Mucosal Melanoma: a Clinically and Biologically Unique Disease Entity

Focused 345 Review

Mucosal Melanoma: A Clinically and Biologically Unique Disease Entity

Richard D. Carvajal, MDa; Sharon A. Spencer, MDb; and William Lydiatt, MD,c New York, New York; Birmingham, Alabama; and Omaha, Nebraska

Key Words of these arise from mucosal surfaces.1–3 Despite a com- Mucosal melanoma, melanoma of the head and neck, anorectal mon cell of origin, MM is a clinical and pathologic entity melanoma, vulvovaginal melanoma, KIT distinct from its more common cutaneous counterpart. MM can arise from any mucosal surface of the body, with Abstract 55% arising from the head and neck, 24% from the ano- Mucosal melanoma (MM) is an aggressive and clinically complex 2 malignancy made more challenging by its relative rarity. Because rectal mucosa, and 18% from the vulvovaginal mucosa. of the rarity of MM as a whole, and because of the unique biology Disease arising from the mucosa of the pharynx, larynx, and clinical challenges of MM arising from each anatomic location, urinary tract, cervix, esophagus, gallbladder, or other mu- understanding of this disease and its optimal management remains cosal sites is less common. limited. The impact of various treatment strategies on disease con- This article summarizes the clinical, pathologic, and trol and survival has been difficult to assess because of the small size of most reported series of MM arising from any one particular site, molecular features, and the diagnostic and therapeutic the retrospective nature of most series, and the lack of a uniform considerations for the management of MM, underscoring comprehensive staging system for this disease. This article summa- the similarities and differences from cutaneous melanoma rizes the clinical, pathologic, and molecular features, and the diag- (CM). Furthermore, the distinct clinical features and man- nostic and therapeutic considerations for the management of MM, agement implications unique to melanoma arising from underscoring the similarities and differences from cutaneous mela- the mucosal surfaces of the head and neck, the anorectal noma. Furthermore, the distinct clinical features and management implications unique to melanoma arising from the mucosal surfaces region, and the female genital tract are highlighted. of the head and neck, the anorectal region, and the female genital tract are highlighted. (JNCCN 2012;10:345–356) Epidemiology and Clinical Features of Mucosal Melanoma The epidemiology and clinical features of MM differ ucosal melanoma (MM) is an aggressive and clini- M significantly from those of CM (Table 1). MM devel- cally complex malignancy made more challenging by its ops at a later age than does CM, with a median age at relative rarity. Of the 68,130 cases of melanoma diag- diagnosis of 70 years compared with 55 years for CM.2 nosed in 2010 in the United States, only 0.8% to 1.8% Although CM is slightly more common in men, with a male-to-female ratio of 1.2:1.0, MM is more commonly

From the aDepartment of Medicine, Melanoma/Sarcoma Service, diagnosed in women, primarily because of the occur- Memorial Sloan-Kettering Cancer Center, New York, New York; rence of melanoma arising within the female genital bDepartment of Radiation Oncology, University of Alabama, Birmingham, Alabama; and cDepartment of Otolaryngology-Head tract, with a male-to-female ratio of 1.0:1.8.2 The in- and Neck Surgery, University of Nebraska and Nebraska Methodist Hospital, Omaha, Nebraska. cidence of CM has been increasing at rate greater than Submitted October 17, 2011; accepted for publication December that of any other cancer in the United States, whereas 12, 2011. the incidence of MM has remained stable over time.4 The authors have disclosed that they have no financial interests, arrangements, or affiliations with the manufacturers of any Although CM is associated with exposure to ultraviolet products discussed in this article or their competitors. Correspondence: Richard D. Carvajal, MD, Memorial Sloan- light, the anatomic location of MM precludes this expo- Kettering Cancer Center, 300 East 66th Street, BAIC1071, New York, sure as a predisposing risk factor, and no clear predispos- NY 10065. E-mail: [email protected] ing risk factors have been identified in MM.

Carvajal et al.

Table 1 Clinical and Pathologic Features of Cutaneous and Mucosal Melanomas Cutaneous Melanoma Mucosal Melanoma Proportion of all melanomas 90% < 2% Demographics Median age at diagnosis 55 y 67 y Male:female ratio 60:40 35:65 Race White 94% 85% Black < 1% 7% Epidemiology Incidence over time Rising Stable Risk factors Ultraviolet radiation Unknown Pathology Multifocality < 5% 20% Amelanotic < 10% Up to 40% Clinical outcomes Advanced stage at diagnosis < 30% > 50% 5-Year overall survival rate 81% 25%

CM is uncommon in blacks, Asians, and Hispan- due to later diagnosis because of lack of clinical sus- ics, with an annual age-adjusted incidence of only 1% picion in the setting of relative rarity, more advanced of that observed in whites. Although the absolute in- disease at initial diagnosis, and the rich lymphovascu- cidence of MM is greatest in whites, the proportion of lar supply of the mucosal surfaces. MM in blacks, Asians, and Hispanics is greater than Mucosal Melanoma of the Head and Neck that observed in whites. Overall, MM does not seem MMHN accounts for 6.3% to 8% of all melanomas to have a racial predilection, except possibly melano- arising in this region, occurring, in decreasing order ma arising from the oral cavity, which may be more of frequency, within the nasal cavity, oral cavity, 2 common in black and Japanese populations. nasal sinuses, pharynx, larynx and upper esopha- When diagnosed early, most CMs are cured with gus.2,3,6 Of MMHNs, 70% to 80% arise in the pa- surgical excision; however, patients with locoregion- ranasal sinuses and nasal cavity, representing 4% ally advanced or metastatic disease fare poorly, with a of all sinonasal tumors.2 Although the exact site median overall survival of 6 to 8 months. Outcomes of origin of these tumors is often difficult to deter- for MM are inferior irrespective of stage at diagno- mine because of the presence of locally advanced sis. Although the 5-year overall survival rate for CM disease, 80% of sinonasal melanomas and approxi- is 80%,5 the rate for MM is only 25% (Figure 1A).2 mately 55% of all MMHNs arise from the nasal cav- Prognosis depends on the primary site of disease (Fig- ity, with the remainder arising from the paranasal ure 1B).2 Patients with disease arising from the head sinuses.6 Within the nasal cavity, the turbinates and neck, anorectal mucosa, and vulvovaginal muco- and nasal wall are most commonly involved, fol- sa have 5-year overall survival rates of 31.7%, 19.8%, lowed by the nasal septum. Among the paranasal and 11.4%, respectively, with survival rates for MM of sinuses, the maxillary sinuses are most commonly the head and neck (MMHN) significantly better than involved, followed by the ethmoid, frontal, and those for MM from the other subsites.2 Locoregional sphenoid sinuses. Oral cavity MM constitutes 0.5% nodal metastasis at presentation is more frequent in of all oral malignancies and approximately 40% of MM than in CM. Although 9% of CMs will present all MMHNs, and most commonly affects the hard with locoregional nodal disease, 21% of MMHNs, palate and upper alveolus.2,6 61% of anorectal mucosal melanomas (ARMMs), and Median age at diagnosis ranges from 60 to 69 23% of vulvovaginal mucosal melanomas (VVMMs) years7; however, oral cavity MM presents at a younger present with involved nodes.2 Patients with nodal dis- age compared with sinonasal melanomas, with many ease have poorer outcomes, with a 5-year survival rate affected individuals younger than 40 years.8 The fre- of 16.4% versus 38.7% for those with negative nodes.2 quency of disease in men and women is equivalent in The inferior outcomes observed in MM may be partly most7,9 but not all reports.8,10 As with other sinona-

Mucosal Melanoma

Cutaneous melanoma: Cutaneous melanoma: Mucosal melanoma: Head and neck Anorectal Vulvovaginal all stages (n = 23,696) stages III and IV (n = 2074) (n = 306) (n = 163) (n = 74) (n = 59)

100 90 A B

90 80

80 70

70 60 60 50 50 40 40

Disease-Speci c Survival % 30 Disease-Specic Survival % 30

20 20

10 10

0 0 1 2345 1 2 34 5

Time Since Diagnosis (y) Time Since Diagnosis (y)

Figure 1 (A) Disease-specific survivals for patients diagnosed with all stages of cutaneous melanoma, stage III or IV cutaneous melanoma, and mucosal melanoma. (B) Disease-specific survivals for patients diagnosed with mucosal melanoma arising from the head and neck, anorectal region, and vulvovaginal region. Data from Chang AE, Karnell LH, Menck HR. The National Cancer Data Base report on cutaneous and noncutaneous melanoma: a summary of 84,836 cases from the past decade. The American College of Surgeons Commission on Cancer and the American Cancer Society. Cancer 1998;83:1664–1678. sal malignancies, inhaled and ingested carcinogens, Anorectal Mucosal Melanoma such as tobacco and formaldehyde, are implicated as ARMM represents 0.4% to 1.6% of all melanomas potential risk factors for the development of sinona- and 1% of all malignancies arising in the anorectal sal MM, although strong evidence is lacking.3 Oral region. One-third of ARMMs are located within the melanosis has been implicated as a possible precursor anal canal, with 42% arising from the rectum and lesion for up to one-third of oral melanoma cases.8 25% with an indeterminate site of origin.13 Unlike The clinical course differs significantly for MM MM arising from other sites, the incidence of ARMM of the oral cavity and sinonasal MM. MM of the oral is increasing.13 ARMM occurs at a median age of 60 to cavity more frequently presents with regional nod- 70 years, with a male-to-female ratio of 0.75.3,13,14 This al involvement, with 25% of cases presenting with finding is confounded by the fact the women are more nodal metastases compared with 6% of sinonasal likely to undergo perineal evaluation as part of rou- melanomas.6,7 Although distant metastases are un- tine examinations. HIV infection may be a risk factor. common at presentation, occurring in 4% to 10% Prognosis is poor, with a 5-year overall survival rate of 15 of cases,6,11 most patients ultimately develop distant 10% to 20% and median survival of 25 months. Up disease. The 5-year overall survival rate for MMHN to 60% of cases will present with locoregional nodal ranges from 17% to 35%6,9,12; however, prognosis var- involvement, with 20% of cases having distant dis- 2,14,16 ies by primary disease site, with 5-year overall surviv- ease at presentation. Perineural invasion is a poor 17 al rates of 15% to 30%, 12%, and 0% to 5% for MM prognostic marker in ARMM. Other prognostic fac- arising from the nasal cavity, oral cavity, and parana- tors may include stage at diagnosis, lymph node in- 15,16 sal sinuses, respectively. Features such as early stage, volvement, and tumor thickness. pigmented lesion, female sex, and young age may Vulvovaginal Mucosal Melanoma be associated with better outcomes.11 Other studies, Melanomas arising from the female urogenital tract however, have found no associations among age, sex, account for up to 7% of melanomas in women.18 site of primary, histologic appearance, or presence of They most commonly arise from the vulva, with nodal disease and prognosis.6 fewer than 5% arising from the vagina. In 20% of

Carvajal et al.

cases, the precise site of origin cannot be identified. Staging of Mucosal Melanoma Vulvar melanoma is the second most common ma- The 7th edition of the AJCC staging system for mela- lignant disease of the vulva, and accounts for 5% to noma is used for CM, with tumor thickness, mitotic 10% of all vulvar malignant disease, with an annual rate, presence of ulceration, and regional nodal status incidence of less than 0.2 per 100,000 women per the dominant prognostic features.5 However, because 19,20 year. Eighty-five percent arise from the labia mi- the histology and implicated prognostic features of nora, clitoris, or the inner, glabrous, nonhairy portion MM differ from those of CM, the AJCC staging sys- of the labia majora, with the remaining 15% arising tem for CM cannot be universally applied to MM. from the outer, nonglabrous, hair-bearing portion of No proposed staging system has been shown to the labia majora.18 Vaginal melanoma accounts for reliably divide cases of MM into low- or high-risk fewer than 3% of all vaginal tumors and constitutes cohorts, regardless of site of origin. In the absence of approximately 1% of all melanomas in women, with a universal MM prognostic staging system, the Bal- an estimated incidence of 0.026 per 100,000 women lantyne staging system can be used.23 Although ini- per year.21,22 Vaginal melanomas more commonly tially devised for MMHN, this system can be applied arise in the lower third of the vagina on the anterior to all MM subsites (Table 2). Because most patients wall.21 Vaginal melanoma occurs most commonly present with localized stage I disease and because of in the sixth to seventh decades of life,21 with vulvar melanoma most commonly affecting postmenopaus- the lack of clear association between nodal involve- al women between 60 and 70 years of age.20 Chronic ment and prognosis in all MM subtypes, the prognos- inflammatory disease, viral infections, and chemical tic utility of this system is limited; however, because irritants are implicated risk factors of VVMM. of its ease of use and broad applicability, it remains a The clinical course differs significantly between commonly accepted staging system for MM. vulvar and vaginal melanoma. Although outcomes Because of the unique constellation of prog- are poor for both diseases, patients with vaginal nostic features implicated for MM arising from melanoma have inferior 5-year overall survival rates, distinct anatomic regions, several additional site- ranging from 5% to 25%, compared with those with specific MM staging systems have been devised. For vulvar melanomas, whose rates are 8% to 55%.4,18,22 MMHN, proposed staging systems have included For vaginal melanoma, tumor size (< 3 vs. ≥ 3 cm) the Prasad microstaging system, which is applicable has been shown to predict survival in some series. for stage I (node-negative) MMHN24; the Thomp- Nodal involvement in vulvar melanoma is of prog- son staging system for sinonasal and nasopharyngeal nostic significance, with 5-years survival rates of 15% MM25; and the AJCC T-staging system for MM of for those with nodal metastasis and 54% for those the nasal cavity and paranasal sinuses (Appendix without.18,19 Ulceration, tumor thickness, and pres- 1).26 In 2010, the AJCC developed a specific stag- ence of mitotic activity of the primary lesion may ing system for MMHN that has been shown to be also be prognostic in this disease.18,19 Additional poor predictive of outcome.27 This system was developed prognostic factors implicated in vulvar melanoma with 2 key factors in mind: 1) the overwhelmingly include advanced age, high mitotic rate, amelanosis, poor prognosis of MMHN, and 2) the notion that emergence from a preexisting nevus, tumor exten- MMHN prognosis is dependent on the local, region- sion to the lateral labia majora, central anatomic al, and distant burden of disease. To reflect the poor location, angioinvasion, DNA nondiploidy, low in- prognosis even in low-volume disease, the most lim- come, and nonwhite race.18,19 ited form of disease is considered stage III disease.

Table 2 Recommendations for Staging Systems by Primary Site of Disease Mucosal Melanoma Site of Origin Recommended Staging System Head and neck AJCC head and neck mucosal melanoma staging system Anorectum Ballantyne staging system Vulva AJCC cutaneous melanoma staging system Vagina Ballantyne staging system

Mucosal Melanoma

Staging ranges from III to IVA, IVB, and IVC (as orbital exenteration for gross orbital involvement, defined in Appendix 1), depending on local extent and radical nasal exenteration for diffuse mucosal and presence of regional and distant disease. There- disease. Evidence suggests that complete endoscopic fore, any evidence of primary MMHN is at least resections may be accomplished with less morbidity stage III disease. MM that does not invade bone, and equivalent local control.34 Thus, advances to en- deep soft tissue, cartilage, or skin is T3. All other able less-destructive and less-deforming endoscopy lesions are T4a unless they invade brain, dura, skull resections may allow for satisfactory local control base, lower cranial nerves (IX, X, XI, XII), mastica- without sacrificing prognosis. This is of particular tor space, carotid artery, prevertebral space, or me- relevance in the setting of multifocal recurrences af- diastinal structures. Stage IV is divided into 3 cat- ter definitive therapy. Oral cavity MM is approached egories, as in other head and neck subsites, to reflect surgically in the same manner as squamous cell car- advanced but not unresectable local disease with or cinoma. Local resection should be accompanied by without regional disease (IVA), highly advanced lo- appropriate reconstruction using free flaps for man- cal disease (IVB), and distant disease (IVC). dibular, soft tissue, or structural deficits. Although no specific staging system has been Historically, surgical management of MM in- developed for ARMM, some series have shown the volved radical procedures, including abdominoperi- Ballantyne staging system to have prognostic sig- neal resection (APR) for ARMM and pelvic exen- nificance.14 Proposed staging systems for VVMM in- teration for VVMM; procedures accompanied by clude Breslow depth,28 Clark level,29 Chung level,30 significant morbidity and functional limitations.15 the International Federation of Gynecology and Ob- Although some series suggest improved local con- stetrics (FIGO) staging systems for vulvar and vagi- trol with these more aggressive surgical procedures, nal carcinoma,31 and the AJCC staging system for available retrospective data suggest no difference CM (Appendix 2).5 Although none of these staging in overall survival with more conservative wide lo- systems are useful predictors of prognosis in vaginal cal excision.14–19,33,35 Because local recurrence occurs melanoma, the 2002 modified AJCC staging system concomitantly with distant metastasis in most cases, for CM was shown to be predictive of recurrence-free improved local control has limited benefit for surviv- survival for vulvar melanoma.32,33 al. Because most patients ultimately develop distant Based on currently available data, it is recom- disease regardless of the primary surgical procedure, mended that the AJCC staging systems for MMHN patient preference and quality of life considerations and CM be used for MMHN and vulvar melanoma, are critical in determining the extent of surgery. respectively (Table 2). Further work is required to Conservative surgery in the form of wide local ex- develop useful staging systems for melanomas arising cision has generally replaced APR for ARMM and from the anorectal and vaginal regions. pelvic exenteration for VVMM. Although regional lymph node status is a critical prognostic factor for intermediate-depth CM, Locoregional Management the implications of nodal involvement in MM are of Mucosal Melanoma less clear and no widely accepted standard of care Surgical Management of the Primary Site and exists for the management of regional lymph nodes Regional Lymph Nodes in this disease. Sentinel lymph node evaluation As in CM, surgery is the most important therapeutic has become a routine procedure in CM, and seems modality for MM, with complete resection providing feasible in MM.36,37 Because of the uncertainty sur- the greatest chance for cure. Because of the lentiginous rounding the prognostic significance of nodal in- growth pattern frequently associated with MM, the volvement in MMHN and ARMM, however, the multifocal nature of this disease, and anatomic con- role of sentinel node evaluation in these diseases is straints, however, wide negative margins can be diffi- undefined. Furthermore, even if the sentinel node is cult to achieve. Despite aggressive surgical resection, positive, the effects of subsequent completion node local recurrence rates are as high as 50% to 90%.10 dissection on overall survival are partly unknown, Sinonasal MM may require craniofacial resec- given the frequent development of metastases inde- tion when disease extends into the cribriform plate, pendent of nodal status.17,20,32 Nevertheless, in the

Carvajal et al.

absence of systemic disease, therapeutic neck dis- increased precision in delivery of radiotherapy, such section followed by radiation therapy in MMHN is as using protons and heavy ions that take advantage likely to improve regional control.26,38 The postop- of higher linear energy transfer. erative morbidity of neck dissection is considerably less compared with that after inguinal or pelvic node dissection, and aggressive management of the lymph Systemic Therapy for Advanced node basins should be performed only after the risks Mucosal Melanoma and benefits are considered. No systemic therapy has been shown to improve Role of Radiation Therapy in Mucosal Melanoma outcomes specifically in patients with advanced Radiosensitivity studies of melanoma cells in vitro MM, and treatment of these patients has generally and in vivo suggest a significant ability of melanoma paralleled the management of advanced cutaneous to repair sublethal damage. Variation, however, ex- disease. Until 2011, the FDA-approved agents avail- ists in radioresistance, as shown in CM and MM cell able for the treatment of advanced melanoma were lines, and recent studies suggest that melanoma is limited to dacarbazine and high-dose interleukin as radiosensitive as other solid tumors.39 Although (IL)-2, with limited clinical benefit associated with surgical resection is recommended as the primary either drug. Approval for each was based on trials therapeutic modality for localized disease, defini- that included primarily patients with CM. Although tive radiotherapy can achieve local control in up data regarding the efficacy of these treatments in to 85% of cases.7,40 Because of the high rate of lo- MM are limited, some retrospective series of various cal recurrence after surgery, particularly in MMHN, systemic regimens suggested that these agents pro- postoperative radiotherapy has become widely used, duced responses in MM equivalent to those observed with some series showing an improvement in locore- in CM. gional control.26,41,42 Most studies failed to identify More recently, 2 additional agents were approved an impact on overall survival with adjuvant radio- for melanoma, with each having the advantage of a therapy, although these analyses are complicated by demonstrable survival advantage over dacarbazine a selection bias for the use of radiotherapy in more and IL-2. Ipilimumab, a fully human IgG1 antibody advanced cases. targeting cytotoxic T lymphocyte–associated anti- Optimal dosing, fractionation schedules, and gen 4, was approved in March 2011 based on a phase clear indications for radiotherapy have yet to be III clinical trial showing an overall survival benefit defined. Use of a dose per fraction higher than 4 to of 3.6 months in previously treated patients with ad- 6 Gy may be required to improve radiation respon- vanced CM compared with a glycoprotein 100.45 A siveness.43 Some retrospective studies noted better subsequent trial showed improved survival with ipili- local control with modification of dose schedule, mumab and dacarbazine versus dacarbazine alone in whereas others note both improved local control previously untreated patients with advanced CM.46 and overall survival.41 An update of the MD Ander- Although no study of ipilimumab has been conduct- son experience suggests that standard fractionation ed specifically in patients with advanced MM, anec- greater than 54 Gy results in superior results com- dotal cases of benefit have been observed. pared with hypofractionated schedules.26 The Ra- Vemurafenib, a small molecule inhibitor with diation Therapy Oncology Group assessed 2 frac- specificity for BRAF harboring a substitution of glu- tion size schemas and randomized 137 patients with tamic acid for valine at position 600 (V600E), was advanced disease to 8 Gy per fraction in 4 fractions approved in August 2011 based on the BRIM3 trial or 2.5 Gy per fraction in 20 fractions. Although no that randomized patients with advanced CM harbor- significant difference was observed, the small num- ing the BRAF V600E mutation to treatment with ber of patients and the advanced stage of disease vemurafenib or dacarbazine.47 Vemurafenib achieved permitted may have obscured any true difference.44 a response rate of 48%, a 63% decrease in hazard of Caution should be taken if considering hypofrac- death, and a 74% decrease in hazard of tumor pro- tionation for MM in areas containing neural tis- gression. Importantly, vemurafenib is only effective sues, where dose per fraction is critical. Ongoing in tumors driven by an activating BRAF mutation. studies are evaluating newer technologies that yield Although BRAF mutations are present in 45%

Mucosal Melanoma to 50% of CM, the prevalence is less than 10% in ized by driving alterations in BRAF or KIT have led MM.48,49 The question of whether vemurafenib in to unprecedented clinical outcomes in patients with MM harboring BRAF mutations will have similar ef- advanced disease. Further refinements of systemic ficacy to that observed in CM harboring these muta- therapy for MM and the potential application of this tions has not been rigorously investigated; however, therapy in the adjuvant setting hold the promise of administration of vemurafenib in this setting is a rea- improving outcomes in this challenging disease. sonable consideration. MMs harbor recurrent mutations and/or ampli- fications of the receptor tyrosine kinase KIT, which References are alterations infrequently observed in CM.48 A tri- 1. Jemal A, Siegel R, Xu J, Ward E. Cancer statistics, 2010. CA al conducted by Memorial Sloan-Kettering Cancer Cancer J Clin 2010;60:277–300. 2. Chang AE, Karnell LH, Menck HR. The National Cancer Data Center screened 295 tumor samples for the presence Base report on cutaneous and noncutaneous melanoma: a summary of KIT mutations and amplification, and observed of 84,836 cases from the past decade. The American College of these alterations in 25% of MM.49 Of the 25 evalu- Surgeons Commission on Cancer and the American Cancer able patients treated with imatinib, an inhibitor of Society. Cancer 1998;83:1664–1678. 3. McLaughlin CC, Wu XC, Jemal A, et al. Incidence of noncutaneous multiple tyrosine kinases, including KIT, 13 had melanomas in the U.S. Cancer 2005;103:1000–1007. MM, 3 of whom experienced a major response. Full 4. Patrick RJ, Fenske NA, Messina JL. Primary mucosal melanoma. J results from a study by Guo et al.50 and interim re- Am Acad Dermatol 2007;56:828–834. sults of a study by Fisher et al.51 also show the efficacy 5. Balch CM, Gershenwald JE, Soong SJ, et al. Final version of of imatinib in select patients with melanoma harbor- 2009 AJCC melanoma staging and classification. J Clin Oncol 2009;27:6199–6206. ing KIT alterations. Responses in the phase II tri- 6. Patel SG, Prasad ML, Escrig M, et al. Primary mucosal malignant als were highly variable, but patterns have emerged melanoma of the head and neck. Head Neck 2002;24:247–257. suggesting that tumors with specific KIT alterations, 7. Gilligan D, Slevin NJ. Radical radiotherapy for 28 cases of particularly in exons 11 and 13, are more likely to mucosal melanoma in the nasal cavity and sinuses. Br J Radiol 1991;64:1147–1150. benefit from KIT inhibition than are those with am- 8. Rapini RP, Golitz LE, Greer RO Jr, et al. Primary malignant plification alone or with KIT alterations affecting melanoma of the oral cavity. A review of 177 cases. Cancer other regions. This suggests that not all KIT altera- 1985;55:1543–1551. tions are equally predictive of benefit to therapeutic 9. Benlyazid A, Thariat J, Temam S, et al. Postoperative radiotherapy inhibition. in head and neck mucosal melanoma: a GETTEC study. Arch Otolaryngol Head Neck Surg 2010;136:1219–1225. Ongoing assessment of specific KIT mutations 10. Lee SP, Shimizu KT, Tran LM, et al. Mucosal melanoma of the head as predictive biomarkers for targeted therapy is criti- and neck: the impact of local control on survival. Laryngoscope cal, and improvement in the ability to identify true 1994;104:121–126. driving mutations in KIT may result in better patient 11. Bachar G, Loh KS, O’Sullivan B, et al. Mucosal melanomas of the head and neck: experience of the Princess Margaret Hospital. Head selection. Results from ongoing work in the identi- Neck 2008;30:1325–1331. fication of primary and secondary resistance to KIT 12. Wu A, Gomez J, Zhung J, et al. Radiotherapy after surgical inhibition, other systemic therapies, and predictive resection for head and neck mucosal melanoma. Am J Clin Oncol biomarkers of response may allow optimization of pa- 2010;33:281–285. tient selection for specific treatments and the devel- 13. Cote TR, Sobin LH. Primary melanomas of the esophagus and anorectum: epidemiologic comparison with melanoma of the skin. opment of effective combination therapies. Melanoma Res 2009;19:58–60. 14. Iddings DM, Fleisig AJ, Chen SL, et al. Practice patterns and outcomes for anorectal melanoma in the USA, reviewing three Conclusions decades of treatment: is more extensive surgical resection beneficial in all patients? Ann Surg Oncol 2010;17:40–44. The increasing understanding of the unique biology 15. Brady MS, Kavolius JP, Quan SH. Anorectal melanoma. A 64-year of MM arising from distinct anatomic regions and experience at Memorial Sloan-Kettering Cancer Center. Dis Colon resultant ongoing refinements in staging and locore- Rectum 1995;38:146–151. gional management over the past several decades has 16. Pessaux P, Pocard M, Elias D, et al. Surgical management of primary led to important improvements in how this disease is anorectal melanoma. Br J Surg 2004;91:1183–1187. 17. Yeh JJ, Shia J, Hwu WJ, et al. The role of abdominoperineal managed. Advances in immunotherapy and the re- resection as surgical therapy for anorectal melanoma. Ann Surg cently identified molecular subsets of MM character- 2006;244:1012–1017.

Carvajal et al.

18. Ragnarsson-Olding BK, Nilsson BR, Kanter-Lewensohn LR, et al. Arch Otolaryngol Head Neck Surg 2009;135:1219–1224. Malignant melanoma of the vulva in a nationwide, 25-year study of 35. Suwandinata FS, Bohle RM, Omwandho CA, et al. Management 219 Swedish females: predictors of survival. Cancer 1999;86:1285– of vulvar melanoma and review of the literature. Eur J Gynaecol 1293. Oncol 2007;28:220–224. 19. Weinstock MA. Malignant melanoma of the vulva and vagina 36. Starek I, Koranda P, Benes P. Sentinel lymph node biopsy: a new in the United States: patterns of incidence and population-based perspective in head and neck mucosal melanoma? Melanoma estimates of survival. Am J Obstet Gynecol 1994;171:1225–1230. research 2006;16:423–427. 20. Sugiyama VE, Chan JK, Shin JY, et al. Vulvar melanoma: 37. Tien HY, McMasters KM, Edwards MJ, Chao C. Sentinel lymph a multivariable analysis of 644 patients. Obstet Gynecol node metastasis in anal melanoma: a case report. Int J Gastrointest 2007;110:296–301. Cancer 2002;32:53–56. 21. Piura B, Rabinovich A, Yanai-Inbar I. Primary malignant melanoma 38. Temam S, Mamelle G, Marandas P, et al. Postoperative radiotherapy of the vagina: case report and review of literature. Eur J Gynaecol for primary mucosal melanoma of the head and neck. Cancer Oncol 2002;23:195–198. 2005;103:313–319. 22. Nakagawa S, Koga K, Kugu K, et al. The evaluation of the sentinel 39. Stevens G, McKay MJ. Dispelling the myths surrounding node successfully conducted in a case of malignant melanoma of radiotherapy for treatment of cutaneous melanoma. Lancet Oncol the vagina. Gynecol Oncol 2002;86:387–389. 2006;7:575–583. 23. Ballantyne AJ. Malignant melanoma of the skin of the head and 40. Yanagi T, Mizoe JE, Hasegawa A, et al. Mucosal malignant neck. An analysis of 405 cases. Am J Surg 1970;120:425–431. melanoma of the head and neck treated by carbon ion radiotherapy. 24. Prasad ML, Patel SG, Huvos AG, et al. Primary mucosal melanoma Int J Radiat Oncol Biol Phys 2009;74:15–20. of the head and neck: a proposal for microstaging localized, Stage I 41. Krengli M, Masini L, Kaanders JH, et al. Radiotherapy in the (lymph node-negative) tumors. Cancer 2004;100:1657–1664. treatment of mucosal melanoma of the upper aerodigestive tract: 25. Thompson LD, Wieneke JA, Miettinen M. Sinonasal tract and analysis of 74 cases. A Rare Cancer Network study. Int J Radiat nasopharyngeal melanomas: a clinicopathologic study of 115 cases Oncol Biol Phys 2006;65:751–759. with a proposed staging system. Am J Surg Pathol 2003;27:594– 42. Ballo MT, Gershenwald JE, Zagars GK, et al. Sphincter-sparing 611. local excision and adjuvant radiation for anal-rectal melanoma. J 26. Moreno MA, Roberts DB, Kupferman ME, et al. Mucosal melanoma Clin Oncol 2002;20:4555–4558. of the nose and paranasal sinuses, a contemporary experience from the M. D. Anderson Cancer Center. Cancer 2010;116:2215–2223. 43. Bentzen SM, Overgaard J, Thames HD, et al. Clinical radiobiology of malignant melanoma. Radiother Oncol 1989;16:169–182. 27. Shuman AG, Light E, Olsen SH, et al. Mucosal melanoma of the head and neck: predictors of prognosis. Arch Otolaryngol Head 44. Sause WT, Cooper JS, Rush S, et al. Fraction size in external beam Neck Surg 2011;137:331–337. radiation therapy in the treatment of melanoma. Int J Radiat Oncol Biol Phys 1991;20:429–432. 28. Breslow A. Thickness, cross-sectional areas and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 1970;172:902– 45. Hodi FS, O’Day SJ, McDermott DF, et al. Improved survival with 908. ipilimumab in patients with metastatic melanoma. N Engl J Med 2010;363:711–723. 29. Clark WH Jr, From L, Bernardino EA, Mihm MC. The histogenesis and biologic behavior of primary human malignant melanomas of 46. Robert C, Thomas L, Bondarenko I, et al. Ipilimumab plus the skin. Cancer Res 1969;29:705–727. dacarbazine for previously untreated metastatic melanoma. N Engl J Med 2011;364:2517–2526. 30. Chung AF, Woodruff JM, Lewis JL Jr. Malignant melanoma of the vulva: a report of 44 cases. Obstet Gynecol 1975;45:638–646. 47. Chapman PB, Hauschild A, Robert C, et al. Improved survival 31. Benedet JL, Bender H, Jones H III, et al. FIGO staging classifications with vemurafenib in melanoma with BRAF V600E mutation. N and clinical practice guidelines in the management of gynecologic Engl J Med 2011;364:2507–1516. cancers. FIGO Committee on Gynecologic Oncology. Int J 48. Curtin JA, Busam K, Pinkel D, Bastian BC. Somatic activation of KIT Gynaecol Obstet 2000;70:209–262. in distinct subtypes of melanoma. J Clin Oncol 2006;24:4340–4346. 32. Phillips GL, Bundy BN, Okagaki T, et al. Malignant melanoma of 49. Carvajal RD, Antonescu CR, Wolchok JD, et al. KIT as a therapeutic the vulva treated by radical hemivulvectomy. A prospective study target in metastatic melanoma. JAMA 2011;305:2327–2334. of the Gynecologic Oncology Group. Cancer 1994;73:2626–2632. 50. Guo J, Si L, Kong Y, et al. Phase II, open-label, single-arm trial of 33. Moxley KM, Fader AN, Rose PG, et al. Malignant melanoma of imatinib mesylate in patients with metastatic melanoma harboring the vulva: an extension of cutaneous melanoma? Gynecol Oncol c-Kit mutation or amplification. J Clin Oncol 2011;29:2904–2909. 2011;122:612–617. 51. Fisher DE, Barnhill R, Hodi FS, et al. Melanoma from bench to 34. Hanna E, DeMonte F, Ibrahim S, et al. Endoscopic resection of bedside: meeting report from the 6th international melanoma sinonasal cancers with and without craniotomy: oncologic results. congress. Pigment Cell Melanoma Res 2010;23:14–26.

Mucosal Melanoma

Appendix 1 Staging Systems for Mucosal Melanoma of the Head and Neck Primary Site Staging System Stages/Levels Stage/Level Definition Any Ballantyne staging Stage I Clinically localized disease 23 system Stage II Regional nodal involvement Stage III Distant metastatic disease Any Prasad microstaging Level I Melanoma in situ or with microinvasion only 24 system Level II Invasion to the lamina propria Level III Deep tissue Invasion Sinonasal Thompson TNM staging Stage I T1N0M0 25 system Stage II T2N0M0 Stage III Any T, N1, M0 Stage IV Any T, Any N, M1 T1 Tumor restricted to one subsite T2 Tumor involves multiple subsites N1 Regional nodal involvement M1 Distant metastatic disease Nasal cavity and AJCC staging system for T1 Tumor restricted to one subsite, with or without ethmoid sinus sinonasal tumors* bony invasion T2 Tumor invading 2 subsites in a single region or involving an adjacent region within the nasoethmoidal complex, with or without bony invasion T3 Tumor invades the medial wall or floor of the orbit, maxillary sinus, palate, or cribriform plate T4a Tumor involves any of the following: anterior orbital contents, skin of nose or cheek, minimal extension to anterior cranial fossa, pterygoid plates, sphenoid or frontal sinuses T4b Tumor involves any of the following: orbital apex, dura, brain, middle cranial fossa, cranial nerves other than V2, nasopharynx, or clivus Maxillary sinus AJCC staging system for T1 Tumor limited to the maxillary sinus mucosa sinonasal tumors* without bony destruction T2 Tumor causing bone destruction, including erosion into the hard palate and/or middle nasal meatus, except extension to posterior maxillary wall and pterygoid plates T3 Tumor invades any of the following: bone of the posterior wall of the maxillary sinus, subcutaneous tissues, floor or medial wall of the orbit, pterygoid fossa, ethmoid sinuses T4a Tumor invades anterior orbital contents, skin of cheek, pterygoid plates, infratemporal fossa, cribiform plate, sphenoid or frontal sinuses T4b Tumor invades any of the following: orbital apex, dura, brain, middle cranial fossa, cranial nerves other than V2, nasopharynx, or clivus

Carvajal et al.

Appendix 1 Staging Systems for Mucosal Melanoma of the Head and Neck (Cont.) Primary Site Staging System Stages/Levels Stage/Level Definition Any AJCC staging system for Stage III T3N0M0 head and neck mucosal Stage IVA T4aN0M0 melanoma* T3 or T4a, N1, M0 Stage IVB T4b, Any N, M0 Stage IVC Any T, Any N, M1 T3 Mucosal disease T4a Moderately advanced disease involving the deep soft tissue, bone, cartilage, or overlying skin T4b Very advanced disease involving brain, dura, skull base, lower cranial nerves (IX, I, XI, XII), masticator space, carotid artery, prevertebral space or mediastinal structures N1 Regional nodal disease present

M1 Distant disease present

*Data from the Edge SB, Byrd DR, Compton CC, et al., eds. AJCC Cancer Staging Manual. 6th ed. New York, NY: Springer; 2002:209–220.

Appendix 2 Staging Systems for Vulvovaginal Melanoma Primary Site Staging System Stages/Levels Stage/Level Definition Any Breslow depth28 < 0.76 mm 0.76–1.5 mm 1.51–2.25 mm 2.26–3 mm > 3 mm Any Clark level29 Level I Confined to epidermis Level II Penetrates basement membrane and extends to papillary dermis Level III Extends to papillary-reticular border Level IV Invades reticular dermis but does not involve the subcutaneous fat Level V Extends into subcutaneous fat Any Chung level30 Level I Confined to epithelium Level II Penetrates basement membrane and invades to ≤ 1 mm Level III Invades to 1–2 mm Level IV Invades to > 2 mm but does not involve the subcutaneous fat Level V Extends into subcutaneous fat Vagina International Stage 0 No evidence of primary tumor Federation of Stage I Tumor confined to vagina Gynecology and Obstetrics staging of Stage II Tumor invades paravaginal tissues but not to pelvic wall 31 vaginal carcinoma Stage III Tumor extends to the pelvic wall Stage IVA Tumor invades mucosa of the bladder or rectum and/or extends beyond the true pelvis Stage IVB Distant metastasis

Mucosal Melanoma

Appendix 2 Staging Systems for Vulvovaginal Melanoma (Cont.) Primary Site Staging System Stages/Levels Stage/Level Definition Vulva International Stage 0 No evidence of primary tumor Federation of Stage IA Lesion ≤ 2 cm, confined to the vulva or perineum and Gynecology and with stromal invasion ≤ 1.0 mm Obstetrics staging of vulvar carcinoma31 Stage IB Lesion > 2 cm or any size with stromal invasion > 1.0 mm, confined to the vulva or perineum Stage II Tumor of any size with extension to adjacent perineal structures (lower/distal one-third urethra, lower/distal one-third vagina, anal involvement) Stage IIIA 1 or 2 regional lymph nodes with the following features: 1 or 2 lymph node metastases, each ≤ 5 mm OR 1 lymph node metastasis ≥ 5 mm Stage IIIB Regional lymph node metastasis with the following features: ≥ 3 lymph node metastases each < 5 mm OR 2 or more lymph node metastases ≥ 5 mm Stage IIIC Lymph node metastasis with extracapsular spread Stage IVA Fixed or ulcerated regional lymph node metastasis Stage IVB Distant metastasis (including pelvic lymph node metastasis) 2002 AJCC staging Stage 0 TisN0M0 system for cutaneous Stage IA T1aN0M0 melanoma* Stage IB T1b or T2a, N0M0 Stage IIA T2b or T3a, N0M0 Stage IIB T3b or T4a, N0M0 Stage IIC T4bN0M0 Stage IIIA T1–4a, N1a or N2a, M0 Stage IIIB T1–4b, N1a or T2a, M0 T1–4a, N1b or N2b or N2c, M0 Stage IIIC T1–4b, N1b or N2b or N2c, M0 Any T, N3, M0 Stage IV Any T, Any N, M1 Tis Melanoma in situ T1a ≤ 1.0 mm without ulceration T1b ≤ 1.0 mm with ulceration or Clark level IV or V T2a 1.01–2.0 mm without ulceration T2b 1.01–2.0 mm with ulceration T3a 2.01–4.0 mm without ulceration T3b 2.01–4.0 mm with ulceration T4a > 4.0 mm without ulceration T4b > 4.0 mm with ulceration N0 No regional metastasis detected N1a 1 lymph node, micrometastases N1b 1 lymph node, macrometastases N2a 2 or 3 lymph nodes, micrometastases N2b 2 or 3 lymph nodes, macrometastases

Carvajal et al.

Appendix 2 Staging Systems for Vulvovaginal Melanoma (Cont.) Primary Site Staging System Stages/Levels Stage/Level Definition Vulva 2002 AJCC staging N2c In-transit met(s)/satellite(s) without metastatic lymph system for cutaneous nodes melanoma* N3 4 or more metastatic lymph nodes, or matted lymph nodes, or in-transit met(s)/satellite(s) with metastatic lymph node(s) M0 No detectable evidence of distant metastases M1a Metastases to skin, subcutaneous, or distant lymph node, normal serum LDH M1b Lung metastases, normal LDH M1c Metastasis to other visceral metastases with a normal LDH, or any distant metastases and an elevated LDH

Abbreviation: LDH, lactate dehydrogenase. *Data from the Edge SB, Byrd DR, Compton CC, et al., eds. AJCC Cancer Staging Manual. 6th ed. New York, NY: Springer; 2002:209–220.