Clinical and.

pathologic distinction between primary and metastatic mucosal melanoma of the head and neck
KATHLEENR. BILLINGS,MD, MARILENEB. WANG, MD, JOELA. SERCARZ,MD, an(] YAO S. FU, MD,

Los Angeles, California

Metastatic mucosal melanoma is extremely rare. Only 0.6% to 9.3% of patients with cutaneous malignant melanoma will have metastases to the mucosa of the upper aerodigestive tract. The records of all patients with mucosal melanoma of the head and neck at the University of California, Los Angeles Medical Center during the past 30 years were reviewed. Patients with primary tumors were separated from those with metastatic involvement from a cutaneous primary site. These two groups were compared for differences in clinical symptoms, histopathologic findings, treatment, and survival characteristics. Frequent sites of metastatic involvement included the base of tongue and nasal cavity. These arose from a variety of cutaneous sites including the trunk and extremities and, in most instances, did not arise until 2 to 7 years after the initial cutaneous lesion. Most of those with metastases to the head and neck mucosa had disseminated disease. The histopathologic distinction between the two groups is described with photomicrographs. Junctional activity in the overlying or adjacent mucosa distinguishes primary mucosal melanoma from metastatic disease, in which the overlying mucosa is usually intact. This difference is useful in determining workup and treatment options. Aggressive surgical resection is suggested in treatment of primary melanomas, whereas surgery is at best palliative in those with metastatic disease. (OTOLARYNGOL HEAD NECK SURG 1995;112:700-6,]

N o tumor disseminates more widely or involves more organs than malignant melanoma. 1 The most frequent sites of metastases from cutaneous melan o m a are the regional lymph nodes, lung, liver, and brain. 2 Metastasis to the mucosa of the u p p e r aerodigestive tract is much less frequent and is often not appreciated clinically. Autopsy studies 1-3 have revealed an incidence of mucosal metastases to the head and neck in 2% to 9.3% of those with cutaneous malignant melanoma. This report evaluates 11 patients identified with metastatic mucosal melan o m a to the upper aerodigestive tract. Special atFrom the Division of Head and Neck Surgery (Drs. Billings, Wang, and Sercarz) and the Department of Pathology(Dr. Fu), University of California, Los Angeles School of Medicine. Presented at the Annual Meeting of the American Academy of Otolaryngology-Head and Neck Surgery, San Diego, Calif., Sept. 18-21, 1994. Received for publication Sept. 20, 1994; accepted Dec. 30, 1994. Reprint requests: Marilene B. Wang, MD, Division of Head and Neck Surgery, UCLA Medical Center, CHS 62-132, 10833 Le Conte Ave., Los Angeles, CA 90024. Copyright 1995 by the American Academy of OtolaryngologyHead and Neck Surgery Foundation, Inc. 0194-5998/95/$3.00 + 0 23/1/63069
700

tention is paid to the clinical course, effect on outcome, and pathologic distinction between primary and metastatic mucosal disease. Primary mucosal malignant m e l a n o m a was the subject of a recent study conducted at the University of California, Los Angeles Medical Center, from which comparisons are m a d e in this series. 4
METHODS AND PATIENTS

A computer search of the pathology records at the U C L A Medical Center from 1963 to 1993 was conducted to identify those patients with mucosal melan o m a of the head and neck. Patients with cutaneous or ocular m e l a n o m a were excluded. An extensive chart review was done to separate primary from metastatic mucosal disease. A total of 11 patients had metastatic lesions. During this time a search by the U C L A T u m o r Registry revealed that 5009 patients with malignant m e l a n o m a were seen at this institution, for an incidence of metastases to the upper aerodigestive tract of 0.2%; autopsy studies were not included. Ten patients were followed up at U C L A for an average of 6.2 years (range, 4 months to 12.6 years). One patient not followed up

Otolaryngology Head and Neck Surgery

Volume 112 Number 6

BILLINGS et al.

701

Table 1. Clinical characterisitics of patients with metastatic mucosal melanoma to the upper aerodigestive tract
Patient no./ Breslow's age (yr)/ Cutaneous Clark's depth sex origin level (mm] Site of mucosal metastasis Years to mucosal metastasis Trealment of mucosal metastasis Years to widespread metastasis

Survival
status

1/31/M 2/38/F 3/47/M 4/67/F 5/46/F 6/77/M 7/46/F 8/30/M 9/64/F 10/30/M 11/62/M

Trunk Extremity Extremity Trunk Trunk Scalp Trunk Trunk Extremity Trunk Neck

II IV Ill III III IV III IV NA NA NA

0.9 3.0 0.77 1.85 1.0 1.6 3.0 NA NA NA NA

Nasopharynx Floor of mouth Posterior pharynx Base of tongue Lateral tongue Base of tongue Nasopharynx Gingiva Base of tongue Nasal cavity Nasal cavity

5.0 7.7 0 6.6 5.0 2.3 3.4 8.3 7.3 3.4 6.0

Chemotherapy, BCG tumor cell vaccine Excision NA Excision, chemotherapy Excision, tumor cell vaccine Excision Excision Chemotherapy, radiation therapy Excision Immunotherapy, radiation therapy Interferon

7.3 4.3 0 3.1 NA 1.0 NA 7.3 4.0 3.8 5.3

DOD, 12.6 yr DOD, 8.9 yr DOD, 4.0 mo DNED, 11.1 yr NA DOD, 2.3 yr DOD, 3.4 yr DOD, 12.6 yr DOD, 7.6 yr DOD, 4.0 yr DOD, 6.6 yr

DOD, Dead of disease;DNED, dead with no evidenceof disease; NA, data not available;BCG, bacillus Calmette-Gu&in.

at UCLA Medical Center was included in evaluation of the time of onset of metastatic disease. The records of each patient were evaluated for age, sex, location of Primary, location of head and neck mucosal metastasis, clinical course, treatment, and outcome. Slides, whenever available, were reviewed by a senior pathologist (Y. S. F.) to document histopathologic features useful in distinguishing metastatic from primary disease. Special note was made of junctional activity, epidermal migration, pigmentation, growth pattern, and degree of differentiation. Chi-squared analysis was performed to determine whether the pathologic differences between metastatic and primary lesions were significant. Not all slides were available for review because of significant earthquake damage to the storage facility on Jan. 17, 1994.
RESULTS

Of the 11 patients with metastatic disease to the upper aerodigestive tract mucosa, 5 were men and 6 were women (Table 1). Average age at diagnosis of cutaneous primary disease was 48.4 years (range, 30 to 77 years). Three patients had metastatic disease to the base of tongue; 3 to oral cavity sites including the floor of mouth, lateral tongue, and gingiva; 2 to the nasopharynx; 2 to the nasal cavity; and 1 to the posterior pharynx. These arose from a variety of cutaneous sites, including the trunk in 6 of 11 (55%), extremities in 3 of 11 (27%), and head and neck in 2 of 11 (18%). Of these cutaneous lesions, information on

Clark's level and Breslow's depth were documented in 8 of 11 patients. Seven of 11 (64%) patients in whom metastatic disease subsequently developed had Clark's level III to IV lesions. Common symptoms at diagnosis included bleeding (4 of 11), mass lesion (3 of 11), pain (2 of 11), nasal obstruction (2 of 11), and dysphagia (2 of 11). Only one patient had mucosal metastatic disease initially, and on average, the remaining patients did not have mucosal metastases until 5.5 years (range, 2.3 to 8.3 years) after the cutaneous disease was diagnosed, as shown in Table 1. Some patients had had metastases to other areas before those involving the head and neck mucosa, including the regional lymph nodes (4), lung (2), skin (2), liver (1), and gastrointestinal tract (1). Seven of 11 (64%) had documented evidence of disseminated metastases within 4 years (range, 0 to 7.3 years) of initial diagnosis. In two patients the presence of metastatic disease other than that to the upper aerodigestive tract mucosa was unknown. Those patients with base of tongue or nasopharyngeal metastases most often~ had multisystem involvement. All patients underwent wide local excision of their primary cutaneous lesion, and two had concomitant regional lymph node dissection. Therapy for the metastatic disease varied from simple excision to immunotherapy protocols and is displayed in Table 1. Six patients underwent simple excision of their mucosal metastasis as initial treatment. Thre e patients received chemotherapy, three immunotherapy, and one bacillus Calmette-Gu6rin injec-

702

BILLINGS et al.

Otolaryngology Head a n d Neck Surgery June 1995

80 70
~,~ 60 Z r~
5O

I
40 p, Z 30 ~" 20

10

I
I
JUNCTIONAL ACTI'ITY

i
EPIDERMAL MIGRATION PIGMENTATION DIFFUSE SHEETS D[PFUSE NESTS EP[THELIOID SPINDLE

GROWTH PATTERN

CELL TYPE

Fig. t . trasted

Characteristic pathologic features of upper aerodigestive tract mucosal metastases c o n to p r i m a r y m u c o s a l m e l a n o m a lesions.

tions in addition to chemotherapy. Two patients subsequently underwent radiation therapy to the site of metastasis. Overall survival was poor: 9 of 10 (90%) died with disease within 6.2 years of their initial diagnosis (range, 4 months to 12.5 years). One patient (patient 4) who was treated with chemotherapy died free of disease 4.7 years after diagnosis of metastases to the base of tongue and liver. Overall, the 5-year survival after onset of metastases was 20%. Pathology review of all available slides revealed characteristic features useful in distinguishing metastatic from primary mucosal melanomas (Fig. 1). Those slides reviewed for patients with primary mucosal melanoma were selected from patients discussed in the prior UCLA series? These patients ranged in age from 17 to 91 years (average, 56 years), and no sex predilection was demonstrated. Most primary lesions arose within the nasal cavity or paranasal sinuses (80%), and the remainder (20%) in the oral cavity or oropharynx. Evaluation of primary mucosal lesions showed evidence of junctional activity in 7 of 16 (44%) of the cases available for review (Fig. 2). Pigmentation was present in 11 of 16, and epidermal migration in 6 of 16. In contrast, metastatic lesions showed no evidence of junctional activity or epidermal migration, and 4 of 5 lesions were amelanotic. Chi-squared analysis revealed a

marginally significant difference (p < 0.1) between the primary and metastatic lesions for these features. No evidence of vascular invasion was found in any primary case, and 4 of 16 showed perineural invasion. It was of interest to note the presence of pigmented melanoma cells in the minor salivary glands of 4 of 16 cases examined. The growth pattern of these lesions was divided almost equally between sheets (44%) and nests (56%). The round, oval, epithelioid cell type predominated in the primary (56%) and metastatic (80%) lesions (Fig. 3). The remainder of the lesions showed a spindle-cell arrangement. All metastatic cases demonstrated band-like lymphocytic infiltrates, and no evidence of vascular invasion was found. Growth pattern was divided between diffuse sheets (80%) and nests (20%). The differences in growth pattern and cell type were not statistically significant on chi-squared analysis.
DISCUSSION

Upper aerodigestive tract mucosal metastases from cutaneous melanoma are rare and have been infrequently discussed in the literature. 13,59In 1973, Mosby et al. 6 reported the first case of known metastatic spread to the gingival mucosa in a 29-year-old patient with widely disseminated mela-

OtolaryngologyH e a d a n d Neck Surgery Volume 112 Number 6 BILLINGSet al.

703

Fig. 2. Primary mucosal melanoma of the lip demonstrating junctional activity in overlying mucosa (small arrow) as well as epidermal migration (large arrow). (Hematoxylin and eosin stain; medium power.)

Fig. 3. Metastatic lesion of base of tongue showing predominance of round, oval, and epithelioid cells and lack of junctional activity in overlying mucosa. (Hematoxylin and eosin stain; high power.]

n o m a who died of the disease within 2 years of the initial diagnosis. Myer et al. 5 described the ninth known case of metastatic m e l a n o m a to the tonsil in a patient with a Clark's level V lesion on

the trunk. Within 3 years of initial diagnosis, the patient had disseminated disease and subsequently died. These and other similar cases illustrate that, although rare, mucosal metastases to the head and

Otolaryngology Head and Neck Surgery 704

BILLINGS et a l .
June 1995

neck are generally part of lethal widespread metastatic disease. In Henderson et al.'s series ~of 8823 patients with disseminated malignant melanoma, mucosal metastases developed in 54, for an incidence of 0.6%. Sixty-two percent of these patients had evidence of disseminated'disease at the time of diagnosis of the mucosal lesion. Autopsy studies of those with disseminated disease have reported an incidence of 2% to 9.3% with mucosal metastases to the head and neck. 13 The overall median survival has been reported at 3 to 6 months in patients with disseminated disease. Patients with a shorter time to development of metastases have a decreased survival compared with those with longer times. 2'H'12 Isolated metastases in malignant melanoma represent less than 1% of patients with metastatic disease and are associated with a significantly better survival. 2 The patients described in this article are unusual in that only one patient initially had disseminated disease including the mucosal metastasis. The remainder did not have mucosal metastases until 5.5 years, on average, after the initial diagnosis. Sixtyfour percent of the patients did have evidence of disseminated metastases at the time of diagnosis of their mucosal lesion, which concurs with the prior study. ~Although a survival rate in this series averaging 7.0 years seems high, disseminated metastases did not develop until 4.0 years after primary diagnosis. This lends support to increased length of time to metastasis improving overall survival in patients with cutaneous melanoma. Presence of a mucosal metastases did not influence overall prognosis because of the already uniformly fatal outcome associated with disseminated disease. Melanin cells can be found in respiratory epithelium, nasal glands, superficial and deep stroma of the septum, and middle and inferior turbinates) 3 Rapini et al. 14found that 37% of patients in whom primary mucosal melanomas developed had preexisting melanosis. The recent UCLA review of primary mucosal melanoma4 identified 35 patients. Eighty percent of these primary lesions were located in the nasal cavity or paranasal sinuses, and 20% in the oral cavity and oropharynx. Common sites within the oral cavity included the palate and gingival mucosa. Metastatic disease occurs in a variety of sites including the tonsil, tongue, nasopharynx, larynx, and lip. In Henderson et al.'s series 1the palate and gingiva appear to be spared by metastatic disease. The base of tongue and oral cavity were involved most frequently in our series. The gingiva was in-

volved in one case, marking this as another case located in this rare site. The larynx was not involved in any of the cases, although laryngeal metastases have b e e n reported.aS Of 22 cases of metastatic mucosal malignant melanoma to the larynx, the supraglottic larynx was most commonly involved. This has been related to the more abundant vascular supply and to the ease of invasion of the ossified cartilages by metastatic tumors. 8 The initial symptoms of those with primary and metastatic lesions are similar and include nasal obstruction, epistaxis, and a mass in the oral cavity. Diagnosis was made by physical examination or endoscopy. Radical surgical resection of primary mucosal melanomas is recommended, for a 5-year survival of 45% in the UCLA study. 4 In contrast, in the presence of widely disseminated disease, most patients in this study group received only local excision, followed in a few cases by radiation therapy, chemotherapy, and immunotherapy. Despite the deadly nature of widespread metastases arising from cutaneous melanomas, primary mucosal melanomas are almost always more lethal than their cutaneous counterpart, with a 5-year survival from 10% to 38% and a relatively high recurrence rate despite early clinical stage at time of diagnosis.15This may be related to the lack of induration or to patient delay due to failure to recognize the nature and significance of pigmentation in the nasal and oral cavities, x6 Certain factors appear to worsen the prognosis, including ulceration, rapid growth, lymph node metastases, bone erosion, multicentric origin, and absence of pigmentation. Like primary disease, metastatic disease is frequently not noticed and may not be discovered until time of autopsy.2,3 This suggests the need for a more thorough head and neck examination in those with disseminated disease so that palliative therapy can be offered for mucosal lesions. Trodahl and Sprague 12 state that in determining primary vs. metastatic mucosal disease the main consideration is involvement of the overlying epithelium. The changes in the junctional layer are the dominant features in the histogenesis of a melanoma in the mucous membranes, as in the skin. 17In metastatic mucosal disease, there is typically intact overlying mucosa. 11 In this study distinct pathologic features were demonstrated that are useful in differentiating primary and metastatic lesions. All metastatic lesions lacked evidence of junctional activity in the overlying mucosa and showed no epidermal migration. This is in contrast to primary lesions, in which 44%

Otolaryngology H e a d and Neck Surgery Volume t t 2 Number 6

BILLINGS et al.

705

and 38% had junctional activity and epidermal migration, respectively. Four of those that did not demonstrate junctional activity had ulcerated surfaces, making this feature difficult to delineate. Pigmentation was more commonly seen in the primary (69%) than in the metastatic lesions (20%). Trodahl and Sprague 12 found that metastases were pigmented in all but one lesion they reported, but the number of cases has been too small to draw firm conclusions. A mixture of cell types was found in both primary and metastatic mucosal disease, including round, oval, and epithelioid- to spindle-shaped cells. This feature is therefore not as useful in distinguishing the two. A unique feature seen in the primary lesions (25%) was the presence of extension of the melanotic pigment into the minor salivary glands. This was not seen in any of the metastatic lesions. When primary mucosal disease is considered, there appears to be no correlation between size, location, or histologic appearance of the tumor and survival. 18 In fact, Clark's classification does not apply to primary lesions because of the absence of histologic landmarks analogous to the papillary or reticular dermis in the mucosa. An attempt was made to correlate tumor thickness of primary mucosal lesions with 3-year survivaU9 This study found that 4 of 6 patients with lesions smaller than 7 mm were disease free compared with 1 of 7 with lesions larger than 7 mm during this time period. Conley and Hamaker 2 noted that the vertical growth phase supervenes more quickly in mucosa, which may explain the frequent extension of melanosis into the minor salivary glands in this series of primary lesions. Conley and Pack 17 also found that the relative incidence of mucosal malignant melanoma in proportion to benign nevi is higher than on the skin and advised removing all pigmented lesions of the mucosal surfaces for diagnostic and prophylactic purposes. The thickness of the primary cutaneous lesion in those with metastatic disease correlates with recurrence rate and survival. 2~All patients in this series had a Breslow's depth greater than 0.76 mm, when reported. The poor survival rate was not surprising in patients with metastases to the upper aerodigestive tract given that these lesions were rarely the sole metastasis. Palliative excision of these metastatic lesions was the most common treatment and is recommended to provide relief of epistaxis, airway obstruction, dysphagia, or other head and neck symptoms. Immunotherapy, chemotherapy, and ra-

diation therapy have all been used as palliative measures in those with disseminated disease, with little impact on overall survival.

CONCLUSION
Metastatic malignant melanoma to the mucosal surfaces of the upper aerodigestive tract is usually part of a widely disseminated disease process and therefore has no impact on the overall survival and prognosis of these patients. Treatment of these lesions is palliative in most instances, and survival is uniformly poor. Pathologic features, in particular the lack of junctional activity in the overlying mucosa, allow the pathologist to more easily distinguish primary from metastatic disease. This is particularly important in primary disease, in which aggressive surgical resection can effect survival. Overall, the metastatic pattern of malignant melanoma is quite variable, and no organ is immune from metastatic disease. We thank George S. Smith, MD, for assistance in the pathology search and Theodore Bell, PhD, for assistance in statistical analysis.
REFERENCES

1. Einhorn LH, Burgess MA, Vallejos C, et al. Prognostic correlations and response to treatment in advanced metastatic malignant melanoma. Cancer Res 1974;34:1995-2004. 2. Patel JK, Didolkar MS, Pickren JW, Moore RH. Metastatic pattern of malignant melanoma: a study of 216 autopsy cases. Am J Surg 1978;135:807-10. 3. Das Gupta T, Brasfield R. Metastatic melanoma: a clinicopathological study. Cancer 1964;17:1323-8. 4. Lee SE Shimizu KT, Tran LM, Juilliard G, Calcaterra TC. Mucosal melanoma of the head and neck: the impact of local control on survival. Laryngoscope 1994;104:121-6. 5. Myer CM, Wood MD, Donegan JO. Metastatic melanoma to the palatine tonsil. Ear Nose Throat J 1983;62:62-4. 6. Mosby EL, Sugg WE, Hiatt WR. Gingival and pharyngeal metastasis from a malignant melanoma: report of a case. Oral Surg Oral Med Oral Pathol 1973;36:6-10. 7. Elidan J, Sela M, Lijovetsky G, Weshler Z. Metastatic maxillary sinus melanoma treated by local excision and brachtherapy. J Otolaryngol 1989;18:293-6. 8. Ferlito A, Caruso G. Secondary malignant melanoma of the larynx. ORL J Otorhinolaryngol Relat Spec 1984;46:117-33. 9. Freeland AP, van Nostrand AWE Jahn AF, Metastases to the larynx. J Otolaryngol 1979;8:448-56. 10. Henderson LT, Robbins KT, Weitzner S. Upper aerodigestive tract metastases in disseminated malignant melanoma. Arch Otolaryngol Head Neck Surg 1986;112:659-63. 11. Nambisan RN, Alexiou G, Reese PA, Karakousis CP. Early metastatic pattern and survival in malignant melanoma. J Surg Oncol 1987;34:248-52. 12. Trodahl JN, Sprague WG. Benign and malignant melanocytic lesions of the oral mucosa. Cancer 1970;25:812-23. 13. Batsakis JG, Regezi JA, Soloman AR, Rice DH. The pathology of head and neck tumors: mucosal melanomas. XIII. Head Neck Surg 1982;4:404-18.

706

BILLINGSet al.

Otolawngology Head and Neck Surgery June 1995

14. Rapini RP, Golitz LE, Greer RO, Krekorian EA, Poulson T. Primary malignant melanoma of the oral cavity: a review of 177 cases. Cancer 1985;55:1543-51. 15. Blatchford SJ, Koopman CF, Coulthard SW. Mucosal melanoma of the head and neck. Laryngoscope 1986;96:92934. 16. Snow GB, van der Waal I. Mucosal melanomas of the head and neck. Otolaryngol Clin North Am 1986;19:537-47. 17. Conley J, Pack GT. Melanoma of the mucous membranes of the head and neck. Arch Otolaryngol 1974;99:315-9.

18. Wenig BM. Atlas of head and neck pathology. Philadelphia: W.B. Saunders Co., 1993:67-72. 19. Trapp TK, Fu YS, Calcaterra TC. Melanoma of the nasal and paranasal sinus mucosa. Arch Otolaryngol Head Neck Surg 1987;113:1086-9. 20. Conley J, Hamaker RC. Melanoma of the head and neck. Laryngoscope 1976;87:760-4. 21. Breslow A. Thickness, cross-sectional area and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 1970;172:902-8.

N O W ! The F I V E - Y E A R (1986-1990) C U M U L A T I V E I N D E X TO O T O L A R Y N G O L O G Y - H E A D AND NECK S U R G E R Y can be purchased from the Publisher for $44.00. This comprehensive 104-page reference guide is a current presentation of all topics included in the Journal from January 1986 through December 1990 (volumes 94-103) - t h e past 10 volumes. It incorporates complete references to more than 640 original articles, abstracts, case reports, letters and editorials. It features 1668 Subject Headings, under which there are 3247 references. Each subject entry lists the complete article title, author(s), volume, page(s), and year of publication. In addition, it includes 3190 Author Entries, which list contributors, along with their respective titles, authorto-author referral, volume, page, and publication. To purchase, call or write: Mosby-Year Book, Inc., 11830 Westline Industrial Dr., St. Louis, Missouri 63146-3318, or telephone F R E E 1-800-453-4351, Subscription Services (in Missouri, call collect at 314-453-4351, Subscription Services). P R E P A Y M E N T REQUIRED. Make checks payable to Mosby-Year Book, Inc. (all payments must be in U.S. funds drawn on a U.S. bank). Price: $44.00 in the U.S.; $50.50 in Canada; and $48.50 in other countries (price includes mailing charges). AVAILABLE