Skin Cancer. Melanoma.
The skin is the largest organ of the body and consists of three layers: epidermis, dermis, and subcutis (i.e., hypodermis). It is a specialized structure with a wide range of functions that include protection from the environment, synthesis of vitamin D, production of a large number of cytokines, antigen presentation, thermoregulation, and sensation of touch and temperature. Several different cell and tissue types, originating from all three embryonic layers, participate in the formation of the three layers of the skin and its associated appendages. These cell and tissue elements can transform to produce a large number of benign and malignant growths. This chapter summarizes the nonmelanoma skin cancers and precancerous conditions.
EPIDEMIOLOGY
Nonmelanoma skin cancers are the most common cancers in the US white population. More than 700,000 new cases are diagnosed annually, according to the data from the American Cancer Society. The mortality rate for the nonmelanoma skin cancers is approximately 2100 per year. Because most skin cancers are diagnosed and treated in private office settings or outpatient clinics, the available statistics are thought to grossly underestimate the actual number of nonmelanoma skin cancer cases.
Basal cell carcinoma is the most frequently diagnosed skin cancer in whites, accounting for approximately 75% to 80% of all reported cases. Squamous cell carcinoma is the second most common skin cancer and is estimated to represent 20% to 25% of all reported skin cancer cases.
Persons with fair skin who sunburn easily are more susceptible to developing basal cell carcinoma on the sun-exposed areas of their skin. Basal cell carcinoma occurs less often in darkly pigmented persons, in whom squamous cell carcinoma is the most common skin cancer. Basal cell carcinoma is seen more frequently in men than in women, and it mostly occurs later in life. However, recent observations indicate an increased incidence of basal cell carcinoma in younger age groups.
Other skin cancers, such as soft tissue sarcomas involving the dermis and subcutis and the adnexal carcinomas, are much less frequent and, although they are encountered in clinical practice, are not as common as basal cell carcinoma or squamous cell carcinoma.
ETIOLOGY AND PATHOGENESIS
The two major factors influencing the development of skin cancers are exposure to ultraviolet radiation and type of skin. Chemical carcinogens have been extensively studied, especially in laboratory animals, as etiologic factors for cutaneous malignancies. Ionizing radiation and primary chronic irritation play major roles in skin cancers. Attention has recently been focused on viruses that may cause skin cancers, specifically human papillomavirus causing skin and mucous membrane carcinomas. Host genetic makeup and host immunity also play roles in the development of skin cancers.
EXPOSURE TO ULTRAVIOLET LIGHT AND SKIN TYPE
Ample evidence supports the combined influence of ultraviolet (UV) light and skin type on the incidence of skin cancer. Exposure to UV radiation in the form of sunlight is overwhelmingly the most important etiologic factor for development of nonmelanoma skin cancer. The wavelengths of 290 to 320 nm (UVB), and to a lesser extent, 320 to 400 nm (UVA) have been implicated. The amount of exposure, the timing of exposure, and the skin type being exposed determine the rate of carcinogenesis. Light-skinned people develop skin cancer much more readily at a given rate of sun exposure than dark-skinned individuals, who are protected by the melanin pigment in their skin. As can be expected, nonmelanoma skin cancers occur most frequently on skin sites that are most often exposed to the sun.
Squamous cell carcinoma usually occurs in the sun-exposed areas of head and neck. Persons with outdoor occupations, such as sailors and farmers, have a higher incidence of skin cancers than those with indoor occupations. Epidemiologic studies worldwide suggest that UV radiation is the most important etiologic agent for skin cancers. The incidence of nonmelanoma skin cancers directly correlates with the proximity to the equator. A quantitative association has been observed between lifetime sun exposure and the risk of developing nonmelanoma skin cancers. In outdoor workers, the most common sites for skin cancers are on the head, neck, and dorsum of the hands, which are the sites of maximal chronic sun exposure. Chronic UV exposure has changed from occupational to a more recreational pattern, and younger and younger persons are being diagnosed with skin cancers. Recent studies have further delineated risk factors for development of UV light-induced skin cancers. There is an increased incidence of skin cancer in
The risk for basal cell carcinoma is increased in light-skinned people, in those who freckled or suffered severe sunburn in childhood. A large increase in the risk for basal cell carcinoma is seen in individuals with increased sun exposure in childhood and adolescence. This relationship was strongest among individuals who burned rather than tanned upon sun exposure. Another study suggested that intense sun exposure delivered intermittently to poorly tanning light-skinned individuals carried more risk than the same dose of sun exposure delivered more continuously over the same total period of time.
Risk for squamous cell carcinoma is also strongly related to UV radiation. Again, it occurs more frequently in light-skinned individuals and is related directly to the amount of solar radiation they receive. Sailors, farmers, and others with outdoor occupations have a higher incidence of squamous cell carcinoma than those with indoor occupations. The temporal relationship to occupational solar exposure was further delineated by a recent study that found a strong trend toward increased risk with increasing occupational solar exposure in the 10 years prior to diagnosis of squamous cell carcinoma. Other forms of UV radiation, such as the one used in conjunction with psoralens for the treatment of psoriasis, also increase the risk of skin cancers, especially that of squamous cell carcinoma.
There is considerable environmental concern about the depletion of the ozone layer by certain chemicals. The ozone layer acts as a strong barrier in absorbing a major portion of UV radiation and preventing it from reaching the earth. The chemicals responsible for the depletion of the ozone layer include chlorofluorocarbons, which are found in aerosol sprays and refrigerators. It is thought that the depletion of ozone over the next few years will rapidly increase the incidence of skin cancers.
Persons with light complexions have a higher chance of developing skin cancers than those with darker skin, who are protected from solar damage by the melanin pigment in the skin. Cutaneous malignancies are rare in dark-skinned persons, but African albinos have a high incidence of skin cancers on sun-exposed areas.
UV light may influence the development and progression of skin cancers by affecting the host immune system. The classic work of Kripke and colleagues showed that UV-induced cancers in mice are highly antigenic and that most are rejected by the host’s immune system after transplantation into a normal, genetically identical animal. However, the primary host in whom the tumor was induced by UV light becomes tolerant to the tumor and allows its rapid growth. Kripke’s work indicates the development of suppressor factors and cells that suppress the host’s immune system and prevent rejection of UV-induced skin cancer. These observations have not yet been confirmed in humans, but it is likely that UV light affects Langerhans cells in human skin, which may alter the host immune system, allowing the development and progression of skin cancer. Recent studies have noted an intriguing finding that may elaborate on Kripke’s work. A tumor-suppressor protein, designated p53, is present in very small amounts for very short periods of time iormal keratinocytes, such that it is normally undetectable. Long-lived mutants of p53, which may have lost the tumor-suppressor activity, have been detected in sun-damaged epidermis adjacent to basal cell carcinomas, which also contain the mutant p53. This suggests that UV damage to keratinocytes may result in mutation of p53 to an ineffective form, allowing carcinogenesis to proceed. Mutant p53 has also been seen in squamous cell carcinomas, suggesting that the above model may have wide application.
CHEMICAL CARCINOGENESIS
Most information about cutaneous carcinogenesis has been obtained from studies using laboratory animals, especially mice. After topical application of a carcinogen at regular intervals, the animals develop multiple squamous papillomas, most of which regress spontaneously. Some of the lesions develop the cytologic criteria of malignant cells, and these tumors may become locally invasive.
The multistage carcinogenesis model developed on mouse skin tumors has provided a useful tool for study of carcinogenesis. Three stages of progression have been identified for chemical carcinogenesis: initiation, promotion, and carcinogenesis. During the initiation phase, the DNA configuration of the cells undergoes some basic changes. The process of initiation may remain unchanged for the life of the tissue; or it may progress to malignancy. The epidermal cells in psoriasis share many features with initiated cells. The initiated cells are usually terminally differentiated but may lose their pattern of differentiation and retain their ability to multiply. For the promoter to be effective and cause malignancies, the cell must have been initiated previously. Most promoters usually cause inflammation and hyperplasia, and their effects are reversible. Promoters can induce tumors only after initiation. Some initiators and promoters are listed in. RasHa and N-ras have been found to be possible initiation events in chemically induced epidermal cancers in mice. Promoter compounds have beeoted to induce increased production of transforming growth factor alpha in mice. [ref: 16] This has been postulated to provide an environment that allows clonal expansion of initiated cells. Progression to malignant transformation has beeoted in initiated cells containing rasHa that were given the activated v-fos oncogene. Cells containing either oncogene alone did not progress to carcinoma, suggesting the need for multiple activational events before progression occurs. As further exciting details come to light it should be kept in mind that these studies were undertaken in special populations of mice, and application to humans should be undertaken with caution.
Tar, which contains polycyclic aromatic hydrocarbons, is an initiator, and it has been used for treating psoriasis. Nitrogen mustard is an accepted treatment for cutaneous T-cell lymphoma. Phorbol esters, which are found in croton oil, are known promoters. Anthralin, used in the treatment of psoriasis, and benzoyl peroxide, used in the treatment of acne, are also known promoters. Long-term use of these agents in humans has not been associated with an increased incidence of malignancy. No chemical carcinogen has been identified that gives rise to basal cell carcinoma or malignant melanoma in animals, but experiments with cutaneous chemical carcinogenesis in laboratory animals cannot be directly applied to humans.
The well-documented cases of scrotal carcinoma in chimney sweepers offer the classic example of chemical carcinogenesis in humans. Arsenic is recognized as a chemical carcinogen. Increased incidence of cancers is reported in localities where there is a high level of arsenic in the drinking water. Medical exposure to arsenic in the form of Fowler’s solution, Donovan’s solution, and Asiatic pills in the treatment of asthma, psoriasis, and syphilis predisposes to the development of arsenical keratosis, skin cancer, and possibly lung cancer.
IONIZING RADIATION
Exposure to ionizing radiation can induce cutaneous malignancies in humans, usually basal cell carcinoma, squamous cell carcinoma, and spindle cell carcinoma. Radiation-induced cancers of the skin have been reported in patients receiving ionizing radiation as therapy. In the past, acne, facial hair, and tinea capitis were treated with x-ray therapy. Patients receiving these types of therapy later developed severe radiodermatitis in the form of skin atrophy, telangiectasia, hypopigmentation, or hyperpigmentation. Some of these patients developed large, invasive, deforming skin cancers. These inappropriate uses of x-ray therapy have been discontinued, but accidental exposure to x-rays and exposure for medical reasons continues to cause radiation dermatitis. Occupational exposure and the resultant radiodermatitis and skin cancers, such as squamous cell carcinoma on the fingers of dentists, is no longer seen. The use of fractionated doses of radiation has reduced the long-term side effects of radiation therapy. Rare cases of squamous cell carcinoma caused by radioactive gold jewelry have also beeoted.
CHRONIC IRRITATION OR INFLAMMATION
Skin cancers can develop in areas of chronic inflammation or irritation. Squamous cell carcinoma has been noted to occur in many skin diseases with a chronic inflammatory or irritant course, such as long-standing granulomas, venereal granulomas, syphilis, lupus vulgaris, leprosy, SLE, chronic ulcers, osteomyelitis sinuses, old burn scars, hidradenitis suppuritiva, poikiloderma congenitale, dystrophic epidermolysis bullosa, and porokeratosis of Mibelli. Squamous cell carcinoma of the oral cavity can be induced by chronic irritation secondary to chewing tobacco or betel nuts.
VIRAL ONCOGENESIS
Many malignant neoplasms are caused by viruses in animals. In humans, such associations have rarely been documented. Human papillomavirus (HPV) has been identified in lesions of verrucous carcinoma, bowenoid papulosis, epidermodysplasia verruciforms, and in situ epidermoid carcinoma. Many papillomavirus subtypes have been identified, and HPV types 5, 8, 14, 16, 17, and 33 are associated with various epidermal carcinomas and carcinoma of the cervix. BPV 6 and 11 have been associated with verrucous carcinoma of the genitals (Buschke-Lowenstein tumor), while HPV 16 has been associated with squamous cell carcinoma of the nail bed. [ref: 39] HPV 5 and HPV 8 have been found in most cases of epidermoplasia verruciformis that have progressed to squamous cell carcinoma. It is possible that other inducing factors are needed to produce the malignant tumor from some of these papillomaviruses. These papillomaviruses appear potentially oncogenic in humans. There are other viruses, such as human T-cell lymphotrophic virus-I (HTLV-I) in leukemia-lymphoma and Epstein-Barr virus in Burkitt’s lymphoma and nasopharyngeal carcinoma, that may be associated with some of the human malignancies, but conclusive information is not available.
CLASSIFICATION OF SKIN TUMORS
To describe the various tumors of the skin in an orderly fashion, the cellular origin and the location in the three layers are followed summarize the tumors of the skin based on their cell of origin, biologic behavior, and location in the three layers of the skin.
BENIGN TUMORS ARISING FROM EPIDERMAL KERATINOCYTES
Benign tumors are numerous and very common. They rarely give rise to malignant tumors, but they may be confused with some of the malignant lesions. It is important for clinicians to be able to recognize and differentiate them from malignant skin tumors.
SEBORRHEIC KERATOSIS
Seborrheic keratosis or basal cell papilloma is a common benign lesion produced by the overgrowth of epidermal keratinocytes. It occurs most often on the trunk, face, and neck of middle-aged and elderly persons. Lesions may be found on the extremities but not on the palms and soles. Essential diagnostic features include multiple domed or flat-topped, round or ovoid, verrucous papules that are brown to black. These slightly raised, sharply demarcated lesions appear to be stuck on the surface of the skin. There may be a genetic tendency toward these lesions, but sun exposure has not be found to be an etiologic factor. A possible role for alpha-transforming growth factor has been evaluated. Histologic features are characteristic and consist of acanthosis, hyperkeratosis, and papillomatosis with keratin horn and pseudocysts. The lesion consists mostly of squamous and basaloid epidermal cells. Histologic types include acanthotic, hyperkeratotic, reticulated (adenoidal), clonal, and irritated (inflamed) lesions. Other types include melanoacanthoma, dermatosis papulosis nigra, stucco keratosis, and seborrheic keratoses of the Leser-Trelat sign. The sign of Leser-Trelat is characterized by a sudden appearance of innumerable new lesions of seborrheic keratosis, which may be a sign of an internal malignancy. Rare transformations of seborrheic keratosis to basal cell carcinoma and squamous cell carcinoma have been reported. Differential diagnosis consists of intraepidermal epitheliorna, verruca vulgaris, dermal nevus, pigmented basal cell carcinoma, well-differentiated squamous cell carcinoma, and malignant melanoma. If histologic confirmation is necessary for differential diagnosis, shave excision is the method of choice. Because of its relatively superficial location, seborrheic keratoses may be removed by curettage and light electrodesiccation or by cryosurgery with liquid nitrogen or carbon dioxide ice.
EPIDERMAL NEVI
Epidermal nevi are benign, congenital, hyperplastic lesions with a smooth or hyperkeratotic surface. These are formed by hyperplasia of epidermal structures, classifying them as hamartomas. They typically appear at birth or in early childhood. Epidermal nevi are uncommon and may be solitary or multiple or form a plaque covering a large area of skin, usually in an asymmetric or linear distribution. The lesions vary in color from pale to deeply pigmented brown to black. Lesions may continue to enlarge through puberty. Clinical variants include nevus verrucoses, linear epidermal nevus, systematized epidermal nevus, ichthyosis hystrix, inflammatory linear epidermal nevus, and epidermal nevus syndrome. Epidermal nevus syndrome or large verrucous epidermal nevus may present with widespread involvement but predominantly on one side of the body. It frequently coexists with central nervous system and skeletal abnormalities. Differential diagnosis includes incontinentia pigment (verrucous stage), lichen striatus, linear lichen planus, and linear porokeratosis. Epidermal nevi remain benign and require treatment only if cosmetically indicated. There have been rare cases of transformation to carcinomas.
CLEAR CELL ACANTHOMA
Clear cell acanthoma appears as a solitary, pink to red, scaly nodule on the lower leg. The lesion is slow growing, sharply delineated, and 1 to
KYRLE’S DISEASE
Kyrle’s disease (i.e., hyperkeratosis follicularis et parafollicularis in cutem penetrans) appears as multiple follicular or perifollicular yellowish brown to brownish red papules and plaques with a central keratinous plug. The lesions are seen mostly on the lower extremities but appear on upper limbs, trunk, soles, and rarely on other areas of the body. It is thought that the condition is produced by abnormal keratinization progressing downward from the epidermis and penetrating from the epidermis into the dermis, forming a keratotic plug that stimulates inflammatory and foreign-body giant cell reactions. The histologic picture includes hyperkeratosis, mainly around hair follicles, lichenoid reactions, and lymphocytic infiltration. Kyrle’s disease and perforating folliculitis are clinically and histologically hard to differentiate. Uremic follicular hyperkeratosis is used to describe lesions that have features of Kyrle’s disease and perforating folliculitis. Unlike the rarity of Kyrle’s disease, this new entity is fairly common and occurs mostly in patients with diabetes and renal failure. Kyrle’s disease is one of the group of perforating dermatoses. These lesions must be differentiated from actinic keratosis and squamous cell carcinoma, and they require biopsy and histologic confirmation. No effective treatment is available, but cryotherapy may be of some value.
CYSTS
Cysts are relatively common lesions occurring in young and middle-aged persons, and they are usually of great concern to patients. Two of the more common forms are epidermal cyst and trichilemmal (i.e., pilar) cyst. Epidermal cysts are single or multiple lesions, mostly found on the face, neck, chest, and back. They arise spontaneously or may form as a result of trauma. A cyst normally enlarges to about 1 to
BIRT-HOGG-DUBE SYNDROME
Birt-Hogg-Dube syndrome is inherited as an autosomal dominant trait and consists of the triad of trichodiscoma, fibrofolliculoma, and acrochordan. The patient has multiple, asymptomatic, flesh-colored papules on the face, mainly around the nose. Histologically, the lesions may show proliferation of the superficial hair follicle (i.e., trichodiscoma) or proliferation of the dermal part of the hair follicle (i.e., fibrofolliculomas) or features of fibrous tags (i.e., acrochordon). Birt-Hogg-Dube syndrome is associated with medullary carcinoma of the thyroid and renal cell carcinoma. Lesions are removed only if cosmetically indicated.
BECKER’S NEVUS
Becker’s pigmented hairy epidermal nevus (i.e., Becker’s melanosis) is a relatively common condition presenting with a patch of hyperpigmented, coalescing macules on the shoulders and upper extremities, mostly in young men. Hypertrichosis is frequently associated with the tan or brown patches. This condition is usually preceded by a sunburn of the involved area. Histologically, the skin appears normal except for an increased number of melanocytes, especially in the dermis, and the increased melanin in the basal layer of the epidermis. The hair follicles appear normal, but the diameter of the hair may be thicker thaormal. Smooth muscle hartomas are occasionally associated with this lesion as well as hypoplasia of underlying structures such as breast tissue or arm bones and coexistence with a connective tissue nevus. A malignant melanoma has cooccurred in the same anatomic site as a Becker’s nevus. There is no need for treatment except if cosmetically indicated.
WARTY DYSKERATOMA
Warty dyskeratoma presents in middle-aged patients as a solitary lesion on the scalp, face, or neck, but it may occur in unexposed skin and the oral mucosa. Multiple lesions have been reported. It occurs as a slightly elevated, circumscribed papule or nodule with a raised border and keratotic umbilicated or pore-like center. The lesion typically grows to about 1 to
PREMALIGNANT TUMORS OF THE EPIDERMIS
Because of the absence of good epidemiologic data, it is difficult to determine with certainty whether a lesion is a true precancerous lesion and gives rise to a malignant condition or if the association is a coincidental finding. An example of this is keratoacanthoma, which is usually a benign, self-healing lesion, but it may predispose the patient to squamous cell carcinoma. The following are conditions that are thought to precede or are associated with cancerous lesions.
ACTINIC OR SOLAR KERATOSIS
Actinic or solar keratoses are common asymptotic lesions seen mostly on sun-exposed areas of light-skinned persons. Those who burn easily, tan poorly, and freckle (type I and type II skin prototypes) are at the greatest risk for development of actinic keratosis. The lesions are commonly multiple and appear on sun-damaged skin as skin-colored to yellow-brown, firm, raised papules with scaly, rough, keratotic surfaces and erythematous bases. Unlike seborrheic keratosis, actinic keratosis appears to arise from within the epidermis, rather than being “stuck on” the skin. Common sites for actinic keratosis are the face, dorsum of the hands, upper chest, upper back, and lower lip. The lesions are typically a few millimeters to
The percentage of patients with actinic keratoses that progress to have squamous cell carcinoma has been reported to be from 1% to 20% in various studies. The tendency of squamous cell carcinoma arising from an actinic keratosis to metastasize is low, 0.5% to 3%, except in squamous cell carcinoma arising from actinic cheilitis on the lower lip, which has been found to metastasize in 11% of cases. If induration, erythema, or erosion is observed, possible progression to squamous cell carcinoma should be suspected. If malignant transformation is suspected, shave excision of the lesion and histologic examination is recommended. Lesions may transform into squamous cell carcinoma, more rapidly in patients who are immunosuppressed or have genetic defects of DNA repair enzymes (e.g., xeroderma pigmentosum). The report from
Differential diagnosis includes seborrheic keratoses, lichenoid keratoses, warts, pigmented basal cell carcinomas, lentigo maligna, malignant melanomas, and early squamous cell carcinomas. Existing lesions are treated by curettage and electrodesiccation or liquid nitrogen cryotherapy. Multiple and widespread lesions may be treated with application of a cream or solution containing 5-fluorouracil (5-FU) for approximately 3 weeks, but this treatment may produce considerable discomfort. Dermabrasion may be an alternative. Some reports indicate spontaneous regression of lesions. Protection against excessive sun exposure by the use of protective clothing and sunscreens helps to prevent actinic keratosis.
CHEMICAL AND OTHER KERATOSES
Other keratoses that may be considered as premalignant cutaneous dysplasia are arsenical keratosis induced by exposure to arsenic, tar keratoses induced by exposure to tars and other polycyclic aromatic hydrocarbons, thermal keratoses induced after prolonged (less than 20 years) exposure to infrared radiation that causes chronic cutaneous thermal damage, chronic radiation keratoses induced by exposure to x-rays, and chronic cicatrix keratoses or scar keratoses that may degenerate to scar carcinoma with metastatic potential. Excision of these lesions is usually the treatment of choice.
CUTANEOUS HORN
Cutaneous horn appears as a protuberant, raised, hard, hyperkeratotic nodule with an erythematous base commonly occurring on the sun-exposed areas of pale-skinned persons. It usually develops due to the underlying dysplasia or frank neoplasm, and it is therefore considered a marker of skin cancer. The histopathology of the underlying lesion includes actinic keratosis, seborrheic keratosis, filiform verruca, trichilemmoma, basal cell carcinoma, squamous cell carcinoma, molluscum contagiosum, or keratoacanthoma. A histopathologic study indicates that 39% of underlying lesions were premalignant and malignant epidermal lesions, and 61% were benign lesions. This study also found that lesions were most likely to have a malignant base if located on the nose, pinna of the ear, or dorsum of the hand. The lesion with a portion of the base should be removed for histologic diagnosis. Final treatment is usually determined by the pathology of the underlying tissue.
RADIATION DERMATITIS
Radiation dermatitis is the term used to describe skin damage due to x-irradiation, whether exposure occurs as a result of occupational hazard or therapeutic treatments. Involved skin is dry, scaly, erythematous, thin, and discolored. Areas of telangiectasia, hyperpigmentation and hypopigmentation, and hyperkeratosis and ulceration are common. Histologically, the damaged area shows atrophy and loss of polarity of epidermal keratinocytes, changes in elastic and collagen bundles of the dermis, and destruction of hair follicles, sebaceous glands, and sweat glands. There is a potential for developing actinic keratosis, basal cell carcinoma, and squamous cell carcinoma on this kind of skin. Management includes removal of any overlying skin lesion for histologic diagnosis and definitive treatment as well as continuous skin care. Soft tissue sarcomas may develop in the irradiated tissue.
BOWEN’S DISEASE
Bowen’s disease or intraepidermal carcinoma in situ is an intraepidermal squamous cell carcinoma that may involve any area of the skin but tends to favor sun-exposed areas of the face, neck, and extremities. Primarily, older persons are affected. In one study, 80% of the patients were older than 60. There is a predominance of Bowen’s disease on the lower limbs in women and on the scalp and ears in men. In one third of the patients, the lesions may be multiple. Bowen’s disease is papulosquamous and appears as a slowly enlarging, sharply defined, round to irregular plaque with a rough, scaly, hyperkeratotic, erythematous surface. Other surface characteristics include pigmentation, fissures, erosion, and ulceration. Similar to in situ epidermoid carcinoma, Bowen’s lesions may progress to invasive squamous cell carcinoma. This transformation to squamous cell carcinoma is reported in 3% to 5% of the cases, with most cases remaining as carcinoma in situ. Histologic features of Bowen’s disease include gross dysplasia of the upper layers of epidermal cells with the basal cell layer relatively normal, individual cell keratinization, and giant keratinocytes with atypical mitoses. The cause of Bowen’s disease is unknown, but UV radiation, HPV-induced epidermodysplasia verruciformis, and history of ingestion of inorganic arsenic (e.g., water or medication) are considered etiologic factors. Bowen’s disease has been seen oonexposed areas of the skin in younger persons without a history of arsenic ingestion and without HPV infections. It has been suggested that patients with Bowen’s disease are more likely to develop other cancers, including basal cell carcinoma, adnexal carcinoma, melanoma, or cancers of the lung, gastrointestinal tract, genitourinary tract, and reticuloendothelial system, but the role of Bowen’s disease as a marker of internal malignancies is unclear. A report from the Armed Forces Institute of Pathology indicates that as many as 80% of cases of Bowen’s disease have other cutaneous and noncutaneous malignancies. Another report suggests that as many as 29% of the patients have internal malignancies. Many of the published reports contain major deficiencies, such as lack of sufficient control groups. In view of the existing controversies, it is recommended that patients with Bowen’s disease have a complete clinical examination combined with the necessary diagnostic work-up. Differential diagnosis includes basal cell carcinoma, psoriasis, extramammary Paget’s disease, and actinic keratosis. Excisional biopsy of the most indurated area is recommended for histologic confirmation of Bowen’s disease and detection of squamous cell carcinoma. Cryotherapy or curettage and desiccation have been recommended, but there is a higher chance of recurrence. In recent years carbon dioxide laser vaporization has been used for lesions located on the hand and finger. Photodynamic therapy has been used for larger lesions or areas that are anatomically difficult to excise. Surgical excision of the entire lesion provides the highest cure rate.
ERYTHROPLASIA OF QUEYRAT
Erythroplasia of Queyrat is a clinicopathologic variant of Bowen’s disease, occurring almost exclusively on the glans penis of uncircumcised, middle-aged and older men. It may occur on the penile shaft, scrotum, and vulva, but it is questionable whether these presentations are Bowen’s disease of the anogenital region. The lesion appears as an asymptomatic, sharply demarcated, bright red, shiny plaque with a smooth, velvety surface. The histologic features are identical to those of Bowen’s disease, especially of the anogenital region. However, the rate of transformation into invasive squamous cell carcinoma is much higher for Bowen’s disease, and the resultant squamous cell carcinoma tends to be more aggressive. Erythroplasia of Queyrat has progressed into squamous cell carcinoma in up to 30% of patients, with metastases in 20% of patients with invasive erythroplasia. It thus has a greater tendency toward invasion and metastases than does Bowen’s disease. Radiotherapy, 5-fluorouracil (5-FU), CO(2) laser, cryotherapy and curettage, and electrodesiccation have been used. One recent study reported temporary regression with low-dose isotretinoin. However, due to its aggressive nature, it is advisable to use surgical excision of the tumor with microscopic control of the margin (Mohs micrographic surgery) and to circumcise the penis.
BOWENOID PAPULOSIS
Bowenoid papulosis is a multicentric anogenital dysplasia that presents as verrucous or lichenoid pink to reddish brown papules, erythematous macules, or leukoplakia-like lesions in young sexually active men and women. Lesions of the oral mucosa have been reported. Although pruritus and inflammation with tenderness and pain are occasionally reported, most patients with bowenoid papulosis have no symptoms. The lesions typically present as inconspicuous, noncondylomatous papules with mostly smooth and only slightly papillomatous surfaces. A homosexual patient with a long-standing history of pruritus and black perianal hyperpigmentation extending out from the anal margin proved to have bowenoid papulosis on skin biopsy. In women, the lesions tend to be bilateral, hyperpigmented, and confluent, involving the labia majora, labia minora, clitoris, perianal region, and inguinal folds; in men, the lesions usually present as small discrete papules on the glans, shaft, and preputium. The lesions are 0.2 to
The clinical features usually differentiate bowenoid papulosis from Bowen’s disease. Bowenoid papulosis usually presents in young, sexually active men and women between 20 and 40 years of age, and the lesions appear as multiple small papules. Bowen’s disease occurs in patients older than 50 years of age and is usually a single plaque. Bowenoid populations normally follow a benign but persistent course over many years. The lesions respond well to local destructive therapy, such as electrodesiccation, laser surgery, cryosurgery, and ablation. [ref: 160] Topical treatment with 5-FU and intramuscular or intralesional interferons have also been used in the treatment of bowenoid papulosis. Recurrences after treatment are common, due to the life-long nature of HPV infections. Patients with this condition, especially women, and their sexual partners are at high risk for anogenital carcinomas and should have frequent evaluations, including Pap smears, because of the high risk for anogenital carcinoma, which is usually a squamous cell carcinoma.
EPIDERMODYSPLASIA VERRUCIFORMIS
Epidermodysplasia verruciformis is a rare, chronic, and often hereditary disease (autosomal recessive, familial, or sporadic) characterized by widespread eruption of flat, wart-like lesions and reddish brown plaques with slightly scaly surfaces and irregular borders. The wart-like lesions are mainly distributed on the hands, feet, and face, sometimes in a linear arrangement, and the pigmented plaques preferentially involve the trunk, neck, and proximal parts of the extremities. Epidermodysplasia verruciformis is an extremely protracted disease that usually begins in infancy or early childhood (5 to 11 years) with various types of warts and plaques. Later, it may progress to form verrucous plaques and nodules or transform to in situ squamous cell carcinoma. The rate of appearance of new lesions varies considerably, with some lesions disappearing in some areas as more appear in other areas. Malignant tumors typically develop in the third or fourth decade of life in approximately one third of the patients. These patients are not inconvenienced by these transformations, which are neither painful nor itchy. Malignant lesions are numerous and continue to progress as noninvasive in situ carcinoma without metastasis. They are locally destructive if not treated. These premalignant lesions have many characteristics of actinic keratoses. Malignant transformation to squamous cell carcinoma in predominantly sun-exposed areas occurs in 30% of patients. In addition, more than 50% of patients with epidermodysplasia verruciformis develop actinic keratoses; they are increased iumber, occur at a younger age, and more frequently become malignant when compared with the general population. In epidermodysplasia verruciformis, the cancers that develop usually occur on the forehead or in areas of trauma. They are slow growing, microinvasive or invasive, and may be locally destructive. They do not metastasize.
The etiologic factors associated with the development of epidermodysplasia verruciformis are impaired cellular immunity, HPV infection, and the genetic makeup of the host. Most patients with epidermodysplasia verruciformis show defects in cell-mediated immunity characterized by anergy to dinitrochlorobenzene and common skin antigens, depressed lymphocyte blastogenic reactivity to mitogens, abnormal T-lymphocyte populations, and decreased number of T lymphocytes. However, the humoral immune system is left intact. HPV 5, 8, 9, 12, 14, 15, 17, and 20 have been isolated from epidermodysplasia verruciformis lesions and are thus felt to contribute to the development of epidermodysplasia verruciformis. HPV types 5 and 8 have consistently been found in benign and malignant lesions of patients with epidermodysplasia verruciformis. HPV appears to take advantage of the immunologic state to induce epidermodysplasia verruciformis. The occurrence of HPV infections and epidermodysplasia verruciformis in immunosuppressed renal allograft recipients, in an immunosuppressed patient with systemic lupus erythematous, and in a patient with Hodgkin’s disease provides support for this hypothesis. Another line of evidence comes from the family members of patients with epidermodysplasia verruciformis, in whom the wart-like lesions started to appear but eventually regressed. However, there are patients with epidermodysplasia verruciformis without immune dysfunction. Some investigators believe that the impaired cellular immunity could be a result of the HPV infection. It has been speculated that patients with epidermodysplasia verruciformis are susceptible to widespread infection with HPV and skin cancers because of a limited, specific immunologic defect, such as a defect in an unidentified suppressor gene that prevents infection with the HPVs associated with epidermodysplasia verruciformis.
On histologic examination, the epidermal changes resemble that of verruca plana but are more dramatic. Affected cells may be laden with keratohyalin granules and are often swollen and irregularly shaped. Dyskeratotic cells may be present in the lower portion of the epidermis.
There is no effective therapy for epidermodysplasia verruciformis, but aromatic retinoids, such as etretinate, may have some beneficial effects. Interferons produce only partial responses during therapy, with lesions returning in the same locations after therapy is discontinued. Surgical removal, electrosurgical approaches, and cryotherapy are used, as is skin grafting of areas with many precancerous or cancerous lesions, such as the forehead, with uninvolved, non-sun-exposed skin. Radiation therapy is contraindicated. Prevention by the use of sunscreens in those with the autosomal recessive gene is essential. It is important for HPV-infected, immunosuppressed patients to have regular dermatologic examinations.
LEUKOPLAKIA
Leukoplakia is a clinical description that denotes white patches or plaques that cannot be rubbed off located on the mucous membranes of the oral, anal, and genital mucosa. About 20% of the patients have histologic findings that consist of epithelial dysplasia and hyperkeratosis of the mucosa. Approximately 15% of the dysplastic cases develop carcinoma in situ, and 3% to 6% develop invasive squamous cell carcinoma within the involved area. Differential diagnosis includes lichen planus, candidiasis, hairy leukoplakia, white sponge nevus, pachyonychia congenita, and dyskeratosis congenita. Treatment consists of excision, 5-FU, laser surgery, and cryosurgery. Some have advocated using 13-cis-retinoic acid, but the potential benefit has not conclusively been shown to justify the risk in its use to control oral leukoplakia. Some have proposed the use of antioxidants and beta-carotene to control leukoplakia. Nearly all agree that abstinence from tobacco and alcohol are important in reducing precancerous leukoplakia of the oral mucosa.
ORGANOID OR SEBACEOUS NEVI
Organoid or sebaceous nevi are skin lesions composed of abnormal numbers of pilosebaceous structures and apocrine glands. These lesions appear as raised, yellowish papules that grow gradually to develop a papillomatous surface. They usually occur on the scalp and are present at birth or appear soon afterwards. Basal cell carcinoma has been reported in as many as 10% of patients, and squamous cell carcinoma occurs rarely. Syringocystadenoma papilliferum, derived from the apocrine gland and basal cell carcinomas may be associated with organoid nevus. Histologically, the lesion consists of large numbers of skin appendages, including sebaceous, eccrine, and apocrine glands, hair follicles, and smooth muscle. Because of these structures, the term organoid nevus appears more appropriate than sebaceous nevus. In syringocystadenoma papilliferuim, cystic lesions of the epidermis are lined by a double layer of columnar epithelium showing decapitation secretion, suggesting apocrine gland origin. After histologic confirmation of the diagnosis, the lesion is usually removed by surgical excision. Ideally, the lesion should be excised before puberty, because the risk of malignancy increases following puberty.
MALIGNANT EPIDERMAL TUMORS
BASAL CELL CARCINOMA
Basal cell carcinoma (BCC) is the most common cancer among whites. It accounts for 77% of the 700,000 new cases of nonmelanoma skin cancer in the
Basal cell carcinomas can be seen in association with several conditions including basal cell nevus syndrome (multiple BCCs, palmar pits, characteristic facies, odontogenic cysts), Bazex’s syndrome (basal cell carcinoma resembling nevi, follicular atrophoderma, anhidrosis, hypotrichosis), xeroderma pigmentosum (premature skin cancers, photosensitivity, freckling, neuronal degeneration, faulty repair of UV radiation damage to DNA), occulocutaneous albinism (skin cancer, nystagmus, photophobia, decreased or absent melanin), linear basal cell nevi, and Rombo syndrome (multiple basal cell carcinomas, trichoepitheliomas that may give rise to basal cell carcinomas, grainy skin, vermicular atrophoderma, hypotrichosis, and acrocyanosis).
The incidence of basal cell carcinoma increases with age, but because of easy access to the sunbelt areas, outdoor recreational activities, and the popularity of suntanning, basal cell carcinoma is increasingly seen in younger people. It is no longer surprising to see basal cell carcinoma in patients in their third or fourth decade of life. Most basal cell carcinomas occur on sun-exposed areas such as the face (Figs. 40.2-9 and 40.2-10), especially the nose, the nasolabial fold, and the inner canthus areas. Solitary basal cell carcinomas are seen in the geographic areas with temperate climate. Multiple tumors and tumors on areas other than the face are seen usually in tropical and equatorial regions. Basal cell carcinoma occurs more commonly in men than in women, with the incidence increasing with age and with latitude.
Clinical Presentation
Typical basal cell carcinoma appears as a slowly growing, shiny, skin-colored to pink, translucent papule with telangiectasias. As the lesion enlarges, it may ulcerate and develop a rolled border and crusted center. The lesion may then regress to a smaller and less visible size. Normally, basal cell carcinomas are reported on only hair-bearing skin, but in cases of basal cell nevus syndrome and linear basal cell nevi, they may be seen on the soles and palms. Several clinicopathologic variants of basal cell carcinoma have been described:
1. Nodular and nodular ulcerative are the most common variants of basal cell carcinoma, and they show central depression and a rolled border; they may bleed with trauma, ulcerate, and form crust, and they are usually located on the head and neck. 2. Pigmented basal cell carcinoma appears nodular with brown to black pigment and must be differentiated from melanoma; it often occurs in darker-complexioned patients.
3. Superficial basal cell carcinoma is flat, erythematous, and scaly and has a thread-like border. It is often found on the trunk, extremities, and neck, and one must rule out psoriasis and eczema.
4. Sclerosing or morphea-type basal cell carcinoma appears as an infiltrated yellowish sclerotic scar-like patch with indistinct margins typically on the face and neck; it lacks telangiectasias, translucency, and a distinct border. Morphea-type basal cell carcinoma may undermine skin and go undetected for an extended period of time. Physical evaluation of the lesion and curretage are poor at defining the extent of this tumor, and Mohs surgery is frequently an appropriate treatment. 5. Cystic basal cell carcinoma often appears as a translucent blue-grey nodule on the face. 6. Fibroepithelioma of Pinkus usually presents as a sessile growth on the lower trunk and thighs.
7. Hyperkeratotic basal cell carcinomas may occur on the head, trunk, and extremities, and these must be differentiated from actinic keratoses, seborrheic keratoses, squamous cell carcinoma, Bowen’s disease, and psoriasis.
8. Wildfire or multicentric basal cell carcinoma occurs on the face as a plaque made of nests of lesions with ulceration and scars.
9. Other variants include giant exophytic basal cell carcinoma and giant pore type basal cell carcinoma. Red dot or halo basal cell carcinoma appear as 1- to 2-mm red papules on the face with or without a halo. Another form of basal cell carcinoma is the so called basosquamous cell carcinoma. This lesion has biologic behavior and pathologic features intermediate between basal cell and squamous cell carcinoma.
Many interesting and unusual cases of basal cell carcinoma have been reported. Pedal basal cell carcinomas have been described, and these are felt to arise from pluripotential epithelial cells, since there are no pilosebaceous or apocrine glands in the feet; it remains speculative whether these pedal basal cell carcinomas are posttraumatic. Basal cell carcinomas have also been reported in association with vaccination and burn scars, nevus sebaceous, nevus flameus, hemangiomas, and a host of other conditions. Linear basal cell carcinoma is a rare form in which lateral spread typically occurs beyond clinically apparent borders. Some question if koebnerization by deep scratching with strornal implantation may occur.
Some basal cell carcinomas may have a hyperkeratotic surface and scaly appearance and may be confused with actinic keratosis, sebaceous keratosis, squamous cell carcinoma, and Bowen’s disease. The morphea-type basal cell carcinoma has a scar-like sclerotic appearance and lacks telangiectasia and translucency, and the usual distinct border is not seen. The lesion is larger and more indurated in palpation than on inspection. Another form of basal cell carcinoma is the so-called basosquamous cell carcinoma. This lesion has biologic behavior and pathologic features intermediate between basal cell and squamous cell carcinomas.
Most basal cell carcinomas are diagnosed when tumors are a few millimeters to 1 to
Most reports indicate the importance of fair complexion, blue eyes, and fair hair in the development of basal cell carcinoma. This is especially true if the people have fair complexions and live in a sunny region, such as Scandinavians who live in
Basal Cell Carcinoma, T1N0M0 (growth pattern)
The association of chronic UV light exposure and development of basal cell carcinoma is well accepted. It is the cumulative effect of UV light over many years that causes the development of basal cell carcinoma. After the necessary level of sun exposure (threshold) is reached, the exposed person continues to develop basal cell carcinomas; the process appears to be irreversible. UVB is the major carcinogen, with UVA playing a lesser role. Ionizing radiation, whether received from the environment or as part of therapeutic modalities, can promote the development of basal cell carcinoma. X-irradiated skin shows evidence of chronic dermatitis several years later and may give rise to basal cell carcinomas. In the 1950s and 1960s, x-irradiation was used for the treatment of acne and hypertrichosis. Many patients who received therapeutic x-rays developed radiation dermatitis and multiple skin cancers of the face 20 to 30 years later.
The chance of developing basal cell carcinomas increases directly in relation to the number of previous basal cell carcinomas. Other conditions that may predispose to the development of basal cell carcinoma include other primary malignancies, especially lymphoreticular types, immunodeficiency states, and trauma such as traumatic injuries or smallpox vaccination. While the risk for basal cell carcinomas is somewhat increased following exposure to inorganic arsenic or industrial tars and following immunosuppression, the relative risk for developing squamous cell carcinomas following these exposures is much higher. Although basal cell carcinomas are most commonly seen on the sun-exposed areas of the skin, many basal cell carcinomas are seen in less exposed areas, such as the nasolabial fold and the inner canthus, suggesting that other factors (e.g., embryonic closure lines) may contribute to the development of the tumor. While the role of p53 and tumor-suppressor gene, in the pathogenesis of basal cell carcinomas has not yet been elucidated, decreased wild-type p53 protein has been found in basal cell carcinomas, squamous cell carcinomas, and keratoacanthomas with a concomitant loss of growth suppression. Histopathology In all forms of basal cell carcinomas, the histologic features include masses of compactly arranged basaloid cells resembling cells in the basal layer of epidermis extending down from the epidermis into the dermis. Basal cell carcinomas arise from immature pluripotential cells of the epidermis or from the outer root sheath of hair follicles, growth is dependent upon the stroma, and basal cell carcinoma may differentiate towards sebaceous, apocrine, or eccrine structures or remain undifferentiated. There is usually a variable dermal stromal reaction with connective tissue proliferation, increased acid mucopolysaccharides, and stromal retraction or peritumoral lacunae. In the sclerosing type of basal cell carcinoma, the cells are more fusiform and are admixed with a more pronounced stromal reaction. In most cases, especially in the nodular, cystic, and adenoid types, there is a clear peripheral palisading of the dark epithelial cells that is characteristic of basal cell carcinoma. The differential diagnosis includes trichoepithelioma, which may also show some palisading of the cells. Pigmented basal cell carcinoma is due to aggregation of melanin within the tumor. Basosquamous cell carcinoma consists of a fibrous stroma admixed with a downward proliferation of epithelial cells that lack palisading and have some features of squamous cell carcinoma and more frequent mitosis.
Metastasizing Basal Cell Carcinoma.
Basal cell carcinoma usually spreads by direct extensions from the primary tumor site and has little tendency to metastasize. Recurrences of basal cell carcinoma usually occur within 4 to 12 months from either aggressive tumor or inadequate treatment. Aggressive behavior in basal cell carcinomas may be forewarned by recurrences, histologically aggressive tumors, location in the H-zone of the central face, size larger than
Management of Basal Cell Carcinoma
Several methods are available for the treatment of basal cell carcinoma including surgical excision, Mohs surgery, electrodesiccation and curretage, cryosurgery, radiation therapy, interferon, laser treatment, and use of sun block and oral retinoinds in high-risk groups. The choice is usually based on the location, type of lesion, and experience of the physician. Smaller lesions can be treated with surgical excision or curretage and desiccation. Radiation therapy, although used less frequently, is the most appropriate treatment for basal cell carcinoma of the eyelid, nose, and lips. Cryosurgery and topical use of 5-FU have also been suggested, but the cure rate is not as high as with other techniques. Tumors with poorly defined margins and recurrent basal cell carcinomas are best treated by microscopically controlled excisional surgery that is known as Mohs micrographic surgery. This technique allows the removed tissue to be mapped in relation to the underlying site. Additional tissue removal is carried out at sites where tumor cells are present microscopically. The treatment is considered complete when there is no tumor found in removed tissue. Cure rates of 90% to 95% are achieved with most methods of treating primary basal cell carcinoma. Mohs surgery has a cure rate of 96% to 99%. Recurrent lesions are larger than 1 to
SQUAMOUS CELL CARCINOMA
Squamous cell carcinoma is a malignant skin cancer arising from epidermal keratinocytes with the potential for metastasis. Squamous cell carcinomas are usually seen in fair-skinned persons who have had excessive sun exposure and developed actinic keratosis. It is more common in men than women. The incidence of squamous cell carcinoma is reported to be 289 per
Pathogenesis
Many factors may contribute to the genesis of squamous cell carcinoma. As in basal cell carcinoma, it is the cumulative amount of lifetime UV exposure that influences the development of squamous cell carcinoma. Fair skinned, red-haired individuals and albinos are at especially high risk for developing squamous cell carcinoma early in life if they are exposed to excessive amounts of UV light, indicating the protective value of pigmentation. Occupational and nonoccupational exposure to sunlight are significant risk factors. It is reported that there is an increased incidence of squamous cell carcinoma in persons who have been treated with PUVA. PUVA treatment has been available in the
The incidence of squamous cell carcinoma is increased among patients who are on immunosuppressive therapy for organ transplantation as well as in patients infected with HIV. In these patients, the ratio of basal cell carcinoma to squamous cell carcinoma is decreased or reversed, suggesting that the immune system may influence the control of early squamous cell carcinoma. Chronic ulceration, chronic sinus disease, chronic inflammation, and scar tissues increase the chance of developing squamous cell carcinoma. Exposure to ionizing radiation can result in the development of squamous cell carcinoma several years later. Certain topical agents used for the treatment of chronic skin conditions are carcinogenic and perhaps cause squamous cell carcinoma. Tar causes skin cancer in laboratory animals, but there is no evidence of tar’s causing skin cancer in humans. Squamous cell carcinoma was considered an occupational hazard among cotton spinners due to contact with the cotton oil. Industrial exposure to chemicals (e.g., cutting oils) is now recognized as a risk factor for squamous cell carcinoma. The historic reports of scrotal malignancies in chimney sweeps was assumed to be due to contact with carcinogenic soot. Arsenic was also recognized to be a carcinogen producing skin cancer and cancers of other organs. In the presteroid era, arsenic was used in the treatment of many diseases. Direct exposure of the skin by industrial workers to hydrocarbons such as paraffin, tar, shell oil, and creosote oil may cause squamous cell carcinoma.
Many HPV types have been found in some squamous cell carcinoma. HPV types 16 and 18 were found in cancers of the cervix. HPV infections are normally found in patients who are not immunosuppressed. Squamous cell carcinoma of mucocutaneous sites usually occurs in patients with a history of heavy alcohol and tobacco use. Lip and mouth lesions are also associated with chewing tobacco and betel nuts. The lower lip is a common location of squamous cell carcinoma associated with solar exposure.
Clinical Presentation
Squamous cell carcinoma usually appears on the areas of the skin that are damaged by sun exposure. The most common sites of squamous cell carcinoma are the face, neck, back, forearm, and dorsum of the hand in elderly white persons with sun-damaged skin. The primary lesion manifests as a red, indurated papule that appears de novo or on an actinic keratosis and expands rapidly, producing a large nodule that eventually ulcerates and metastasizes to a local draining lymph node. Bowen’s disease, cutaneous horn, chondrodermatitis nodularis helicis, chronic ulcers, scar tissues, and radiodermatitis may be precursor sites of squamous cell carcinomas and basal cell carcinomas. The patient may have several primary lesions developing at the same time or in a rapid succession. Tumors arising from sun-damaged skin have relatively lower rates of metastasis, but tumors arising on the mucocutaneous surface or those arising from skin with previous tissue changes (e.g., scars of burns, sites of trauma) tend to be more aggressive, invade locally, and metastasize rapidly.
Squamous Cell Carcinoma, T4N0M0
Squamous cell carcinoma of mucocutaneous sites usually occurs in patients with a history of heavy smoking and heavy alcohol intake. These lesions may arise from an area of leukoplakia or an indurated or ulcerated plaque and metastasize rapidly. Chewing tobacco and betel nuts can influence the development of oral squamous cell carcinoma. Verrucous carcinoma may occur in the oral cavity or sole of the foot and may present as a persistent, firm, or vegetating plaque on these areas that enlarges slowly, invades locally, and rarely metastasizes.
Squamous Cell Carcinoma of lower eyelid with invasion in bulbar conjunctiva,
T4N0M0
Giant condylomata of Buschke-Lowenstein usually presents as warty lesions on male genitalia. This condition is associated with HPV. Squamous cell carcinoma may develop in a preexisting condition of Buschke-Lowenstein condylomata. Epithelioma cuniculaturn usually presents as a small ulcer with peripheral hyperkeratosis on the soles of the feet. This lesion is difficult to eradicate.
Diagnosis, Classification, and Prognosis
Diagnosis of squamous cell carcinoma ultimately requires histologic conformation. The characteristic features of squamous cell carcinoma include proliferation of atypical squamous cells invading the underlying dermis producing an irregular, disorganized architecture. Individually keratinized cells and horn cysts are present. In verrucous carcinoma, giant condyloma of Buschke-Lowenstein, and epithelioma cuniculatum, there is hyperplasia of the mucous membrane epidermis but no evidence of invasion into the dermis. Mitotic figures are infrequent.
Attempts have been made to classify squamous cell carcinoma based on the percentage of the undifferentiated cells and to relate this to the prognosis. There are four grades of squamous cell carcinoma. Grade I has fewer than 25% undifferentiated cells, and grade IV has more than 75%. The higher the number of undifferentiated cells, the worse the prognosis. Differential diagnosis of squamous cell carcinoma includes keratoacanthoma, seborrheic keratosis, pyogenic granuloma spindle cell melanoma, and soft tissue sarcoma with spindle cells. Antikeratin antibodies can be used to identify and confirm the keratinocyte origin of a given tumor. This is even possible in formalin-fixed, paraffin-embedded specimens. Antibodies against S-100 proteins or NKIC3 indicate a possible melanoma.
Management
Most practitioners use similar treatment modalities for basal cell carcinoma and squamous cell carcinoma. Squamous cell carcinoma requires a more aggressive approach, such as wider excision. Tumors on the mucous membranes must be excised with good margins. Mohs micrographic surgery may be used. Close follow-up for recurrences and metastatic spread is highly recommended. In most cases, surgery is preferred to radiation therapy.
The selection of a treatment modality for squamous cell carcinoma depends on several factors: size, shape, and location of lesion, depth of infiltrate, desired cosmetic result, and physician experience. Excision with adequate margins, radiation, and electrodesiccation are the most frequently used. Radiation treatment in verrucous carcinoma is, however, probably contraindicated. Radiation or chemotherapy may be used in local or distant metastasis. Cryosurgery has been used with reasonable cure rates by experienced physicians.
Methods of treatment currently under development or in early clinical use include laser radiation, photodynamic therapy, and local immunotherapy. Trials involving interferon-alpha and humaatural leukocyte interferon have shown good response rates in both basal cell carcinomas and squamous cell carcinomas.
Squamous Cell Carcinoma of Cheek, T4N0M0
Results after half course of gamma-therapy, 45 Gy
Surgery approach:
Skin Squamous Cell Cancer of Neck, T4N0M0. Postoperative wound is temporary covered with pig skin xenografts
Granulation wound after removing of xenografts. Skin autografting.
Complete recovery. 12 months later.
Screening and Prevention
Approximately 700,000 nonmelanoma skin cancers are diagnosed and treated in the
An additional factor contributing to the increasing incidence of skin cancer is the ballooning number of elderly people who are at increased risk for skin cancers due to a lifetime of UV radiation exposure. Theoretically, if everyone were fully protected from UV radiation beginning today, basal cell carcinomas and squamous cell carcinomas secondary to solar exposure would continue to appear for several generations due to damage already perpetrated. Currently, the most logical and effective approach is to identify those at risk for skin cancer, provide early diagnostic guidelines, and educate the public about the need for regular skin cancer screening by an experienced physician. Excessive sun damage can be prevented with commercially available sunscreens. A sun protective factor (SPF) of at least 15 is recommended in the general population. Although sunscreen use has not yet been proven to decrease the incidence of skin cancer, it has been shown to decrease the incidence of precancerous actinic keratoses. Increased pigmentation in light-skinned persons could be an alternative approach. Several laboratories are experimenting with a melanocyte-stimulating hormone analogue that would produce greater melanization of the skin after topical or systemic use.
Another possible approach is the reversal of actinic damage before the development of skin cancers. It is thought that UV-induced photoaging is the precursor of actinic keratoses that are precursors of nonmelanoma skin cancers. By reversing photoaging, it may be possible to decrease, prevent, or delay the development of these nonmelanoma skin cancers. Studies have demonstrated that with use of sunscreens, the number of actinic keratoses in patients are decreased.
CANCER-ASSOCIATED GENODERMATOSES
Several inherited dermatoses are associated with the development of cutaneous malignancies. These include xeroderma pigmentosum, nevoid basal cell carcinoma, familial dysplastic nevus syndrome, and multiple self-healing epitheliorna of Ferguson-Smith.
XERODERMA PIGMENTOSUM
Xeroderma pigmentosum is a rare autosomal recessive disease that occurs in approximately
The underlying abnormality is a defect in the first steps of nucleotide excision repair. UV radiation induces damage to DNA, known as photoproducts, which iormal persons are excised and replaced. In xeroderma pigmentosum, the rate of enzymatic removal of these photoproducts is substantially decreased, so that the occurrence of spontaneous mutations is increased. [ref: 283] The faulty enzyme may be an endonuclease in some cases or an accessory proteieeded to initiate the repair process in others.
Xeroderma pigmentosum is a heterogenous disease. Seven complementation groups, A through G, have been identified. The sensitivity to UV radiation and the degree of neurologic disease vary between each group. Approximately 20% of xeroderma pigmentosum patients do not fall into one of the complementation groups. These people, known as xeroderma pigmentosum variants, are clinically indistinguishable from classic xeroderma pigmentosum. However, the nucleotide excision repair system is not defective in these patients. It is hypothesized that these patients may have a postreplication repair defect. Although the clinical picture is very characteristic in older children, a young child may present with only sun sensitivity. In this case, it is necessary to rule out the possibility of porphyria with the proper porphyrin studies. The clinical diagnosis of xeroderma pigmentosum can be confirmed by DNA repair study of fibroblast culture. Management centers on avoidance and protection from UV light exposure and adequate treatment of precancerous and cancerous skin lesions. The use of topical 5-FU or cryotherapy for multiple lesions is beneficial. Radiation therapy is not recommended. Oral retinoids have been successful in the prevention of new skin cancers, but withdrawal of therapy results in reversal of this effect. The severity of the disease correlates closely with the residual repair capacity. Avoidance of the sun and use of total sunblocks as early in life as possible should be initiated. With the use of sunscreens and other protection mechanisms, these patients live much longer, and they usually die of neurologic involvement and infection rather than UV-induced cancer. Neurologic involvement and progressive mental dysfunction are seen in more than 50% of those who live longer.
NEVOID BASAL CELL CARCINOMA SYNDROME
Basal cell nevus syndrome is inherited as an autosomal dominant gene with complete penetrance and variable expression. It is a relatively common genodermatosis, but 30% to 50% of the patients have no family history, suggesting a high rate of spontaneous mutation. The gene defect is mapped to chromosome 9q, but the abnormal gene product has not yet been delineated. Clinical features are seen in skin, bone, optic tissue, and the central nervous system. Skin lesions consist of basal cell carcinoma that have atypical features resembling skin tags or pink or flesh-colored small papules and palmar and plantar pits. Bone cysts are seen in the mandible, and abnormalities of the ribs and vertebrae occur. The ophthalmic evidence includes coloboma or cataracts. Other manifestations include intracranial calcification, ocular hypertelorism, enlarged occipitofrontal head circumference, macrocephaly, and increased risk of developing certain tumors, especially medulloblastoma, meningioma, and ovarian fibromas.
It is thought that the expression of the autosomal dominant gene requires solar radiation. Skin lesions usually develop in large numbers between puberty and age 35. They may become nodular or ulcerative and locally aggressive and cause significant destruction of normal tissue. After basal cell carcinoma appears, various modes of therapy should be used to remove the tumors completely. Radiation therapy is not recommended, since there is evidence that it may lead to the development of additional basal cell carcinomas as well as other malignancies.
Cutaneous Melanoma
Cutaneous melanoma is becoming a more common disease. In 1995, an estimated 32,100 individuals developed melanoma and 7200 died of the disease in the
Fortunately, most new patients are diagnosed early in the disease’s clinical course, when it can be cured with simple surgical treatment. While some subtlety can be involved in the diagnosis of “thin” melanomas, over 90% of intermediate and thick melanomas can be recognized as malignant by experienced observers; therefore, it behooves each physician to know the clinical characteristics of melanoma so that biopsies can be performed on suspicious moles or skin lesions as early as possible. Histologic verification and microstaging with an accurate tumor thickness are essential before embarking on treatment. The options for therapy range from very conservative surgical treatment for early lesions to more radical approaches for biologically aggressive melanomas. Judgment, experience, and knowledge of the prognostic factors are essential for choosing the most appropriate treatment for individual patients.
CLINICAL CHARACTERISTICS
HIGH-RISK POPULATIONS
Investigators have identified risk factors for the development of melanoma and mathematic models have been used so that “high-risk” populations can be identified for potential screening programs. Counting the number of raised nevi on the arms has been suggested as a simple self-screening test to identify persons requiring physician examination. Prospective randomized controlled trials should assess the feasibility of this procedure so that models can be produced for the population screened. These models can then be applied to statewide and national programs. For example, MacKie and colleagues followed 116 patients with 3 or more clinically atypical nevi for at least 5 years. Patients were examined and photographed every 3 to 6 months and nevi undergoing change were excised for histologic diagnosis. Among 85 patients with no personal or family history of melanoma, 5 invasive melanomas developed during 583 person-years of follow-up. The expected number in this population was 0.05 and the calculated relative risk was 92 (P < .001). There was a similarly increased risk among 24 patients with atypical nevi and a family or personal history of melanoma, with a relative risk reported to be 91. By comparison, no second melanoma developed among 25 patients with previous melanoma and a normal nevus pattern during 213 person-years of similarly intensive follow-up. The risk of melanoma was highest among seven patients with atypical nevi and a family history of melanoma (relative risk, 444). The median thickness of surveillance-detected melanomas was
A patient with a past history of melanoma has a higher risk of developing a second primary melanoma than an individual in the general population has of developing a melanoma. [ref: 9] The risk of developing a second and third melanoma varies from 3% to 7% in different series, a 900 times higher risk than that of the general population. A patient who has multiple dysplastic nevi or has a familial form of melanoma has an even greater risk of developing multiple primary melanomas than do other patients. Familial melanomas are uncommon but have been well documented, and individuals in such families constitute an identifiable high-risk group. Between 4% and 10% of patients describe a history of melanoma among their first-degree relatives. Clark and colleagues described an autosomal dominant hereditary occurrence of melanoma, originally termed the B-K mole syndrome and now referred to as the dysplastic nevus syndrome, familial type (FAM-M). Patients with this syndrome typically have between 10 and 100 pigmented lesions, predominantly on the trunk, buttocks, or lower extremities. Genetic factors may also play an important role in the predisposition to melanoma, perhaps by genes controlling some aspect of immune response to melanoma antigens. Melanoma-prone patients appear to have increased frequencies of certain blood, complement, and human leukocyte antigen (HLA) phenotypes.
SIGNS AND SYMPTOMS
Melanomas can be located anywhere on the body, but most commonly occur on the lower extremities in women and the back (between the scapulae) in men. Several articles have described and illustrated some of the clinical characteristics of melanoma. Typical features of intermediate and thick cutaneous melanoma include (1) variegation in color; (2) an irregular raised surface; (3) an irregular perimeter with indentations; and (4) ulceration of the surface epithelium.
Although melanomas may have a variety of clinical appearances, the common denominator is their changing nature. Therefore, any pigmented lesion that undergoes a change in size, configuration, or color should be considered a melanoma, and an excisional biopsy should be performed. Recently, emphasis has been placed on the early diagnosis of “thin” melanoma: lesions that are malignant but curable with simple surgical excision. This increased emphasis is due to the ability to identify high-risk populations and the proliferation of screening programs for the disease, both on a national and regional basis. The diagnosis of “thin” melanomas may be subtle, but physicians and primary care providers need to familiarize themselves with early signs of malignant melanoma. These include the ABCD’s of early diagnosis with A denoting asymmetry of the lesion, B border irregularity, C color variation, and D a diameter greater than
GROWTH PATTERNS
A convenient way to categorize melanomas is by their growth patterns. These growth patterns represent distinct pathologic entities and have unique clinical features that should be identifiable by the experienced clinician. Histologic confirmation is obviously essential before making any definitive treatment plans. The four major growth patterns are superficial spreading melanomas, nodular melanomas, lentigo maligna melanomas, and acral lentiginous melanomas.
Superficial Spreading Melanoma
Superficial spreading melanoma (SSM) constitutes the majority of melanomas (about 70% in most series). The lesions generally arise in a preexisting nevus. A history of slowly evolving change of the precursor lesion over 1 to 5 years is not uncommon, with more rapid growth developing months before diagnosis. SSMs can occur at any age after puberty. A typical SSM first appears as a deeply pigmented area in a brown junctional nevus. The lesion may take on a lacy appearance.
Often there are patches of regression recognizable by an amelanotic area. Early in its evolution, an SSM is generally a flat lesion. It may develop an irregular surface, usually asymmetrically, depending on the vertical growth phase that occurs as it enlarges. As the lesion grows, the surface may become glossy. There is characteristic notching or indentation of the perimeter, especially as the SSM enlarges.
Nodular Melanoma
Nodular melanoma (NM) is the second most common growth pattern in most series (15% to 30% of patients). NMs are more aggressive tumors and usually develop more rapidly than SSMs. They can occur at any age (but usually in middle age and are most common on the trunk or head and neck). Men tend to have more NMs than women, whereas the opposite is true for SSMs. NMs are usually
Nodular melanomas are generally darker than SSMs, more uniform in coloration, and more raised or dome shaped. The typical NM is a blue-black lesion that often resembles a blood blister or hemangioma. It may have other shadings of red, gray, or purple. About 5% of NMs lack pigment altogether (i.e., are amelanotic) and have a fleshy appearance. NMs are often symmetric, but sometimes they appear as irregularly shaped plaques. They lack the radial (horizontal) growth phase so typical of the other growth patterns and therefore have discrete, sharply demarcated borders, often with irregular perimeters. NMs that are polypoid, with a stalk or cauliflower appearance, are particularly aggressive.
Lentigo Maligna Melanoma
Lentigo maligna melanoma (LMM) appears to be a separate entity from melanomas with the other growth patterns, because LMMs do not have the same propensity to metastasize. LMMs constitute a small percentage of melanomas (usually 4% to 10%) and are typically located on the face in older caucasian women. Usually LMMs have been present for long periods of time (5 to 15 years). They are generally large (more than
Almost all are located on the face or neck, although a few may occur on the back of the hands or the lower legs. They are typically tan lesions with differing shades of brown. Irregular mottling or flecking may appear as the lesions enlarge, with areas of dark brown or black in some parts and areas of regression in others. LMMs can have extremely convoluted borders with prominent notching and indentation, which generally represent areas of regression. The diagnosis of LMM requires the presence of sun-related changes in both the epidermis and dermis.
Acral Lentiginous Melanoma
Acral lentiginous melanoma (ALM) characteristically occurs on the palms or soles or beneath the nail beds. However, not all plantar or volar melanomas are ALMs: a minority are SSMs or NMs. ALMs occur in only 2% to 8% of caucasian melanoma patients but in a substantially higher proportion (35% to 60%) of dark-skinned patients, such as blacks, Asians, and Hispanics.
The majority of ALMs are located on the sole of the foot. They are generally large, with an average diameter of about
Acral lentigines melanoma also include subungual melanoma, an infrequently seen presentation of cutaneous melanoma. It develops in only 2% to 3% of caucasian patients but in a higher proportion of dark-skinned patients. It occurs equally in men and women and is most often diagnosed in older patients (median age: 55 to 65 years). Over three-quarters of subungual melanomas involve either the great toe or the thumb. The most common sign of an early subungual melanoma is a brown to black discoloration under the nail bed.
SCREENING FOR MELANOMA
Malignant melanoma is theoretically conducive to screening because a detectable preclinical phase corresponds to the noninvasive radial growth phase that may last months or even years. This clinically apparent radial phase consists of horizontal growth of the melanoma without the potential for metastases. Since the radial growth phase may last for years, the changes may be found during interval screening examinations. Another line of evidence to suggest the efficacy of screening and early removal of “atypical” pigmented lesions is the finding that if patients are kept free of “atypical” moles, they are also kept free of melanoma. Cohen has recently reviewed his 14-year experience with prophylactic skin nevus removal. Seventy-five of 250 patients with a past history of melanoma underwent prophylactic removal of multiple skievi and 28% of the lesions showed some atypia. Twelve melanomas were discovered and it was estimated from previous reports of prevention of melanoma with dysplastic nevus removal that four to six additional melanomas were prevented by excising precursor lesions. During the same time period, another 112 patients without a history of melanoma underwent prophylactic lesion removal. Three cases of melanoma were found and it was estimated that another three to five cases were prevented with the removal of the precursor lesion. An average of 2.7 lesions were removed among all screened patients. No melanoma-related deaths were reported during this 14 year period.
Another fact that supports melanoma screening is that there is an “intermediate” end point for efficacy: the tumor thickness at diagnosis. Multiple regression analyses have repeatedly confirmed tumor thickness to be the most powerful prognostic factor for stage 1 and 2 melanoma. Tumor thickness helps the clinician estimate the chance of occult nodal and systemic metastases, the need for nodal staging, and the prognosis for the individual patient. There is a high correlation between tumor thickness and survival. Estimated 5-year survival rate for persons diagnosed with melanoma less than
This phenomenon is illustrated by data from the
Studies from other parts of the world confirm the increase in survival noted with the diagnosis of thinner melanomas. In
Screening for melanoma in the
From 1985 to 1993, this volunteer effort has grown rapidly, providing more than 650,000 free screenings. Screenings have averaged nearly 100,000 persons per year since 1990.
Mass media have been involved in the educational component of this demonstration project. For example, in May 1992, 238 television stations in more than 150
Screening for Melanoma (Continued)
The free screening programs are generally conducted in community hospitals, physician offices, area workplaces, and local shopping malls. Volunteer screening dermatologists provide educational materials and visually examine the skin for cancer (melanoma, basal cell carcinoma, squamous cell carcinoma) or precancerous lesions (actinic keratosis, atypical mole/dysplastic nevus). No biopsies or diagnostic studies are performed. Screening participants found to have lesions suspected of being cancer or precancerous lesions are asked to consult a dermatologist or their own family physician after the screening to obtain definitive diagnosis and appropriate treatment. AAD staff encourage persons to seek follow-up medical care and provide names of physicians in the area. All persons across the country with suspected melanoma receive telephone calls and letters to ensure rapid follow-up.
The results of the national screening program in the
Thirty percent of the screened population had an abnormal examination and 0.3% had melanoma. An appropriate self-selection took place: 86% of the registrants had one of the risk factors for melanoma and 78% had two. The screened population had a predominance of white, female college-educated people, and these investigators have questioned how to attract older white men of low socioeconomic status. Rigel reported on a follow-up study from the 1986
As the first step to achieve comprehensive ascertainment of those screened nationwide, the AAD used several follow-up systems for people with a screening diagnosis of melanoma in the 1989 to 1993 programs. A centralized follow-up program has now successfully contacted 97% of people in the 1992 to 1993 screened population with a suspected diagnosis of melanoma at the time of screening. Preliminary results find that AAD skin cancer screenings appear to detect early melanoma (almost 99% of all screen-detected melanomas are stage I and II), with fewer advanced melanomas identified compared to the 1990 SEER registry. Specifically, of those persons with a screening diagnosis of melanoma attending the 1992 to 1993 AAD national programs, 257 melanomas have been confirmed, with all but four cases being diagnosed at a localized stage of disease (American Joint Committee on Cancer [AJCC] stage I and II). Only 18 of the 257 melanomas (7.0%) discovered during this time have had thickness greater than
ACS/AAD screening programs for melanoma/skin cancer in the
During the next few years, further studies should examine issues including:
The benefit of educational campaigns (e.g., can one measure an increase in high-risk persons seeking skin cancer examinations from nondermatologist physicians?) Melanoma mortality in intensely screened states versus states without intensive screening. Proportions of high risk persons (i.e., men 50 years of age or over) attending screening programs. Rigorous analysis of these and other issues should help determine the ultimate role of education and screening in reducing melanoma mortality. The ideal study for skin cancer screening would be a randomized prospective trial. No such randomized trial currently exists in the
PRIMARY PREVENTION OF MALIGNANT MELANOMA
Primary prevention of malignant melanoma is a longer-term exercise. From what is known about the growth kinetics of malignant melanoma, it is likely that trends of decreasing melanoma mortality and decreasing tumor thickness might be seen within 3 to 5 years of mounting a public education campaign aimed at early detection. However, the latent interval between receipt of an insult on the skin and development of melanoma may be longer than 20 years. Programs in many countries currently educate the public on safe sun exposure, on the assumption that excessive exposure to natural ultraviolet radiation is the most important causative agent in developing malignant melanoma. Epidemiologic studies strongly support this hypothesis, with increasing evidence that sunlight exposure in early childhood is a significant risk factor for subsequent development of malignant melanoma. The exact wavelength and action spectrum for the development of malignant melanoma are, however, not yet established.
Because of the recently recognized importance of avoiding excessive sun exposure in early life, emphasis also includes the education of young mothers and school-aged children. A wealth of educational material is now available for primary school children in a variety of many languages. Educational materials stress safe sun approaches including avoidance of noonday sun, the use of shade (such as trees or sun umbrellas), protective clothing (e.g., hats and t-shirts) and frequent application of a sunscreen with a high sun protection factor (SPF).
The field of assessment of primary prevention of melanoma is a new one, pioneered by Drs. Robin Marks and David Hill in
However, there is still considerable room for improvement in the public’s knowledge of dangerous sun exposure. For example, a survey of 22,000 individuals in the
Effecting structural change is the next strategy that will be employed in
Assessment of the efficacy of primary prevention campaigns is a long-term activity, involving rigorous monitoring of incidence and mortality trends. With increasing availability of leisure time, and the possible effects of a fall in ozone levels in the Northern Hemisphere, it is possible that incidence rates will continue their upward trend of the last 20 years. The importance of this observation has recently been recognized with the awarding of the 1995 Nobel Prize in Chemistry to three researchers for their work in atmospheric chemistry, particularly concerning the formation and decomposition of ozone and the ozone layer. Primary prevention activities also refer to the avoidance of excessive exposure to artificial ultraviolet radiation. In the northern
Although UVA radiation is more efficient at producing a tan, the tan that develops is inefficient for subsequent protection against burning. Short-term problems such as eye injuries (UVA is predominantly absorbed by the lens of the eye) and erythema are common, and long-term problems such as cataract development, skin photoaging, and skin cancer are expected to surface with time. It appears that the training of operators and warning of patrons is inadequate.
Education and legislation are addressing these concerns. The Federal Trade Commission has recently ruled that patrons need to use eyewear and health risk notices need to be placed on the tanning units. In addition, tanning salons cannot advertise the ability to provide a “safe” tan or any other health benefit associated with the activity. In
Targeting primary prevention messages to the high-risk population has been recommended. In
The total vertical height of a melanoma is the single most important prognostic factor in stage I and II melanoma. [ref: 80] It is a quantitative parameter that can define subsets of patients with different survival rates (P < .00001). Numerous groups of thickness subsets have been analyzed for their prognostic value (e.g., smaller than
Level of Invasion
There is an inverse correlation between increasing level of melanoma invasion and survival. The level of invasion is a significant prognostic factor by single-factor analysis (P < .00001), and it will differentiate patients at various risks for metastases. A direct comparison of the level of invasion and thickness microstaging methods was made by subdividing each level of invasion into thickness categories. Within levels III, IV, and V, gradations of thickness influenced survival. Converse relationships were not observed when analyzing sets of melanoma thickness subdivided by levels of invasion. For example, the 5-year survival rates for patients with level III, IV, or V lesions that measured 1.5 to
Ulceration
Ulcerated melanomas appear to be more aggressive lesions biologically, since they invade through the epidermis rather than pushing it upwards. Thus, the presence of ulceration in microscopic sections of melanoma was a significant adverse determinant of survival (P < .00001) in the UAB-SMU study. [ref: 85] Stage I and II melanoma patients with ulceration had a 10-year survival rate of 50%, whereas those without ulceration had a 79% 10-year survival rate (P < .0001). Even after correction for tumor thickness in prognostic models, the variable ulceration continues to add prognostic information.
Other Prognostic Factors Based on the Primary Tumor
Clinical variables such as age, gender, race, and primary site; and pathologic factors such as histologic type, mitotic rate, tumor ploidy, S-phase fraction, DNA content, lymphatic or vascular invasion, host response, and regression have all been analyzed in multiple regression analyses. Occasionally these factors will be important in the mathematic model that predicts prognosis in individual populations. [ref: 89,90] However, once tumor thickness and ulceration are controlled for, many if not most of the other prognostic factors lose much of their significance noted in a multivariate analysis.
Multifactorial Analysis
Clinical and pathologic parameters can be simultaneously compared for their prognostic strength using multifactorial analysis. The influence of these factors on survival was examined in the UAB and SMU population using a cohort of 4568 AJC stage I and II patients for whom information was available for all prognostic factors being analyzed. There were three dominant factors predicting survival: (1) thickness of the melanoma, (2) melanoma ulceration (presence or absence), and (3) anatomic location (upper extremity, lower extremity, trunk, or head and neck). Four other variables also correlated to a lesser extent with survival in certain subgroups of patients: (1) the type of initial surgical management (primary excision alone versus excision plus elective lymph node dissection), (2) pathologic stage (I and II versus III), (3) level of invasion, and (4) gender. Most other major studies utilizing multifactorial analysis have established tumor thickness to be the most important prognostic factor in their series of stage I and II melanoma patients. [ref: 88] Ulceration was likewise found to be another strong predictor of survival in other patient series. [ref: 89-91] Some of these series confirmed that anatomic location of the primary lesion was another major predictor of survival. Primary site location may be more predictive of survival in specific sites of the body, such as melanomas that occur beneath the nail bed.
Other factors analyzed by multifactorial analysis that emerged as important variables overall or within selected subgroups in other series included gender, race, [ref: 83] and age of the patient, as well as tumor features such as lymphocytic infiltration, [ref: 91] tumor diameter, [ref: 91] cell type, microscopic satellites, [ref: 96] mitotic activity, [ref: 88] and level of invasion. The timing of the biopsy prior to first definitive treatment was also found to have a significant influence on survival in one series. Flow cytometry analyses of DNA content (aneuploidy versus diploidy) correlated with survival in several recent studies.
The impression that melanoma arising during a pregnancy is associated with virulent disease and poor survival is derived from case reports and clinical studies of small numbers of patients. Wong and associates demonstrated that the clinical course of 66 patients with stage I and II melanoma diagnosed during a pregnancy was no different than that of the 619 nonpregnant controls. In a larger series from
Other studies have addressed the impact of a subsequent pregnancy after diagnosis on the clinical course of patients with melanoma. Reintgen and associates have found that the disease-free interval did not differ between 43 patients who had a pregnancy within 5 years of their diagnosis of melanoma and those women who never became pregnant during their clinical course.
METASTATIC MELANOMA IN REGIONAL LYMPH NODES
(AJCC STAGE III)
Once patients develop metastatic melanoma to their regional nodes, prognostic factors based on the primary melanoma contribute very little to the prognostic model. Number of Metastatic Nodes There was a direct correlation betweeumber of metastatic nodes and survival in the UAB and SMU database. Patients with one metastatic node had a better survival rate than patients with two metastatic nodes or more. Data from patients with different numbers of nodal metastases were analyzed for survival differences. The greatest differences in this series were between patients with one metastatic node, patients with two to four nodes, and patients with five or more nodes (Fig. 41.2-8). Thirty-seven percent of patients had one metastatic node, 38% had two to four, and 25% had five or more. Their 3-year survival rates were 40%, 26%, and 15%, respectively (P < .001). Ten-year survival rates demonstrated that only patients with one positive node had a reasonable prospect of cure (40% were alive at 10 years), whereas only about 15% of patients with two or more metastatic nodes were alive at 10 years.
The number of metastatic nodes was first shown to be of prognostic significance in a multifactorial analysis by Cohen and associates. Later, both Day [ref: 104] and Callery and associates demonstrated that tumor thickness and the number of metastatic nodes were dominant independent variables in stage III patients. Cascinelli and colleagues identified the extent of nodal metastases (i.e., confined to or invading through the lymph node capsule) and the number of metastatic nodes as the most significant factors in a multifactorial analysis of 530 stage III patients.
Molecular Biology Techniques for the Staging of Melanoma
For most solid tumors, including melanoma, the most powerful predictor of survival is the presence or absence of lymph node metastases. The presence of lymph node metastases decreases the 5-year survival of patients approximately 40% compared to those that have no evidence of nodal metastases. Much time, effort, and expense are placed on identifying prognostic factors based on the primary tumor, while not enough emphasis is given to identifying which patients really have signs of micrometastatic disease in their nodal basins. For instance, there are currently 26 prognostic factors (Table 41.2-3) for melanoma based on variables from the primary tumor. Yet in multiple regression analysis performed on many collected populations in the literature, the lymph node status of the patient with melanoma is the most powerful factor for predicting recurrence and survival. Primary tumor variables such as Breslow thickness, ulceration, primary site, and gender may add to the prognostic model, but only after nodal status is considered. It is also clear that routine histologic examination of the regional lymph nodes, an examination that typically involves making one or two sections of the central area of the node and staining with a standard hematoxylin and eosin stain, examines less than 1% of the submitted material and will miss micrometastatic disease. The sensitivity of this examination is finding one abnormal melanoma cell in a background of 10**4 normal lymphocytes. If serial sectioning and immunohistochemical staining are added, the yield of positive dissections may double and the sensitivity becomes identifying one abnormal melanoma cell in a background of 10**5 normal lymphocytes. Serial sectioning and immunohistochemical staining techniques have been available for years, yet have not been incorporated into the everyday practice of the pathologist for examining nodes, because of the time and expense involved.
New technology that enables the surgeon to map the cutaneous lymphatic flow from the primary tumor and identify the sentinel lymph node (SLN, see below) in the regional basin could contribute to better nodal staging of the melanoma patient. This procedure, as initially proposed by Morton and colleagues [ref: 107,108] has shown that the SLN is the first site of metastatic disease. If the SLN is negative, the remainder of the lymph nodes in the basin should also be negative. Selective lymphadenectomy also allows for the detailed examination of the SLN, since it is an examination of one or two nodes. This advance allows the pathologist to section the node serially and use immunohistochemistry to look for micrometastatic disease. Nevertheless, 25% of the histologic node-negative stage I and II melanoma patients will experience recurrence and die of their disease within 5 years of diagnosis, suggesting that some of these patients have missed nodal micrometastases or these patients have hematogenous metastases. A more sensitive method was needed to identify accurately the presence or absence of metastatic disease in the node.
A study was initiated by investigators at
The assay was refined using the combination of reverse transcription and double round polymerase chain reaction (RT-PCR) on preparations from lymph nodes. The amplified samples were examined on a 2% agarose gel and tyrosinase cDNA was seen as a 207 base-pair (bp) fragment. SLNs from 124 patients with primary melanomas greater than
Other investigators have also used molecular biology techniques to improve the staging of the melanoma patient. Dale and colleagues used an RT-PCR technique to detect occult circulating tumor cells in the blood of melanoma patients. Four different markers were used to improve the sensitivity and specificity of the occult metastases assay. Preliminary studies of this multiple marker RT-PCR assay, which utilized the four melanoma markers tyrosinase, MAGE-3, Muc-18, and p97 showed the presence of all four markers in 10 melanoma cell lines and none detected in the blood of 14 normal volunteers. The assay was used to detect circulating tumor cells in 74 melanoma patients of various clinical stage. The pattern of marker detection was as follows: tyrosinase (59%), MAGE-3 (9%), Muc-18 (66%), and p97 (65%). For all AJCC stages, the ability to detect circulating tumor cells was significantly higher (P = .025) in the 53 patients alive with disease than the 21 disease-free patients.
For all patients alive with disease, as the clinical stage increased from I to IV there was a corresponding increase in the ability of the assays to detect circulating melanoma cells. It was concluded that the detection of circulating occult tumor cells could provide a tumor marker for the early detection of metastatic or recurrent disease or evaluate the response to specific therapeutic modalities.
The staging of the melanoma patient is becoming more important with the recent publication of a multicenter, prospective randomized trial that purports to show a benefit for the adjuvant treatment of T4 or stage III melanoma patients with interferon alpha (IFN-alpha). Adjuvant therapy should be applied early when tumor burden is minimal and in a selective fashion so that only those patients with a proven benefit are exposed to the toxicities and expense of the adjuvant therapies. The accurate staging of the melanoma patient will identify the subgroup of patients who have the most to benefit from IFN-alpha.
METASTATIC MELANOMA AT DISTANT SITES (AJCC STAGE IV)
The data for 200 patients with distant metastases treated at UAB were analyzed for predictive factors affecting survival rates.
Site of Distant Metastases
The locations of distant metastases were an important prognostic factor when examined by single-factor analysis (P = .0001). The skin, subcutaneous tissues, and distant lymph nodes were the most common first sites of relapse, which occurred in 59% of patients. In 23% of patients, nonvisceral metastases at these sites were the sole manifestation of disease (14% for skin, 5% for subcutaneous sites, and 4% for distant lymph nodes). The median survival duration for the entire patient group was 7 months, with 25% alive at 1 year. There was no difference in survival among patients with metastases at any combination of these three sites. The next most common site of first relapse was the lungs (36% of patients). Patients with isolated lung metastases had the longest median survival duration (11 months) of patients with metastases at distant sites. The brain, liver, and bone were the next most common sites of first relapse. The median duration of survival for these patients was very poor, ranging from 2 to 6 months, with a 1-year survival rate of only 8% to 10%.
The presence of visceral metastases had an overriding influence on survival, since those patients with combined visceral and nonvisceral metastases had the same poor prognoses as those with visceral metastases alone.
SURGICAL MARGINS OF EXCISION
Local control of a primary melanoma requires wide, local excision (WLE) of the tumor or biopsy site with a margin of normal-appearing skin. The reason for a WLE is that approximately 5% of the primary melanomas will have a satellite focus of melanoma separated from the main lesion. Until recently, the routine surgical approach was to excise all primary melanomas with a 3- to 5-cm margin and apply a split-thickness skin graft to the defect. However, it has become increasingly clear that the risk of local recurrence correlates more with the tumor thickness than with the margins of surgical excision. It therefore seems more rational to excise melanomas using surgical margins that vary according to tumor thickness and ulceration, since these factors correlate best with the risk for local recurrence. The earliest lesion is a melanoma in situ. This is a noninvasive tumor that does not metastasize but is capable of recurring locally. Although the natural history of these noninvasive lesions is not completely understood, there is a risk of local recurrence (either as an in situ melanoma or an invasive melanoma) if they are not reexcised after biopsy. It is therefore recommended that the biopsy site of an in situ melanoma be excised, usually with a 0.5 to
Wide local excision of skin melanoma with wound plastic.
A randomized prospective study conducted by the Intergroup Melanoma Committee has evaluated 2-cm versus 4-cm radial margins of excision for intermediate-thickness melanomas (1 to
ELECTIVE LYMPH NODE DISSECTION
RATIONALE. The issue of ELND is probably one of the most important controversies in the management of patients with melanoma. The debate is finally coming to a close since the results of the Intergroup Melanoma Surgical Trial have shown significant improvement in subgroups of patients based on age and thickness (see later) and because of the increasing applicability of intraoperative lymphatic mapping. Two randomized prospective studies involving only extremity melanomas did not demonstrate any survival advantage for ELND, while three nonrandomized studies involving melanomas from all anatomic sites showed a statistically significant improvement in survival for the subgroup of intermediate thickness melanomas. While there is unanimous opinion that all melanoma patients do not need ELND, there is still a continuing debate that centers around two issues: (1) Is it possible to identify accurately a subgroup of melanoma patients at high risk for microscopic regional nodal metastases and a low enough risk of occult systemic metastases to justify a regional node procedure? (2) What is the optimal timing of dissection (immediate versus delayed) if such a high-risk group can be identified? Prospective randomized studies have recently been completed to resolve this issue.
SELECTION OF PATIENTS
Prognostic Factors. Tumor thickness provides a quantitative estimate of the risk for occult metastatic melanoma at regional and distant sites. In fact, melanoma thickness is the most important but not the sole guide for selecting patients who might benefit from ELND. The major advantage of using tumor thickness for these surgical decisions is that it can provide a quantitative estimate of the risk for occult metastatic melanoma in both regional and distant sites. Thin melanomas (less than
Identifying the Regional Lymph Node. Basins at Risk for Metastases.
Since melanomas located on the trunk and on the head and neck area have unpredictable lymphatic drainage, it is difficult to decide which nodal basin is at risk for metastatic disease. In many patients, this problem has been overcome by performing a radionuclide cutaneous scan that can accurately define the location of nodes that are the primary drainage site for a melanoma located anywhere on the body. Lymphoscintigraphy, a nuclear medicine test that involves a radiocolloid injection around the primary site and a scan of the regional basins, has been used to redefine cutaneous lymphatic flow from classic anatomic descriptions. [ref: 153] The understanding of cutaneous lymphatic flow remains grounded in the work performed by Sappey in 1874. By the injection of mercury into the skin of cadavers, he was able to show lines of demarcation running along the midline, and from just above the umbilicus curving slowly backward to the second lumbar vertebra. Lesions within
Lymphoscintigraphy, a dynamic study performed in live patients, has shown that the watershed areas of the body are much expanded compared to Sappey’s original description and one does not see straightforward lymphatic flow until a melanoma is located
Lymphoscintigraphy is an extremely accurate method of identifying all nodal basins at risk for metastatic disease, and unless it is performed preoperatively, the ELND or sentinel node biopsy may be misdirected in up to 50% of the cases. The advantage of lymphoscintigraphy over classic anatomic guidelines is the scan’s reflection of functional and not just structural anatomy. The results from the MCC and the SMU show that lymphoscintigraphy can be used as a road map so that the surgeon does not perform too little or too aggressive surgery. The results of these series also cast into doubt the conclusions of studies of ELND in which preoperative lymphoscintigraphy was not used as part of the protocol. Of the recently completed prospective randomized trials evaluating the efficacy of ELND, only the Intergroup Melanoma Study required lymphoscintigraphy prior to the ELND.
Either all of the regional nodes should be removed or a policy of monitoring multiple nodal sites at risk should be followed. An ELND of two nodal basins (e.g., bilateral axillary dissection) for trunk melanomas or ipsilateral anterior and posterior neck dissections for head and neck melanoma may be warranted in select cases, but removing more than two nodal basins or performing a bilateral cervical dissection as ELNDs is never indicated. A bilateral inguinal dissection is usually not performed electively because of its attendant side effects, especially lymphedema of the extremities and genitalia.
Intraoperative Lymphatic Mapping and Sentinel Node Biopsy
A new procedure has been developed to assess the status of the regional lymph nodes more accurately and decrease the morbidity and expense of a complete ELND. The technique termed intraoperative lymphatic mapping and selective lymphadenectomy relies on the concept that regions of the skin have specific patterns of lymphatic drainage not only to the regional lymphatic basin, but also to a specific lymph node (SLN) in the basin. Morton initially proposed the technique [ref: 107,108] using a vital blue dye method and showed that in animals and initial human trials, the SLN is the first node in the lymphatic basin into which the primary melanoma drains. He showed that the SLN histology reflected the histologic status of the remainder of the nodal basin, so that complete nodal staging could be obtained with an SLN biopsy.
These data have been confirmed by four other institutions, including the MCC, MD
With the addition of intraoperative radiolymphoscintigraphy to vital blue dye lymphatic mapping, [ref: 192] the SLN localization becomes easier and more widely applicable. In the initial study [ref: 192] from MCC comparing the vital blue dye and radiocolloid mapping technique, 450 microCuries of technetium-labeled sulfur colloid was mixed with the standard isosulfan blue and injected at the site of the primary cutaneous melanoma. The nuclear probe (Neoprobe,
This study consisted of 106 consecutive patients with cutaneous melanoma greater than
An NCI-sponsored prospective randomized trial is currently being performed that randomizes patients to receive either WLE of the primary melanoma site or WLE and SLN biopsy of the regional lymphatic basins at risk for metastases. The endpoint of the study is whether this surgical strategy can extend the survival of the melanoma patient. The study is different from the previous randomized trials addressing the efficacy of ELND, since only a percentage of the patients with melanomas greater than
It is crucial for the surgeon who performs intraoperative mapping to have adequate nuclear medicine and pathology support. Lymphatic mapping requires the close collaboration of nuclear medicine, surgery, and pathology in order to perform the procedure accurately. The preoperative lymphoscintigraphy is a road map for the surgeon and is used for four distinct reasons in planning the surgical procedure:
1. To identify all nodal basins at risk for metastatic disease.
2. To identify any intransit areas of nodal collections thatcan be tattooed by the nuclear medicine colleague for later harvesting. Intransit nodal areas occur in 5% of the melanoma population and, by definition, are the SLN.
3. To identify the location of the SLN in relation to the rest of the nodes in the basin.
4. To estimate the number of SLN in the regional basin that will need to be harvested.
Intraoperative vital blue dye and radiocolloid mapping with the Neoprobe can identify the blue-staining afferent lymphatic draining into the blue-staining and hot SLN. Both intraoperative mapping techniques are needed and complementary since primary sites near regional basins are difficult to map with the radiocolloid alone because of enormous radioactivity at the injection site (primary site). Likewise, radiocolloid mapping adds to the blue dye technique by the ability to locate the so-called “hot” spot in the regional basin through the skin, lessening the need for extensive flaps and dissections, and providing a measure that all the SLN have been removed from the basin (activity in basin at background levels after the removal of the SLNs). The SLN can then be harvested, allowing the pathologist time to make a detailed examination of the SLN with multiple sections, immunohistochemical staining with melanoma-specific monoclonal antibodies, HMB-45 and S-100, and perhaps RT-PCR determination. This strategy will invariably increase the yield of positive dissections and more accurately stage the patient with melanoma.
The question then becomes what constitutes the standard of surgical care for the patient with melanomas greater than
ADJUVANT THERAPY FOR THE TREATMENT OF HIGH RISK FOR RECURRENT STAGE II AND III DISEASE
ROLE OF ADJUVANT RADIATION THERAPY
The role of radiotherapy as a surgical adjuvant after therapeutic node dissection or as an alternative to ELND in the regional treatment of patients with intermediate to thick melanomas has not been clearly defined. The rationale for its consideration in this context is as follows. ELND, while effectively reducing regional recurrence rates, carries varying degrees of morbidity and does not offer any survival advantage in patients with thick primary tumors. [ref: 176] On the other hand, therapeutic dissection of pathologically involved nodes is associated with a local recurrence rate of up to 50% in patients with head and neck melanomas. [ref: 155,195] In view of the management problems associated with uncontrolled local/regional disease and the extent of elective dissections required for scalp and facial primary sites, a study was initiated at M.D. Anderson Cancer Center in 1983 to evaluate the role of radiotherapy in the treatment of clinically uninvolved lymph drainage areas in patients at high risk of nodal metastases, or as an adjunct to surgery in patients undergoing therapeutic nodal dissections.
The preliminary results of this trial, indicating an apparent advantage in terms of local regional recurrence for adjuvant radiotherapy, were published by Ang and colleagues in 1990. [ref: 196] The following is an update of the study with a median follow-up of 38 months. Through July, 1994, 224 patients (171 men and 52 women) were enrolled. Their ages ranged from 16 to 89 years (median: 55 years). Patients were organized into three groups. Group I consisted of 118 previously untreated patients who presented with primary lesions either greater than
Group II consisted of 39 patients with previously untreated disease who presented with clinically positive lymphadenopathy. These patients mostly received postoperative radiotherapy (30 Gy D(max) delivered in five fractions over 2.5 weeks); the remainder received preoperative treatment (2 Gy D(max) delivered in four fractions over 2 weeks). These patients achieved a 5-year local regional control rate of 92% and 5-year survival of 41%, with survival being inversely proportional to the number of pathologically involved lymph nodes.
Group III consisted of 67 patients who presented with recurrent regional and/or local disease, but without evidence of distant metastases. These patients were treated in the same manner as group II patients. Their 5-year local regional control rate was 88% and survival 45%. With the addition of adjuvant radiotherapy, the local-regional control rate of group II and III patients was not affected by the number of positive nodes or extracapsular extension.
These results demonstrate that adjuvant radiotherapy, either alone in clinically node-negative patients or as a surgical adjuvant in pathologically node-positive patients, can achieve local regional control in excess of 85%. This is substantially better than rates previously reported with surgery alone in comparable patients. The radiotherapy schedule used was not associated with any significant morbidity. In total, 3 of the 224 patients treated to date have sustained mild to moderate sequelae. There has beeo severe complications. Although these data are impressive, proof of a therapeutic benefit from adjuvant radiotherapy can be obtained only from a prospective randomized trial. Such a trial, based on this pilot study, has been activated by the Radiation Therapy Oncology Group (RTOG).
ROLE OF ADJUVANT SYSTEMIC THERAPY FOR PATIENTS
WITH MELANOMA AT HIGH RISK FOR RECURRENCE
Interferon
Advances in melanoma treatment over the past decade have evolved primarily from more detailed knowledge about prognostic factors of primary and metastatic lesions. Within the larger group of melanoma patients who undergo potentially curative treatment by surgical resection, subgroups can be identified who are at high risk for recurrence and for development of systemic metastases. Patients with melanomas 1.51 to
Toxicity of this therapy is significant, with nearly ubiquitous flulike symptoms of moderate to severe degree, and the need for dose delay or attenuation in more than one-third of patients during induction, and again during the subcutaneous maintenance phase of treatment. The completion of 1 year of therapy was feasible in the majority (74%) of patients who did not experience progression. Proper attention to hematologic and hepatic function has been sufficient to avoid lethal hepatic toxicities observed early during the application of this intensive protocol. The ECOG has led an intergroup confirmatory sequel study (E1690/SWOG9111/CALGB9190) that completed accrual in mid-1995 and will require up to several years of maturity for evaluation. This protocol was designed not only to corroborate E1684 but also to test the influence of a lower dose of IFN-alpha2b, 3 MU/d given subcutaneously three times weekly for 2 years. It closed in 1995 with accrual of 642 patients. This Intergroup Trial differed from the initial trial E1684 in that lymph node dissection was no longer required for patients with T4 (greater than
In an accompanying editorial on the E1684 ECOG trial, [ref: 198] Balch and Buzaid questioned the role of adjuvant IFN-alpha in a high risk for recurrence patient whose tumor is greater than
Only a small subset of patients who receive high-dose IFN-alpha2b therapy actually derived benefit from it. Serum markers of tumors may provide a key to the more selective application of adjuvant therapy. Tyrosinase (a key enzyme in melanin biosynthesis whose gene is actively expressed only in melanocytes, melanoma cells, and Schwann cells) may serve as a marker of early dissemination of disease to allow more selective use of adjuvant therapy in the future. Because melanocytes and melanoma cells do not usually circulate in the blood, the detection of tyrosinase mRNA by reverse transcriptase-polymerase chain reaction (RT-PCR) in blood is considered a potential indication of the presence of circulating melanoma cells. Patients who have a positive RT-PCR before starting therapy that subsequently becomes negative may be those who benefit from the therapy. [ref: 198] The North Central Cancer Treatment Group (NCCTG 83-7052) evaluated the adjuvant role of IFN-alpha2a given at 20 MU/IM three times weekly for 3 months, versus observation in patients with 1.69-mm Breslow depth T3+ or T4, and N1 patients (subsets of AJCC stages IIA or IIB and III). An analysis of 260 evaluable patients who entered this trial demonstrates no significant prolongation of survival, or of relapse-free interval overall. Subset evaluation of patients in stage II as opposed to those in stage III participating in this trial reveals a potential impact in the latter, by Cox analysis.
The WHO Melanoma Program Trial 16 has evaluated the efficacy of a lower, less-toxic dosage of IFN-alpha2a, given 3 MU/d subcutaneously three times weekly for 3 years versus observation. Of 444 patients who entered this trial, a majority exhibited extracapsular extranodal involvement, a pathologic variable that made patients ineligible for the trials of the US Cooperative groups previously noted. The analysis of WHO Trial 16 has been reported preliminarily at 22 months of follow-up, and in a subset analysis an interaction between age and gender and treatment was found. This interim analysis of subsets, defined by the presence or absence of extranodal extension, has suggested an influence of this therapy upon intranodal disease, and lack of any trend to benefit among patients with extracapsular extension. Subsequent reports of interim analyses of this trial at intervals up to 39 months median follow-up have suggested the absence of a significant impact upon either relapse-free or overall survival.
In summary, the analyses of E1684, NCCTG 83-7052, and WHO 16, taken together, argue that IFN-alpha2b at higher dosages, delivered intravenously, may be necessary for the benefit observed in E1684. Equivalent dosages of IFN-alpha2a administered by the IM route for shorter periods have been less effective, and lower dosages of IFN-alpha2a administered by the subcutaneous route for longer periods have been ineffective to date. The approval of IFN-alpha2b for adjuvant therapy within 2 months of surgery of high-risk node-positive and deep primary melanoma is the first adjuvant therapy approved for melanoma, and the first new agent approved for therapy of this disease in any stage or setting, since dacarbazine.
The evaluation of newer and potentially more effective biologic agents in the adjuvant setting has been pursued in multiple trials. SWOG 86-42 is a trial of IFN-alpha administered at dosages projected to be the optimal immunomodulatory dosages (0.2 mg/d sc qod for 1 year). IFN-alpha is among the most potent immunomodulators yet tested, and one for which initial hopes for adjuvant and metastatic disease applications were elevated even beyond IFN-alpha2. As previously noted, the therapeutic activity of IFN gamma in the advanced disease setting has beeegligible. The recently published SWOG 86-42 trial demonstrates the lack of therapeutic benefit for this agent at the dosage tested, in either intermediate-risk stage II or stage III resected melanoma. Of concern is that the initial report of this trial suggested a potentially adverse impact of treatment that has not yet been confirmed. In any case, the unequivocal failure of IFN-gamma to improve relapse-free or overall survival in the adjuvant setting argues that the effects of the dose, route, and schedule of IFN-gamma and the monocyte activation and NK activation associated with IFN-alpha administered as in this trial are not sufficient to alter the outcome of this disease. [ref: 203] The pleiotropic effects of IFN-alpha2b, which account for its therapeutic benefit in melanoma, have yet to be proven. In vivo activities that may be important range from a direct antiproliferative and cytotoxic action to immunopotentiation of T- or B-cell host responses, dendritic cell antigen presentation, anti-angiogenesis factor, and tumor-restricted antigen expression. These are being evaluated prospectively in the context of the recently completed intergroup trial (ECOG 1690/SWOG 91-11, CALGB 91-90).
Immunotherapy
Tumor vaccines have a long history, but the search for effective methods to induce active immunity against tumors has been difficult. Repeated attempts have been made to inhibit the outgrowth of tumors and influence the natural history of melanoma by immunization against tumor cells or extracts of tumor cells. More recently, purified antigens recognized by the host antibody response, as well as the epitopes defined by T-cell responses, have been identified and prepared for use as vaccines. In addition, genetically modified melanoma cells, recombinant vaccines, and antiidiotype antibodies that represent the antibody-recognized antigens of melanoma have entered into clinical trials. Evidence is mounting that vaccination can induce immune responses to melanoma. Efforts to vaccinate patients against the antigens of melanoma can be classified categorically into trials of allogeneic tumor cell immunization, autologous tumor cell immunization, and immunization against synthetic chemically defined antigens.
The last category can be subclassified into approaches directed at the induction of antibody responses, such as the gangliosides noted below, and approaches directed at the induction of a T-cell response to antigens that may include either proteins or peptides. Formal phase III trials have been initiated to evaluate the potential efficacy of relatively few tumor vaccines as adjuvant therapy in high-risk patients.
At this point, it remains unclear which approaches will be most effective. The multitude of different ongoing trials listed indicate that there is no consensus on vaccine strategies. Early clinical trials of each of the categories of vaccines have been conducted in patients with metastatic disease, for reasons of safety evaluation. Despite early concerns, no reports of tumor enhancement have been documented with the whole-cell vaccination trials conducted to date. Antitumor responses have been recorded in a small fraction of subjects with metastatic disease, receiving varied programs of tumor cell immunization against allogeneic tumor cells incorporating two of three cell lines of hapten-coupled autologous [autochthonous] tumor cells. [ref: 205,206] Most subsequent clinical studies have been performed in patients after resection of regional lymph node metastases or high-risk primary lesions. A more interesting observation is that patients who develop evidence of immune responses have improved disease-free survival or overall survival. This intriguing correlation does not directly imply that vaccines improve the clinical course (for instance, patients who a priori have a better prognosis may be more likely to develop an immune response to vaccination), but is consistent with the hypothesis that an immune response induced by vaccination may have antitumor benefit. A number of problems exist in the construction of tumor vaccines, including the weak immunogenicity of tumor antigens, heterogeneity of tumor-restricted antigen expression in tumors, and the ability of tumors to escape any immune responses that may be induced naturally, or therapeutically. [ref: 208] Most melanoma antigens on tumor cells are not tumor-restricted or specific (i.e., present only on cancer cells) but are shared with certaiormal cells.
Active specific vaccination trials in melanoma (as opposed to nonspecific vaccination with BCG) have progressed from the use of unmodified, irradiated allogeneic or autologous melanoma cells for immunization, either alone or in combination with BCG and other nonspecific immunomodulators. It has been difficult to demonstrate any benefit or specific immune responses in the course of these trials. [ref: 210,211] One exception has been immune responses to gangliosides, the major acidic glycolipid of melanoma cells, in some patients immunized with selected allogeneic melanoma cells.
Efforts have been made to improve immunogenicity of allogeneic tumor cells by treating with the enzyme neuraminidase to remove sialic acid residues from the cell surface. [ref: 214] Strategies have been developed to augment the immune response to tumor antigens by infecting tumor cells with nonpathogenic viruses. Viral proteins presented on the cell membrane of tumors have been shown to augment the immunogenicity of tumor antigens. [ref: 215] Several centers have investigated the immunogenicity of lysates derived from tumor cells infected with
The allogeneic tumor cell vaccine reported first by Mitchell and colleagues, admixed with the immunologic adjuvant Detox (Ribi Immunochem Res, Hamilton, MT) has been studied in a second large multicenter randomized trial that is enrolling moderate-risk T3 N0 patients, and will be completed in 1996. This study has been enlarged from its original goal to 600 patients to increase the power of the trial, and its ability to detect more subtle movements in relapse-free survival. In addition, the understanding of the nature of melanoma antigens recognized has moved forward rapidly, including the discovery of a number of peptide antigens whose recognition is restricted by the HLA-A2 allele, which was absent in the tumor cell lines employed in this vaccine trial. Another approach to increasing immune response has been to alter tumor cells by binding haptens to autologous tumor cells. [ref: 217] Pilot single-institution data with haptenated autologous tumor cells have been intriguing but, to date, this type of vaccine has not been evaluated in a randomized controlled or multicenter trial, owing in part to the difficulty of manipulating autologous tumor cells for the vaccine. [ref: 218]
A variety of efforts are also underway to immunize melanoma patients with purified or partially purified preparations of potentially immunogenic molecules from melanoma cells. The rationale for this approach comes from the identification of antigens on melanoma cells that can be recognized by antibodies in sera or lymphocytes from patients with melanoma. A large experience using partially purified cultured allogeneic tumor cell antigens has been the stimulus for the initiation of controlled trials of one vaccine preparation derived from shed antigens of cultured melanoma cell lines. Gangliosides have been shown to induce antibody responses after immunization. Of particular interest is that the ganglioside GM2 induces IgM antibody responses in more than 80% of patients immunized with GM2 and BCG following cyclophosphamide treatment at low dose as an immunomodulator. A prospective randomized study evaluated the role of vaccination for 6 months against GM2 with BCG, in comparison to BCG alone, using low-dose cyclophosphamide pretreatment in both groups.
Serologic studies have shown a favorable prognostic significance for the presence of anti-GM2 ganglioside antibodies, suggesting that the induction of antibodies against gangliosides may confer protection from relapse and mortality due to melanoma. However, native autologous antibody reactivity against GM2 is infrequently detected in patients with melanoma. A variety of vaccination programs utilizing intact autologous and allogeneic cells, and cell fractions derived from allogeneic melanoma cell lines, have been employed to increase immune responses to autologous melanoma cells. The most promising results from randomized controlled trials of vaccination have emerged from studies of purified ganglioside GM2. Livingston and colleagues have performed a series of trials employing autologous tumor cells, and more recently purified ganglioside preparations, with various immunologic adjuvants. The combination of BCG given together with ganglioside GM2 has been shown to induce the formation of increased titers of IgM antibodies in the majority of patients immunized (33 of 44).
A randomized adjuvant trial of GM2/BCG given following surgery for stage III melanoma (AJCC staging) has been conducted in 122 patients at Memorial-Sloan-Kettering Cancer Center (MSKCC) between 1987 and 1988. This trial, reported at more than 5 years of follow-up, has demonstrated the induction of IgM anti-GM2 antibodies in a majority of patients who were immunized, and the presence of native antibodies against GM2 in a small (fewer than 5%) subset of unvaccinated patients in the control group. Antibody response against GM2 was again demonstrated to be a highly significant and favorable prognostic factor. In addition, vaccination was associated with a prolongation of relapse-free survival of borderline statistical significance, confounded by the occurrence of native anti-GM2 antibody responses among six patients in the total, five of whom were entered into the control group. Analysis of the efficacy of immunization among seronegative entrants into this protocol revealed a significant improvement of relapse-free survival associated with GM2/BCG vaccination (P = .02).
To further augment the immunogenicity of GM2, a variety of measures have been explored. Covalent attachment of GM2 to carrier proteins, such as keyhole limpet hemocyanin (KLH), has improved the titer and durability of the IgM response seen with BCG and GM2 and induced antibody of the IgG isotype for the first time. A phase-III evaluation of the GM2-KLH vaccine with the QS21 adjuvant has been planned in the ECOG and US Intergroup mechanism, as well as in
NOTICE.
The education materials were taken from “CANCER: Principles & Practice of Oncology. 5th Edition.
|
|
|
|
|
Head and Neck Cancers
Tumors of the Nasal Cavity and Paranasal Sinuses, Nasopharynx,
Oral Cavity, and Oropharynx
The analysis of cancers within the upper aerodigestive tract reveals a heterogeneity of neoplastic processes. Each bears its own unique set of epidemiologic, anatomic, pathologic, and treatment considerations. This chapter will review such considerations based on four anatomically defined regions: the nasal cavity and paranasal sinuses; the nasopharynx; oral cavity; and oropharynx, respectively. However, there exist general principles regarding these cancers which may be considered here. Such principles involve anatomy (i.e., primarily the anatomy of regional lymph nodes within the head and neck), pathology, staging and screening, as well as general principles of treatment involving either single modality or multimodality therapy that are relevant to all sites.
ANATOMY An understanding of the regional lymph node anatomy is critical to the care of head and neck cancer patients. There are several major lymphatic chains in the neck containing nearly 200 lymph nodes that run parallel to the jugular veins, spinal accessory nerve, and facial artery and into the submandibular triangle. To facilitate communication regarding cervical lymph node anatomy, the regions of the neck have been characterized by levels. Level I includes nodes within the submental triangle and the submandibular triangle. The submental triangle extends from the midline anteriorly to the anterior belly of the digastric muscle posteriorly. Its third border is formed by the hyoid bone inferiorly. The submandibular triangle is bounded by the mandible superiorly. The anterior and the posterior belly of the digastric muscle complete the triangle. Level II includes the jugular nodes extending from the subdigastric area down to the carotid bifurcation and the nodes surrounding the spinal accessory nerve from the jugular foramen to the posterior border of the sternocleidomastoid muscle. It includes the lymph nodes in the upper posterior cervical triangle above the entrance of the spinal accessory nerve into this triangle.
Level III represents the nodal area principally along the jugular vein between the carotid and its bifurcation, the posterior border of the sternocleidomastoid muscle, and the omohyoid muscle. Level IV constitutes nodal areas below the omohyoid muscle above the level of the clavicle and between the carotid vessels anteriorly and the omohyoid muscle posteriorly. Level V represents nodes in the posterior cervical triangle. Its borders are formed by the posterior edge of the sternocleidomastoid muscle, the level of the entrance of the spinal accessory nerve, the trapezius muscle, and the posterior belly of the omohyoid muscle. Specific sites within the aerodigestive tract have a predetermined drainage pattern. A knowledge of this pattern will aid in diagnosis. It will also impact on therapy. Drainage patterns will be addressed in each of the anatomic subsites detailed in this chapter.
PATHOLOGY The predominant lesion within these anatomically defined regions is squamous cell carcinoma. Squamous cell carcinoma can be categorized into three classic differentiations. Well differentiated disease shows greater than 75% keratinization; moderately differentiated disease contributes to the bulk of squamous cell carcinoma and is characterized by 25% to 75% keratinization; poorly differentiated disease demonstrates less than 25% keratinization. Other variants of squamous cell carcinoma include verrucous carcinoma, sarcomatoid squamous cell carcinoma, and lymphoepithelioma. Additional pathologic criteria of squamous cell carcinoma, which are believed clinically relevant, were developed by Jacobsson and others. [ref: 4-7] This includes the number of mitoses, presence of vascular invasion, size of nuclei, degree of inflammatory infiltrate, and pushing or infiltrating borders (Table 29.2-1). Premalignancy A series of pathologic changes from premalignant disease to frank malignancy can occur in the oral cavity. Among the premalignant diseases are leukoplakia, erythoplakia, hyperplasia, and dysplasia. Each of these types has a propensity for malignant transformation. [ref: 8] Histopathologic assessment of leukoplakia reveals hyperparakeratosis, which is variably associated with an underlying epithelial hyperplasia. Leukoplakia without underlying dysplastic changes is rarely associated with progression to malignancy (i.e., less than 5% probability of malignant changes). Erythroplakia is a condition within the oral cavity and pharynx characterized by red superficial patches adjacent to normal mucosa. Distinct from leukoplakic lesions as identified above, erythroplakia is commonly associated with underlying epithelial dysplasia. Likewise, it can be associated with carcinoma-in-situ to frank malignancy in nearly 40% of lesions.
Dysplasia as compared to the above two clinical descriptives is a true histopathologic term, which is characterized by several morphologic changes including the presence of mitoses, pleomorphism, and prominent nucleoli. When dysplasia involves the entire thickness of the mucosa, it is commonly referred to as carcinoma in-situ. Dysplasia has been associated with a subsequent risk of progression to frank malignancy ranging from 15% to 30% of cases. [ref: 12,13] Significant advances have occurred over the last several years in defining phenotypic and genotypic characteristics of oral premalignancy. Premalignant lesions can be characterized by their increased proliferative capacity. For instance, Coltrera and colleagues [ref: 14] have demonstrated that premalignant lesions of the oral cavity demonstrate increased expression of proliferating cell nuclear antigen. Other studies have demonstrated abnormalities in differentiation pathways, primarily in the aberrant expression of cytokeratin content. [ref: 15,16] Premalignant lesions of the upper aerodigestive tract have been shown to express genotypic abnormalities as well. This includes aneuploidy, increased DNA index, and specific mutational abnormalities. [ref: 17-30] Mutations involving the p53 gene have perhaps been the most extensively characterized of the aberrant genetic events involving the premalignant state of the upper aerodigestive tract.
STAGING AND SCREENING Staging The role and current status of staging for head and neck cancer is undergoing continuous analysis. Though standard staging has been defined, there is a growing debate as to what the primary function of staging should be. The American Joint Committee on Cancer Staging (AJCC), however, has described the principal goal of staging as a means of defining the natural history of disease. Additional goals include the ability to judge therapeutic results between various centers as well as a means of defining patient prognosis. Revisions to standard staging have continually been offered, but today it is based on the tumor-nodes-metastasis (TNM) classification. T stage represents extent of primary disease. N represents the extent of regional lymph node metastasis, and M is a measure of distant metastasis. The AJCC staging system, used for classifying TNM status, is periodically revised. The current clinical staging system, although based principally on physical examination, has also incorporated specific radiographic observations of disease status. Thus, invasion through cortical bone by an oral cavity tumor will upstage a T2 or T3 lesion to T4 (i.e., from stage II or III to stage IV). Radiographic assessment of cervical lymph node metastases has not been integrated into clinical staging. The benefits of these diagnostic techniques beyond that provided by standard physical examination are under investigation. Most would consider the combination of radiographic and clinical assessment to be more accurate than either one alone. Furthermore, growing emphasis is being placed on the advantages of one imaging technique versus another (i.e., the relative merit of magnetic resonance imaging versus computed tomography). Summarizes the relative value of these two techniques in head and neck imaging. The criteria for T staging within the upper aerodigestive tract differs dependent on the primary site. N staging and M staging, however, are uniform and will therefore be considered here.
REGIONAL LYMPH NODES (N) NX: Regional lymph nodes cannot be assessed N0: No regional lymph node metastasis N1: Metastasis in a single ipsilateral lymph node,
DISTANT METASTASIS (M) MX: Presence of distant metastasis cannot be assessed M0: No distant metastasis M1: Distant metastasis Table 29.2-3 represents the most current stage classification as defined by the AJCC. [ref: 31] Screening for Primary Cancers The significance of screening is emphasized by reports that note that most oral cancer lesions could have been detected several months earlier by appropriate dental examination. [ref: 39,40] However, although it is intuitive that early detection of disease would benefit an individual patient, firm evidence to support routine screening of head and neck cancer has not been demonstrated. Indeed, the report of the United States Preventive Task Force does not recommend such a strategy for assymptomatic persons. [ref: 41] Several considerations may account for the pessimism. Head and neck cancer is a chronic disease process. It may exist in a subclinical state for a prolonged period of time and elude detection. Furthermore, head and neck cancer is a disease of patients who are in the fifth and sixth decades of life. These individuals visit the dentist infrequently and far less frequently than they visit a physician. [ref: 42,43] Some studies have suggested that dentists are more effective than physicians in routinely performing a complete mouth examination and detecting early stage oral cancer. [ref: 44] Thus, an effective approach to the screening of head and neck cancer may be in the education of primary care providers in high-risk areas on effective clinical examination. Indeed, education designs have been established and implemented. [ref: 45] Current limitations have been identified and relate to physician time constraints for prevention activities generated by other health care demands within the practice setting. The most extensively studied screening procedure has been oral exfoliative cytology, which has been recommended by the American Dental Association. [ref: 46] A high proportion of false-negative examinations have beeoted with this procedure. [ref: 47] Those individuals who are most at risk (i.e., individuals with high tobacco and alcohol consumption patterns) are least likely to voluntarily participate in such efforts. This has been emphasized by a recent report of the Metropolitan Detroit Oral Cancer Screening Control Program in which 5679 individuals were evaluated in a 27-month period. [ref: 48] Eighteen cancers were identified. High-risk populations were, however, felt not to be benefited. Another screening method frequently utilized has been toluidine blue staining of aerodigestive mucosa. [ref: 49,50] Toluidine blue is a metachromatic nuclear stain that is taken up by dysplastic and cancerous epithelium. Mashberg [ref: 50] has extensively detailed the appropriate use of toluidine blue and has noted false positive results to approximate 9% and false negative result approaching 5%. Its value is that it is quick, inexpensive, and noninvasive. Standardized physical examination is the best means of detecting lesions of the upper aerodigestive tract. The care and thoroughness of the examiner are paramount. [ref: 51,52] The need for careful assessment was emphasized by Jesse and colleagues [ref: 51] in their report on patients who present with regional metastatic disease from a presumed unknown primary. Many individuals were subsequently found to have a primary cancer with careful examination. Host Susceptibility and Screening Perhaps the most effective means of improving screening strategies is through the better identification of individuals at increased risk of developing disease. More intensive follow-up can be undertaken. Likewise, the use of potentially invasive or costly procedures can be reserved for those who would most benefit. The risk of performing such diagnostic tests on those who are unlikely to benefit would also be minimized. In that regard, the relative risk of disease as a function of duration and intensity of tobacco exposure has been well defined. The confounding influence of alcohol on tobacco-induced cancers has been extensively investigated. New developments in genetics and molecular biology, however, are now providing information as to which smoker may be at increased risk of developing head and neck cancer. Summaries of our current understanding of genetic susceptibility to head and neck cancer development have been provided. [ref: 53-55] Susceptibility factors can be placed into broadly-defined categorical groupings including genetic polymorphisms involving carcinogen-metabolizing enzymes, heritable characteristics associated with race and gender, HLA phenotypes, cancer family syndromes, as well as DNA repair deficiencies. When considering polymorphisms involving carcinogen metabolizing enzymes, one must account for innate host systems that serve to activate both carcinogens and enzyme systems that act to protect the host. In the latter category, it is the glutathione S-transferases and the acetyl transferases that have been most extensively investigated regarding the risk of head and neck cancer. Considerable variation in the general population exists in the individual capacity of an individual to express these protective carcinogen-detoxifying enzymes. Those persons who have deficient activity would be at greatest risk. Although the data are somewhat sparse at this time, several epidemiologic studies have been conducted that support this contention. [ref: 56,57] It is relevant that epidemiologic studies are defining an increased incidence of disease among certain populations within the
TREATMENT Pretreatment Considerations The comprehensive care of the head and neck cancer patient begins with pretreatment considerations including the assessment of general medical conditions, nutritional status, dental health, and the appropriate choice of medical therapies designed to minimize treatment-related complications. It is beyond the scope of this chapter to detail the numerous associated medical illnesses that are typically identified in the head and neck cancer patient. Given the prolonged history of tobacco and alcohol abuse that can be typically identified, diseases involving the pulmonary, cardiovascular, and digestive systems are common. There remain, however, important considerations that should be stressed. These include the significance of pretreatment dental care, nutritional support, the impact of therapy on the elderly patient, and the choice of preoperative medications (i.e., antibiotics for the patient about to undergo major surgical procedures). The standard of care today for the head and neck cancer patient is the reduction of oral diseases prior to initiation of treatment. [ref: 73-75] Periodontal diseases, infections, and caries are common in this patient population. This can lead to loss of integrity of the gingival-crevicular tissues. Left unchecked, significant morbidity can result to structural elements of the oral cavity in the face of aggressive therapy. Following initial evaluation by the oncologic team, dentulous patients should be referred to dental colleagues for appropriate oral hygiene. Pretreatment radiographic dental surveys should identify caries and periapical lesions. Other factors that should be considered include defective restorations, ill-fitting prostheses, and impacted molars. It is generally considered prudent to perform necessary dental rehabilitation, including extractions, approximately 2 weeks prior to the initiation of any radiotherapy. This allows for the appropriate healing of extraction sites and mucosal coverage of exposed bone. To minimize delays in the initiation of radiation, dental care can be performed at the time of surgical resection in a patient who is to undergo multimadality therapy. The assessment of nutritional status and the choice of pretreatment nutritional regimens are more controversial. Several authors report the common finding of malnutrition in the head and neck cancer patient. [ref: 76-78] Severe malnutrition has been identified in over 25% of the patients. [ref: 76] Furthermore, the presence of severe malnutrition was associated with increased operative morbidity. [ref: 78] This has led to efforts to appropriately quantitate nutritional status through the use of documentation of pretreatment weight loss, the measurement of triceps skin fold thickness, and the inclusion of various laboratory measures such as plasma protein levels and the creatinine/height index. To date there has beeo conclusive randomized trial of pretherapy nutritional restoration to minimize treatment morbidity. However, Goodwin and colleagues [ref: 76] have stressed that in the severely malnourished patient, attention should be given to a 2-week pretreatment course of nutritional support. Such attempts can most often be achieved through enteral means with the placement of either a nasogastric tube or a percutaneously-placed gastrostomy tube. The care of the elderly patient represents a commonly occuring dilemma. [ref: 79-81] The tendency to deny a patient optimal treatment because of his or her advanced years should be avoided. In a study by McGuirt and Davis, [ref: 80] operative mortality was 4% in 217 patients greater than 65 years of age and not significantly different than those less than 65 years. In a subset of patients over the age of 75 and with stage III or IV disease, however, the mortality rate increased to 6%. A prospective case-control study by Kowalski and colleagues [ref: 79] on elderly patients undergoing head and neck surgery failed to identify any increased frequency of postoperative complications or mortality as compared to younger patients. The choice of treatment should not be predicated on the age of the patient. Rather, it is the patient’s general medical condition that remains the most critical consideration regardless of age. Increasing attention over the last 5 years has been given to the significance of continued tobacco use in the head and neck population. The data would support the notion that continued tobacco use following diagnosis of the index cancer will lead to an adverse patient outcome. This relates to not only the more obvious problem of progressive cardiopulmonary disease, but considerations related to head and neck cancer progression as well. Day and colleagues [ref: 82] have recently provided more information regarding the risk of second primary malignancies in patients who continue to smoke following treatment of their index cancer. The risk of second primary malignancies was significantly higher in the smoking population as compared to those who achieved smoking cessation. This difference was apparent only after 5 years following initial treatment. In another study by Browman and colleagues [ref: 83] continued tobacco use was associated with a decreased likelihood to respond to primary therapy. This latter study represents an initial report and is limited by its small population size. It does, however, raise an important consideration in the overall care of these patients, namely, the systematic approach toward achieving smoking cessation. In that regard, several studies have recently identified characteristics of patients who are likely to continue smoking habits. [ref: 84-86] Interestingly, Ostroff and colleagues [ref: 86] have recently reported that it is the patients with the best outlook who seem to be the most recalcitrant. Patients with early staged disease continued to use tobacco at higher rates than patients with higher staged disease. Strategies are being explored to effectively support the patient through this critical period. [ref: 84] The hallmark of any approach should include surgeon-delivered advice, as well as the judicious use of nicotine replacement. Finally, head and neck surgical oncologists should critically assess the need for supportive therapies in the patient who is to undergo operative procedures. This has principally related to the use of the choice of perioperative antibiotics. The use of prophylactic antibiotics can only be supported for those individuals undergoing clean-contaminated surgery (i.e., when it is anticipated that the aerodigestive tract is to be entered, or in those circumstances in which there is frank contamination). Postoperative infectious complications may occur in up to 30% of the patients. Several studies have addressed risk factors for infectious complications and found duration of the operative procedure, blood loss, and complexity of the reconstructive procedure to be high-risk variables. [ref: 87-89] Common microbials isolated from infected wounds include both aerobes and anaerobes with Bacteroides fragilis, Escherichia coli, beta-hemolytic Streptococcus, Staphylococcus, and Pseudomonas being frequently identified. Perioperative antibiotics should be started prior to the operative procedure and should continue for 72 hours postoperatively depending on the likelihood of infection. Antibiotic regimens should allow for broad coverage. Regimens include combination sulbactam and ampicillin or metronidazle combined with a cephalosporin. General Principles of Surgery In the execution of effective surgical management, the single most significant principle is the adequate preoperative assessment of disease extent. Precise and methodical physical examination of the patient is paramount. Such examination allows for the assessment of adequate extent of surgical excision, which remains for most cancers the fundamental tenet for achieving cure. An extension of adequate preoperative assessment is optimal intraoperative exposure of disease. The surgeon should consider appropriate means to achieve operative exposure. The choice of incision and the ability to mobilize surrounding anatomical structures to achieve exposure will be considered later in the text for each anatomical subsite. Additionally, exposure is facilitated by careful hemostasis. Besides allowing for better operative exposure, minimizing blood loss prevents potential sequelae associated with blood transfusion. Weber and colleagues [ref: 90] have recently reported expected blood loss for various surgical procedures involving cancers of the upper aerodigestive tract. Electrocautery dissection had been utilized by Weber and colleagues, which may explain the relatively infrequent need for blood transfusion. Electrocautery dissection has been adopted by many experienced surgeons as the preferred extirpative technique. Additional methods of surgical excision have included the use of the Mohs technique and laser ablation. [ref: 91-93] These techniques, however, cannot be considered standard surgical procedure at this time. There exist general principles of surgery involving regional lymph nodes. The standard in the surgical control of cervical metastases by which various procedures are judged is the radical neck dissection. The radical neck dissection involves complete removal of the lymphatic pathways within the neck. To assure complete extirpation, anatomical structures including the sternocleidomastoid muscle, spinal accessory nerve, and jugular vein are routinely sacrificed. Recent developments in the management of cervical lymph node disease involve more conservative surgical procedures. [ref: 3,94-96] These procedures differ from the classic radical neck dissection principally in the sparing of specific anatomical structures (i.e., principally the spinal accessory nerve and the sternocleidomastoid muscle). Table 29.2-4 provides a classification of currently used selective neck dissection and details removed lymph node regions. Controversy exists as to the value of elective neck dissection in the face of clinically N0 disease. Arguments in favor of such procedures are based on the finding that greater than 20% of clinically negative necks harbor histopathologic evidence of metastases. [ref: 97-99] Whether resection of this disease improves survival beyond that afforded by delayed neck dissection when disease becomes clinically evident cannot be stated with certainty. In one of the few prospective trials of elective neck dissection, Vandenbrouck and colleagues [ref: 100] failed to find survival benefit in oral cavity cancer patients randomized to receive elective dissection. The study is limited by the small number of individuals entered. More conclusive studies are in order. In those individuals who are not likely to receive careful postoperative assessment, it is recommended that elective neck dissection be performed. General Principles of Radiation Therapy For early-stage disease, both radiation and surgery are frequently curative and can produce similar rates of cure. Selection of treatment must be individualized to each patient and must consider issues such as cosmetic and functional outcome, quality of life, speed with which treatment can be completed, sequelae of each modality, patient reliability, risk of subsequent cancers, and capacity of salvage therapy should there be a recurrence. For advanced head and neck cancer, surgery and radiation are frequently combined. Tupchong and colleagues [ref: 101] reported the RTOG experience comparing preoperative radiation therapy (5000 cGy) versus postoperative radiation therapy (6000 cGy) for supraglottic and hypopharynx cancers. Locoregional control was significantly better for the postoperative radiation group. The results were most significant for the supraglottic cancers. Marcial and colleagues [ref: 102] found that 5000 cGy given preoperatively did not adversely affect the complication rate compared to patients who had surgery alone for advanced but resectable lesions. Nowadays, if surgery is planned for an otherwise resectable lesion, there is no indication for preoperative radiation therapy. Definitive surgery should be performed, and postoperative radiation is used. In planning treatment for advanced lesions, it is essential to separately consider the primary site and the neck and to integrate a management strategy that considers both locoregional control and quality of life. For example, an early T stage lesion with N2 or greater neck disease is a stage IV lesion. However, it is likely that the primary site can be managed by radiation alone, and the neck should receive combination therapy with surgery and radiation. Such a patient can be treated with definitive radiation therapy to the primary and neck, followed by neck dissection. This strategy maximizes locoregional control, while attempting to optimize functional outcome and quality of life. Similarly, some T3 to T4 patients may be managed by definitive radiation therapy, including brachytherapy, reserving surgery for salvage. This is especially true for oropharynx lesions, where major surgery has significant impact on speech and swallowing. Therefore, it must be emphasized that treatment of advanced, resectable disease must be individualized; whereas most patients may benefit from surgery, followed by postoperative radiation, many may be treated by definitive irradiation, with or without a neck dissection, in an attempt to obtain maximal locoregional control and quality of life. In general, when making decisions about postoperative radiation, it is important to consider the local site as well as the regional volume. In terms of the local site, reasons to add postoperative radiation include inadequate margins, significant local invasion, and a large tumor in the T3 or T4 category. Studies have demonstrated that adequate doses of well-delivered postoperative radiation therapy can sterilize cancer-positive surgical margins. [ref: 103,104] Regional issues that determine the need for postoperative radiation therapy include the presence of involved lymph nodes at one or more levels, extracapsular extension, the risk of contralateral nodal metastases requiring elective irradiation to the contralateral neck, risk of disease in undissected nodal areas such as the retropharyngeal region, and risk of disease in the lower neck. Huang and colleagues [ref: 105] identified 125 patients who underwent radical head and neck operations, and were found to have either extracapsular lymph node extension in the neck, positive resection margins, or both. Surgery alone was used in 71 patients, whereas 54 received postoperative radiation therapy. Treatment was based on physician preference, not tumor characteristics, as one group of surgeons at their institution preferred not to refer patients for postoperative radiation. The two groups were quite comparable except for the delivery of postoperative radiation. The 5-year, locoregional control and adjusted survival were significantly better in patients who received postoperative radiation. This finding was true for those with extracapsular extension as well as those with positive margins. For those with both, locoregional control was 68% in the postoperative radiation group, and 0% (P = .001) in the surgery alone group. Survival was 72% versus 41% (P = .001) for the postoperative radiation patients versus the surgery alone group. Peters and colleagues [ref: 106] reported an important prospective randomized trial from the
SINGLE-AGENT CHEMOTHERAPY TRIALS. The most active agents and their response rates are listed in Table 29.2-7 (120-143) and include methotrexate, bleomycin, cisplatin, 5-fluorouracil (5-FU), and paclitaxel. Methotrexate. Methotrexate is the standard palliative therapy for recurrent squamous cell carcinoma of the head and neck. The standard dose for initiation is 40 mg/m**2/wk to be escalated weekly by 10 mg/m**2 increments to 60 mg/m**2/wk or until dose-limiting toxicity or an objective response is reached. Therapy with this drug is relatively nontoxic, inexpensive, and convenient. Higher doses of methotrexate in single-arm studies were shown to produce higher response rates. [ref: 120-123] Five randomized trials have showo significant difference in survival rates between higher doses of methotrexate with leucovorin (as much as 5000 mg) and standard-dose methotrexate. [ref: 121-126] Bleomycin. Bleomycin has been studied extensively as a single agent and in combination in recurrent-metastatic squamous cell carcinoma of the head and neck. Response rates as a single agent vary from 6% to 45%, with a pooled average of 21%. [ref: 127,128] Because of its distinct toxicity (skin, mucous membranes, and lung but no significant myelosuppression), it has an advantage in combination chemotherapy. Bleomycin can be given in full dosage in combination with other agents with a different spectrum of toxicity. Continuous infusion of bleomycin produces less pulmonary toxicity than bolus injection. Cisplatin. Cisplatin is perhaps the most important chemotherapeutic agent in squamous cell carcinoma of the head and neck. Most of the studies have used a dose of 80 to 100 mg/m**2 every 3 to 4 weeks as standard dose. Response rates have ranged from 14% to 41%, with a pooled average of 28%. [ref: 127,128] Whether there is a dose-response relation is not yet proved. Single-agent cisplatin in doses of up to 200 mg/m**2 produced higher response rates in pilot trials, [ref: 129,130] but a randomized trial comparing 60 mg/m**2 doses with 120 mg/m**2 doses found no difference in response or survival. [ref: 131] 5-Fluorouracil (5-FU). 5-FU was studied initially in recurrent squamous cell carcinoma of the head and neck as second or third line chemotherapy to be used after other drugs failed. Response rates ranged from 0% to 33%, with an average of only 15%. [ref: 127,128] In these studies, 5-FU was usually given as an intravenous bolus daily for 5 days or weekly. The dose-limiting toxicity of this method of administration was myelosuppression. In the 1970s, 5-FU was studied as a prolonged infusion for 96 to 120 hours, usually in a dose of 1000 mg/m**2/day. The dose-limiting toxicity was found to be mucositis; myelosuppression was significantly lower. When 5-FU was administered as an infusion instead of a bolus, it was found to have increased activity in squamous cell carcinoma of the head and neck, synergistic interaction with cisplatin, and enhanced cytotoxicity with modulators such as leucovorin. Taxanes. The taxanes paclitaxel and docetaxel represent a new class of compounds that bind to the beta-subunit of tubulin, induce the formation of stable microtubule bundles, and inhibit microtubule depolymerization. In vitro, docetaxel is the more potent analogue. Docetaxel appears to be schedule independent, whereas paclitaxel appears to be more effective with prolonged exposure. [ref: 144-147] These two agents are undergoing intensive study in patients with head and neck cancer. Trials conducted in the
COMBINATION CHEMOTHERAPY TRIALS. Combination chemotherapy regimens have been developed for treating recurrent squamous cell carcinoma of the head and neck in an effort to improve response rates and survival. Over the last two decades, numerous phase II trials of methotrexate or cisplatin-based regimens have been published. [ref: 127,128] Most contain small numbers of patients and often the results suggest greater efficacy than would be expected with single agent cisplatin or methotrexate. In the early 1980s, researchers at
RANDOMIZED TRIALS OF SINGLE AND COMBINATION CHEMOTHERAPY REGIMENS. Two trials directly compared the single agents methotrexate and cisplatin. [ref: 179,180] No differences in response rate or survival were observed. Randomized trials comparing combination regimens to either single agent methotrexate or cisplatin are shown in Table 29.2-9. [ref: 181-192] Survival benefit was reported in 3 trials. [ref: 184,187,189] Morton and colleagues [ref: 184] reported median survivals of 4.2 and 4 months for patients treated with cisplatin or cisplatin plus bleomycin, respectively, compared to a 2-month median survival for a no treatment control group.
BIOLOGIC THERAPY FOR RECURRENT SQUAMOUS CELL HEAD AND NECK CANCER. Patients with head and neck cancer characteristically have moderate to severe depression of cellular and hormonal immunity, which may be reduced even further by surgery or radiation therapy. Early clinical trials adding nonspecific immunoadjuvant therapy (BCG or levamisole) were negative. Cytokines (interferon [FN] alpha and gamma interleukin-2 [IL-2], and combination IL-2 and IFN) have been used in squamous cell head and neck cancer with disappointing results. [ref: 198] Combinations of interferon alpha-2a and the single agents cisplatin and 5-FU have also been disappointing, with no improvement in either response rates or response duration and substantial increase in toxicity. [ref: 199] Several human squamous cell head and neck cell lines express high-affinity IL-2 receptors and are growth inhibited by IL-
NEOADJUVANT (INDUCTION) CHEMOTHERAPY. Chemotherapy may be given before or after ablative surgery or radiation therapy. The main advantage of induction chemotherapy (i.e., chemotherapy before surgery or radiation therapy) is the preservation of organ function. Induction chemotherapy eliminates the pharmacologic sanctuary problems (i.e., poor vascularity leading to poor concentrations of chemotherapy) after surgery or radiation therapy. Initial chemotherapy results in delivery of chemotherapy to the best possible host, resulting in increased compliance, responsiveness, and tolerance of higher doses. Downstaging of the primary and regional node disease has successfully allowed for organ preservation and a reduction in distant micrometastases. Nonrandomized Trials of Neoadjuvant Chemotherapy. Many neoadjuvant trials have been reported in which the single agents methotrexate, bleomycin, or cisplatin were used. [ref: 204-212] The CR rate with single agents was low (5%). In the late 1970s, a combination of cisplatin and bleomycin used by Randolph and associates [ref: 213] obtained a 71% response rate with a 19% CR rate. Many other investigators since then have used the regimen and have confirmed the high overall response rate, but reported lower CR rates. [ref: 214-218] To achieve higher response rates, investigators added methotrexate [ref: 219-226] or vinca alkaloids [ref: 227-233] to the cisplatin and bleomycin regimen. The CR rate improved from 7% with cisplatin and bleomycin to 16% with cisplatin, bleomycin, and methotrexate and to 20% with cisplatin and bleomycin and vinca alkaloid. The next major advance occurred in the early 1980s when
POSTOPERATIVE ADJUVANT TREATMENT Adjuvant Chemotherapy. Chemotherapy administered after a patient has been rendered disease free with surgery and/or radiation has been evaluated in three large multicenter randomized trials. [ref: 269-271] The rationale for posttreatment adjuvant chemotherapy rather thaeoadjuvant chemotherapy is three-fold. First, for resectable disease definitive surgery is not delayed. Second, wheeoadjuvant chemotherapy is successful, tumor margins are blurred and the extent of required surgery may be uncertain. Third, in pilot trials of neoadjuvant chemotherapy, up to 20% of patients refused surgery once response was achieved and their symptoms abated. The sequence of surgery, followed by adjuvant chemotherapy and then radiation, was a strategy developed to avoid the disadvantages of neoadjuvant chemotherapy. The Radiation Therapy Oncology Group tested this strategy using three cycles of cisplatin and infusional 5-FU adjuvant chemotherapy. [ref: 269] From 1985 to 1990, 499 patients with stage III and IV resected squamous cell carcinoma of the oral cavity, oropharynx, hypopharynx, and larynx were randomized through six cooperative groups. Patients were excluded if margins of resection were positive. Randomization was to immediate radiation or adjuvant chemotherapy, followed by radiation. Patients were stratified into high- and low-risk groups with additional RT (10 Gy) given to high-risk patients (close surgical margins less than
CONCURRENT CHEMORADIOTHERAPY Single Agents and Radiotherapy. The purpose of simultaneous chemoradiotherapy is to increase local-regional control and to prevent distant metastases. The possible synergistic effect has been explained by supposing that the drugs (e.g., bleomycin, cisplatin, and others) interfere with cell repair after sublethal or potentially lethal damage or with tumor cell synchronization. [ref: 282,283] Many clinical trials testing simultaneous chemoradiotherapy have been conducted since the 1960s (Table 29.2-13). [ref: 287,290-308] All the agents active in head and neck cancer have been combined with radiation therapy to enhance its effect and possibly prevent distant metastases. Methotrexate. Methotrexate has been used with concomitant radiation therapy based on its high single-agent activity in head and neck cancer. Two randomized trials have been published with opposing results. Condit [ref: 293] reported no improvement in the combined chemotherapy and radiation therapy group, whereas Gupta and colleagues [ref: 294] reported a significantly better primary control and survival, especially in oropharyngeal cancers. In the study by Gupta and colleagues, there was more cutaneous and mucosal toxicity in the chemotherapy and radiation therapy group, but no patients had their treatment interrupted or required nasogastric tube or intravenous feedings. [ref: 294] There was no increase in late toxicity. Bleomycin. Bleomycin has been combined with radiation therapy frequently in head and neck cancer. Many studies of bleomycin and radiation therapy report CR rates in the range of 38% to 79%. [ref: 282] There are at least nine randomized trials [ref: 295-303] comparing bleomycin and radiation therapy with radiation therapy alone. Three of the trials showed a benefit in response rate or survival or both. [ref: 295,296,302] However, the other five trials, including the large European Organization for Research on Treatment of Cancer (EORTC) randomized trial, did not reveal any significant benefit in the bleomycin and radiation therapy arm. [ref: 303] Most of the studies did report increased acute toxicity, especially mucositis and skin reaction. 5-Fluorouracil. Many studies have used 5-FU with concurrent radiation therapy. Three randomized trials have been reported. [ref: 304-306] In a study from
ORAL CAVITY EPIDEMIOLOGY Oral cavity cancer represents a multiplicity of diseases. Epidemiology as it relates to each of these disease processes likewise differs. However, given that squamous cell carcinoma represents the preponderance of cancers that occur in this region, greater attention will be focused on its etiologic determinants. It is estimated that 30,000 new cases of oral cavity cancer will have occurred in 1991. [ref: 508,509] The relationship between tobacco exposure and disease development has been clearly demonstrated. [ref: 510,511] A clear dose-response relationship has been identified with a greater risk being directly proportional to intensity and duration of exposure. Alcohol has been identified as a coagent, most probably through a topical effect. [ref: 512] The mucosal areas, which are exposed to prolonged contact with alcohol, are at greatest risk of cancer development. Readers are referred to reviews on mechanisms of tobacco-induced carcinogenesis for in-depth understanding. [ref: 513] Likewise, reviews regarding the role of alcohol in cancer development are available. [ref: 512,514] Cigarette smoking cannot be considered the sole etiologic agent for oral cavity cancer. This fact is made evident by the observation that over 50 million individuals in the
ANATOMY The anatomy of the oral cavity will be covered with each specific site.
STAGING T staging for oral cavity cancer applies to all subsites within the oral cavity unless otherwise stated. For the oral cavity, the definition of T staging is as follows:
PRIMARY TUMOR (T) TX: Primary tumor cannot be assessed T0: No existence of primary tumor Tis: Carcinoma in situ T1: Tumor
PATHOLOGY The most common cancer within the oral cavity is squamous cell carcinoma. Additionally, cancers can arise from minor salivary glands; these latter cancers include adenoid cystic carcinoma, mucoepidermoid carcinoma, and adenocarcinoma. Rare soft tissue neoplasms include mucosal melanoma, plasmacytoma, and soft tissue sarcomas. Also found within the oral cavity are cancers arising from bone, including osteosarcomas. There also exist neoplastic lesions that are not truly malignant disorders of bone growth, such as ameloblastoma. These lesions, however, have a propensity for local expansion and destruction. The principles of sound oncologic surgery and radiation thus apply to these latter processes.
NATURAL HISTORY Earliest changes associated with squamous cell carcinoma are associated with erythema and slight mucosal surface irregularities. Many times a punctate lesion will also be identified. As disease progresses, several growth patterns emerge that can be typically characterized as exophytic or infiltrative. The latter is more characteristically associated with destruction of surrounding anatomical structures. The exophytic lesions have a less aggressive growth pattern. It should be realized that both patterns are capable of producing metastasis and disease progression. Therapy should be planned accordingly. Characteristics of the disease will be reflected in certain histopathologic criteria. When considering tumor differentiation, it has been reported that poorly-differentiated disease will have a greater propensity for metastasis than well-differentiated disease. This has, however, not been universally accepted. Jacobsson’s criteria as outlined in Table 29.2-2 will likewise reflect the natural history of the disease. More specific information regarding natural history will be covered in each anatomical subsite. LIP Epidemiology Carcinoma of the lip is second only to skin cancer as a site of neoplasia within the head and neck region. It is noted to occur in approximately 3600 cases per year (i.e., 1.8 persons per 100,000 population annually). [ref: 508] The majority of these lesions occur on the lower lip and 95% occur in males. A principal etiologic factor, similar to other upper aerodigestive cancer, has been the use of tobacco, including both pipes and cigars. [ref: 528] Sun exposure has also been incriminated and may represent the most significant factor. The latter fact is of potential relevance given the increase incidence of other skin cancers as well as lip cancer. Patients genetically susceptible to skin cancers following sun exposure, (i.e., patients with xeroderma pigmentosum) are likewise susceptible to lip cancer. [ref: 516] Such an observation emphasizes ultraviolet radiation as an etiologic agent. Disease has likewise beeoted in renal and homograft recipients, implicating immune suppression as a determinant. [ref: 529] Anatomy The lip is composed of the orbicularis oris muscle and is delineated by the junction of the vermillion border with the skin. Blood supply and sensory nerve supply are by means of the labial artery (a branch of the facial artery) and by cranial nerve V, respectively. The primary lymph node drainage is to levels I and II. Pathology The principal cancer involving the lip is squamous cell carcinoma. Other lesions include basal cell carcinoma. Rarely, minor salivary gland cancers can occur. Natural History Patients most frequently present with either an exophytic or ulcerative lesion of the lower lip. Occasionally these lesions are associated with bleeding and pain. The latter symptom, however, is a late feature of the disease. These lesions are typically slow growing lesions. With progression there may be associated numbness of the skin of the chin secondary to involvement of the mental nerve, a branch of the third division of cranial nerve V. Furthermore, progression of disease along the mental nerve may extend into the mental foramen of the mandible. Such involvement leads to enlargement of the foramen with bone destruction and widening of the inferior alveolar canal. A Panorex examination of the mandible is recommended as part of each diagnostic evaluation. Lymphatic spread occurs relatively infrequently in lip cancer; approximately 5% to 10% of patients will develop evidence of nodal involvement. [ref: 530-532] Lymph node spread is typically to submandibular nodes or submental nodes. Lesions in the midline may spread bilaterally. The incidence of metastases has been related to histologic grading, with high grade lesions being at greatest risk. The upper lip tends to metastasize earlier than the lower lip. Upper lesions will metastasize to periparotid nodes (i.e., preauricular nodes), in addition to submandibular nodes. The prognosis from lip cancer principally depends on the size of the primary tumor. [ref: 533-535] T1 lip cancers have a 5-year survival of 90%. T2 survival is 84%. With evidence of lymph node metastases, survival will decrease to 50%. Perineural invasion represents a bad prognostic sign. [ref: 536] Likewise, prognosis appears worse in younger adults. [ref: 537] Tumors have a greater tendency to metastatic spread in these latter patients. Treatment
EARLY DISEASE. Surgery and radiotherapy are the mainstay of therapy. Dysplasias and carcinoma in situ can be handled by lip shave (i.e., vermilionectomy with advancement of a mucosal flap). Those lesions that involve less than 30% of the lip can be resected with a V excision and primary closure of resulting defects. For larger lesions, transposition flaps are required for reconstruction. Undoubtedly, the challenge in the surgical management of lip cancer resides in the best means of reconstruction.
Oral competence remains the primary goal. Those lesions that require resection of 30% to 50% of the lip can be best handled with a transposition flap drawn from the uninvolved opposing lip. This reconstruction technique is termed Abbe-Estlander. [ref: 538-540] A detailed description of the Abbe-Estlander flap has been provided elsewhere. [ref: 538] When the near total lip is involved (i.e., 50% to 75%), the Karapandzic advancement flap can be utilized. [ref: 541] This has the benefit of providing a competent oral sphincter with an associated neurovascular integrity. The problem with the Karapandzic reconstruction is that the reconstructed lip is tight and significantly foreshortened. Other methods have been devised. [ref: 542-544] These vary from simple nasolabial flaps based inferiorly or superiorly to more formal fan-type flaps such as the Gillies flap and Webster cheek advancement flap. [ref: 544] Most T1-3 squamous cell carcinomas of the lip can be managed by either radiation therapy or surgery. The choice of radiation or surgery may depend on the size and location of disease. If the lesion is quite small and can be easily excised without functional sequelae, surgery would be the chosen treatment. Lesions involving commissures can be irradiated, without the functional sequelae of surgery. However, involvement of the commissure under such circumstances is rare. Brachytherapy alone can be used for early T1 and small T2 lesions. Temporary implantation with iridium-192 is the treatment of choice. Doses in the 6000 cGy range are usually adequate, with the dose rate being in the 40 to 60 cGy per hour range. Figures 29.2-10a, 29.2-10b, and 29.2-10c show a clinical example of the brachytherapy procedure. Similarly, external beam can be used either alone or with an implant for T3 lesions. Given the infrequency with which early cancers spread to regional lymph nodes, elective treatment of the neck is not necessarily required.
ADVANCED DISEASE. Stages III and IV lip disease are optimally managed with combined surgery and postoperative radiation therapy. Reconstructive options are as described above. Doses in the 6000 to 6300 cGy range, delivered at 180 to 200 cGy per fractions over 6 to 7 weeks, is preferred. If the patient has lymph node metastases in the neck, a neck dissection would be done along with the resection of the primary site. The postoperative radiation therapy would then be delivered to both the neck and the primary site. Even if the patient were N0, one should still utilize elective radiation therapy or elective node dissection as part of the management, given the increased risk of microscopic lymph node metastases in these patients. For patients with T(1-3) disease who have had an operation, sometimes the radiation oncologist is faced with a positive margin of resection. This can either be managed with brachytherapy alone, or localized superficial external beam irradiation, with similar doses and techniques as when radiation therapy alone is used. Results of Treatment The radiation therapy results are similar to the reported results of surgical management. [ref: 545] Heller and Shah [ref: 533] reported approximately 90% local control for T1, T2, and T3 lesions treated with surgery alone. A significant problem in surgical management is local recurrence, which may approach 40% for T3 and T4 lesions. [ref: 531] Fitzpatrick [ref: 546] has reported the
FLOOR OF MOUTH Epidemiology The annual incidence of cancers of the floor of mouth is 0.6 cases per
EARLY DISEASE. Treatment of floor of mouth cancers has been principally surgical resection but may either be surgery or radiation alone. As is true for cancers of the alveolar ridge, superficial involvement of the mandible can be handled by marginal mandibulectomy (Fig. 29.2-12). When radiation is used for early disease, it has been shown that results are improved when at least a portion of the treatment is delivered by an interstitial implant. [ref: 566,567] Interstitial implant alone can also be used. [ref: 568] Lesions that abut or are tethered to the periosteum of the mandible are not good candidates for primary radiotherapy. Implants against the mandible can lead to osteonecrosis. The treatment of the neck for early cancer of the floor of mouth is controversial. Most would advocate elective neck treatment for clinically N0 disease. The value of this approach as compared to observation with neck dissection and radiation therapy being performed for clinically developing disease remains unproven. Some have advocated performing neck dissection dependent on the thickness of the primary lesion (i.e., more than
ADVANCED DISEASE. For advanced lesions, combined therapy of surgery and radiation is the treatment of choice. Surgical resection generally entails partial glossectomy and segmental mandibulectomy. Identification of the inferior alveolar nerve and frozen section histopathologic assessment should be performed during the operation.
This is to ensure that disease has not extended beyond surgical margins by perineural spread. Resection for most advanced lesions will require removal of the entire thickness of the floor of mouth. New reconstructive techniques have greatly facilitated rehabilitation following surgical excision of advanced tumors. Techniques include myocutaneous flaps as well as osteomyocutaneous free flaps with microvessel anastomoses. [ref: 569] Elective and/or therapeutic neck dissections are considered necessary in each case. Bilateral neck dissections are indicated for those lesions that approach or cross the midline. Postoperative radiation entails doses in the range of 6000 to 6300 cGy at the primary site. In instances of positive surgical margins, our policy is to treat the area of positive margins to 6300 cGy. Patients are treated with opposed lateral fields for the primary site and upper neck, which junctions to a low neck field at approximately the thyroid notch. A midline block is used for the low neck field to protect the larynx and spinal cord. The dose to the low neck is usually 5000 cGy given in 5 weeks. The upper neck and primary site are generally treated to 4500 cGy in 5 weeks, after which a spinal cord block is placed. The primary site and upper neck are generally taken to 5400 cGy, after which a boost to the primary site and the involved region of the neck is taken to approximately 6300 cGy. Electron fields are used posterior to the spinal cord block, as needed, to bring the posterior neck to the appropriate dose. Results of Treatment Local recurrences after surgical resection of T1 and T2 floor of mouth cancers are noted io less than 10% of the patient population. [ref: 561,562,570] As tumors increase in size, the pattern for failure becomes predominately a regional problem. Nearly 40% of failures will be solely within regional cervical lymph nodes. [ref: 561] Mazeron and colleagues [ref: 568] have reported a large recent radiotherapy series. The majority of patients were treated to 65 Gy with Iridium-192 brachytherapy alone. Local control was 94% for T1, N0, and 74% for T2, N0 lesions and was dependent on size of lesion as well as the presence or absence of gingival extension. Wang and colleagues [ref: 571] have reported excellent results with the use of intraoral cone electron boost and no brachytherapy. The daily dose of radiation via the cone is frequently greater than the conventional 180 to 200 cGy range. Local control was obtained in all 13 patients with T1 lesions, and in 19 of 20 (95%) with T2 lesions. Fu and colleagues [ref: 572] have published an extensive radiation therapy experience with floor of mouth cancer. When implant was either the only treatment or a part of the treatment, local failure occurred in only 2% (1 of 39 patients) with T1 lesions, 7% (4 of 54 patients) with T2 lesions, and 14% (5 of 35 patients) with T3 lesions. The use of primary radiation may be associated with improved functional outcome as compared to surgery, but this requires more investigation. [ref: 573] When examining results of treatment for advanced disease, a retrospective review of the results at the Memorial Sloan-Kettering Cancer Center [ref: 104] highlights the fact that local control is improved with combined surgery and radiotherapy as compared to surgery alone, for patients with stage III and IV head and neck lesions. In a prospective randomized trial, Kramer and colleagues [ref: 574] compared preoperative radiation (5000 cGy) plus surgery, surgery plus postoperative radiation (6000 cGy), and radiation therapy alone (6500 to 7000 cGy) with surgical salvage for patients with certain stage II, and all stage III and IV squamous cancers of the oral cavity and oropharynx. There was no difference in the 4-year actuarial survival, or locoregional control, regardless of treatment type. However, this study did not stratify the results by anatomic subsite within the oral cavity and oropharynx. A recent report from
EARLY DISEASE. It is generally considered that disease control rates for early disease when using either surgery or radiation therapy are equivalent depending on treatment bias. Early stage I and II lesions can usually be removed intraorally. Excision usually entails an hemiglossectomy. Special attention to surgical margins should be exercised, since disease may spread along muscle bundles beyond that expected by clinical assessment. Most T1 lesions can be managed with brachytherapy alone. This generally consists of an iridium-192 implant. Although needles are still used by some groups, this is really an old technique that should be avoided. It introduces unnecessary exposure to the physicians, nurses, and other staff, and it does not allow the physician to optimize the dose distribution of the brachytherapy procedure. Iridium-192 is inserted via after loading catheters. The catheters themselves are placed in the operating room under general anesthesia. The iridium-192 is loaded 1 to 2 days postoperatively. Localization films are taken, and computerized dosimetry is performed. The usual dose rate is in the 40 to 60 cGy per hour range, and the usual total dose is 6000 to 7000 cGy. The patient wears a tongue prosthesis during the dwell time of the implant to protect the hard and soft palates as much as possible. Because the lesion increases in size when using radiation as primary therapy, it is preferred to combine external beam irradiation with implant. First, the external beam can be used as elective neck irradiation simultaneous with irradiation to the tongue. The implant then serves as the boost to the tongue. Second, the external beam allows a wider margin of tongue to be treated than does the implant. In these situations, it is typical to treat the primary site and the neck to doses in the 5000 cGy range, followed by a 2000- to 3000-cGy implant boost to the tongue. For N0 patients, this treatment program manages both the primary site and the neck. For those patients with palpable neck nodes, a neck dissection can be performed at the same anesthesia as the implant, thereby completing the treatment to the primary site and the neck. This can usually be done about 3 weeks after the completion of the external beam irradiation. Radiation therapy is certainly suitable for most T1 lesions. For T2 and T3 lesions, it is most suitable for those tumors that are exophytic or have minimal infiltration. Tumors that are deeply infiltrative are preferably managed with a primary surgical approach, usually with postoperative radiation therapy.
ADVANCED DISEASE. The surgical management of more extensive lesions requires either a mandibulotomy or a lingual releasing procedure to gain access to disease. The latter procedure entails removal of neck contents prior to primary cancer resection. The tongue is delivered into the neck by releasing musculature attachments posteriorly and mucosal attachments within the oral cavity. Large lesions with mandibular involvement will require composite resection. The term composite resection refers to the removal of tissue involving multiple anatomically defined structures, one of which includes mandible (Fig. 29.2-13). Typically, it refers to resection of a portion of tongue, floor of mouth, and segment of mandible. For patients who require postoperative radiation therapy to the primary site alone, this can frequently be done with brachytherapy. This is especially true for the smaller lesions. The decision of how to deliver this irradiation is integrated with the management of the neck. We prefer to use neck dissection as part of the management of all deeply infiltrative or advanced tongue lesions. For the N0 patients, this generally means a staging procedure or a functional neck dissection. For patients who have involved lymph nodes, this means either a radical neck dissection or one of its modifications. Results of Therapy Decroix and Ghossein [ref: 582] from the Curie Institute in Paris have reported on over 600 patients treated with primary radiation therapy for T1, T2, or T3 squamous cell carcinoma of the oral tongue. Although the majority of patients had implants alone, a large cohort had combined external beam irradiation plus implant. Almost all patients at this center received radiation therapy as their primary treatment. Primary control was obtained at 86% for T1, 80% for T2, and 68% for T3. These data compare quite favorably with the results obtained with partial glossectomy. Pernot and colleagues [ref: 583] have recently reported a series of 448 patients with brachytherapy alone with or without neck dissection (181 patients) or combined external beam plus brachytherapy (267 patients) for oral tongue cancer. The 5-year local control was 93% for T1, 65% for T2, and 49% for T3 lesions. For T2 lesions managed by brachytherapy alone, local control was 90% versus 50% for those managed by external beam plus implant. These data emphasize the importance of using brachytherapy as a major part of the radiation program, but patient selection factors clearly play a role as well. The 5-year overall survival for T1, T2, and T3 lesions (all N stages) was 69%, 41%, and 25%, respectively. Severe complications were uncommon. Whereas 15% experienced grade 1 soft tissue injury and 3% had grade 1 bone necrosis, only 1% and 2% had grade 3 soft tissue and bone complications, respectively. Spiro and colleagues [ref: 584-586] have reviewed the
REHABILITATION OF THE ORAL CANCER PATIENT Over the last decade there has developed an increased effort to address the rehabilitation of the oral cancer patient, the major considerations reflected in speech and swallowing disorders. [ref: 601-606] There is no question as to the impact of surgical resection of anatomical components of the oral cavity including the mandible, tongue, and other soft tissue components. The ability to preserve function through appropriate reconstructive measures is becoming increasingly apparent. Likewise, rehabilitation efforts have been enhanced in recent years by improved quantitative assessments of functional outcomes, as well as through improving rehabilitation techniques. No medical center or treating physician can truly be considered as providing state-of-the-art therapy unless the are prepared to systematically address these issues.
Tumors of the Larynx and Hypopharynx
LARYNX Considering that cancer occurs in the larynx 14 times less frequently than in the lung, 15 times less frequently than in the breast, 16 times less frequently than in the prostate gland, and 9 times less frequently than in the colon, the number of publications on laryngeal cancer that appear in the North American literature seems excessive. This considerable body of writing probably reflects the perceived importance of this disease relative to its potential impact on people’s communication skills. Among oncologists, there seems to exist a new attitude that is characterized by a keener concern for quality of life and death; curing the cancer at any cost is no longer accepted casually, and more than ever before, a premium is now placed on return to a productive and useful lifestyle after cancer treatment. Nowhere in oncology is this changed attitude more vividly demonstrated than in the treatment of laryngeal cancer. It would seem that the mere threat to a patient’s vocal organ is associated with profound psychological overtones. Investigations continue, therefore, into the methods of conservation laryngeal surgery, different radiation therapy strategies, and combined chemotherapy/radiation therapy protocols designed for larynx preservation. [ref: 1-3] Although the cure rates of the various laryngeal malignancies have not changed dramatically during recent years, [ref: 4] the treatment options and the sequencing of those options have, and a higher percentage of laryngeal cancer patients are retaining their larynx in the process.
EPIDEMIOLOGY AND ETIOLOGY Even though considerable differences between countries exist in the incidence of larynx cancer, its distribution within each country is consistent. For example, the disease most commonly affects middle-aged or older men who have smoked tobacco [ref: 5,6] and have drunk alcohol. [ref: 7] Laryngeal cancer rarely occurs in people who have done neither. In the
SURGICAL AND DEVELOPMENTAL ANATOMY The larynx is a uniquely complicated organ that is strategically located so that significant alteration of its anatomy by either surgery or cancer can have a noticeable impact on vocal, digestive, and respiratory physiology. The organ consists of three subsites, which are the glottis (paired true vocal cords), the supraglottis, and the subglottis. Because of different embryologic development and different lymphatic patterns that are subsite specific, discussing larynx cancers without specific reference to the exact location(s) within that structure invites inaccuracies in staging and miscalculations in treatment planning. The larynx consists of four cartilages: the cricoid, the epiglottis, the paired arytenoids, and the shield-like thyroid cartilage. Suspended within the endolarynx are the mobile true vocal cords, which are collectively known as the glottis.
That portion above the glottis, the supraglottis, consists of the false vocal cords, the epiglottis, and the aryepiglottic folds. These folds form the junction with the hypopharynx. The medial wall of the aryepiglottic fold is within the endolarynx, and its lateral wall makes up the medial wall of the adjacent pyriform sinus (Fig. 29.3-1). Those lesions that arise on the rime of the aryepiglottic folds, therefore, have been appropriately referred to as marginal cancers, because they bridge the junction between the larynx and the hypopharynx. Those marginal lesions that extend predominantly into the endolarynx behave more like supraglottic cancers, whereas those lesions that spill into the pyriform sinus tend to follow the natural history of the hypopharynx. The subglottis is that portion of the larynx between the underedge of the true vocal cords and the cephalic border of the cricoid cartilage. [ref: 30] The true vocal cords are a marvel of engineering and are attached anteriorly to the inner lamina of the thyroid cartilage and posteriorly to the arytenoids. The muscles of the vocal cords are complex in their activity, and the relation between them and the overlying mucosa is critical to voice production. Any loss of mucosal mobility relative to the underlying muscle, such as that produced by surgery or, to a lesser extent, by radiation therapy, alters the voice. An appreciation of this fundamental fact is an important component in the selection of treatment of vocal cord cancer. The lining of the endolarynx consists of respiratory epithelium except on the vibratory edges of the true vocal cords, which typically are lined with pseudostratified squamous epithelium. The paired arytenoid cartilages each sit on the cephalic rim of the cricoid cartilage and rotate in a relatively horizontal axis around a central pivot point. Each arytenoid is attached anteriorly to a true vocal cord, and the clockwise and counter-clockwise rotation of these cartilages pulls the respective vocal cord attachment with it, causing abduction and adduction of those structures. Any or all of the muscles that are responsible for arytenoid rotation and also the branches of the recurrent laryngeal nerve fibers that innervate them can be damaged by invading cancer. The posterolateral aspect of the larynx is particularly vulnerable to the invasion of cancer because of the adjacency of the medial wall of the pyriform sinus. When cancers of this part of the hypopharynx extend through the mucosa, they gain direct access to the important laryngeal compartment known as the paraglottic space, which leads to all parts of the endolarynx, including the vocal muscles and the preepiglottic space (Fig. 29.3-2). Treatment options for such a tumor are altered significantly because of the paraglottic space involvement. Tumors that invade the endolaryngeal muscles or the nerve fibers that innervate them usually create a noticeable effect on vocal cord motion. Of all the findings on laryngeal examination during cancer evaluation, the state of endolaryngeal mobility is one of the most important. This fact that has been substantiated by the separate designation that the American Joint Committee on Cancer (AJCC) has ascribed to the immobile vocal cord in the staging categorization of this disease. [ref: 31] Another type of motion alteration pertains to the anatomic relation between the vocal cord musculature and the overlying mucosa. This relatively recent knowledge has enhanced our understanding of the pathogenesis and treatment of early glottic cancer. [ref: 32] The vibratory mechanism that produces the voice is due to the mobility of the mucous membrane overlying the musculature of the vocal cord. The free edge of the true vocal cord consists of a pseudostratified squamous epithelium, under which is a lamina propria of fibroelastic and gelatinous consistency. This arrangement allows a sliding motion of the mucous membrane that creates a mucosal wave, the fluidity of which is a direct reflection of the freedom of that layer from the underlying muscle. Any surface cancer that invades through the basement membrane, such as any cancer deeper than carcinoma in situ, affects the mucosal wave by creating a tethering effect. These subtle differences are hardly appreciated by routine laryngeal examination but are obvious with laryngostroboscopic evaluation. [ref: 33] An appreciation of these subtleties translates into the practical matter of determining whether to radiate or microscopically excise certain minimal vocal cord cancers. Because of the different embryologic origins of the supraglottic from the glottic and subglottic larynx, and also because of the independent lymphatic drainage patterns from each of these subsites, the larynx can be thought of as a compartmentalized structure. These features are important influences in determining the spread of various cancers within that organ. [ref: 34] The lymphatics of the supraglottic larynx are profuse, and the frequency of metastasis associated with cancers of this subsite reflects that fact. [ref: 35] Lymphatic spread from the epiglottis is to the false cords, and these channels are directed bilaterally. The drainage from the false cords and the remainder of the supraglottic larynx is lateral and superior, and these channels exit the larynx bilaterally through the thyrohyoid membrane. They then proceed to the adjacent deep cervical nodes. The lymphatics of the infraglottic larynx drain laterally and inferiorly, out of the cricothyroid membrane into the lower deep cervical lymph nodes. The true vocal cords, on the other hand, are unique because they possess little or no lymphatic drainage. [ref: 34] From a lymphatic drainage standpoint, the left half of the larynx is essentially independent from the right and the supraglottic larynx is independent from the structures below. These facts are clinically demonstrated; in their early stages, supraglottic cancers have little affinity for extension into the lower structures, and those beginning below do not tend to extend cephalad. [ref: 36,37] Knowledge of this unique pathogenesis has substantial impact on our ability to predict metastasis into various parts of the neck and on our planning of the various partial laryngectomies, which are those conservation operations that allow removal of laryngeal parts while preserving vocal and swallowing functions. Additionally, radiation therapy planning, especially for occult cervical metastasis, is predicated on a thorough knowledge of these and other drainage tendencies of laryngeal cancers.
PATHOLOGY, PATHOGENESIS, AND NATURAL HISTORY General Considerations A variety of malignancies, most of which are primary to the larynx, and others that are metastatic from other sites have been reported. A comprehensive classification is shown in Table 29.3-2. More than 95% of all primary laryngeal malignancies are squamous cell carcinomas, with the remainder being sarcomas, adenocarcinomas, neuroendocrine tumors, and other types. [ref: 38] It should be noted that knowledge and recognition of the neuroendocrine family of malignancies has changed considerably during recent years, and the exact percentage of these within the overall population of larynx cancers is unknown. In the past, certain tumors were vaguely classified as poorly differentiated malignancies, when they actually were neuroendocrine in origin. Modern techniques of immunohistochemical and morphologic analysis will almost certainly lead to the recognition and accrual of more of these tumors in the future. [ref: 39,40] Separate consideration must be given to the spectrum of premalignant squamous lesions, carcinoma in situ, and superficially invasive carcinomas. To discuss the epithelial changes that precede and probably lead to carcinoma of the larynx is of considerable importance, because it is with this group of lesions that cancer prevention and conservative management are most effective. As our knowledge of this subject has increased, so too has our sophistication in applying the minimal techniques necessary to achieve excellent cure rates in these disorders. The value of this philosophy of treatment minimalism is that it leads to the least physiologic change. Investigators have studied the occurrence of aberrant squamous epithelium in various areas of the larynx, and there seems to be a correlation between that metaplasia and the predilection for carcinogenesis in those respective sites; [ref: 41] however, because only lesions that began on the true vocal cords produce early symptoms and signs, the opportunity to treat minimal disease is largely limited to that structure, a fact that leads to spectacular cure rates for lesions of that site.
The mucosal changes that lead to cancer take years to develop, and that evolution probably follows a consistent pattern. Most laryngeal squamous cell carcinomas result from prolonged exposure to recognized carcinogens that stimulate mucosal hyperplasia and metaplasia. Some of these changes are associated with keratosis, and others are not. In some situations, epithelial atypia or dysplasia may exist, the degree of which probably determines whether a lesion is destined to become malignant. [ref: 42-44] In one large study by Slamniker, 3% of those patients who demonstrated vocal cord keratosis without atypia and 7% with mild atypia developed invasive carcinoma [ref: 45]; however, in those patients with moderate and severe atypia, 18% and 24%, respectively, developed carcinoma. Another study by Hjslet and colleagues showed a similar probability of cancer evolution in the group with less dysplasia and a strikingly higher probability in those patients with severe atypia. [ref: 46] In addition to the morphologic appearance of mucosal alteration, DNA changes seem to show a correlation between cancer potential and cellular aneuploidy. In a study by Munck-Wirland and coworkers, for instance, all of those patients with dysplastic laryngeal lesions that later went on to become carcinoma demonstrated an aneuploid DNA pattern. [ref: 47] These surface lesions, whether premalignant or not, have an inconsistent gross appearance. Some of these lesions are white, and others are hyperemic. Many investigators believe the risk for cancer development is substantially higher in lesions that are soft and red in appearance. [ref: 44] Without histologic study, even the most experienced diagnostician cannot consistently predict the presence of cancer or the likelihood of its evolution in any of these surface lesions. Any given spot within a lesion does not necessarily represent the balance of that lesion. The facts that carcinoma in situ is often surrounded by dysplastic epithelium and that many areas of invasive carcinoma are surrounded by zones of carcinoma in situ and dysplastic epithelium [ref: 48] lend credence to the concept that each of these morphologic categories of epithelial disturbance is but part of a dynamic spectrum of disorders, each probably related and representing different stages of the same process. This rationale means that dysplasia leads to carcinoma in situ, which leads to invasive carcinoma. It is unknown whether those lesions that have achieved the status of carcinoma continue to grow at the same rate as they did during their premalignant state or whether their growth is accelerated. The growth of a cancer through the basement membrane into the lamina propria constitutes the transition from carcinoma in situ to microinvasive carcinoma, and accompanying this is a tethering of vocal cord mucosal motion. [ref: 33] Failure to appreciate these subtle changes can result in the employment of suboptimal treatment. For example, high failure rates that have been reported with mucosal stripping in carcinoma in situ patients almost certainly represent underestimation of these lesions. [ref: 49] Some of those lesions that had been classified in the prestroboscopic era as carcinoma in situ probably contained areas of invasive carcinoma, and the stripping left behind foci of cancer that resulted in recurrence. Actually, pure vocal cord carcinoma in situ is probably an unusual occurrence. The gross appearance of a given laryngeal lesion is suggestive of its general type. Squamous cell carcinomas originate within the mucous membrane and are exophytic or ulcerative, are of surface origin (Fig. 29.3-3), and are frequently adjacent to or surrounded by mucosal keratoses. Neuroendocrine cancers and tumors metastatic to the larynx are usually submucosal and, as such, do not resemble lesions of surface origin. Metastatic lesions of various types and neuroendocrine tumors are seen throughout the various subsites within the larynx, although the latter group shows a predilection for the supraglottic area. The distribution of squamous cell carcinoma within the various laryngeal subsites varies between different countries, a fact that reflects the different social habits within those cultures. For example, in the
SUPRAGLOTTIS. Lesions of the supraglottic larynx tend to spread locally. If they begin on the epiglottis, they can extend onto the false vocal cords and into the ventricle. Inferior extension beyond the ventricle is initially thwarted, but as growth continues these cancers can penetrate into the paraglottic space from which they gain full access to the length of the endolarynx. These cancers often exit the paraglottic space cephalad and caudad to enter directly into the neck. Most supraglottic lesions arise on the epiglottis, with fewer being seen on the false vocal cords and aryepiglottic folds. Those lesions that occur on the suprahyoid or upper part of the epiglottis are more often exophytic, whereas those that occur on the lower portion of that structure are likely to be endophytic or ulcerative. [ref: 50,51] The characteristic of endophytic growth is especially significant in this particular area of the epiglottis, because there are foramina here that lead directly through the cartilage into the preepiglottic space, which is a compartment that leads to the base of the tongue. What would appear to be a localized tumor in the endolarynx, therefore, can actually involve considerable unrecognized extralaryngeal extension. [ref: 52] Tumors are confined initially to the preepiglottic space by the ligamentous boundaries of that compartment, but once those barriers are overcome, the loosely arranged skeletal muscle fibers of the tongue provide no restriction to further tumor extension. [ref: 61] Modern imaging, especially magnetic resonance (MRI), has greatly improved the ability to recognize tumor extension into the preepiglottic space and base of tongue. Those lesions that occur on the laryngeal surface of the epiglottis are capable of invading and destroying that structure. Supraglottic cancers, on the other hand, almost never destroy the thyroid cartilage. [ref: 62,63] This feature has an influence on the design of treatment plans; for example, an ossified and invaded thyroid cartilage poses a substantial problem for surgeons attempting to perform partial laryngectomy and also for radiation oncologists attempting to deliver curative therapy. Aryepiglottic fold cancers are somewhat different in their behavior, following more the tendencies of the pyriform sinus lesions, that is, spreading in a more diffuse fashion and metastasizing more frequently than their endolaryngeal counterparts. The particularly ominous natural history of these lesions probably relates as much to the more abundant and multidirectional lymphatic drainage of the area as it does to individual cellular peculiarities. [ref: 52] Because of the profuse lymphatic network of the area, supraglottic carcinomas metastasize frequently to the cervical lymph nodes, and failure of treatment is usually a result of metastasis rather than local disease. [ref: 35,64-66] The incidence of patients with clinically positive lymph nodes at the time of diagnosis is 23% to 50% for all supraglottic sites and stages combined. [ref: 51,67-71] A substantial number of those patients with clinically negative necks turn out to have histologic disease if a neck dissection is done or, if left untreated, to convert to clinically positive necks. [ref: 64,65] In supraglottic cancers, the probability of cervical metastasis and the probability of delayed contralateral metastasis increase in direct proportion to the size of the primary (i.e., the T stage). [ref: 50,72,73] Lindberg reported impressive overall metastatic rates with various supraglottic carcinomas: T1 had 63%; T2 had 70%; T3 had 79%; and T4 had 73%. [ref: 35] In that group of patients with supraglottic lesions that present with a clinically positive cervical node
To a large extent, growth characteristics and the natural history of glottic carcinomas are determined by the unique anatomy of the true vocal cords. First, the sparcity of the lymphatic drainage of the true vocal cords in all areas other than the posterior commissure makes metastasis of early lesions extremely unlikely. Second, the elastic layers (conus elasticus) within the larynx often divert cancers that begin on the free edge of the vocal cord and continue into the underlying vocalis muscle and paraglottic space, which is an inferolateral pathway that leads out of the larynx through the cricothyroid space.
With penetration into the underlying tissues, all degrees of motion impairment, from subtle mucous membrane stiffness to frank fixation of the vocal cord, can follow. That increasing impairment of motion has a telling effect on local control and survival data, a fact that is reflected in AJCC staging designations. Much discussion continues about mobility change and its therapeutic implications, and it is in the group of glottic cancers that demonstrate this that the clinical judgment of the physician is most tested. The final anatomic factor unique to the glottis that influences the growth pattern of certain cancers is the anterior commissure ligament that forms the bridge between the anterior ends of the true vocal cords. This structure lies immediately against the inner lamina of the thyroid cartilage, and its presence initially retards penetration of cancers of that area, often causing their diversion upward onto the epiglottis or downward onto the cricothyroid membrane. From there, these lesions can escape the larynx into the anterior neck. If the cancer overcomes the ligamentous barrier at the anterior commissure, the cartilage is penetrated. [ref: 62,80] This event is particularly likely in thyroid cartilages that are ossified, and when this does occur, there are substantial therapeutic implications that compromise the radiotherapist and dictate certain surgical approaches. [ref: 81,82]
SUBGLOTTIS. Carcinomas of the subglottic larynx are unusual, making up only about 1% to 8% of all laryngeal cancers. [ref: 83] These lesions tend to be poorly differentiated and often demonstrate an infiltrative growth pattern unrestricted by tissue barriers. These tumors are, therefore, frequently circumferential and can extend down the trachea. The incidence of cervical metastasis in this group of cancers is reported to be 20% to 30%, but that figure is somewhat obscured by the fact that the primary drainage pattern of these lesions is to the less detectable pretracheal and paratracheal nodes. The actual incidence of metastasis may, therefore, be significantly higher. [ref: 84,85] Unusual and Rare Neoplasms The pathology and pathogenesis of verrucous carcinoma is unique and deserves special consideration. This unusual tumor is poorly understood, and its origin, classification, and response to treatment are controversial. [ref: 86,87] Verrucous carcinoma is described as a distinct neoplastic entity of squamous origin that occurs in the oral cavity, larynx, esophagus, and nose and on the genitalia. [ref: 88-90] Some authorities have suggested the human papilloma virus as its cause. [ref: 92] Even though there are views to the contrary, most investigators consider verrucous carcinoma to be an entity unto itself. [ref: 91] Just because some tumors originally thought to be verrucous carcinoma are discovered to have features of squamous cell carcinoma and can metastasize does not, in their opinion, justify combining the two diagnoses; actually, they think that such tumors were always low-grade squamous cell carcinomas rather than verrucous carcinoma. [ref: 90] Other investigators, although conceding verrucous carcinoma to be unique, believe it to be only a variant of well-differentiated squamous cell carcinoma. Different authors believe that because verrucous carcinomas neither fulfill the histologic and cytologic criteria of malignancy nor possess the capability to metastasize, they should be renamed verrucous acanthoma rather than carcinoma. [ref: 93] When this lesion does occur in the larynx, it usually is on the true vocal cord, where it grows slowly and can cause significant local destruction by expanding gradually. Even though these lesions often destroy cartilage, they do not tend to metastasize; aggressiveness is directed locally. Verrucous carcinoma is consistently difficult to diagnose, even when the clinical index of suspicion is high. This observation relates to the fact that these tumors microscopically demonstrate an exuberant and keratinizing hyperplasia that is benign by pure histologic and cytologic criteria. [ref: 94] The diagnosis is largely a clinical one and is most effectively achieved by concern between pathologist and surgeon, but usually only after multiple biopsies have been taken. This tumor is typically a slow-growing but relentless mass, exophytic and warty in appearance, and broad based at its interface with the mucosa (Fig. 29.3-4). Its surface is ofteecrotic and infected, and the associated inflammation of adjacent tissues can be remarkable. This tendency to cause inflammation can erroneously influence treatment planning. For example, the patient with verrucous carcinoma can demonstrate enlarged adjacent cervical lymph nodes that are worrisome when in fact the adenopathy is only secondary to the inflammatory process. Although this finding has been described in other aerodigestive tumors, [ref: 95] the mere presence of lymphadenopathy in the primary drainage area of an impressive primary tumor is worrisome, no matter how benign looking its histology. In such a circumstance, clinical judgment is enhanced greatly by modern imaging and cytopathologic techniques. This discussion of the nature of verrucous carcinoma has substantial therapeutic implications, especially when the lesion occurs in the larynx. Essentially, squamous cell carcinoma is a radiosensitive cancer, a fact that provides treatment options to the physician. On the other hand, verrucous carcinoma seems to be somewhat radioresistant, whether found in the mouth or the larynx. [ref: 86] Additionally, there is anecdotal information suggesting radiation induced dedifferentiation into anaplastic cancer in these lesions. This transformation seems to occur in fewer than 10% of verrucous carcinoma [ref: 67,96-99] and may involve alteration of the DNA that facilitates the integration of the human papillomavirus into host cells. [ref: 92] Both the concept of radiation resistance and the transformation into anaplastic cancers are vigorously disputed. [ref: 90,100-102] The neuroendocrine family of tumors represents an evolving data base. This is true largely because recent diagnostic techniques have allowed pathologists specifically to label as neuroendocrine a variety of previously undefined cancers. Almost certainly, an immunohistochemical reexamination of laryngeal cancers previously diagnosed as atypical or undifferentiated malignancies would result in the reclassification of many as neuroendocrine tumors. The small cell tumors look and act much like their counterpart oat cell lung lesions and as such generally are managed by chemotherapy and radiation therapy. [ref: 3,39,103] Surgical procedures do not seem to enhance the likelihood of survival in patients with these tumors. [ref: 104-108] The other neuroendocrine tumors that occur in the larynx — carcinoids and paragangliomas — are rare and are best managed surgically. [ref: 109-111] Cartilaginous malignancies, [ref: 112,113] adenocarcinomas, [ref: 114] sarcomas, [ref: 115] malignant fibrous histiocytomas, [ref: 116,117] plasmacytomas, [ref: 118,119] granular cell tumors, [ref: 120] and primary lymphomas [ref: 121] have all been reported but are rare. Primary melanomas of the larynx are equally rare; of all of the larynx cancers reported from
DIAGNOSIS AND EVALUATION Cancers of the supraglottic larynx usually do not produce early symptoms or signs, and it is common for the first hint of such a cancer to be cervical adenopathy. When symptoms do occur, they are often subtle; pain perceived in the primary site or in the ear (otalgia), a scratchy sensation when swallowing, or merely an alteration of one’s tolerance for hot or cold foods may be all that is noticeable. Airway alteration, hoarseness, or a tendency to aspirate liquids are all produced by more advanced lesions. Cancers of the glottis, on the other hand, are often detected early in the course of the disease because even a slight alteration of the vibratory surface of the true vocal cord(s) produces voice change. Smokers are often hoarse, however, and such alteration of the voice may not alarm them. Anyone with a voice change that does not resolve within several weeks should have a laryngeal examination. It is unusual for glottic cancer patients to seek medical attention because of cervical adenopathy. In such lesions, metastasis generally occurs late in the course of the disease, long after the early warning signals. Fortunately, subglottic cancers are uncommon, but when they occur, they fail to produce early symptoms; therefore, the disease is often advanced by the time of diagnosis. Most larynx cancers are squamous carcinomas and, as such, are surface lesions. Most are visible with routine laryngeal inspection, but a small percentage are obscure. The modern generation of flexible endoscopes (Fig. 29.3-5) has provided to a broad range of physicians the capability to examine the larynx; and with the improved optical resolution of these instruments, the overall process of screening and follow-up after treatment has been enhanced by this technology. Importantly, these methods allow the occasional laryngeal examiner to see areas that previously have been visually inaccessible. It is essential that the larynx be examined in the awake patient who is sitting upright. Direct laryngoscopy should be reserved for biopsy and a more detailed tumor mapping (Fig. 29.3-6). Even when done under local anesthesia, the introduction of a direct laryngoscope distorts the natural position and the relaxed motion of the larynx and, by doing so, tends to disguise subtle motion changes that are important in staging of these tumors. Certain subtleties of contour, such as bulging and tethering, are visually not appreciable during direct laryngoscopy. The earliest stage of invasive glottic carcinoma through mucosa into the underlying lamina propria is visible as a tethering of the mucous membrane that is designed to slide over the underlying structures (i.e., a loss of the membrane wave). [ref: 32] The gross abductive capabilities of the vocal cord may be intact, but the early invasive character of a lesion can be appreciated when the stroboscope is employed and demonstrates this restrictive feature. As the process of invasion continues into the vocalis muscle, the actual lateral excursion of the vocal cord is limited and eventually lost. The ability of the clinician to view and interpret this scenario is critical to the sophisticated management of larynx cancer in general and vocal cord cancer in particular. Essentially, lesions of the true vocal cord that do not transgress the basement membrane (i.e., surface lesions) do not cause tethering of the membrane, and those that enter the underlying lamina propria do. [ref: 33] Benign lesions and even carcinoma in situ, therefore, may look extensive topographically, but their lack of depth is revealed by appropriate diagnostic technology. Even though contemporary methods of staging tend to emphasize the bulk and topographic size of tumors rather than depth, recent thinking has begun to focus more on this third dimension. As data are accumulated, more emphasis will be placed on this important matter. Imaging should not be relied upon to detect early larynx cancer, because the routine methods of physical examination are far more suitable. The primary care physician should not, therefore, initially resort to CT or MRI when a laryngeal cancer symptom persists. Instead, that patient should be examined by someone skilled in the appropriate techniques. Once a lesion is discovered, the evaluation of its depth, bulk, and cartilage invasion and the status of the regional lymph nodes are enhanced by CT or MRI. It is not clear which of the two imaging methods is better for larynx cancer evaluation. Both methods have certain advantages over the other, and both are of limited usefulness in evaluating the radiated larynx. CT of the larynx is most effectively achieved in the axial plane, and because of this, the images show well the lateral tumor extension and the relation of that extension to cervical nodal disease. The axial projection is effective in demonstrating the important paraglottic space. CT effectively demonstrates the vertical extension of tumor, especially in the subglottic and anterior commissure areas. MRI offers the advantages of multiplane visualization of the larynx, and therefore, is especially valuable in evaluating the preepiglottic space and the adjacent base of the tongue. Invasion of laryngeal cartilage is important in treatment planning. Determining whether this feature exists has always been difficult because of the inconsistency of the ossification that occurs in the laryngeal framework. Generally, cartilage is vulnerable to tumor invasion in those areas where it is ossified and somewhat resistant where it is not. In fact, healthy, nonossified cartilage provides a considerable natural barrier to cancer invasion. Writings by Castelijns and associates [ref: 123,124] and by Towler and Young [ref: 125] have suggested that MRI is the method of choice for delineating the important finding of cartilage invasion. Other investigators would dispute this. One study correlated MRI findings of cartilage invasion with the effectiveness of radiation treatment, and in so doing, the authors found a surprising number of small glottic lesions with foci of cartilage invasion. [ref: 126] Significantly, it was from this group that most of the radiation failures of the series occurred. Other imaging methods such as tomography and laryngography have been surpassed by more elaborate technology and are only of historic interest.
STAGING The AJCC last updated larynx staging in 1983, and that version is presented in Table 29.3-3. [ref: 31] Staging provides a commonality of language that is essential for effective outcomes analysis. The larynx is a complex structure because it involves many anatomic and physiologic factors that impact on performance and, therefore, on staging. Although it is essential that pathologic findings always be compared with preoperative analysis, it should be remembered that the staging referred to and that reported by the AJCC is a clinical one, which is based on performance. The accuracy of clinical staging is periodically updated on the basis of better recognition of performance. For example, Pillsbury and Kirchner studied this question by comparing whole-organ sections of nonradiated larynges and compared the actual pathologic findings with the preoperative staging. [ref: 127] They found that 40% had been categorized incorrectly, most inaccuracies being attributable to understaging. Most commonly, the depth of invasion had been underestimated, and the frequency of cartilage invasion was much higher than had been realized previously. Certainly, as imaging technology improves, so will the ability to stage more accurately. As clinicians employ treatment protocols that use neoadjuvant chemotherapy and in which the assessment of complete versus partial response is required, modern imaging hopefully will enhance the accuracy of that assessment. Survival in larynx cancer decreases in a linear fashion with increasing stage. The most remarkable change is between stages II and III, the zone that generally represents the occurrence of cervical metastasis.
TREATMENT AND SURVIVAL Supraglottis Because the supraglottic larynx is composed of multiple sites, referring to it as one unit is not always accurate when discussing treatment results. Because all of the subsites are intimately related and because the supraglottis is continuous with its neighboring hypopharynx, glottic larynx, and oropharynx, it can be difficult to determine the exact site of origin for many larger cancers. For example, when one encounters a lesion that involves the pharyngoepiglottic fold, aryepiglottic fold, and pyriform sinus, it can be difficult to know whether this is a primary hypopharynx cancer extending superiorly or a primary supraglottic cancer extending inferiorly. Unlike glottic cancer, in which cervical metastasis is relatively uncommon for early-stage disease, the probability of nodal spread in all supraglottic lesions is substantial. [ref: 35,50] There is a significant probability of contralateral metastasis in these lesions, and that increases with increasing primary size.
This is especially true for epiglottic lesions, which make up most of the supraglottic carcinomas. It is essential to recognize that the site of treatment failure in supraglottic cancer is usually the neck; therefore, treatment strategies require neck management for virtually all lesions. For the N0 neck, this implies selective neck dissection or elective radiation, and for patients with clinically positive neck or necks, this implies neck dissection, therapeutic radiation, or a combination of both. Early-stage primary disease is highly curable by partial laryngeal surgery or by radiation therapy. More advanced lesions, on the other hand, usually are treated with combined modalities and often require total laryngectomy. Wang and associates have reported an extensive experience with radiating early lesions dating back to 1973. [ref: 128] They have reported 5-year actuarial disease-free survival of 73% and 50% for T1 and T2 lesions, respectively. When surgical salvage was added, this survival increased to 80% and 58%, respectively. A total of 92% of T1 patients and 86% of T2 patients survived with their larynx. Mendenhall and colleagues reported local control of 100% of T1 and 81% of T2 lesions. [ref: 129] When surgical salvage was added, local control for T2 cancers increased to 88%. Importantly, this study revealed no significant differences in local control by anatomic subsite. These data were recently fortified when Mendenhall and his group reported on 209 patients with a minimum 2-year follow-up after radiation therapy, with or without neck dissection. Local control was 100% for T1 lesions and 83% for T2 lesions. In the T2 group, local control was 80% for once daily radiation versus 90% for twice daily. [ref: 130] Olofsson and coworkers reported a 47% 3-year survival for T1 and T2 lesions. [ref: 131] Lederman reported a 53% control rate for T1 and 64% for T2 tumors at 5 years. [ref: 132] Fletcher and associates reported local control in 94% of T1 and 89% of T2 lesions, with no difference noted by anatomic subsite within the supraglottis. [ref: 133] Vermund compared the 5-year survival for primary surgery with primary radiation for T1, N0 to T4, N0 supraglottic carcinoma. [ref: 134] The results for T1, N0 to T3, N0 disease were identical for both groups, but there was an advantage (56% versus 14%) for T4, N0 treated by primary surgery. Those smaller lesions of the supraglottis (i.e., T1 and T2) are equally well treated with surgical procedures that remove only the upper portion of the larynx. The so-called supraglottic laryngectomy is physiologic and allows retention of vocal and swallowing functions. Because of the unique lymphatic drainage patterns of the organ and the presence of certaiatural anatomic barriers to tumor spread, this operation is oncologically sound, yielding the same local control rates as achieved by total laryngectomy in comparable lesions. [ref: 27,51,135] Some authors described a 68% 5-year control rate with supraglottic laryngectomy [ref: 74]; Ogura and Biller reported an 85% 3-year control for epiglottis cancers treated with supraglottic laryngectomy, but this decreased to 71% with extension onto the false cords. [ref: 135] Because of the complications associated with persistent swelling, swallowing difficulty, and wound healing, supraglottic laryngectomy usually is not recommended in patients who have had full-course radiation therapy. The decision to radiate primarily a supraglottic laryngeal lesion should be made with the realization that the backup operation for failure of therapy is almost always a total laryngectomy. Management of the neck is critical to successful therapy of supraglottic cancer. Levendag and coworkers studied elective surgical management of the neck in a group of patients with stages I and II supraglottic carcinomas treated with surgery alone at the
Various studies report dismal survival statistics with radiation therapy management alone; Siirala and Pavolainen reported a 20% 5-year survival [ref: 138]; Vermund reported a 32% 5-year survival [ref: 134]; and Wang and colleagues reported a 5-year actuarial survival of 37% for T3 and 23% for T4 lesions. [ref: 139] Mendenhall and colleagues reported local control in 61% and 33% of radiated T3 and T4 lesions, respectively. [ref: 129] These numbers increase to 83% and 66% with surgical salvage. There was no difference in local control in patients suitable for a supraglottic laryngectomy compared with those who were not suitable for this procedure. Supraglottis (Continued) (1 of 1) There are a variety of factors that can help select a patient for primary radiation therapy, even if the patient has an advanced-stage lesion. For example, all T3 lesions are not alike. Freeman and coworkers [ref: 140] evaluated the local control after definitive radiation therapy as a function of the extent of preepiglottic space invasion and a CT-derived measurement of the primary tumor volume. For lesions involving less than 25% of the preepiglottic space, local control was 75% versus 60% when more extensive involvement exists. In addition, for T3 lesions whose volume is less than 6 cm**2, local control was 82% versus 55% for lesions at least 6cm**2. This trend is sustained in a more recent report. [ref: 141] This information suggests that additional factors other than stage should be considered before excluding definitive radiation therapy as a treatment option in T3 lesions. It also suggests that tumor bulk, rather than T-stage designation, is a better predictor of local control. It should also be noted that T3 designation by virtue of preepiglottic space invasion is different from the lesions that achieve T3 status because of vocal cord invasion and immobilization. This latter circumstance is, in fact, often associated with unrecognized cartilage invasion, and as such, these are really T4 cancers, and the performance is probably different from that of the preepiglottic space group. [ref: 62] An important advance has been the development of chemotherapy-radiation therapy programs for larynx preservation for patients who would otherwise require total laryngectomy. The concepts are discussed in some detail later in this chapter, and apply to supraglottic cancers as well. If total laryngectomy is required, postoperative radiation is usually given to the primary site and necks. For those with subglottic extension, the stoma is included in the field. It is not the mission of this text to describe elaborately the various surgical techniques that are used to manage supraglottic laryngeal cancer; however, the student of this disease and its management should have at least a summary knowledge of the precise methods known collectively as conservation laryngeal surgery]. The compartmentalization of the larynx and the directional drainage patterns of the lymph channels within it provide the surgeon with a unique setting for removing that portion of the larynx above the true vocal cords and, with proper reconstruction, to allow retention of the swallowing and vocal functions of the patient. Because of the consistent drainage patterns of the area cancers, the so-called supraglottic laryngectomy is associated with the same cure rates as would be produced by a total laryngectomy in the same lesion. Essentially, this procedure is a horizontially directed hemilaryngectomy in which the surgeon removes the upper half of the thyroid cartilage and the contents within it, namely the false vocal cords, the epiglottis, and the aryepiglottic folds. The edge of the thyroid cartilage is brought up to and approximated to the transected base of the tongue. Because the motor nerve supply of the vocal cords comes from below (recurrent laryngeal nerves), the important glottic function of abduction and adduction are retained; because of this, voice and the important functions of glottic closure are retained. The supraglottic laryngectomy is, however, physiologically challenging, and patients with chronic pulmonary disease do not tolerate the aspiration that can follow. Essentially, this elegant technique is oncologically sound in appropriate tumors, but certain patients are not good candidates for its implementation. The correct use of the supraglottic laryngectomy is accomplished only by surgeons properly trained in this complex methodology and who have the experience to apply the right methods in the right situation. A succinct discussion of the method of selection for conservation procedures and which patients are suitable for them was developed by Sessions and Parish. [ref: 142] Although there have been certain chemoradiotherapy options popularized since that discussion was published, the fundamental principles are the same and can be applied to the current philosophy. As those alternative schema such as the larynx preservation protocols are implemented more frequently, fewer supraglottic laryngectomies will be done, and the number of surgeons accomplished in these methods will diminish. While the idea of saving more larynges without compromising cure is worthwhile, it is worrisome that those subtle skills necessary for achieving excellent functional results with the family of partial laryngectomies is to some extent being lost by the current generation of head and neck surgeons. There are a variety of conservation surgery procedures that are applied to variations of laryngeal cancer, and while a description of each is beyond the scope of this text, a classification of these operations should be helpful to the reader in placing this important surgical methodology into the proper perspective. Such a classification is the following: 1. Hemilaryngectomy a. Horizontal hemilaryngectomy (supraglottic) b. Vertical hemilaryngectomy i. Lateral hemilaryngectomy ii. Frontal hemilaryngectomy 2. Cordectomy 3. Partial laryngopharyngectomy Elaborate details of radiation techniques for the treatment of supraglottic carcinoma are not intended in this writing, but certain principles and concepts are essential and will be described. Patients are generally treated with opposed lateral fields for the primary site and upper neck, and a separate anterior field for the lower neck (including the stoma in appropriate patients). The spinal cord is restricted to 45 Gy. Electrons are used to boost the posterior neck, when appropriate, after the spinal cord has been protected. For T1 and T2 lesions, a total dose of 65 to 68 Gy is used, while more advanced lesions receive at least 70 Gy. Patients receive 1.8 to 2.0 Gy per fraction, continuous course treatment. Generally, either Co-60 or a low energy linear accelerator is used. Care should be taken to ensure dose homogeneity throughout the larynx, usually via a larynx compensator. A variety of fractionation programs have been used. The most traditional is a once daily, continuous course program. Several twice daily (BID) programs have been used, but there is no firm data to suggest their superiority to the traditional, once daily programs. Wang [ref: 128] reported a regimen of 1.6 Gy BID to approximately 38.4 Gy, a 2-week break, and continuation to approximately 64 to 67 Gy. Local control was 66% for T3-4 lesions managed with BID radiation versus 33% for patients treated once daily. These data are retrospective and must be confirmed in a randomized trial before this approach is considered superior. Mendenhall [ref: 129] attempted to compare BID to once daily radiation. Their program involved 1.2 Gy BID, continuous course, to doses generally in the 74.4 to 76.8-Gy range. There is clearly no benefit to BID treatment for early-stage disease. There are too few patients with advanced disease to make definitive judgements. Glottis Carcinoma in situ (CIS) of the true vocal cord is highly curable. It can be cured with equal efficiency by microexcision, laser vaporization, [ref: 143] or radiation therapy. [ref: 144,145] Although pure carcinoma in situ lesions are unusual, there is a frequent association between carcinoma in situ and invasive carcinoma. Those series that have been reported in which there were numerous recurrences after stripping of vocal cord carcinoma in situ lesions almost certainly consisted of a heterogenous group that included lesions containing areas of invasive cancer. True carcinoma in situ, by definition, remains superficial to the basement membrane, and if mucosal excision techniques are confined to that group of patients, the cure rate should be the same as the best that can be achieved by radiation therapy. The advantages of radiation therapy are that the voice is probably better than with surgery, it is a more definitive treatment for invasive cancer that may exist within the lesion, there is no requirement for general anesthesia, and when treatment fails, a conservation surgical procedure can salvage most patients. The advantage of surgery, whether it be microexcision or laser vaporization, is that it is simpler, and radiation therapy is held in reserve for future use. As long as the integrity of the submucosa is not violated, the voice results with microexcision or laser vaporization are probably comparable with that achieved with radiation. All things considered, if the diagnosis of pure carcinoma in situ of the vocal cord or vocal cords is fairly certain, microexcision is probably the treatment of choice. So called “vocal cord stripping” often has a crippling effect on voice and is to be discouraged. The more precise microexcision, on the other hand, is followed by minimal impact on vocal function. Additionally, microincision has the advantage over both laser vaporization and radiation therapy of providing the pathologist with a specimen for complete histologic examination, perhaps revealing small areas of microinvasion that would have otherwise been unappreciated. Radiation therapy is used either as the definitive treatment of CIS when microexcision is not possible or is refused, or as a salvage treatment for patients with recurrent CIS following prior surgical procedures. [ref: 146,147] This is not an issue, however, on which there is complete agreement. There are those who feel that endoscopic cordectomy is the superior method of addressing this problem, both from a curative and from an economic standpoint. [ref: 148] On one hand, function after radiation for CIS of the vocal cord is better than in cordectomy, but considering the value of saving the radiation option for potential need later, there is a case to be made for the surgical option. Importantly, if a surgical option is chosen, microexcision of these lesions must be associated with a thorough pathologic evaluation of the specimen to minimize the possibility of microinvasive areas going undetected. Cordectomy, on the other hand, although associated with a substantial compromise of vocal function, tends to avoid this pitfall. We tend to rely heavily on the preoperative evaluation by videostroboscopy [ref: 33] to estimate the depth of the surface cancer. If we suspect the lesion is truly confined superficial to the basement membrane, we do microexcision and examine the specimen carefully for microinvasion. If there is evidence of microinvasion, the patient usually receives radiation. On the other hand, if the stroboscopic analysis suggests restricted mucus membrane motion and possible invasion, we generally begin with radiation. We do not favor cordectomy because of the poor vocalization results associated with this procedure. Many of these so-called carcinoma in situ lesions have probably been understaged and are riddled with areas of superficial invasive carcinoma. Also, certain carcinoma in situ lesions, such as those on the anterior commissure, in the subglottis, or some that extend into the laryngeal ventricle, do not lend themselves to these surgical methods; in these, radiation is probably a better means of treatment. For early glottic cancer (T1 or T2), excellent local control is achieved by radiation or partial laryngectomy. The relative advantage of radiation over surgery relates to function. With radiation, the voice is unquestionably better. After this treatment, the voice is normal or near normal most of the time. Harrison and coworkers have analyzed the impact on vocal quality by radiation therapy for T1 and T2 glottic lesions, and their conclusion corroborated this tenet. [ref: 149] In contrast, all patients are hoarse to a varying degree after hemilaryngectomy or cordectomy. Hemilaryngectomy can be used successfully for salvage in many patients who fail radiation therapy. [ref: 150-152] In a carefully selected subset of those patients in whom the initial treatment is radiation therapy, there is a second line of defense against glottic cancer that does not involve sacrifice of the larynx. The actual survival results for primary surgery for T1 glottic cancer are 84% to 98% with laryngofissure and cordectomy. [ref: 152-155] Various series have reported similar results with hemilaryngectomy, which is considered to be a better operation, oncologically and functionally, than a cordectomy. Essentially, the 5-year survival rates for primary surgery and primary radiation for T1 lesions are comparable. [ref: 156-160] Local control obtained in this same stage glottic group using conservation surgical procedures is reported to be 78% by Kirchner and Owen [ref: 67] and 87% by Ogura and associates. [ref: 161] Results obtained for comparable (T1) lesions treated by radiation therapy show local control of 91% by Harwood [ref: 144] and 93% by Pellitteri and colleagues. [ref: 160] As is true with any treatment, the local control rate decreases with increasing tumor bulk. This finding is especially true for radiation therapy of early glottic cancer. Dickens and coworkers have reported the results for early glottic tumors of various sizes and extent. [ref: 162] The lesions were categorized by the type of surgical procedure that would have beeecessary had surgery been used. This type of analysis provides an excellent basis for comparing the results of surgery and radiation. The local control with radiation alone in patients suitable for cordectomy was 97%, and in patients who needed hemilaryngectomy it was 94%. In both categories, local control increased to 100% when surgical salvage was added. On the contrary, for patients managed by radiation therapy who would have required total laryngectomy, local control was only 65%, but it increased to 91% when surgical salvage was added. Extension of these glottic carcinomas onto the anterior and posterior aspects of the larynx lessens the local control rates achieved with radiation therapy. [ref: 163] The surgical procedure required for salvage usually was a total laryngectomy rather than a hemilaryngectomy. Specifically, with anterior commissure involvement, Sessions and associates reported a 5-year survival of 74% for T1 and T2 lesions [ref: 164] and usually found that survival and recurrence rates of anterior commissure lesions correlated with the size and stage of the tumor. [ref: 165] Results with radiation therapy for early-stage glottic carcinomas that involve the anterior commissure are reported by Olofsson and colleagues [ref: 166] to be 80% survival at 5 years (including those recurrences salvaged by surgery) and Kirchner and Fischer, [ref: 80] who reported a local control rate of 85%. Those anterior commissure lesions that are thin and of low volume and that do not have substantial subglottic extension probably are treated with equal efficiency by partial laryngectomy or radiation therapy. As lesions become more advanced, the natural barrier of the anterior commissure ligament is overcome and the thyroid cartilage is invaded [ref: 167]; therefore, radiation therapy becomes less appealing than surgery as the front-line treatment. Most tumors involving the anterior commissure occur as a result of spread from the true vocal cord. Lesions actually arising in the anterior commissure are unusual, making up 1% to 2% of glottic cancers. [ref: 165,167] Glottis (Continued) (1 of 4) In most centers, most T1 glottic cancers are treated with radiation therapy, and partial or total laryngectomy is used as a salvage operation in those patients who fail radiation. The management of T2 lesions is more complicated, because this is a somewhat heterogenous group. Surgical management usually consists of vertical hemilaryngectomy and is associated with 3-year survival rates of 83% [ref: 164] and 82% [ref: 161] in two major series. Primary radiation therapy with surgical salvage yields a net 5-year survival rate of 92% in the series of Pellitteri and associates, [ref: 160] 72% in Wang’s series, [ref: 157] 90% in Fletcher and colleagues’ series, [ref: 159] and 72% in Jorgensen’s series. [ref: 158] Because of the heterogeneity of the T2 group, Wang has suggested subdividing these lesions into those with normal mobility (T2A) and those with impaired mobility (T2B). [ref: 157] He showed that local control was obtained in 86% of the former and 63% of the latter when primary radiation was used. Similar observations were made by Harwood, [ref: 144] whose series yielded 77% and 51% local control rates for T2A and T2B lesions, respectively. Essentially, T2B lesions seem to behave more like T3 than T2 lesions. Radiation therapy usually is recommended as the primary treatment modality for T2 lesions with no vocal cord mobility impairment (T2A); however, in those lesions that demonstrate impairment of motion (T2B), hemilaryngectomy usually is preferred. These criteria are variable, however, depending on tumor bulk, and in those T2B lesions with less bulky lesions, radiation can be employed. It should be noted also that when vocal cord motion is restricted by actual surface tumor bulk, rather than by invasion of the underlying muscle, radiation is often effective. The overall 5-year survival rate for T1 glottic carcinomas is 82% to 96%, and for T2 lesions it is 51% to 85%. [ref: 168,170] Those conservation surgical procedures that can be applied to the management of glottic cancer are time honored and tested, and when used in the properly selected case of laryngeal cancer, they consistently yield excellent results functionally and oncologically. The most commonly employed of these procedures is the vertical hemilaryngectomy (see the classification of conservation laryngeal procedures outlined earlier) in which the surgeon bisects the larynx and, to a varying degree, removes a portion or all of the true and false vocal cord along with the respective half of the thyroid cartilage. Because most of the lesions for which this operation is employed are located on the anterior two thirds of the vocal cord, the most posterior resection line usually is in front of the arytenoid cartilage. In those circumstances in which the cancer extends onto the posterior larynx, this cartilage can be resected (Fig. 29.3-7). By using the perichondrium from the external surface of the half of the thyroid cartilage that has been removed, the operated side of the larynx heals in the midline, forming a firm buttress (pseudocord) against which the opposite and normal true vocal cord vibrates. Although a variety of radiation therapy techniques for early-stage lesions have been reported, the most commonly used technique is opposed lateral fields. The patient is simulated in the supine position with maximal head extension. A head cast is made to ensure immobilization, and small portals are designed that treat the larynx alone. Field size is an important prognostic factor. The local recurrence rate is substantially higher for those patients whose field size is 25 cm**2 or larger. Actually, it would seem that field size is best at 5.5 x
LARYNGEAL REHABILITATION The philosophy that emphasizes a maximal rehabilitative effort to return the patient to a functional role in society is exemplified vividly in the efforts currently being expended on vocal restoration after total laryngectomy. When laryngectomy is done, the normal method of vocal communication is lost. Since the first total laryngectomy by Billroth, [ref: 204] an ideal method of artificially providing voice has been sought. [ref: 205-214] Until recently, most laryngectomy patients have had to rely on mechanical vibratory devices, esophageal speech, or written communication. Only about 20% to 40% of patients master esophageal speech. [ref: 215] It is those postlaryngectomy patients with ineffective or no esophageal speech who are least likely to reenter their social and family environments as full participants.
Even laryngectomy patients who have mastered esophageal speech are hampered by certain mechanical problems and stigmas that create difficulty with their ability to communicate well. Even when mastered, esophageal speech produces only short sentences or monotonal words of limited volume, and its production requires swallowing techniques that are often accompanied by facial grimacing, distortions, and embarrassment. The electrovibratory devices are helpful for intermittent use during the immediate postoperative period and as a back-up system, but they are of limited value because one hand is occupied during their use (Fig. 29.3-9). With these instruments, the voice is mechanical and often distracting, and it is difficult for the user to be heard in a noisy environment. The state of the art of vocal rehabilitation involves the employment of air shunting from the trachea to the neogullet, thereby setting up a vibratory column that produces noise that comes out of the mouth where it is articulated as speech. This is accomplished by a tracheoesophageal puncture (TEP) and placement of a silicone valve-like device that is structured to allow air into the neogullet but not to allow food or liquids out (Fig. 29.3-10). What is produced is a form of esophageal speech and voice, but one that is vastly superior to the traditional method of swallowing and burping. The sounds produced are less monotonal, the volume of air is limited only by the inspiratory lung volume capacity (just as normal speech is) rather than by how much air is swallowed and burped, and the product has a more fundamental resemblance to normal voice than other methods. After the establishment of an appropriate fistula and the placement of the TEP prosthesis, a tracheostoma valve and external housing device is used (see Fig. 29.3-10). This device allows normal breathing but shunts the air into the underlying TEP prosthesis. With the increased airflow of voice production, air is routed into the esophagus and up the neogullet that vibrates and causes sound that comes out of the mouth as articulated speech. When an external valve is placed over the stoma with the TEP in place, the final product is a laryngectomized person who usually has exceptional speech with normal articulation and who can dress with clothing that allows complete coverage of the device. [ref: 216] Patients can return to work and a relatively normal social life with a functional method of communication. This method effectively achieves voice in about 80% to 90% of patients. [ref: 217] There are other surgical methods of vocal rehabilitation, [ref: 218,219] but none with the extensive record of clinical application and success achieved by the tracheoesophageal phonation as developed and popularized by Singer and Blom. [ref: 216] The TEP procedure can be done with equal results at the time of laryngectomy (primary TEP) or as a secondary procedure after radiation. The matter of alaryngeal rehabilitation should be undertaken in concert between the head and neck surgeon and the speech pathologist. The patient who has been rushed to laryngectomy without adequate preoperative preparation (i.e., education, counseling) is usually more difficult to rehabilitate. On the other hand, the patient who is well prepared usually achieves successful speech rehabilitation. It is our approach to take the patient methodically through a series of experiences after the determination is made that a laryngectomy is necessary. The laryngectomy patient should have multiple communication options; therefore, the authors endeavor to teach them how to use the electrovibratory device (see Fig. 29.3-9), esophageal speech, and TEP phonation (in a selected subset). The speech pathologist is involved at the outset, giving an in-depth description of the anatomy, the operation, and the proposed plan for rehabilitation. The team relies heavily on video demonstrations of the whole subject and on patient volunteers who have had similar procedures. At this point, an attempt is made to incorporate family members in the instructional sessions, because a constructive support system contributes to a successful outcome. Additionally, the patient is taught how to use the electrovibrating device before surgery so that he or she will be able to communicate during convalescence in the hospital. Contacts are also made with the American Cancer Society and various laryngectomy clubs. During this preoperative educational period, the initial training for esophageal speech is begun. After laryngectomy, the speech pathologist visits the patient in the hospital and encourages the use of the mechanical vibrating devices. The patient’s perception of a continuum of rehabilitative effort is well served by this comprehensive approach in which the various members of the team communicate and function toward preestablished goals. There is an ongoing controversy regarding primary versus secondary TEP. The authors favor the latter; however, others believe the procedure should be done at the time of the laryngectomy. The primary approach shortens the time of rehabilitation and lessens overall expense, and according to the advocates of this approach, the patients have more immediate positive reinforcement in the critical perioperative period. On the other hand, the advocates of the secondary placement of the TEP argue that by delaying the procedure, more time is allowed for the patient to adapt to and learn to care for the new laryngectomy stoma, the wound is allowed to mature, postoperative radiation therapy is completed, and the patient is allowed to begin learning skills of esophageal speech. The authors believe that the motivation to learn esophageal speech is lessened when a TEP is in place. If the patient is properly prepared psychologically and educationally, this orderly step-by-step approach to rehabilitation is more efficient. Finally, healing is a dynamic phenomenon, and the precise placement and angle of the prosthesis are better ensured by a secondary or delayed placement. There are circumstances that make patient selection critical to the success of rehabilitation with TEP. To function at a practical level of phonation, the patient must have the cognitive skills and the dexterity necessary to care for the stoma and the TEP site. Patients who have had laryngopharyngectomies with reconstruction of the gullet have a somewhat unpredictable tracheoesophageal voice rehabilitation. Those patients who undergo free jejunal interposition procedures have a somewhat better and more predictable voice than those who have had a gastric transposition for reconstruction. The experience in voice rehabilitation in both of these groups is limited, and the vocal results achieved to date must be viewed as preliminary. Whether employed primarily or secondarily, the enormous contribution made to the laryngectomy patient by the current methods of TEP phonation should not be underrated. There is a considerable experience currently with this method, and as the plastics technology continues to improve, the process of this type of vocal rehabilitation should be enhanced. It is hoped that the reader appreciates that the current concern among oncologists for quality of life is reflected by the considerable effort being devoted to saving the larynx with conservation operations, innovative radiation strategies, larynx-sparing chemoradiotherapy protocols, and, when these methods fail and laryngectomy is required, a vigorous and thoughtful effort to communication rehabilitation.
Prepared by Prof. Igor Galaychuk, MD,
2011
