of disease, positive resected margin, or perineural or perivascular invasion are considered to be in a high-risk category. Ang and colleagues noted a 5-year actuarial distant relapse rate of 33% (versus 3% in the low-risk group).[161] A comparable level of risk for distant relapses in various chemoradiotherapy series with patients treated nonsurgically has been described ranging from 15% to 30%. [135] [178] [212] Laramore and colleagues reported the results of a randomized trial conducted through the Intergroup (0034), randomizing patients after surgical resection to either three cycles of cisplatin and 5-fluorouracil chemotherapy followed by postoperative radiotherapy (CT/RT) or postoperative radiotherapy alone (RT). Patients were stratified as having either low-risk or high-risk treatment volumes depending on whether the surgical margin was greater than or equal to 5 mm, there was extracapsular nodal extension, and/or there was carcinoma in situ at the surgical margins. Radiation doses of 50 to 54 Gy were given to low-risk volumes and 60 Gy were given to high-risk volumes. A total of 442 patients were analyzable with no difference noted in the overall survival (4-year actuarial survival rate was 44% on the radiotherapy arm and 48% on the chemotherapy plus radiotherapy arm [P = not significant]), 4-year disease-free survival (38% versus 46%, respectively), and 4-year local-regional control rates (29% versus 26%, respectively). The overall incidence of distant metastases, however, was 23% on the radiotherapy arm, compared with 15% on the chemotherapy-radiotherapy arm (P = 0.03), again confirming activity as noted in trials employing neoadjuvant chemotherapy, but with no improvement in overall survival as noted in the recently updated meta-analysis reported by Pignon and colleagues.[181] A limited number of other randomized trials have been conducted. These include a small randomized trial of adjuvant chemotherapy for oral cavity lesions reported in abstract only that favored the control arm,[327] and a second trial also limited to the oral cavity that did not demonstrate any improvement in disease-free survival or overall survival.[328]
CHEMOPREVENTION.
Retinoids have been increasingly used in the treatment of oral leukoplakia and dysplasia of the head and neck since the 1960s. Studies have been published using retinoids in the treatment of head and neck cancers in conjunction with interferon. 13-Cis-retinoic acid also has been looked at in the adjuvant setting. More recently, retinoids have been evaluated in the “preventive” setting after definitive therapy.
A phase I/II study looked at cis-retinoic acid, cisplatin, and ifosfamide in patients with advanced or recurrent HNSCC.[329] Patients were given cisplatin at 20 mg/m2 per day for 5 days every 3 weeks, withcis-retinoic acid at 0.5 mg/kg orally for 5 days per week and dose-escalating ifosfamide at 1000 to 1500 mg/m2. A response rate of 72% was reported, with median time to progression of 10.4 months and overall survival time of 13 months.
The combination of retinoids and interferons has synergistic effects in modulating proliferation, differentiation, and apoptosis. A German study evaluated 30 patients after treatment for stage IV HNSCC in which adjuvant “chemopreventive” cis-retinoic acid and interferon were administered for 6 months.[330] The dose of cis-retinoic acid was 0.5 mg/kg per day given orally, and the dose of interferon was 3 million IU per week given subcutaneously. Sixteen patients remained disease free 1 year after definitive treatment. Associated side effects were weight loss, flushing, cachexia, worsening xerostomia, and dysphagia from cis-retinoic acid. Interferon side effects were reported as pyrexia and hematologic changes.
Based on the preceding data, a phase II study was conducted in this country using cis-retinoic acid, alpha-tocopherol, and interferon as adjuvant therapy in patients receiving definitive treatment for locally advanced HNSCC.[331] Three million units of interferon were given subcutaneously three times weekly with alpha-tocopherol at 1200 IU per day orally and cis-retinoic acid at 50 mg/m2 per day orally for 12 months. Forty-five patients were enrolled; 38 completed the 1-year trial. On follow-up evaluation at a median of 24 months, the local-regional failure rate was 9%, with 5% failing distantly. Median survival at 2 years was reported as 84%. This adjuvant regimen is being tested at the phase III level.
Celecoxib is a novel compound that specifically inhibits the inducible form of the enzyme cyclooxygenase (COX-2) (prostaglandin G/H synthase). Celecoxib is an oral anti-inflammatory agent indicated for the treatment of rheumatoid arthritis and osteoarthritis. Nonsteroidal anti-inflammatory drugs (NSAIDs) and related drugs such as COX-2 inhibitors are attractive candidates for prevention based on recent epidemiologic and case-control studies suggesting that the risk of several malignancies such as colon, esophagus, gastric, and bladder is reduced in chronic NSAID users. [332] [333] [334] [335] [336] Many tumors, both human and animal, express elevated levels of COX-2 compared with normal tissue. [337] [338] This elevation also is noted in premalignant lesions.[339] In the head and neck area, COX-2 overexpression is seen in oral leukoplakia as well as in squamous cell carcinomas. Increased levels of COX-2 can contribute to carcinogenesis by modulating xenobiotic metabolism, apoptosis, immune surveillance, and angiogenesis. In animal models, selective COX-2 inhibitors suppress the formation of tumors, including tongue cancer. Selective COX-2 inhibitors also can suppress the growth and metastasis of established tumors and enhance the anticancer activity of both radiation and chemotherapy agents. Celecoxib is now being evaluated for efficacy and safety as an adjunct in the prevention of cancer and the prevention of recurrence and metastasis after therapy. Thus, COX-2 inhibition may be a promising strategy to prevent and treat HNSCC.
Targeted Therapy and Novel Approaches
Novel biologic agents have been developed to target multiple specific regions of cancer cells. Protein tyrosine kinases are major components of cell signaling pathways. Various subfamilies of these kinases include receptors for the epidermal growth factor (EGFRs), platelet-derived growth factor (PDGF), vascular endothelial growth factor (VEGF), fibroblast growth factor, and hepatocyte growth factor. EGFR is one of four receptors involved in cellular proliferation, differentiation, and survival and is widely expressed in many malignant tissues. EGFR inhibitors such as anti-EGFR monoclonal antibodies, tyrosine kinase inhibitors, ligand conjugates, and antisense oligonucleotides have received significant attention in patients with HNSCC, because the EGFR commonly is overexpressed in 80% to 90% of patients.
EPIDERMAL GROWTH FACTOR RECEPTOR.
EGFR–erb-B1 is part of the erb-B family of receptor tyrosine kinases, which includes erb-B2/Her2/neu, erb-B3/Her3, and erb-B4/Her4. [340] [341] EGFR is composed of three domains: an extracellular ligand-binding domain, a transmembrane lipophilic region, and an intracellular protein tyrosine kinase domain. [342] [343] Endogenous ligands to EGFR include EGF, TGF-a, and heparin-binding EGF. When activated, phosphorylation of the intracellular tyrosine residues results in a cascade of protein phosporylations, resulting in turn in the activation of various downstream signal transduction pathways, including ras/MAP kinase, phosphatidylinositol-3 kinase, and STAT-3. The signal transduction pathway can lead to cell proliferation, tumor growth, and progression of invasion and metastasis signals. [344] [345] Based on its overexpression in many cell types (particularly HNSCC cell lines) and the fact that EGFR-based signals can mediate resistance to chemotherapy and radiotherapy, it has been hypothesized that inhibition of EGFR may result in a synergistic antitumor effect.
Numerous EGFR inhibitors have been evaluated, including anti-EGFR monoclonal antibodies, tyrosine kinase inhibitors, ligand conjugates, immunoconjugates, and antisense oligonucleotides. Small molecules such as the tyrosine kinase inhibitors target intracellular tyrosine kinase signaling and inhibit EGFR; antibodies are more directed at the extracellular domain.
IMC-C225 (ImClone Systems, Somerville, New Jersey) is a monoclonal antibody targeting the EGFR and has been studied in several tumor types. In vitro, this antibody appears to enhance the antitumor activity of chemotherapy such as cisplatin (CDDP) and doxorubicin as well as the radiosensitivity of HNSCC cell lines.[346] Similar in vitro data have shown C225 to enhance radiosensitivity. [347] [348]
Recent evidence now supports a role for EGFR to mediate a cytoprotective stress response to ionizing radiation. Therapeutically relevant doses of ionizing radiation have been demonstrated to increase expression of EGFR[349] and cytosolic release of TGF-μ from its membrane-bound form,[350] modulate an activated EGFR autophosphorylation profile, [350] [351] [352] [353] with increased mitogenic signals involving the MAPK pathway with cellular proliferation, [352] [354] and protect from radiation-induced cell death.[350] In a series of experiments with the radioresistant vulva squamous carcinoma cell line A431, inhibition of EGFR with the tyrosine kinase inhibitor tyrphostin AG1478 was associated with reduced MAPK activation and reduced EGFR-mediated tumor cell proliferation.[352] When these findings are taken collectively, it is intriguing to postulate that such a response may serve to mediate the clinical phenomenon of accelerated tumor repopulation during radiotherapy for HNSCC.[152] This hypothesis may be particularly relevant in HNSCC in light of the significant body of clinical data supporting the notion of tumor clonogen repopulation contributing to radiation failure. [179] [238]
Early phase I trials of C225 alone and in combination with CDDP were reported. Fifty-two patients were initially treated with dose-escalation of C225 as well as CDDP weekly.[355] C225 was dose escalated to 200 to 400 mg/m2, but CDDP dosing was found to be optimal only at 60 mg/m2, owing to toxicity at higher doses. One patient experienced a humoral response. Toxic manifestations included acneiform rashes, gastrointestinal distress, seborrheic dermatitis, flushing, asthenia, and transaminitis. This led to subsequent studies in the metastatic setting.
In a phase I trial, 12 patients with metastatic HNSCC with high levels of EGFR expression were given three doses of C225 with 100 mg/m2 of CDDP every 3 weeks.[356] Weekly maintenance of C225 was given at 250 mg/m2. Responses were seen in 67% of the 12 patients enrolled, with minimum toxicity. The specifics for a follow-up study of the use of C225 plus CDDP in patients with recurrent HNSCC were presented at the American Association for Cancer Research (AACR) in 2001. Sixty-three patients were evaluated and given a loading dose of 400 mg/m2 of C225, followed by weekly maintenance at 250 mg/m2, plus CDDP at 75 or 100 mg/m2. The overall response was 24%. Use of C225 by itself has produced response rates of 10%, which is encouraging. ECOG recently compared single-dose CDDP and CDDP plus C225 in a doubleblind randomized trial in patients with previously untreated recurrent or metastatic HNSCC, with 35% to 40% stable disease in both treatment groups.[357]
Anti-EGFR antibodies have radiosensitization properties in cell cultures and preclinical animal studies. A phase I study with C225 and radiation in 16 patients with locally advanced disease (13 with stage IV) was recently reported with an overall clinical response rate of 100% and a complete response rate of 87%.[358] The impressive complete response rate has been interpreted to be consistent with radiosensitization. A multi-institutional phase III trial comparing radiation therapy alone with radiation therapy plus C225 in locally advanced HNSCC has recently closed to accrual with results currently being analyzed.
TYROSINE KINASE INHIBITORS.
ZD1839 (Iressa, AstraZeneca Pharmaceuticals LP, Wilmington, Delaware) is a selective EGFR tyrosine kinase inhibitor, recently approved by the U.S. Food and Drug Administration (FDA) as palliative therapy in non-small-cell lung cancer. Preclinically, ZD1839 potentiates the antitumor and apoptotic effects of several cytotoxic agents, including CDDP and taxanes. [359] [360] In combination with radiation, ZD1839 shows dose-dependent inhibition of cellular proliferation in human SCC cell lines. ZD1839 also can inhibit tumor angiogenesis in tumor zenograft models in vivo.[348]
Four phase I trials using ZD1839 as monotherapy mostly in patients with prostate, lung, colorectal, ovarian, and head and neck cancer have reported modest toxicity such as diarrhea, nausea and vomiting, transient transaminitis, and rash. Most patients had been heavily pretreated and most were lung cancer patients. The results of the non-small-cell lung cancer monotherapy trial were presented. The overall response rate in 208 patients was 53% and the median progression-free survival rate was 84 days.[361] This study, as well as several other trials, led to recent FDA approval.
The role of ZD1839 as a radiosensitizer is the subject of an ongoing NCI-sponsored multi-institutional trial in combination with a delayed concomitant boost accelerated radiotherapy schedule and with concurrent chemoradiotherapy. The chemoradiotherapy incorporates a weekly cisplatin schedule. The use of oral tyrosine kinase inhibitors may be particularly attractive for its flexible daily dosing schedules. It is hypothesized that this may permit continued arrest of tumor repopulation during any unplanned radiotherapy treatment interruptions. Accordingly, this clinical trial seeks to continue daily dosing of ZD1839 during these unplanned interruptions and during the weekends in light of the provocative positive results with an accelerated schedule that delivered radiotherapy through the weekends.[238]
A phase II trial in recurrent or metastatic squamous cell carcinoma of the head and neck also was recently published.[362] Fifty-two patients who previously received treatment with only one other strategy were allowed to enroll. Patients were given the higher dose of 500 mg per day orally; 250 mg per day is the approved dose based on phase I data in non-small-cell lung cancer. Patients were able to tolerate this drug through feeding tubes. The observed response rate was 10.6%. Median time to progression and overall survival time were 3.4 and 8.1 months, respectively.
OSI-774 (Genetech, San Francisco, California) also is an orally active quinazoline and potent selective inhibitor of EGFR tyrosine kinase. This results in cell cycle arrest at G1. Phase I studies with OSI-774 were well tolerated, again with toxicities including diarrhea, skin rash, and gastrointestinal upset. OSI-774 in a dose of 150 mg per day has been tested in patients with HNSCC refractory to chemotherapy. In a phase II trial in 114 patients, 13% of patients had a partial response and 29% exhibited disease stabilization.[363] Toxicities were as previously noted and included acneiform rash, diarrhea, nausea, vomiting, headache, and fatigue.
RAS INHIBITORS.
It has been suggested that oral cavity cancers have a 27% mutation rate in H-ras.[364] Farnesyl transferase inhibitors (FTIs) inhibit a critical enzymatic step in the post-translational modification of ras, allowing the constitutive expression of mutated ras genes. Several FTIs are available in the study setting: R115777 (Janssen Pharmaceutica); BMS214662 (Bristol-Meyers Squibb); and SCH66336 (Schering Plough, Kenilworth, New Jersey).
R115777 is a nonpeptidomimetic orally available FTI. Clinically, it can be used alone or in combination with chemotherapy. Toxicity has been minimal. Reversible myelosuppression is the most common dose-limiting toxicity, as well as fatigue, nausea, renal dysfunction, and peripheral neuropathy. A phase I trial using R115777 with irinotecan was reported.[365] Another trial with docetaxel is being conducted, as is a third with gemcitibine. [366] [367]
BMS214662 has preferential cytotoxicity against nonproliferating cells. Combination chemotherapy from preclinical human colon cancer cell lines has exhibited synergy with paclitaxel, irinotecan, gemcitabine, and an epothilone analog, BMS247550.[368] Phase I studies are ongoing in advanced solid tumors, including using this drug with CDDP at 75 mg/m2 every 3 weeks.
p53 TARGETS.
As discussed elsewhere in this chapter, p53 mutations occur in 45% to 70% of patients with HNSCC and are associated with continual tobacco and alcohol use. [29] [30] p53 is a multifunctional protein that can be induced by DNA damage and plays a significant role in the detection and repair of damaged DNA. P53 can induce apoptosis in severely damaged cells and has been associated with both carcinogenesis and poor prognosis in many cancers, including HNSCC.
ONYX-015 is an E1B-55 kD gene-deleted replication-selective adenovirus that replicates and causes cytopathogenicity in certain cancer cell lines. [369] [370] Selective intratumoral replication and tumor-selective tissue destruction of ONYX-015 have been demonstrated in phase I and II trials in patients with refractory or recurrent HNSCC. [371] [372] Clinical benefit was seen in 15% of patients. A phase II multicenter trial of intratumoral ONYX-015 in combination with CDDP and 5-fluorouracil in patients with recurrent HNSCC was reported.[373] Forty patients received injections, 30 for 5 consecutive days and 10 twice daily for 2 weeks. Responses and stable disease were noted in both groups. Pain was noteworthy in the twice-daily injection group, but otherwise this was well tolerated.
Thus, novel modalities involving molecular targets are actively being investigated. Cytotoxic agents have limited efficacy as demonstrated throughout many sections of this chapter, so targeted therapies may improve treatment in the setting of recurrent and metastatic disease and, eventually, for definitive treatments.
Site-Specific Treatment Considerations
Nasopharynx
The nasopharynx is a cuboidal structure bounded by the sphenoid bone superiorly, the posterior choanae anteriorly, the clivus and the first two cervical vertebrae posteriorly, and the soft palate inferiorly. The eustachian tube enters through the lateral wall, with the posterior portion of the tube being cartilaginous and forming the portion of the lateral nasopharyngeal wall known as the torus tubarus. Just posterior to this is the fossa of Rosenmuller.
The vast majority of malignancies in the nasopharynx are epithelial neoplasms and arise from the lateral wall, particularly from the fossa of Rosenmuller. Local spread may include extension anteriorly through the submucosa including the nasal cavity, laterally and superiorly through the foramen lacerum with cranial nerve involvement, and inferiorly into the oropharynx. Extension into the cavernous sinus commonly results in a sixth cranial nerve palsy. Two cranial nerve syndromes have been characterized. The petrosphenoidal syndrome describes involvement of the third, fourth, fifth, and sixth cranial nerves. The retroparotidian syndrome describes involvement of cranial nerves IX, X, XI, and XII. Metastatic spread to the adjacent upper cervical lymph nodes and to the retropharyngeal lymph nodes that are located in the retropharyngeal space that lies between the lateral border of the posterior nasopharyngeal wall and medial to the carotid artery may extend inferiorly to the level of the hyoid bone.
Both the tumor stage and the histologic grade of epithelial malignancies are prognostic. The World Health Organization (WHO) identifies three histopathologic types: type 1, differentiated; type 2, nondifferentiated; and type 3, undifferentiated or lymphoepithelial, which highlights the presence of numerous infiltrating lymphocytes. The presence of keratin is an adverse prognostic feature for local control and overall survival. Other malignancies may include lymphoma, plasmacytomas, melanomas, and, in the pediatric population, juvenile angiofibromas and rhabdomyosarcomas. Approximately 60% to 90% of patients with nasopharyngeal cancer present with palpable adenopathy and up to 50% of patients with involved nodes have bilateral disease. [374] [375] [376] Patients with adenopathy at the mastoid tip require exclusion of a malignancy in the nasopharynx due to the characteristic lymphatic drainage pattern from the retropharyngeal space. Staging follows the sixth edition of the AJCC TNM criteria. No modifications have been recommended, including subdivision of the T4 stage that is new to this sixth edition.[137] Both MRI and CT scans are complementary in this tumor site, with the former favored in most cases and the latter beneficial when bone invasion or destruction has occurred.
Because of the anatomic location of the nasopharynx, surgical resection typically has not been recommended owing to the inherent surgical complication rates with surgery in this area, including the inability to achieve tumor-free margins. Accordingly, radiation therapy is the treatment of choice. Fortunately, NPCs are both sensitive and responsive to radiotherapy and chemotherapy, the two principal treatment modalities that are used. For early-stage disease, radiotherapy alone may be used, with excellent results reported, with 3-year overall survival rates ranging from 70% to 100% [374] [377] [378] [379]and 65% to 100% [374] [377] [378] [379] for stage I and II disease, respectively, based on the AJCC 1997 staging system. Typical local control rates to current treatments have been summarized by Lee[380]: approximately 80% (72% to 90%) for T1 and approximately 70% (5% to 100%) for T2. Even with locally confined disease (T1), generous initial radiotherapy margins are advised.
Unfortunately, a majority of patients present with locally advanced disease. Local control for T3–4 disease may be expected to be approximately 50% (39% to 72%).[380] For locally advanced-stage III and IV disease, concurrent chemoradiotherapy has emerged as the standard treatment option following from the results of the Intergroup 0099 (IG0099) study reported by al-Sarraf and coworkers.[381] These investigators reported the results for 147 evaluable patients of 193 registered randomized to receive radiotherapy alone or radiotherapy with concomitant cisplatin (100 mg/m2 given intravenously on days 1, 22, and 43) followed by adjuvant chemotherapy with cisplatin 80 mg/m2 on day 1 and fluorouracil 1000 mg/m2 per day on days 1 to 4, administered every 4 weeks for 3 courses. An improvement in the 3-year progression-free survival rate (69% versus 24%; P < 0.001) and overall survival (76% versus 46%; P = 0.001) was observed on interim analysis, prompting premature study closure. These results are not without debate, however.[380]
Several investigators have questioned whether this positive study may have resulted from inferior outcomes in the control arm rather than a therapeutic effect. Chow and coworkers from the Princess Margaret Hospital recently published the results of a large institutional series using radiotherapy alone administered in a homogeneous manner to 172 patients with advanced-stage disease, demonstrating 5-year disease-free survival and overall survival rates of 48% and 62%, respectively.[382] Although these investigators acknowledge that direct comparisons with the control arm of IG0099 have limited validity, these observations made during a similar time period as the IG0099 provide a context within which the positive results of the IG0099 should be interpreted. In a similar report by Cooper and colleagues, 86 patients with locally advanced disease treated with radiotherapy alone had a 3-year actuarial disease-free survival and overall survival rates of 43% and 61%, respectively.[383] These investigators also observed that in 35 patients undergoing the IG0099 protocol, the projected 3-year disease-free survival and overall survival rates were 63% and 93%, respectively, suggesting benefit. Recently, Cheng and colleagues reported the results for 107 patients with NPC treated with concurrent 5-fluorouracil and cisplatin (weeks 1 and 6) and radiotherapy followed by two cycles of adjuvant 5-fluorouracil and cisplatin. The 5-year overall survival rate, disease-free survival rate, and local-regional control rate were 84.1%, 74.4%, and 89.8%, respectively. The 3-year overall survival rates for stage II, III, and IV were 100%, 92.8%, and 69.4% (P = 0.0002), and the 3-year disease-free survival rates were 96.9%, 87.7%, and 51.9% (P = 0.0001). Although a confirmatory randomized trial is noted to be under way,[384] the current consensus opinion continues to recommend concurrent chemoradiotherapy. These results further suggest that Asian patients also may benefit from concurrent chemoradiotherapy, which has been questioned owing to geographic differences in the histologic subtype between North America and Asia. A pilot trial of a modified schedule of the IG0099 regimen (reduced cisplatin dose) has been demonstrated to be safe and has given rise to a phase III trial to confirm the generalizability of these results to the Asian population.[384]
With concurrent chemoradiotherapy, toxicities clearly are increased. In addition, the late toxic effects with concurrent chemoradiotherapy have not been well described. Increased acute toxic effects include not only mucosal toxicity, further compounded by the large volume of normal mucosa that is irradiated, but also the asthenia and emetogenic side effects of concurrent high-dose cisplatin. As Cooper describes,[383] the observations by Cheng and colleagues may suggest that other alternative regimens may be equally if not more effective. The impact of these toxicities is important to consider in patient management, because treatment interruptions during radiotherapy also have been demonstrated to result in inferior local-regional control and disease-free survival rates.[385] These investigators estimated that local-regional relapses increased 3.3% per day of treatment interruption. The timing of the interruption, including interruptions early in the course of treatment, appeared to be equally detrimental. Although it is unclear if NPC tumor kinetics may be different with concurrent administration of chemotherapy, close attention to the toxicities of treatment and causes for interruption during a course of radiotherapy may be prudent.
One attractive strategy to minimize toxicities associated with concurrent chemoradiotherapy has been to employ lower but more frequent doses of chemotherapy. Chan and colleagues recently reported the results of a randomized trial of concurrent weekly cisplatin (40 mg/m2) with daily fractionated radiotherapy in 350 patients.[386] Although no improvement was observed overall for the primary endpoint of progression-free survival, subgroup analysis demonstrated benefit in Ho's T3 stage owing to an improved time to first distant failure. The treatment was well tolerated. The results of a second randomized trial of concurrent 5-fluorouracil and cisplatin, however, showed a significant improvement in progression-free survival.[387]
Despite the improved outcome that appears to have been realized with concurrent chemoradiotherapy for locally advanced NPC, a proportion of patients may continue to have persistent disease that is slow to respond. The therapeutic options recommended have been further irradiation and observation. A limited number of studies have reported on the role of an implant in patients with persistent disease after standard therapy. [388] [389] [390] [391] [392] [393] These studies have suggested that further irradiation in early-stage disease that is amenable to intracavitary and interstitial techniques may result in local control rates comparable to those achieved in patients demonstrating a prompt complete response.[389] Hence, dose escalation may be adequate in compensating for tumors demonstrating a low radioresponsiveness. The optimal dose schedule remains to be determined, with both 60-Gy LDR and 22.5- to 25-Gy HDR schedules reported. Stereotactic radiosurgery has been used with preliminary data, suggesting that up to 70% of patients with organ-confined recurrences may achieve local control for up to 2 years.[394] Late toxicity does not appear to be increased. In light of the poorer local control rates and survival rates in patients managed for local recurrences, a brachytherapy implant or stereotactic radiosurgery may be considered in the management of patients demonstrating persistent disease.
Neck control for NPC is unique for its increased sensitivity to radiotherapy compared with the typical HNSCC.[395] In the clinically negative neck, elective irradiation is recommended owing to the high probability of subclinical nodal metastases, with a risk of 40% neck relapse if untreated, which is associated with a significantly higher incidence of distant failure (21% versus 6%) despite successful nodal salvage. [396] [397] With appropriate doses of radiation (50 Gy), the probability of relapse is less than 5%.[398] In the clinically positive neck, doses of 60 Gy or greater are typically recommended; average regional control rates of 90% (range, 86% to 96%) and 75% (range, 71% to 87%) might be expected for neck nodes 3 cm or less and greater than 6 cm, respectively.[380] Accordingly, a planned neck dissection typically is not recommended but may be appropriate. When a residual abnormality remains, a neck dissection typically is recommended. This approach may be complicated, however, by the recognition that in some NPC lesions, gross disease response can be slow at both the primary site and in the neck. As well, the integration of a neck dissection can create conflicts with regard to the prompt initiation of adjuvant chemotherapy. In light of the association between local-regional relapses and the subsequent risk of distant metastases,[397] a neck dissection continues to be a prudent recommendation at this time.
Various strategies have been used in attempts to improve upon the survival rates, particularly for locally advanced disease. The first is to optimize local-regional control, not only because this is a major pattern of relapse in this group of patients but also for the potential impact on the risk of distant metastases. The second is the integration of various chemotherapy agents with definitive radiotherapy. Because a dose-response relationship appears to exist for local control,[399] and the overall treatment time[385] also appears to have an adverse impact on treatment outcome, several institutions have reported their results using altered fractionation for locally advanced NPC with[400] or without concurrent chemotherapy.[401] Jian and colleagues used the same chemotherapy regimen as that described by Cheng and colleagues, with cisplatin and 5-fluorouracil on weeks 1 and 6, with concurrent hyperfractionated radiotherapy to a total dose of 74.4 Gy, in 48 patients. With a median follow-up period of 57 months, 3-year local-regional control rate was 93%, the disease-free survival rate was 71%, and the overall survival rate was 72%. In particular, patients with T4 disease had a 3-year local-regional control rate of 91%, disease-free survival of 62%, and an overall survival of 63%. The major acute toxicity was grade 3 mucositis in 73% and grade 2 weight loss in 31% of patients. These investigators concluded that the treatment was well tolerated, with 88% of patients completing their radiation treatment within 8 weeks.[400] These promising results require further validation of the results and definition of the increased toxicity risks, particularly in light of the unacceptable toxicities observed in other non-nasopharyngeal concurrent chemoradiotherapy series employing altered fractionation.[326]
Various randomized trials incorporating various sequences of chemotherapy and radiotherapy have been reported. To date, a number of randomized trials have failed to demonstrate a benefit with either neoadjuvant [402] [403] [404] [405] [406] or adjuvant chemotherapy. [404] [407] [408] These results are consistent with a recent large patient-based meta-analysis.[181] Accordingly, the use of adjuvant chemotherapy, particularly with definitive chemoradiotherapy, should be judiciously applied.
Because NPCs are radiosensitive, locally recurrent carcinomas may be amenable to reirradiation. The risk of late complications, including soft tissue and brain necrosis and neuropathies, is highly dose- and volume-dependent with reirradiation, so the general strategy has been to incorporate conformal radiotherapy techniques for the boost component of the reirradiation regimen. As with treatment for persistent disease, this may take the form of either brachytherapy or stereotactic radiosurgery boost.[409] In a series of over 891 patients undergoing reirradiation, the extent of the recurrence was prognostic. Overall, approximately 30% achieved local disease control with local control best seen when a repeat radiation dose of 60 Gy or greater was delivered.[410] The selection of small EBRT fraction size and the use of a brachytherapy implant was associated with a reduced risk of late complications. Several other institutional series have reported sustained local control rates of 20% to 60%, with the variability due to the extent of initial disease presentation and at recurrence and the dose of reirradiation. [388] [390] [391] [392] In selected series treating only disease confined to the nasopharynx mucosa amenable to management by either intracavitary or interstitial implant, sustained local control rates of 50 to 60% may be realized. These series also have demonstrated, however, a significant risk of developing late radiation-related complications including soft tissue and bone necrosis, trismus, fistula formation, and neurologic complications such as radiation myelitis and temporal lobe necrosis.
Salvage surgery is now being performed in Taiwan for NPC in selected patients. Large prospective studies are needed to determine efficacy, but small retrospective studies have demonstrated feasibility and success in local control. [411] [412] Hsu and colleagues studied 60 patients who underwent salvage surgery, and showed that the results of surgical resection in terms of local control and overall survival were slightly better than those in patients undergoing high-dose reirradiation for local relapse, with fewer late complications.[412] These investigators favored salvage surgery for rT1–2 and limited rT3 disease, owing to the lower complication rate. This therapeutic option may be considered for selected lesions and in appropriate centers with expertise in this technique, as its relative efficacy to high-dose reirradiation remains to be defined.
Locally recurrent and metastatic nasopharyngeal cancers generally remain chemosensitive. Many patients have been previously treated with combined modality strategies involving a platinum agent and 5-fluorouracil. Several phase II trials have evaluated carboplatin and paclitaxel in this setting. Carboplatin at an area under the curve (AUC) value of 7 with 3-hour infusional paclitaxel at 200 mg/m2 revealed an overall response rate of 57% in the metastatic setting.[413] Another phase II study investigated carboplatin at AUC 6 with 135 mg/m2 paclitaxel infused over over 3 hours, with an overall response rate of 59%.[414] Two other phase II studies, again using similar dosing schedules of carboplatin AUC 6 or 5.5 with 175 mg/m2 paclitaxel, demonstrated 75% and 25% response rates, with median overall survival rates of 12 and 9.5 months, respectively. [415] [416]
Paranasal Sinus and Nose
Malignant tumors of the sinonasal tract are relatively rare, constituting approximately 3% of upper respiratory tract cancers. Accordingly, the scientific literature is limited to multiple small retrospective reports describing the treatment and outcome of these patients. This has limited progress in defining the optimal management for these malignancies. Although squamous cell carcinoma is the most common, several other types of epithelial tumors are less commonly found, including melanoma, adenocarcinoma, adenoid cystic carcinoma, and esthesioneuroblastoma.[417] Nonepithelial malignancies arising in the sinonasal tract include sarcomas and lymphoma.
Among the different subsites within the paranasal sinuses, squamous cell carcinomas occur most commonly in the maxillary sinus.[418] The second most common location is the nasal cavity. The ethmoid sinus is more frequently the site of adenocarcinoma or esthesioneuroblastoma. Cancer arises rarely in the frontal and sphenoid sinuses.[418] The prognosis with these lesions is based on staging classification, as well as on their relationship to Ohngren's line. This theoretical plane extends from the medial canthus of the eye to the angle of the mandible. Tumors anteromedial to this plane are thought to have a considerably better prognosis. Staging follows the sixth-edition AJCC staging criteria. The nasoethmoid complex was added as a tumor site with subdivision of the T4 stage for both maxillary sinus and nasal cavity and ethmoid sinus.[137]
Most malignancies are advanced at presentation and commonly involve one or more adjacent structures ( Fig. 72-17 ). Orbital invasion often occurs early with cancers of the maxilla and of the ethmoid sinuses, whereas it often is a late event for nasal cavity tumors. Malignancies beginning in the anterolateral infrastructure of the maxilla often erode through the inferolateral wall and extend into the oral cavity, with involvement of the maxillary gingival or the adjacent gingivobuccal sulcus. In general, the risk of cervical nodal metastases is low unless the tumor has progressed to involve mucosal surfaces with abundant lymphatics such as the oral cavity.
|
Figure 72-17 Routes of spread of cancer of the paranasal sinuses. Solid arrows, maxillary sinus cancer; dashed arrows, ethmoid sinus cancer. |
Management of paranasal sinus malignancies is primarily surgical, with adjunctive irradiation and possibly chemotherapy for advanced lesions. [419] [420] For maxillary sinus tumors, a partial or total maxillectomy is required to excise the tumor with disease-free margins, depending on its location. The maxilla can be accessed through a variety of approaches. For smaller, medially based tumors, a medial maxillectomy can be performed using a mid-face degloving approach, in which incisions are made under the lip. For larger lesions, well-placed skin incisions in the nasal crease and upper lip often are required for access. Reconstruction in these cases usually involves a skin graft or acellular dermal graft to reline the mucosal surface, as well as a dental appliance to recreate the hard palate.
Tumors that involve the ethmoid sinuses frequently require a craniofacial resection for surgical access because of the proximity to the skull base. This procedure requires not only an anterior approach to the sphenoethmoid area but also a craniotomy by a skilled neurosurgeon to address the skull base and dura. Orbital exenteration must be considered if the tumor involves the periorbital fat or extraocular muscles. The surgical indications for this vary, however, with an inclination toward eye conservation and an evolving consensus that bone erosion is not an absolute indication. The decision is an intraoperative decision.
When surgical resection is not feasible, either medically or surgically, definitive radiotherapy may be used. It also has been favored as an organ preservation strategy to avoid an orbital exenteration. In this capacity, it may be used definitively, reserving surgery for salvage,[421] or as preoperative radiation as a strategy to downstage the tumor. In view of the proximity of these cancers to many critical normal structures, new radiotherapy techniques including IMRT, stereotactic radiosurgery, or fractionated stereotactic radiotherapy are recommended because they are likely to improve the therapeutic ratio. These techniques are an important consideration in light of evidence suggesting a potential dose-response relationship, with doses greater than 65 Gy recommended.[422] Other technical considerations include the use of computer image fusion software that can permit the use of MRI and the superiority of delineating soft tissue disease during the radiotherapy treatment planning process. Although the current literature does not necessarily support the potential incremental therapeutic efficacy, this is due to the relative rarity of this disease and not to the absence of any effect. As well, the value of concurrent chemotherapy as a radiosensitizer, including early interest in the use of intra-arterial chemotherapy, remains to be established.[423] The value of elective management of the neck remains unclear. Because this often increases the treatment morbidity, it may be omitted. Where considered, most nodal metastases occur in the level I and II regions.
The overall 5-year survival rate for patients with maxillary sinus cancer is 30% to 50%. [422] [424] [425] Cervical metastases occur in less than 10% of maxillary sinus cancers.[426] Therefore, a prophylactic neck dissection is not indicated in the N0 neck. Cervical metastases are associated with a very poor prognosis, with 5-year survival rates less than 10%.
Oral Cavity
The oral cavity ( Fig. 72-18 ) is composed of the lip, anterior two thirds of the tongue (oral tongue), floor of mouth, buccal mucosa, gingiva, hard palate, and retromolar trigone. The floor of the mouth is bounded by the lower alveolar ridge anteriorly and laterally and the ventral tongue surface and anterior tonsillar pillar posteriorly. The oral tongue lies anterior to the circumvallate papillae. The buccal mucosa overlies the buccinator muscle, is bounded superiorly and inferiorly by the gingiva, and extends posteriorly to the retromolar trigone. The gingiva is the soft tissue overlying the alveolar ridges of the mandible and maxilla. The hard and soft palates form the roof of the mouth. The retromolar trigone mucosa overlies the mandibular ramus and is bounded anteriorly by the buccal mucosa and posteriorly by the anterior tonsillar pillar. As elsewhere in the head and neck, squamous cell cancer is the most common type, except in the hard palate, where most tumors originate in the minor salivary glands. For staging of cancers of the oral cavity, the AJCC classification system is used ( Table 72-8 ). The principle management approach is surgical resection followed by postoperative radiotherapy. Carcinomas of the oral cavity have an adverse prognosis and often require postoperative radiotherapy.[162]
|
Figure 72-18 Anatomy of oral cavity and oropharynx. A, Open-mouth view. B, Tongue elevated, showing floor of mouth. C, Sagittal view. |
Table 72-8 -- American Joint Committee on Cancer Tumor Classification System for Staging Cancer of the Oral Cavity
|
Stage |
Characteristics |
|
Tis |
Carcinoma in situ |
|
T1 |
Tumor 2 cm or less in greatest dimension |
|
T2 |
Tumor more than 2 cm but not more than 4 cm in greatest dimension |
|
Tumor more than 4 cm in greatest dimension |
|
|
T3 |
Tumor invades adjacent structures (e.g., through cortical bone, into deep [extrinsic] muscle of tongue, maxillary sinus, skin) |
From the American Joint Committee on Cancer: Manual on Staging of Cancer, 4th ed. Philadelphia, JB Lippincott, 1992, with permission.
LIP.
Surgical management of squamous cell carcinoma of the lip is complex, due to the challenges in reconstruction of this unique part of the body. Although secondary to the complete resection of the lesion, the preservation of speech, oral competence, and cosmesis must be considered. Surgical therapy is considered to be equal in effectiveness to irradiation for the treatment of early, T1, or T2 lesions. [427] [428] [429] [430] Small lesions of the lip can be managed surgically by means of wide local excision and primary closure. The morbidity associated with this approach often is less than would be experienced with irradiation. With larger lesions that require resection of more than one half of the lip, local flap reconstruction will be necessary. These flaps involve the mobilization of remaining lip tissue, or even the use of tissue from the opposite lip.
Metatases from lip carcinoma is relatively rare, with the incidence reported at 12% or less. [431] [432] In the N0 neck, occult metastases are estimated to occur in 5% to 10% of cases. Therefore, elective neck dissection is not routinely performed in the N0 neck. Neck dissections are generally performed when cervical metastases are clinically or radiographically apparent.
Predictably, the results of these procedures depend on the extent of disease. The 5-year survival rate for T1 and T2 cancers of the lip is greater than 90%, whether treated with surgery or irradiation.[428] For larger cancers, especially those that have metastasized, the cure rates are in the range of 40% to 50%. [428] [433] Thus, for stage III and IV cancers, a combined approach with surgery and postoperative radiation is indicated. For smaller lesions, surgery and irradiation are equally effective; however, surgery often is recommended because of shorter treatment time and excellent rehabilitation with minimal morbidity.
BUCCAL MUCOSA.
Malignant tumors arising from the buccal mucosa are rare, and early lesions are treated primarily with surgery. Although surgical management of T1 buccal mucosa lesions can be managed with transoral wide local excision, larger tumors may require more complex resections. Extension to the mandible or maxilla may lead to a partial mandibulectomy or maxillectomy. Reconstructive options for smaller lesions include primary closure, fat grafting, or a split-thickness skin graft. Larger defects would require local mucosal flaps, myocutanteous rotational flaps, or free flaps, depending on the size and extent of the lesion. Neck dissections are indicated only for clinically positive cervical metastases. The treatment algorithm for carcinoma of the buccal mucosa is similar to that for other sites in the oral cavity, with surgery indicated for early lesions and combined therapy with surgery and irradiation for advanced lesions.
ORAL TONGUE.
Oral tongue carcinomas represent approximately 25% of oral cavity carcinomas. These lesions are characterized by early infiltration into the underlying tongue musculature with an early and high risk for regional metastases. Surgical resection has been favored, because EBRT results alone have been disappointing. [434] [435] Although it has been shown that irradiation and surgery are equally effective in treating early lesions,[436] a recent review of 332 patients revealed that disease-free survival was better with surgery alone than irradiation alone.[435] In general, surgery is often the treatment of choice for early lesions because of the ease of surgical access, excellent reconstruction options, and quick treatment time. For stage III and IV lesions, a combined treatment approach is favored. For select lesions, various series describe comparable results with the use of EBRT and a brachytherapy implant as definitive therapy.
Surgical management for squamous cell carcinoma of the oral tongue can take a variety of forms depending on the size and location of the lesion. Small T1 or T2 tumors are treated with a partial glossectomy accomplished through a transoral approach. Larger lesions may require a total or near-total glossectomy, often requiring a mandibulotomy or a cervical pull-through procedure. T4 lesions that involve the mandible require composite resection including either a marginal or segmental mandibular resection.
Reconstruction of the surgical defects also depends on their size and location. After resection of smaller lesions, allowing healing through secondary intention may best preserve function. For larger defects resulting from resection of T2 or T3 defects, healing is facilitated by primary closure, or split-thickness skin graft. For near-total or total glossectomy defects, a pedicled myocutaneous flap or free-tissue transfer is needed for reconstruction. For those patients with mandibular involvement, reconstruction options vary with the location of the defect. Mandibular defects that are large and are located near the mandibular symphysis often require vascularized composite free flaps to restore mandibular continuity with acceptable function and cosmesis.
Surgical or radiation treatment of the neck is indicated in most cases of oral tongue squamous cell carcinoma. The rates of occult cervical metastases exceed 30% for lesions T2 and greater. [277] [278] [433] [437] [438] Therefore, for the N0 neck, a selective neck dissection is indicated in T2–T4 lesions, or T1 lesions with depth of invasion greater than 3 to 4 mm. [139] [438] [439] [440] Patients with clinical or radiographic evidence of cervical metastasis may require more radical procedures depending upon the number and size of nodal involvement.
A significant body of literature exists supporting a role for brachytherapy in the management of selected oral tongue carcinomas that tend to be well defined and thus probably less infiltrative. Brachytherapy has been used alone or in combination with EBRT, demonstrating local control efficacy. The largest experience, comprising more than 600 patients, from the Curie Institute reported local control rates of 86%, 78%, and 71% for T1, T2, and T3 lesions, respectively.[277] Early T1 and T2 lesions were treated with temporary interstitial LDR 192Ir implants alone, delivering 70 Gy in 6 to 9 days. Larger T2 and T3 lesions were treated with EBRT (50 to 55 Gy), followed by an implant (20 to 30 Gy). Other investigators have demonstrated comparable results, demonstrating a high rate of local control.[280] [296] [441] [442] [443] [444] [445] Several series reported local control rates of 90% or greater for very selected lesions often amenable to a single-plane implant alone with a lesion thickness of less than 1 cm.[296] [441] Mazeron and colleagues reported their series of 121 patients with T1 or T2N0 tumors treated with 60 to 70 Gy by the Paris system.[442] The crude local control rates for T1, T2a (2.1 to 3 cm), and T2b (3.1 to 4 cm) reported were 86%, 89%, and 74%, respectively. The dose prescribed was found to be significantly associated with the risk of local control, with doses less than 65 Gy associated with a fivefold risk of relapse. Selection by the growth pattern has been shown to influence the 5-year local control rates, with 85%, 79%, and 45% reported for superficial, exophytic, and infiltrative lesions, respectively.[445] Implant of the oral tongue has been associated with a 10% to 20% risk of mild to moderate self-limiting soft tissue ulceration and a low risk (less than 10%) of mandibular osteoradionecrosis in experienced hands. Custom lead-embedded mandibular prostheses and spacers are recommended and have been demonstrated to reduce the risk of bone complications.[441] A predicted 5-year probability of osteoradionecrosis of 38% was reduced to 4% with the use of a spacer. When implant brachytherapy is used in combination with EBRT, the overall treatment time and the proportion of dose delivered with the implant may be important.[446]
Several studies suggest that treating selected early-stage T1 and T2 node negative lesions with HDR brachytherapy alone may facilitate treatment delivery. [296] [444] [447] One promising schedule comes from a small randomized trial of 29 patients comparing LDR brachytherapy (70 Gy over 4 to 9 days) with an HDR schedule (60 Gy in 10 fractions of 6 Gy per fraction delivered twice daily over 6 days) for a selected group of patients with T1 or T2N0 squamous cell carcinoma of the lateral oral tongue.[444] The lesions had a thickness of 10 mm or less, allowing treatment with a single-plane HDR implant with the dose prescribed at 0.5 cm from the reference plane. The 1-year local control rates were 86% and 100% (P = 0.157) in the LDR (N = 15) and HDR (N = 14) groups, respectively. The 2-year local control rate was identical, although the median follow-up period of 24 months (10 to 32 months) limits this observation. One soft tissue ulceration and one bone exposure complication arose in the HDR arm, although a prosthetic spacer was not used in the latter case. Leung and colleagues have reported preliminary results in 8 patients subjected to the same HDR treatment schedule, showing a 100% local control rate with a median follow-up period of 26 months.[447] Alhough these results are promising, the short follow-up coupled with the significant risk of a false-negative error limits any definitive conclusions regarding the generalized application of HDR brachytherapy in place of standard LDR implants for the oral tongue.
In a subset of patients with a close or positive surgical excision margin and no indication for neck irradiation, treatment with a brachytherapy implant alone (192Ir LDR to 60 Gy) has been used. This obviates the risk of further major surgery or EBRT-related toxicities, including xerostomia. A promising mature local control rate of 89% has been observed in a small retrospective series.[263] Similar results have been reported in other studies. [448] [449]
FLOOR OF MOUTH.
Most floor of mouth cancers are amenable to surgical treatment. T1 and T2 cancers that do not involve the mandible are often treated with wide local excision with 1 cm margins. In contrast to lip cancers, postoperative loss of speech and swallowing function is not as prevalent with small floor of mouth tumors. Therefore, reconstruction can be performed simply with primary closure, secondary intention, skin graft, or an acellular dermis graft. Although not always necessary, it is recommended that the patient undergo preoperative extraction of any decaying teeth close to the lesion.
For those tumors that approach the mandible, a more complex procedure is indicated. Tumors that involve in periosteum of the mandible require a marginal mandibulectomy, where the top half of the involved bone is removed. The overall continuity of the mandible remains intact. T4 tumors that have invaded the cortex of the mandible require a segmental resection of the involved bone. Reconstruction mandates either composite free-flap reconstruction or no reconstruction if the defect is laterally based.
Surgical management of the neck must take into consideration that occult metastases occur frequently in floor of mouth cancers. The incidence is between 23% and 35% of lesions. [277] [278] [450] [451]Therefore, most investigators recommend treatment of the node-negative neck in all patienrs with T2 or higher-grade lesions. [450] [451] [452] For T1 lesions, tumors greater than 2 to 4 mm in thickness have been found to have a high rate of metastasis, and treatment also is recommended. [453] [454] If surgical rather than radiation treatment is pursued, a selective neck dissection that includes the first-echelon nodes (level I and upper level II regions of the neck) should be performed. Unless the lesion is clearly unilateral, a bilateral neck dissection is recommended. Treatment of floor of mouth primaries generally calls for a supraomohyoid neck dissection with resection of levels I, II, and III. Level IV should be included in performing a dissection for cancers of the oral tongue and oropharynx.[455] Levels I and V rarely are involved with laryngeal cancer; therefore, a lateral neck dissection of levels II, III, and IV is the procedure of choice for these lesions.[222]
The cure rates with surgical therapy for floor of mouth cancers have been quoted as 95% for stage I cancers and 86% for stage II cancers.[456] Another study in which the vast majority of patients were treated surgically also demonstrated cure rates of 80% or greater for early lesions.[457] The cure rates for similarly staged cancers treated with irradiation are reported by one study as 88% and 47% for T1 and T2 lesions, respectively.[265] Another study showed, however, that surgery and radiation are equally effective in treating early lesions, and that stage III and stage IV lesions should be treated with combined therapy.[458] In general, the early lesions of the floor of mouth are best managed surgically unless a contraindication exists. Combined therapy is necessary for larger lesions.
HARD PALATE.
The type of surgical management indicated for lesions originating in the hard palate depends largely on the presence of bone involvement. With those lesions that do not involve the periosteum, the tumor can be excised without the underlying bone with an adequate mucosal margin. Tumors invading the periosteum, however, require full-thickness resection of the involved bone. Larger tumors may require partial or total maxillectomy. Reconstruction usually involves a skin graft and a dental prosthesis for large palatal defects. Neck metastases are rare and should be managed surgically with a neck dissection. The role of primary irradiation is limited in this disease, but postoperative irradiation is indicated for advanced lesions.
Oropharynx
Although radiation therapy plays a large role in the treatment of squamous cell carcinoma of the oropharynx, surgery often is indicated in combination with radiation for larger T3 and T4 lesions.[459]Although site-specific considerations are recognized, certain principles apply to surgery in the oropharynx whether the lesion is in the tonsil, base of tongue, or soft palate. Selected T1 tumors can be approached transorally. A transoral approach, however, may not provide the access necessary to excise larger lesions completely and safely. In these cases, an anterior or lateral mandibulotomy often is used to gain access. Mandibular involvement by oropharyngeal lesions necessitates marginal or segmental mandibular resection. A neck dissection can be performed in continuity with or separately from the primary tumor.
The goal of reconstruction in this region is to minimize the severe disfigurement and functional compromise that may result from oropharyngeal procedures. Reconstruction options for surgical defects of the oropharynx depend on the size and location of the lesion. Defects from small T1 lesions can be repaired via secondary intention, primary closure, or a split-thickness skin graft. Reconstruction options for larger lesions include a skin graft, a tongue flap, a myocutaneous flap, and free-tissue transfer. Reconstruction after segmental mandibular resection preferably includes osseous free-tissue transfer to replace the excised bone. If a marginal or small segmental resection is performed, however, the bone does not need to be replaced. The complications associated with surgical treatment of oropharyngeal tumors are similar to those encountered with the treatment of oral cavity tumors. Because the tongue base is critical to swallowing function, dysphagia and aspiration are frequent complications after tongue base resection; these problems can be managed with aggressive swallowing rehabilitation. Site-specific surgical and nonsurgical considerations within the oropharynx also are recognized.
TONSIL.
Although T1 and T2 tonsil lesions generally are best treated with irradiation, large T2, T3 and T4 lesions are often treated with combined therapy. If the mandible is not involved, the procedure of choice is a radical tonsillectomy that includes the tonsil, the tonsillar pillars, and a portion of the underlying muscle. If the mandible is involved, a composite resection including a segmental mandibulectomy is required.
Some authors argue that early tonsil lesions are best treated with surgery. A recently published small study of 18 patients, cited a 5-year survival rate of 92% for surgically treated patients with T1 or T2 cancers of the tonsil.[460] Similar results are found in patients managed with irradiation alone, with less functional morbidity, however. [461] [462] In addition, the use of radiotherapy alone for early tonsil lesions allows the retropharyngeal lymph nodes to be included in the treatment plan, particularly for progressively more posterior lesions and those involving the posterior pillar. Although early lesions are best treated with irradiation, advanced-stage III and IV cancers require combined surgical therapy with radiotherapy for best results. [459] [463] The efficacy of this approach relative to an increasing role for concurrent chemoradiotherapy in advanced oropharyngeal carcinomas including tonsil carcinomas is unknown. With increasing involvement of the soft palate and of the pharyngeal wall, however, surgical resection becomes less favored (although often preferred owing to concomitant bulky tumor) because of the functional consequences and increased risk of complications, respectively. For more advanced lesions, with increasing posterior disease extension, the risk of parapharyngeal involvement and retropharyngeal lymph node metastases increases, necessitating radiotherapy as either definitive therapy or in the postoperative setting. When surgical resection is not indicated, concurrent chemoradiotherapy is emerging as a favored definitive strategy on the basis of recent randomized trials and meta-analyses.[181]This treatment approach is limited by the uncertainty with regard to the optimal chemoradiotherapy schedule and the potential for these results not to be generalizable to the tonsil site specifically. It also is recognized that salvage surgery, when chemoradiotherapy is used for potential organ preservation indications, is limited by the proportion of patients with recurrences able to undergo salvage surgery and then the proportion of those in whom salvage surgery can be expected to be successful.
Despite these concerns, several prospective studies have demonstrated high local-regional control rates (80% to 90%) when oropharyngeal and tonsil sites were the predominant tumor sites. It is not clear whether these results are superior to local-regional control rates that have been reported for treatment with radiotherapy alone. Several institutional series have outlined the results that may be expected with fractionated radiotherapy alone. It is recognized, however, that the fractionation schedule may in turn influence the results.
To address these concerns, Withers and colleagues reported the results of a remarkable collaborative retrospective multi-institutional study of carcinoma of the tonsil fossa treated with radiotherapy alone as part of the Patterns of Fractionation Study.[153] A total of 676 patients from 9 participating institutions provided sufficient treatment variability in fractionation schedules resulting from institutional treatment policies to permit study. Several noteworthy observations were made. With the exception of T1 disease, decreased local control rates were observed with the presence of clinically evident nodal disease across each T stage. Cox regression modeling demonstrated that T stage, N stage, total dose, and the overall treatment time were independent significant factors. Although the optimal fractionation schedule could not be identified, a nonsignificant reduction in local relapse was observed with altered fractionated schedules consistent with the observations of RTOG 90–03.[179] Modeling of the relationship between local tumor control and the overall treatment time suggested that an accelerated growth rate may occur at approximately day 30 but did not significantly improve on the basic model of a constant tumor growth rate throughout the treatment duration. Whether or not these tumor kinetics differ with perturbations from the use of concurrent chemotherapy is unclear but should remain respected.
Hence, when radiotherapy is used as definitive therapy alone, an altered fractionation schedule is preferred, with particular attention paid to minimize any treatment interruptions at any time during treatment. Several reports provide a basis for selection of lesions appropriate for homolateral irradiation. [464] [465] [466] The advantages of homolateral irradiation include reduced acute toxicities, less risk of treatment interruptions, reduced dose to the contralateral parotid, and less risk of complete xerostomia. O'Sullivan and colleagues reported the results of a large retrospective analysis of data for 228 patients who received daily fractionated radiotherapy alone, with mature follow-up.[464] Based on an institutional policy of homolateral irradiation for lesions that did not cross midline structures, these investigators were able to demonstrate a spectrum of risk of contralateral neck relapses to guide treatment selection. Patients at low risk (less than 5%) for contralateral neck relapse include those with T1 or T2 lateralized lesions with involvement of even the lateral two thirds of the soft palate or lateral one third of the base of the tongue. With involvement of these structures, however, judicious use of ipsilateral irradiation will be required.
SOFT PALATE.
Soft palate lesions that are large may require a partial maxillectomy for complete excision. In general, soft palate tumors are amenable to definitive radiotherapy for cure. The fields should cover the draining lymphatics and the retropharyngeal lymph nodes. With this modality, palatal function usually can be preserved. Reconstruction of this defect may include a skin graft along with a prefabricated prosthesis to maintain swallowing function. Irradiation has been shown to be effective in controlling early lesions, whereas combined-modality therapy is needed for stage III and IV lesions. [467] [468] [469]
Cervical metastases should be treated with surgery, or irradiation, or both. Neck dissections can be performed before or after radiation therapy. The decision to perform surgery or irradiation first depends on how the primary is to be treated, and on the size and extent of the neck metastasis. If the metastasis is of sufficient size to require a radical neck dissection, or if it encases the carotid artery, primary treatment with irradiation is preferred to shrink the neck mass to a more resectable size. Bilateral treatment should be considered in large oropharyngeal tumors or those that cross the midline.
Even with a clinically negative neck, treatment by surgery or irradiation is indicated when the primary malignancy originates from the oropharynx. The risk of occult disease is 30% or greater, and these nodes can be found anywhere in levels I to IV of the neck. [470] [471] [472] Therefore, if an elective neck dissection is performed rather than radiation therapy, a selective neck dissection exploring these levels is indicated. For midline lesions, bilateral treatment of the neck must be performed.
BASE OF TONGUE.
Treatment options for carcinomas of the base of tongue may include surgery or radiotherapy-based strategies. Limited comparative studies have been published to guide the treatment decision-making process. In view of the functional impact of therapy to this site, the feasibility of functional organ preservation has received increasing consideration, because comparable local-regional control and survival rates have been suggested. In this regard, although radiotherapy typically has been preferred, several surgical issues are important to consider.
With regard to the tongue base, additional considerations relate to the size and location of the lesion. If a small tumor lies posteriorly and inferiorly in the base of tongue, a transhyoid approach through the neck can be considered rather than a mandible splitting approach. For small tumors that lie laterally in the tongue base or on the pharyngeal walls, a lateral pharyngotomy can be considered. For larger tumors, resection of a significant portion may result in chronic aspiration because the tongue base is critical to swallowing function. Recurrent pulmonary infections in the elderly or immunocompromised can be life threatening. Thus, a laryngectomy often is considered when more than half of the tongue base is to be removed, especially in high-risk patients. If the larynx is preserved, postoperative functional results can be optimized with swallowing therapy or a laryngeal suspension procedure.[473]
For base of tongue tumors, the data show that early tumors are best treated by radiotherapy or surgery while patients with advanced lesions should receive combined treatment. In one study of 173 patients, early primary tumors treated with surgery or radiotherapy gave a control rate of 83% (5 of 6 tumors) and 89% (40 of 45 tumors), respectively. For advanced primary tumors, definitive radiotherapy produced a local control rate of 55% (42 of 76 tumors), compared with 79% (23 of 29 tumors) for surgery and postoperative radiotherapy.[474] Radiotherapy is preferred for early lesions because of the decreased treatment morbidity compared with base of tongue surgery.
In this regard, data exist to demonstrate that the local control rates for EBRT followed by a brachytherapy implant are superior to those achieved with EBRT alone.[284] One retrospective review demonstrated comparable local control rates between EBRT with an implant and surgery, both of which were superior to EBRT alone.[287] Several independent investigators have consistently demonstrated that a brachytherapy implant boost (20 to 30 Gy) after EBRT (45 to 55 Gy) is associated with effective local control rates. No significant functional deficits with this organ-preserving strategy have been reported when several quality of life domains were studied, even for advanced lesions. [291] [292] The most extensive experience has been with temporary interstitial LDR 192Ir implants. Mature local control rates of 85% or greater may be expected for T1 and T2 lesions and 80% to 85% with T3 lesions.[281] Similar results have been reported in other series. [282] [283] [284] [285] [286] [287] [288] [475] [476] In general, selection of the more advanced lesions for treatment has been based on favorable exophytic growth patterns. The experience with a brachytherapy implant in T4 lesions remains limited. [281] [283] [284] [288] [475] Concurrent chemoradiotherapy techniques are favored, particularly for the more advanced T3 and T4 lesions. Puthawala and colleagues reported a mature crude local control rate of 67% with more relapses observed in patients with more advanced neck disease[288] and used a higher brachytherapy boost dose of 30 to 40 Gy, compared with 20 to 25 Gy for T1 and T2 lesions. The value of an implant for T4 lesions appears promising, but this approach requires further evaluation, particularly when combined with combination chemoradiotherapy, because late swallowing complications may compromise the benefits of an organ-preserving treatment strategy.
Larynx
From the standpoint of staging and treatment, the larynx is divided into supraglottic, glottic, and subglottic regions ( Table 72-9 ; Fig. 72-19 ). The supraglottic region is composed of the epiglottis, arytenoid cartilages, aryepiglottic folds, false cords, and laryngeal ventricles. The glottic larynx includes the true vocal cords as well as the anterior and posterior commissures. The subglottic region extends to the inferior edge of the cricoid cartilage. Transglottic tumors involve the glottic level as well as another site within the larynx.
Table 72-9 -- American Joint Committee on Cancer Classification for Staging Primary Laryngeal Cancer
|
Stage |
Characteristics |
|
SUPRAGLOTTIS |
|
|
T1 |
Tumor limited to one subsite of supraglottis, with normal vocal cord mobility |
|
T2 |
Tumor invades more than one subsite of supraglottis or glottis, with normal vocal cord mobility |
|
T3 |
Tumor limited to larynx with vocal cord fixation and/or invades postcricoid area, medial wall of pyriform sinus, or pre-epiglottic tissues |
|
T4 |
Tumor invades thyroid cartilage and/or extends to other tissue beyond the larynx (e.g., to oropharynx, soft tissue of neck) |
|
GLOTTIS |
|
|
T1 |
Tumor limited to vocal cord(s) (may involve anterior or posterior commissures) with normal mobility |
|
T1a |
Tumor limited to one vocal cord |
|
T1b |
Tumor involves both vocal cords |
|
T2 |
Tumor extends to supraglottis or subglottis, or both, with or without impaired vocal cord mobility |
|
T3 |
Tumor limited to the larynx with vocal cord fixation |
|
T4 |
Tumor invades through thyroid cartilage and/or extends to other tissue beyond the larynx (e.g., oropharynx, soft tissue of neck) |
|
SUBGLOTTIS |
|
|
T1 |
Tumor limited to the subglottis |
|
T2 |
Tumor extends to vocal cord(s) with normal or impaired mobility |
|
T3 |
Tumor limited to larynx with vocal cord fixation |
|
T4 |
Tumor invades through cricoid or thyroid cartilage and/or extends to other tissues beyond the larynx (e.g., oropharynx, soft tissues of neck) |
From the American Joint Committee on Cancer: Manual for Staging of Cancer, 4th ed. Philadelphia, JB Lippincott, 1992, with permission.
|
Figure 72-19 Anatomy of larynx and hypopharynx: Sagittal view. (From Grégoire V, Coche E, Cosnard G, et al: Selection and delineation of lymph node target volumes in head and neck conformal radiotherapy. Proposal for standardizing terminology and procedure based on the surgical experience. Radiother Oncol 2000;56:135–150.) |
SUPRAGLOTTIC LARYNX.
With regard to early lesions, surgical management of supraglottic tumors varies depending on the exact location of the lesion. T1 tumors of the suprahyoid epiglottis are readily managed endoscopically by CO2 laser excision. Tumors of the infrahyoid epiglottis are not amenable to this type of resection because of the possibility of pre-epiglottic space invasion.[477] An open procedure is more oncologically sound in these cases.
Surgical treatment of early lesions of the false vocal folds often requires a supraglottic laryngectomy, which spares the true vocal folds. If a supraglottic tumor extends to the true cords, the patient may be a candidate for a supracricoid laryngectomy, which spares at least one arytenoid cartilage but not the true cords. Patients who undergo this procedure generally have excellent functional results with regard to speech and swallowing. [478] [479]
Surgical treatment of T4 lesions requires a total laryngectomy in most cases. This procedure involves the resection of the entire larynx, including the epiglottis, the true and false vocal cords, the thyroid cartilage, one lobe of the thyroid gland, and the involved mucosa of the hypopharynx or base of tongue. The remaining pharyngeal mucosa is then closed, either primarily or with additional tissue from a free or rotational flap. The long-term functional swallowing results for this procedure generally are excellent, especially if the pharynx was closed primarily. [480] [481] [482] Speech rehabilitation is performed with a number of devices and techniques, including esophageal speech, an electrolarynx, or a tracheoesophageal puncture device.
Owing to the high rate of occult metastasis, bilateral treatment of the neck is recommended in most cases of squamous cell carcinoma of the supraglottis. In the clinically node-negative neck, surgical treatment includes a selective neck dissection including levels II, III, and IV, because these are the most likely locations for occult metastases. [221] [483]
The decision to treat T1, T2, and T3 lesions with irradiation or surgery can be difficult. Recent studies in the surgical literature show good 5-year survival rates with either modality. [484] [485] [486] [487]Irradiation with surgical salvage may be the best option for patients who have pulmonary disease who are at a high risk for severe complications from aspiration.
Salvage surgery often entails a total laryngectomy. Patients who are surgical candidates for a partial laryngectomy retain glottic function without compromising oncologic outcome.[488] The status of the cervical nodes also must be taken into consideration in deciding on treatment modality. Surgery certainly is indicated when large metastatic nodes are present, because this allows for combined therapy. Surgery or irradiation is sufficient to treat the node-negative neck.
GLOTTIC LARYNX.
Early lesions of the true vocal cords that are not treated with irradiation can be treated surgically in a number of ways. Carcinomas in situ may be managed endoscopically with vocal cord stripping or with the CO2 laser. T1 lesions of the true vocal cords may be treated surgically with cordectomy or CO2 laser excision. T2 lesions often require open procedures for adequate surgical resection, although radiotherapy is an equally effective option. Operations geared toward organ preservation such as a vertical hemilaryngectomy or a supracricoid laryngectomy should be considered before total laryngectomy for T2 lesions. T3 and T4 lesions of the glottis almost always require a total laryngectomy if surgical management is chosen, although selected patients with T3 cancers are candidates for a supracricoid hemilaryngectomy.[489]
Management of the neck in glottic carcinoma differs from supraglottic carcinoma because the risk of occult metastasis is less, ranging from 3% to 21%. [490] [491] [492] [493] Observation is the treatment of choice for the neck in T1 and T2 primary glottic cancers, but some authors recommend treatment of the N0 neck in T3 and T4 cancers.[494] If surgical management is chosen, a selective neck dissection of levels II to IV is the procedure of choice. In the node-positive neck, a selective or modified radical neck dissection may be indicated, depending on the size and location of the metastasis.
The oncologic results with treatment of T1 lesions of the glottis are good. Endoscopic cordectomy is equal in effectiveness to radiotherapy for most lesions, with cure rates greater than 90%. [495] [496]Several factors should determine which modality to use. Patient concerns and health are of paramount importance. It must be made clear that although irradiation often can be used for salvage after surgery, the salvage procedure indicated after failed radiation therapy may be a total laryngectomy. Lesions of the anterior commissure, although classified as T1, are not as effectively treated with irradiation and are difficult to excise endoscopically. [497] [498] An open partial laryngeal procedure such as a hemilaryngectomy or supracricoid laryngectomy can be considered, depending on the extent of the lesion.
T2 lesions can be managed with radiation or surgery equally well. The 5-year local control rate has been estimated to be greater than 80% after either primary surgical therapy or radiation therapy. [489] [496] [499] [500] [501] [502] It is more likely, however, that an open partial laryngectomy approach rather than a simpler endoscopic approach will be needed for adequate surgical treatment. [489] [496] [501] Therefore, it becomes very important to choose surgical candidates carefully with regard to pulmonary function, intelligence, motivation, and home situation.
For advanced-stage tumors, it has been shown that survival is similar between patients managed with an organ preservation protocol and those managed with surgery. [503] [504] It should be noted that in the landmark Veterans Affairs study, although 66% of patients in the organ preservation arm retained a functional larynx, 39% were tracheostomy-dependent. Patients should be offered chemotherapy and radiation therapy as an alternative to total laryngectomy, however. It recently has been shown that salvage laryngectomy after organ pres ervation therapy is associated with acceptable morbidity, and that survival after the surgery was not affected by the initial organ preservation treatment.[185]
SUBGLOTTIC LARYNX.
Primary cancers originating in the subglottis are relatively rare.[505] Because conservative endoscopic procedures would fail to clear the tumor, management of these tumors most often requires total laryngectomy. A paratracheal lymph node dissection and ipsilateral thyroidectomy also should be considered. Subglottic extension of T3 or T4 subglottic cancers should be managed similarly. Postoperative radiation therapy should be added to decrease the risk of a stomal recurrence.
HYPOPHARYNX.
Surgical management of hypopharyngeal cancers is particularly challenging because the mucosa of the hypopharynx is vital to swallowing function. Thus, surgical extirpation of large tumors often requires complex reconstructions to minimize postoperative dysphagia.
The surgical options vary with the size and precise location of the tumor within the hypopharynx. Selected small tumors of the pyriform sinuses can be managed using conservation procedures that maintain laryngeal function. These procedures include the partial laryngopharyngectomy and supracricoid hemilaryngopharyngectomy. [506] [507] To be a candidate for these procedures, a patient must have a small lesion that does not involve surrounding structures or impair vocal cord motion. In addition, the patient must have good pulmonary function and motivation to undergo rigorous swallowing rehabilitation.
For most tumors of the pyriform sinus staged as T2 and larger, a total laryngectomy with partial pharyngectomy is indicated. This procedure can carry significant morbidity, especially when performed as salvage after failed radiation therapy. To reconstruct the pharynx, surgical options include myocutaneous flaps such as the pectoralis flap, fasciocutaneous flaps such as the deltopectoral flap, and free-tissue transfer from the forearm or jejunum. A gastric transposition procedure, facilitated by a general surgeon in the operating room, also is used in cases in which a circumferential pharyngeal defect is present.
Tumors of the posterior pharyngeal wall are more accessible anatomically, making surgical management less complex than for pyriform sinus cancers. Most pharyngeal wall cancers can be resected directly, without laryngectomy, through a suprahyoid or lateral pharyngotomy approach. Reconstruction is often performed using a split-thickness skin graft or acellular dermal graft. Larger tumors may require myocutaneous or free flap reconstruction.
Cancers that involve the postcricoid region of the hypopharynx require an extensive procedure if they are to be managed surgically. A total laryngectomy, partial pharyngectomy, and cervical esophagectomy usually are required. A total esophagectomy may be indicated, depending on the inferior extent of the lesion.
Occult metastases from hypopharyngeal cancers are present in greater than 30% of N0 necks. [490] [508] Therefore, selective neck dissection of levels II, III, and IV is indicated, even in patients with tumors staged N0. In addition, if the tumor approaches the midline, bilateral neck dissections should be performed. [508] [509]
The outcome for patients with hypopharyngeal cancer is poor, especially if the tumor originates in the pyriform sinus. Marks and colleagues estimated the 5-year survival rate to be 14% for patients with advanced pyriform sinus cancers.[510] In the same retrospective study, it was demonstrated that surgery alone was significantly superior to radiotherapy, with or without chemotherapy, for treatment of the disease. Another retrospective study showed that surgery is superior to combined chemotherapy and radiation therapy, although the difference was not statistically significant.[511] A prospective study showed that organ preservation is superior in terms of survival, but again, the difference was not statistically significant.[192] The treatment of advanced hypopharyngeal cancer must be individualized for each patient.
MAJOR AND MINOR SALIVARY GLANDS.
Surgical treatment is the mainstay of management of salivary gland cancer, whether arising from the parotid, submandibular, sublingual, or minor salivary glands. Surgery for parotid gland malignancy is perhaps the most challenging because of the location of the facial nerve coursing between the superficial and deep lobes. Most parotid tumors are located in the superficial lobe of the gland and therefore are treated with a superficial parotidectomy through a transcervical approach, sparing the facial nerve. Those tumors within the deep lobe of the gland require a total parotidectomy. The facial nerve is spared in these cases if it is not involved with the malignancy. For deep lobe malignancies, additional access via a submandibular or, rarely for massive tumors, a mandible splitting approach.
Submandibular gland tumors usually are contained within the gland, so resection is limited to the submandibular triangle. The marginal mandibular, lingual, and hypoglossal nerves should be spared. If spread to the surrounding tissues has occurred, these structures, as well as surrounding bone, floor of mouth, and skin, may need to be removed surgically. The extent of surgery for malignancies of the sublingual glands and minor salivary glands depends on the size and location of the tumor within the oral cavity.
Combined therapy usually is recommended for high-grade malignancies of the salivary gland such as high-grade mucoepidermoid carcinoma. Postoperative irradiation has been shown to improve local-regional control in several studies. [512] [513] Neck dissections are indicated in the event of known cervical metastasis.
Unknown Primary
The treatment for unknown primary squamous carcinoma may include neck dissection followed by postoperative radiotherapy alone or radiotherapy followed by a neck dissection. The relative efficacy of these two strategies has not been evaluated. No established role for the addition of concurrent chemotherapy has been recognized at this time, although this remains an attractive therapeutic strategy. The increased mucosal toxicity, including the potential for long-term swallowing dysfunction, however, must be balanced against unclear benefits including regional control and survival. Currently, irradiation may include the neck alone, or elective mucosal irradiation of potential primary sites may be performed. Weir and colleagues noted no survival benefits with elective mucosal irradiation,[514] although Reddy and colleagues continued to recommend elective mucosal irradiation owing to reduced risk of contralateral nodal and primary relapses.[515] The latter depends on the location of the nodal metastases. For level II nodal metastases, this may require elective irradiation of the oropharynx and laryngeal structures.
Recurrent Head and Neck Squamous Cell Carcinoma and Second Head and Neck Squamous Cell Carcinoma
The optimal management for recurrent non-nasopharyngeal HNSCC and second primary HNSCC remains to be defined and is complicated by retrospective series reporting on outcomes for heterogeneous groups of patients. Reirradiation of the head and neck is possible, however, with increased but potentially acceptable toxicities. Various investigators have studied combinations of chemotherapy and radiotherapy in hopes of realizing significant radiosensitization such that the total required for reirradiation is reduced. The results have been summarized by Kao and associates.[218] Unfortunately, only a select few demonstrate benefit with survival prolongation. Higher total doses appear to be important, with a reirradiation dose of 60 Gy typically recommended.[516] Accordingly, the limited potential for gains must be balanced with an increased risk of complications, including soft tissue necrosis and neurologic damage. Currently, reirradiation with EBRT and chemotherapy remains limited to clinical trials at centers with ongoing systematic experience with particular protocols. Various conformal radiotherapy techniques may facilitate safer reirradiation. These include recent techniques such as IMRT. The use of brachytherapy implant for reirradiation, however, can play an important role in improving the therapeutic ratio, because full-dose irradiation is required for any potential for cure.
Surgery is preferred and offered for resectable lesions in the absence of unacceptable functional and cosmetic sequelae. In this setting, management of microscopic residual disease with radiotherapy often is necessary. A planned introduction of nonirradiated tissue flaps with coordination of the implant placement and of the wound reconstruction can reduce the risk of wound complications. Both pedicled myocutaneous flaps[517] and microvascular free flaps have been described.[518] Nevertheless, complication rates of 20% to 50%, including a risk of carotid rupture during placement of the implant in the neck, have been reported. Factors other than the radiation may be contributing to this increased risk of complications. These considerations are tempered by local control rates of 44% to 80%, suggesting a limited therapeutic ratio. [519] [520] [521] [522] Hence, patient selection is particularly important with this indication.
A brachytherapy implant alone may be used for treatment of selected second primary HNSCCs. Peiffert and colleagues reported the results with use of a 192Ir implant alone for 73 patients with tonsil carcinoma treated at the Centre Alexis Vautrin.[295] A majority of the patients presented with early node-negative disease; a median dose of 60 Gy was prescribed according to the Paris dosimetry system. The 5-year actuarial local control rates for T1N0 and T2N0 lesions were 80% and 67%, respectively. Acceptable grade 2 self-resolving complications were observed in 10 cases (13%), with a majority of the complications being soft tissue necrosis due to a dose greater than 60 Gy. This appeared to be increased when compared with the same institution's series when tonsil implants were prescribed in unirradiated tissues.[272] The 5-year actuarial disease-specific survival and overall survival rates were 64% and 30%, respectively. No long-term survivors were observed, reflecting the increased risk of other malignancies and other alcohol- and smoking-related comorbid conditions in this patient population.
Other retrospective series have reported the results of reirradiation with an implant, but include both patients with a second primary HNSCC and recurrent HNSCC, confounding the overall analyses. Treatment of recurrent HNSCC may be expected to give lower local control rates as a result of the presence of more radioresistant clonogens, an observation consistent with clinical data. [293] [294] [523]Langlois and colleagues reported the reirradiation results for a larger and heterogeneous group consisting of 123 patients with T1 to T3 disease treated with a 192Ir implant to a mean dose of 62 Gy at the Centre Alexis Vautrin.[294] A 5-year actuarial local control rate of 59% was observed. Local control correlated with tumor size less than 3 cm, second primary (versus recurrent lesion), dose greater than 60 Gy, and tumor site (oral cavity favorable compared with oropharynx). Only a 5-year actuarial survival rate of 24% was realized, with local control achieved, and a time interval between reirradiation of more than 2 years associated with a better prognosis. Mucosal necrosis was observed in 28 cases (23%). Stevens and colleagues also noted favorable local control rates with second primary HNSCC, a reirradiation dose of 60 Gy or greater for second primary HNSCC, a treatment interval longer than 1 year for recurrent lesions, and the use of an implant in addition to EBRT.[523] Mazeron and colleagues reported a 5-year actuarial local control rate of 69% for 70 patients with oropharyngeal carcinomas re-irradiated with a 192Ir implant alone delivering a mean dose of 60 Gy with the Paris system.[524]Similarly, tumor site (glossotonsillar sulcus and base of tongue being unfavorable) and tumor size (larger than 2 cm) adversely influenced the local control rate, although larger lesions occurred more frequently in the base of tongue. Regional nodal irradiation was not intentionally treated, with only 7 of 69 (10%) developing nodal relapses. Soft tissue necrosis was the main complication (seen in 27%), was self-resolving in 13 of 14 patients, and appeared to be increased when a large lesion was implanted.
Levandag reported the results for 73 patients with either second primary HNSCC or recurrent HNSCC comparing EBRT reirradiation with implant with or without EBRT reirradiation (18 patients), although significant heterogeneity existed in the treatment applied in each of the patient cohorts, precluding definitive conclusions.[525] Selection bias was minimized because the two cohorts represented sequential treatment periods resulting from an institutional policy change. Crude local control rates of 29% and 50% were reported, respectively, suggesting improved control rates with an implant though this was confounded by the higher mean dose delivered in the implant cohort compared with the EBRT cohort. The high rate of temporary mucosa ulcerations (in 13 of 18 versus 9 of 55) probably was related to the implant technique used. To reduce the increased risk of mucosal ulceration associated with an implant, Housset and colleagues employed a planned treatment interruption, delivering the intended dose using two separate implants with a source shift.[293] A total of 55 patients with both recurrent and second primary base of tongue squamous cell carcinomas underwent this implant protocol; 31 patients received a single-course implant delivering 60 Gy and 24 patients received a split-course implant delivering 35 Gy and then 30 Gy after a 1-month interruption. A significant reduction in the risk of mucosal necrosis was observed with the introduction of a treatment interruption (16% versus 43%), although a trend toward a lower crude local control rate (37.5% versus 52%) suggests that such a strategy warrants further investigation.
In summary, selective application of a brachytherapy implant for reirradiation may be appropriate in view of the poor prognosis associated with this patient population. Favorable local control rates may be expected in patients with second primary HNSCC, small tumor sizes, tumor sites other than the base of the tongue, and the delivery of 60 Gy or greater. For recurrent HNSCC, treatment intervals of longer than 1 year appear to select for lesions that are less radioresistant. Soft tissue complications appear to be increased, with a frequency ranging from 20% to 30%; a majority of such complications are self-resolving.
SUMMARY AND FUTURE DIRECTIONS
A paradigm shift has occurred in the management of squamous cell carcinoma of the upper aerodigestive tract over the past 25 years. In the 1970s and 1980s, Fletcher and others demonstrated the ability to preserve function through the use of definitive radiation therapy for patients with cancers of the larynx and oropharynx. In the mid-1980s, Wolf and Hong introduced the concept of laryngeal preservation by combining induction chemotherapy and definitive radiotherapy for patients who demonstrated a significant response to the induction phase of treatment. For the first time, organ preservation was achieved without a deleterious effect on survival. After this favorable report by the Veterans Affairs Cooperative Group, treatment intensification trials were designed to increase the rate of organ preservation and to improve survival rates. Concomitant chemotherapy and radiation therapy were demonstrated to provide a survival benefit over neoadjuvant chemotherapy and irradiation, albeit with increased toxicity. The RTOG implemented a three-arm trial for laryngeal preservation in 1992 to test this concept. Preliminary results from this trial demonstrated a higher rate of organ preservation among patients receiving concomitant chemotherapy than among those receiving induction chemotherapy and radiation therapy or radiation alone. Despite an increased rate of laryngeal preservation, however, treatment intensification did not translate into improved survival.
Current therapeutic approaches have been designed to improve local-regional cancer control and survival through further treatment intensification. Phase III trials are evaluating different chemotherapeutic agents combined with altered fractionated radiation schemes and take advantage of the radiosensitizing effects of chemotherapy when administered concurrently with radiation. Although the outcome from these trials will not be reported for some time, the toxicity associated with aggressive combination therapy is of increasing concern. Anatomic organ preservation through nonsurgical means has been convincingly demonstrated; nevertheless outcome data that demonstrate effective functional preservation of the organ are lacking. Assessment of organ function and quality of life are now integrated into these trials and will provide longitudinal data on pre- and post-treatment organ function and how these relate to the patient's perceived quality of life. In the future, the success of anatomic as well as functional organ preservation will be clarified.
Further treatment intensification is likely to be associated with unacceptable toxicity and will require new approaches to ameliorate the side effects of cytotoxic therapy. Toxicity scoring schemes are being developed and implemented to provide a more quantitative assessment of early and late treatment side effects. Mucosal injury, an early treatment effect, and fibrosis, a late effect, will constrain further attempts to intensify treatment. Mucositis contributes to nutritional deficits and may lead to circumferential scarring and stenosis of the pharynx. Fibrosis contributes to laryngeal and pharyngeal dysfunction through loss of neuromuscular coordination and muscle function. IMRT to limit exposure of normal tissues and agents to mitigate these side effects are being studied in clinical trials. To date, none of the systemic agents intended to lessen toxic side effects has convincingly demonstrated effectiveness.
Although current therapeutic approaches destroy neoplastic cells, collateral damage to normal tissues is responsible for the toxic side effects and long-term functional impairment. Future directions may come from continued technological advancements including organ-preserving surgical techniques and the use of more precise radiotherapy techniques such as IMRT. The greatest potential for gain, however, comes from capitalizing on the molecular mechanisms specific to cancer cells that are responsible for tumor progression. The hope is for more cancer-specific therapy without the increased toxicity. The results from clinical trials conducted with several of these agents alone support the notion that there may be more specific tumor targeting, as the toxicity profiles have to date been very modest. Activity also has been modest, however, which has lead to the use of surrogate measures to determine if in fact these agents are inhibiting the molecular targets in vivo. Although the head and neck site lends itself to tissue biopsies for these assays, the hope is for the continued development of functional imaging to provide these answers. In view of the modest activity with these agents alone, the challenge of the moment is to develop novel strategies for combining these targeted biologic agents with cytotoxic drugs and radiation therapy in a rational manner. This will depend on successful translation of knowledge gained from laboratory investigations, particularly about the molecular mechanisms of treatment resistance. Today's oncologists are at the threshold of realizing a long-sought goal of effective cancer control through the use of targeted therapy that will optimize local-regional control and enhance survival within the context of acceptable toxicity.
Copyright © 2008 Elsevier Inc. All rights reserved. - www.mdconsult.com
Abeloff: Abeloff's Clinical Oncology, 4th ed.
Copyright © 2008 Churchill Livingstone, An Imprint of Elsevier
REFERENCES