Originally published as Ann Surg Oncol Early Release 10.1245/ASO.2004.04.024 on December 8, 2003
Annals of Surgical Oncology 11:34-39 (2004)
© 2004 Society of Surgical Oncology
Primary Vaginal Melanoma: A Critical Analysis of Therapy
Thomas J. Miner, MD,
Ruby Delgado, MD,
Jonathan Zeisler, MD,
Klaus Busam, MD,
Kaled Alektiar, MD,
Richard Barakat, MD and
Elizabeth Poynor, MD
From the Departments of Surgery (TJM, RB, EP), Pathology (RD, KB), and Radiation Oncology (KA), Memorial Sloan-Kettering Cancer Center, New York, NY.
Correspondence: Address correspondence and reprint requests to: Richard Barakat, MD, Memorial Sloan-Kettering Cancer Center, 1275 York Ave., MRI-1026, New York, NY 10021; Fax: 212-717-3214; E-mail: gynbreast{at}mskcc.org
 |
ABSTRACT
|
|---|
Background: Primary vaginal melanoma is a rare and highly malignant disease. The impact of therapy on outcomes is poorly understood.
Methods: Records of all patients treated for primary vaginal melanoma at Memorial Sloan-Kettering Cancer Center from 1977 to 2001 were reviewed. Survival analysis was performed based on appropriate patient, tumor, and treatment variables. Pathologic materials were reviewed to confirm the original diagnosis and examine appropriate clinicopathologic features.
Results: Thirty-five women were treated for vaginal melanoma; the primary treatment selected was surgical for 69% (24) and radiotherapy for 31% (11) of the patients. Surgical removal of the tumor was achieved in 92% (22) of the 24 patients selected for surgical therapy. At operation, radical excision with en bloc removal of involved pelvic organs was performed in 50% (12) of the 24 patients, a wide excision was performed in 42% (10), and a total vaginectomy was performed in 8% (2). Elective pelvic lymph node dissection was performed in 74% (26) of the 35 cases. Lymph node metastasis was found in only 8% (2) of these 26 patients. The overall median survival was 20 months. Primary surgical therapy was associated with longer overall survival (25 vs. 13 months; P = .039). Recurrence-free survival was not associated with the extent of surgery. None of the examined clinicopathologic features were associated with survival differences.
Conclusions: The prognosis is poor for patients with primary vaginal melanoma. Improved clinical outcomes were associated with surgical removal of gross disease whenever possible. Because of the low rate of lymph node metastasis, elective pelvic lymph node dissection is not obligatory. In cases of surgically unresectable disease, primary radiation therapy is indicated.
Key Words: Melanoma • Outcomes • Treatment • Tumors • Vaginal
|
INTRODUCTION
|
|---|
Noncutaneous melanoma is an uncommon form of melanoma that develops in areas of the body not exposed to ultraviolet radiation. A review of 84,836 cases of melanoma from the National Cancer Data Base that occurred between 1985 and 1994 showed that 91.2% of the melanomas were cutaneous, 5.2% were ocular, 1.3% were mucosal, and 2.2% involved unknown primaries.1 In women, 1.6% of melanomas are genital; the vulva is the most frequent site of involvement (70%), followed by the vagina (21%) and cervix (9%). Malignant melanoma of the vagina is extremely rare and accounts for fewer than 0.3% of all melanomas in women and fewer than 3% of primary vaginal tumors.2,3 The presumed precursor of malignant melanoma, the melanocyte, is embryologically derived from neural crest cells and can be found in the basal portion of the vaginal epidermis in 3% of normal adult females. It is believed that during the migration of melanocytes, some of these cells come to rest aberrantly in the vaginal mucosa and serve as the potential site of malignant melanoma.4
Vaginal melanoma is a highly lethal disease and has the worst prognosis of all vaginal malignancies.3 Regardless of the therapy chosen, results of treatment have been poor, with reported 5-year survival rates ranging from zero to 25%.2,3,5 Although most reports favor surgical therapy for vaginal melanoma, some investigators have argued that extensive operations are not warranted because most patients will die of their disease regardless of therapy. In the absence of other, more effective therapies, however, it is important to consider the value of surgery for this disease. With this in mind, we reviewed our experience with primary vaginal melanoma with the goal of improving patient selection for aggressive or conservative management.
|
MATERIALS AND METHODS
|
|---|
The Memorial Sloan-Kettering Cancer Center gynecology database and disease management systems database were searched for the period 1977 to 2001 for patients whose final diagnosis was vaginal melanoma. Approval to perform the study was obtained from the Institutional Review Board of the Memorial Sloan-Kettering Cancer Center. Melanomas metastatic to this site or lesions that were extending from the cervix or vulva at the time of presentation were excluded. Only patients with primary malignant melanomas of the vagina were selected for this study. Treatment was based on the discretion of the attending surgeon.
Original pathologic materials were reviewed to confirm the original diagnosis. Gross features were obtained from pathology reports. Samples were evaluated for the presence of ulceration, pigmentation, multifocality, melanoma in situ, vascular invasion, perineural invasion, melanosis, cellular phenotype and pleomorphism (graded 1–4), tumor thickness, and mitoses per square millimeter.
Individual medical records were reviewed retrospectively to obtain demographic, clinical, treatment, recurrence, and survival data. In this study, vaginal melanomas were staged both by the thickness of the primary tumor and with the International Federation of Gynecology and Obstetrics (FIGO) system for vaginal carcinomas. The length of follow-up was determined from the patient’s last available medical record. Cases censored for survival analysis were confirmed by checking patient identifiers against death records.
Data were analyzed with SAS statistical software (release 4.0, SAS Institute, Cary, NC). Data were expressed as percentages in cases of categorical variables and as medians in cases of continuous variables. Means were compared with Student’s t-test and frequencies were compared with the 
2 test when appropriate. Survival curves were constructed by the Kaplan-Meier method. The univariate associates between clinical variables and survival were examined by the log-rank test. P values less than .05 were considered significant.
|
RESULTS
|
|---|
Thirty-five women were treated for primary vaginal melanoma at the Memorial Sloan-Kettering Cancer Center over a 24-year period. The patients’ ages ranged from 36 to 87 years (mean, 62 years). All cases presented clinically as a lesion in the vagina. The most common presenting symptom was vaginal bleeding in 23 patients (66%). One patient (3%) presented with pain. Malignant melanoma was identified on routine screening examination in the remaining 11 patients (31%).
Patient characteristics and clinicopathologic information are summarized in Table 1. The clinical stage at presentation was not associated with advanced patient age (P = .24). The majority of the melanomas were thick: 77% (27) of the 35 lesions had a depth of greater than or equal to 4 mm. For eight subjects the minimum depth was determined from incisional biopsy specimens and most likely was less than the actual thickness of the lesion. Sufficient material for a more complete histopathologic analysis (Table 2) was available from 83% (29) of the 35 subjects.
View this table:
[in this window]
[in a new window]
|
TABLE 2. Histologic features of patients treated for primary vaginal melanoma. Sufficient pathologic material for complete analysis was available in 29 patients
|
|
Table 3 summarizes the treatment received by patients in this series. Primary surgical therapy was performed on 69% (24) of the 35 patients. Although resection of all gross disease was achieved following 92% (22) of the 24 operations, pathologic analysis revealed positive microscopic margins in 71% (17 of 24). Radiation therapy was given to 59% (10) of the 17 patients with positive margins. Radiation therapy was selected as the primary treatment for the remaining 31% of patients (11 of 35). In 55% (6) of the 11 patients who received primary radiation therapy, the tumor was locally advanced and believed to be surgically unresectable; the remaining 5 (45%) had early stage (I and II) disease with no identifiable features precluding surgical therapy.
Elective staging of pelvic lymph nodes was performed in 74% (26) of the 35 cases. Four patients also underwent sentinel lymph node biopsy. Two patients were found to have inguinal lymph node metastasis, initially identified by physical examination in one patient and by sentinel lymph node biopsy in the other. Further internal iliac nodal disease was identified in the patient with palpable lymphadenopathy. Some form of adjuvant therapy was given in 74% of the cases (26 of 35). Treatment included adjuvant interferon 
-2b in 20% of cases (7 of 35) or a vaccine in 20% (7 of 35).
The overall median survival of the 35 patients in this series was 20 months. As seen in Fig. 1, primary surgical therapy was associated with longer survival than primary radiation therapy (median, 25 months vs. 13 months; P = .039). Longer overall survival was associated with primary surgical therapy, even when we excluded those patients with known, presumably more advanced, surgically unresectable disease (median, 25 months vs. 9 months; P = .006). No examined histopathologic feature was associated significantly with overall survival.
Although tumor thickness is commonly used as a marker for cutaneous melanoma, the preponderance of thick (T4) lesions in this small study population makes analysis based on tumor depth unreliable. Long-term survival was uncommon: 5-year and 10-year survival was achieved in only 11% (4) and 6% (2) of the 35 cases, respectively. Both of the 10-year survivors were treated initially with radical surgery associated with removal of gross disease and microscopically positive margins. These patients had local recurrences, one at 80 months and the other at 119 months, and underwent complete gross resection of tumor with wide local excision and pelvic exenteration, respectively. Although local control was achieved, each patient died with distant disease. Primary treatment of the two remaining 5-year survivors was surgical as well. Local control was maintained until death in a patient who underwent wide excision. The remaining patient died with local recurrence following a radical procedure. Both patients had positive microscopic margins following gross resection of disease during the initial operation.
For those 22 patients who were able to undergo complete resection of disease, the median recurrence-free survival was 12 months. Recurrent disease was identified in 82% (18) of these 22 patients. All four patients censored during analysis of recurrence-free survival had shown no evidence of disease 12 months after primary surgical therapy. As seen in Fig. 2, the extent of operation (radical resection vs. wide excision) was not associated with differences in recurrence-free survival (median, 12 months vs. 10 months; P = .53). The presence of microscopically positive margins was not associated with diminished recurrence-free survival (P = .78). At the completion of follow-up, 82% of patients (18 of 22) had evidence of recurrent disease. The site of first recurrence in those 18 patients was distant in 12 (67%) and local in 6 (33%). In those patients who had local disease at the time of first recurrence, 67% (4 of 6) had adjuvant radiation therapy following the operative procedure.
|
DISCUSSION
|
|---|
The proper and effective treatment of vaginal melanoma is poorly defined, but recommendations include (individually or in combination) conservative wide local excision, radical surgery, irradiation, and chemotherapy. Decisions regarding therapy are difficult, as one must balance the value of often radical and deforming operations against a poor anticipated outcome.
Few data are available on which to base recommendations. Because of the rarity of this disease, the literature includes only a few small series and case reports. It has been estimated that only about 200 cases have been reported in the English-language literature.6 As summarized in Table 4, the current literature is limited by the absence of any prospective data, the large time span over which cases have been reported, incomplete description of the treatment modalities and elements of patient selection, and the tendency for reports to combine patients with vulvar and vaginal melanoma, the former associated with a much better prognosis.3,4,6–10 Although the current study is subject to the same problems inherent to any retrospective study, to our knowledge it represents the largest reported series on primary vaginal melanoma. It is our hope that the clear definition of our group of patients will enable an improved understanding of how to best manage this disease and will promote the generation of hypotheses about factors related to survival.
Vaginal melanoma is associated with a poor clinical outcome for the majority of patients. The paucity of long-term survivors and the median survival of 20 months seen in this study confirm previous reports demonstrating limited survival of patients with vaginal melanoma. Poor prognosis is not unique to vaginal melanoma, however. Like other noncutaneous melanomas, vaginal melanomas usually remain undetected until late in the natural history of the disease. As with other mucosal melanomas, the chance for long-term survival is limited. In mucosal lesions of the head and neck, cure rates of 15% have been demonstrated.11 Similarly, anorectal melanoma has been associated with cure in only approximately 10% of cases.12,13 As is true for other mucosal melanomas, surgery remains the only potentially curative treatment for a limited number of patients.
Although the prognosis for most patients in this study was poor, long-term survival (more than 5 years) was observed only among patients whose tumors were completely surgically resected. Longer overall survival in this study was associated with performance of surgery rather than primary radiation therapy for potentially resectable disease. Petru et al.3 suggested that primary radiation therapy (at doses higher than given to our patients) could serve as an alternative to surgery for patients with lesions less than 3 cm in diameter.12 The findings from the present study, however, are consistent with other reports that favor a surgical approach to vaginal melanoma.3,8–11 Radiation therapy has been associated with 5-year survival in a minority of patients but has been advocated for long-term local control.3,10,14,15 Although melanoma is frequently thought of as a radioresistant tumor, radiation therapy appeared to provide good and durable local control in the limited number of patients with surgically unresectable disease in this series. Potential prognostic factors such as age, stage, depth of invasion, pigmentation, ulceration, and adjuvant therapy did not correlate with patient outcome. Over the 24-year period of this study, there was no a priori treatment triage system. Thus, the influence of patient selection factors, therapeutic intent (palliative vs. curative), and adjuvant therapy cannot be fully determined by this retrospective analysis.
Although the high rate of locoregional failure after local excision of vaginal melanomas has led many authors to recommend more extensive operative procedures, other reports have stated that the extent of resection was not associated with survival.7,9 The types of operations performed in this series, similarly, were not associated with differences in either overall or recurrence-free survival. Gross tumor excision, however, was a common component of most of the operations performed. Pathologically negative margins of resection were not commonly obtained, perhaps because of associated melanoma in situ or diffuse vaginal melanosis, and were not associated with decreased recurrence-free survival. Because of the retrospective nature of this report, factors associated with choice of operative procedure cannot be reliably determined. Although surgical therapy was associated with prolonged survival and excision was associated with a high rate of microscopically positive margins, a policy requiring exenterative procedures in all patients with primary vaginal melanoma is not recommended. In this study, in two-thirds of patients the first site of recurrence was distant. This suggests that perhaps the burden of distant disease outweighs the clinical significance of microscopic margins and that distant disease decreases the need for procedures more radical than gross tumor removal.
The risk of systemic recurrence of melanoma after resection has stimulated an intensive search for effective adjuvant therapies. In large, prospective, randomized trials, numerous agents (bacillus Calmette-Guerin [BCG], Corynebacterium parvum, dacarbazine-based chemotherapy, megestrol acetate, vitamin A, and interferon-
have been shown to have no efficacy in the adjuvant setting after resection of high-risk melanoma. The value of agents such as adjuvant interferon -2b remains controversial. The relative inactivity of conventional agents and the significant toxicity of biologic therapies have led to the study of immunotherapy for melanoma. Significant advances have been made in the understanding of the biology of vaccine therapy, and several agents are undergoing clinical evaluation in phase III trials.16
The role of elective lymph node dissection in vaginal melanoma is controversial as well. Proponents suggest that it is a useful staging procedure and that removal of metastatic nodal disease before it becomes clinically apparent may prevent systemic spread in some patients. Proponents of observation argue that elective lymph node dissection subjects patients to a potentially morbid procedure without benefit and that no prospective, randomized trial to date has shown a survival benefit after routine elective lymph node dissection.16 Although a procedure to stage the pelvic lymph nodes was commonly performed in our patients, this procedure did not enable identification of unrecognized lymph node metastases in any subject. Patients with lymph node metastasis were identified with physical examination and sentinel lymph node biopsy. Because of the rarity of nodal metastasis and the morbidity associated with pelvic lymphadenectomy, routine staging of pelvic lymph nodes should not be considered an obligatory component of the surgical management of vaginal melanoma. Cobellis et al.6 have also concluded recently that lymphadenectomy is unjustified for vaginal melanoma. Sentinel lymph node biopsy has been adopted as a less morbid way to determine the pathologic status of regional lymph nodes. It not only improves the accuracy of staging and prognosis but also allows for selection of patients for therapeutic lymph node dissection of nodal metastasis. In addition, it identifies a homogeneous patient population for enrollment in clinical trials and evaluations of novel adjuvant therapy agents.17 Sentinel lymph node biopsy has recently been described by Abramova et al.18 as a rational way to stage lymph nodes in patients with vaginal melanoma. The early experience of these authors with two patients with vaginal melanoma parallels our experience with four patients. Although this procedure is slightly more technically demanding than routine sentinel lymph node mapping for cutaneous melanoma, this operation can identify the lymphatic drainage without subjecting patients to the associated morbidity of complete lymphadenectomy.
In conclusion, primary vaginal melanoma remains a rare and lethal disease. Because of the rich lymphatic and vascular supply of the vagina, both local and distant recurrence is common in this biologically aggressive malignancy. Although the majority of patients with vaginal melanoma will die of their disease, surgical therapy remains the only chance for long-term survival and is associated with increased overall survival. The goal of operative intervention should be complete resection of gross disease. Patients with surgically unresectable disease should receive primary radiation therapy, with the intent of local control. As in other areas of the body, elective lymph node dissection for melanoma is not indicated, and the use of sentinel lymph node biopsy may allow a more rational way to stage this disease. Because of the rarity of this disease, multicenter trials will be required to enroll enough patients to study vaginal melanomas more effectively. Because vaginal melanoma appears to behave aggressively, like other mucosal melanomas, perhaps it should not be included in trials involving patients with cutaneous melanomas. Inclusion of patients with noncutaneous melanoma as a distinct high-risk group is suitable for clinical trials of adjuvant therapy. Sentinel node biopsy may allow further stratification by effectively identifying patients with nodal metastasis.
|
FOOTNOTES
|
|---|
The prognosis for patients with primary vaginal melanoma is poor. Improved clinical outcomes were associated with surgical removal of gross disease where possible.
Received for publication April 23, 2003.
Accepted for publication October 6, 2003.
|
REFERENCES
|
|---|
- Chang AE, Karnell LH, Menck HR. The National Cancer Database report on cutaneous and non cutaneous melanoma. Cancer 1998; 83: 1664–78.[CrossRef][Medline]
- Weinstock M. Malignant melanomas of the vulva and vagina in the United States: patterns of incidence and population-based estimates of survival. Am J Obstet Gynecol 1994; 171: 1225–30.[Medline]
- Petru E, Nagele F, Czerwenka K, et al. Primary malignant melanoma of the vagina: long-term remission following radiation therapy. Gynecol Oncol 1998; 70: 23–6.[Medline]
- Irvin WP, Bliss SA, Rice LW, Taylor PT, Andersen WA. Malignant melanoma of the vagina and locoregional control. Gynecol Oncol 1998; 71: 476–80.[CrossRef][Medline]
- Ragnarsson-Olding B, Johansson H, Ringborg U. Malignant melanoma of the vulva and vagina: trends in incidence, age distribution, and long-term survival among 245 consecutive cases in Sweden 1960–1984. Cancer 1993; 71: 1893–7.[CrossRef][Medline]
- Cobellis L, Calabrese E, Stefanon B, Raspagliesi F. Malignant melanoma of the vagina. Eur J Gynecol Oncol 2000; 21: 295–7.[Medline]
- Bonner JA, Perez-Tamayo C, Reid GC, Roberts JA, Morley GW. The management of vaginal melanoma. Cancer 1988; 62: 2066–72.[Medline]
- Reid GC, Schmidt RW, Roberts JA, Hopkins MP, Barrett RJ, Morely GW. Primary melanoma of the vagina: a clinicopathologic analysis. Obstet Gynecol 1989; 74: 190–9.[Abstract/Free Full Text]
- Van Nostrand KM, Lucci JA, Schell M, Berman ML, Manetta A, DiSaia PR. Primary vaginal melanoma: improved survival. Gynecol Oncol 1995; 55: 234–7.[CrossRef]
- Konstadoulakis MM, Ricaniadis N, Driscoll DL, Karakousis CP. Malignant melanoma of the female genital system. Eur J Gynecol Oncol 1994; 20: 141–5.
- Conley J, Pack JT. Melanoma of the mucous membranes of the head and neck. Otolaryngology 1974; 99: 315–9.
- Wanebo HJ, Woodruff JM, Farr GH, Quan SH. Anorectal melanoma. Cancer 1981; 47: 1891–5.[CrossRef][Medline]
- Brady MS, Kavolius JP, Quan SH. Anorectal melanoma: a 64 year experience at Memorial Sloan-Lettering Cancer Center. Dis Colon Rectum 1995; 38: 146–51.[CrossRef][Medline]
- Ariel IM. Five year cure of a primary malignant melanoma of the vagina by local radioactive isotope therapy. Am J Obstet Gynecol 1961; 82: 405–7.[Medline]
- Chung A, Casey M, Flannery J, Woodruff J, Lewis J. Malignant melanoma of the vagina: report of 19 cases. Obstet Gynecol 1980; 55: 720–7.[Abstract/Free Full Text]
- Reaves ME, Coit DG. Melanoma: a multidisciplinary approach for the general surgeon. Surg Clin N Am 2000; 80: 2–21.
- McMasters KM, Reitngen DS, Ross MI, et al. Sentinel lymph node biopsy for melanoma: controversy despite widespread agreement. J Clin Oncol 2001; 19: 2851–5.[Abstract/Free Full Text]
- Abramova L, Parekh J, Irvin WP, et al. Sentinel node biopsy in vulvar and vaginal melanoma: presentation of six cases and a literature review. Ann Surg Oncol 2002; 9: 840–6.[Abstract/Free Full Text]
TOP
ABSTRACT
INTRODUCTION
