It is estimated that 80,270 cancers of the female reproductive tract will be diagnosed this year in the United States. Of these cancers, primary vaginal cancer is the least frequent, expected to account for only 2,160 cases, or less than 3%.1 The majority of primary vaginal cancers are of squamous histology. Malignant melanoma (Fig. 1) accounts for only about 5% of all vaginal cancers.2 In fact, the overall incidence of vaginal melanoma is exceedingly rare, at only 0.46 cases per 1 million women per year.3
Given that only just more than 100 new cases of vaginal melanoma are seen each year in the United States, it is not surprising that there are sparse published data on this disease. The published literature on vaginal melanoma consists of case reports and small case series (15 or fewer cases)4–7 and reports of population-based studies.2,3,8,9 Unfortunately, case reports and small case series do not permit any conclusions to be drawn, and population-based studies may be limited by issues of data quality. For example, in one study of all subtypes of vaginal cancer using Surveillance, Epidemiology, and End Results program data, 52% of cases abstracted had no documentation of tumor size and another 33% had no documentation of tumor grade.9 In another study on vaginal cancer that used data from the National Cancer Data Base, the authors analyzed and reported survival of women with vaginal melanoma based on Clark's level, but only 31% of patients had that end point reported.2
The 5-year overall survival rate for women with vaginal melanoma is only 18%.10 This is considerably lower than the 5-year overall survival rate for patients with melanoma of the vulva (47%) and strikingly lower than the 5-year overall survival rate for women with cutaneous melanoma (81%).10 Given the limited information on vaginal melanoma published thus far in the literature, we performed this study to estimate the clinical and pathologic features of vaginal melanoma and to determine predictors of outcome in patients with this disease.
MATERIALS AND METHODS
This study was conducted with approval from the Institutional Review Board at The University of Texas MD Anderson Cancer Center. We reviewed the records of all patients who presented to MD Anderson with a new diagnosis of untreated primary malignant melanoma of the vagina from January 1, 1980, through December 31, 2009. Only patients with clinical stage I melanoma confined to the vagina as determined by clinical and radiographic examinations were included. Patients were excluded if they had a nonmelanomatous lesion (eg, squamous or adenocarcinoma lesion); if they had cervical or vulvar involvement (by definition considered a cervical or vulvar primary tumor); if they had undergone any therapy for vaginal melanoma other than excisional punch biopsy; if they presented for a second opinion only and received their primary therapy elsewhere; if they did not have follow-up at our institution; if they had recurrent or metastatic disease at presentation; or if they had a second primary tumor diagnosed within 5 years of their vaginal melanoma diagnosis.
Clinical records were reviewed for this study. Surgical pathology reports were reviewed to obtain gross and histopathologic data. All histopathologic diagnoses were confirmed by gynecologic pathologists at MD Anderson. Depth of invasion was measured in millimeters (Chung's modification of Breslow's levels), from the outermost epithelial layer to the deepest point of invasion. Demographic and clinical information, including age at the time of diagnosis, race or ethnicity, body mass index, presenting symptoms, and tumor characteristics, were obtained from medical records, as was information about treatment and follow-up. Tumor size was evaluated both as a continuous variable and as a binary variable (smaller than 3 cm, or 3 cm or larger) based on previous reports that tumors 3 cm or larger were associated with worse outcomes.7,11
Treatment modalities were assigned to one of three categories: pelvic exenteration, wide excision, and nonsurgical (primary radiation therapy, chemotherapy, or both). Most patients in the wide-excision group underwent a wide radical excision. However, circumferential 2-cm margins were at times difficult to obtain because of the anatomic location of the tumor. Therefore, some patients had wide local excisions (margins less than 2 cm) or mixed components of local and radical excisions. Patient in the “nonsurgical” group were judged by their attending physician to not be medically fit for surgery because of age, comorbidities, or poor performance status. After completion of therapy, patients were typically followed with clinical examinations every 3 months for the first year, every 4 months for the second year, every 6 months for the 3rd through 5th years, and then annually. A chest radiograph was obtained annually. Other imaging was ordered as clinically indicated.
Overall survival and progression-free survival were the primary end points evaluated. When information on survival was not in the medical record, an attempt was made to obtain death certificate information. If such information could not be obtained, data on the patient after her last contact were censored. Similarly, when information on recurrence was not in the medical record, data on the patient after her last contact were censored. Overall survival and progression-free survival were calculated from the date of the surgical diagnosis.
Data were analyzed using one-way analysis of variance and χ2 analyses. Overall and progression-free survivals were calculated using the method described by Kaplan and Meier. Log-rank test was used to compare survival curves. Missing data were coded as “unknown,” and those data points were excluded from the analysis. All data were analyzed using SPSS 17 for Windows.
Between 1980 and 2009, 37 women presented to MD Anderson with malignant melanoma clinically and radiographically isolated to the vagina. Patient characteristics are summarized in Table 1. The median age at presentation was 60.6 years. Eighty-four percent of patients were white; only 3% were African American. Vaginal bleeding was the most common presenting symptom. Eleven percent of patients were asymptomatic at presentation and had the melanoma discovered during a routine gynecologic examination.
At presentation, the median tumor size was 3.0 cm (range 0.4–5 cm), and the median depth of invasion was 7 mm (range 1–21 mm). Only two patients had tumor invading less than 2 mm. Twenty-nine patients (79%) had a single lesion; the remaining eight patients (21%) had multifocal disease. Twenty-four patients (65%) had lesions located in the distal third of the vagina or introitus, ten (26%) had lesions located in the proximal third, and two (6%) had lesions located in the middle third. The other patient (3%) had disease in all three of these anatomic locations.
Initial treatment varied: 28 women (76%) underwent a wide excision. Five women (14%) were taken to the operating room for pelvic exenteration, but one of these patients had the procedure aborted because of intraoperative findings consistent with metastatic disease. Table 2 provides details of the four patients who underwent exenterative procedures. Four women (10%) were felt not to be surgical candidates and received radiation therapy, chemotherapy, or both radiation therapy and chemotherapy. Of the 33 women treated surgically, 27 (83%) had negative margins, including all four women who had pelvic exenteration. Twenty-four women had surgical resection of lymph nodes. Six (25%) of these patients had metastatic disease in the lymph nodes. Four of these patients (66%) had positive nodes in the pelvis, and the other two (33%) had positive inguinal nodes. In these six patients, five were noted to have a positive node at the time of their wide radical excision and lymph node dissection. Three of these women received adjuvant radiation therapy and two opted for no further therapy. All five suffered disease recurrence at a median of 5.5 months (range 0–8 months). Sites of recurrence were local in one patient, the liver in one patient, the lung in one patient and simultaneous lung and liver recurrences in two patients. The sixth patient had initial radiation therapy followed by an anterior exenteration where a positive pelvic node was found on final pathology. This patient was lost to follow-up 5 months after exentertative surgery.
Of the 20 patients who received adjuvant radiation therapy, 10 were treated using a hypofractionated schedule of 30 Gy in five fractions given over 2.5 weeks. In all cases, treatment was delivered to the vagina and perivaginal tissues only or to the distal pelvis, including the vaginal and distal iliac or inguinal nodes as appropriate. An additional two patients were treated using a modified hypofractionated schedule of 25 Gy in five fractions plus brachytherapy. The remaining patients received either standard pelvic radiation by external beam (three patients), intensity-modulated radiation therapy to the vagina and pelvic nodes (one patient), and postoperative brachytherapy (three patients). Finally, one patient received preoperative brachytherapy before surgical excision by means of a radical hysterectomy and upper vaginectomy. When examining management of newly diagnosed stage I vaginal melanoma by decade, 100% (seven of seven) of patients were treated surgically in the 1980s, 81% (9 of 11) were treated surgically in the 1990s, and 84% (16 of 19) were treated surgically in the 2000s. Owing to small cell values, χ2 was not calculable.
At a median follow-up time of 17.4 months, 33 women (89%) had had recurrence. Of these 33 women, seven (22%) had local recurrence only (ie, in vagina or vulva), 20 (63%) had distant recurrence only, and five (15%) had both local and distant recurrence. One patient had an unknown site of recurrence. Table 3 details the sites of first recurrence. All 11 (100%) women who underwent wide local or radical excision without adjuvant therapy had recurrence. Of these 11 women, five (45%) had local recurrence only, six (55%) had distant recurrence only, and none had both local and distant recurrence. Fifteen (94%) of the 16 women who underwent wide local or radical excision followed by adjuvant radiation therapy had recurrence. Of these 15 women, one (7%) had local recurrence only, 11 (73%) had distant recurrence only, and two (13%) had both local and distant recurrence. The remaining patient had recurrence at an unknown location. One patient received chemotherapy only after undergoing wide radical excision. She had recurrence at a distant site.
Median progression-free survival (Fig. 2A) was 11.4 months (95% confidence interval [CI] 6.0–16.8 months), and median overall survival (Fig. 2B) was 19.1 months (95% CI 8.9–29.4 months). The 2- and 5- year progression-free survival rates were 19.1% and 9.5%, respectively. The 2- and 5-year overall survival rates were 46.4% and 20.0%, respectively. For women who underwent wide local or radical excision, the median survival was 24.3 months, compared with 34.4 months for women who underwent pelvic exenteration and 8.7 months for women who had radiation therapy, chemotherapy, or radiation therapy and chemotherapy without surgery. Median survival for the radiation therapy and chemotherapy group was significantly shorter than that for the two surgical approaches (P=.01). There was no significant difference in survival between the two surgical approaches. For patients who had wide local or radical excision, those who received adjuvant radiation had a median survival of 29.4 months, compared with 16.1 months for those who did not (P=.46). Median survival by other pathologic factors is shown in Table 4. The only pathologic finding that was significantly associated with survival was nodal status: median survival was 7.8 months for patients with positive nodes, compared with 30.0 months for patients with negative nodes (P<.001). Details on the four patients who did not experience recurrence during follow-up are presented in Table 5.
In this report, we found a 5-year survival rate of 20%, which is similar to what has previously been reported in large database studies.7,10 Patients who were medically fit and had lesions amenable to surgical resection lived longer than those who were not able to undergo surgery. Although we were able to achieve negative margins in a majority of patients, achieving negative margins in patients with vaginal melanoma can often prove difficult given the anatomic location of many of these lesions. The literature on cutaneous melanoma supports a 1-cm margin for melanomas 2 mm thick and less, and a 2-cm margin for melanomas more than 2 mm thick,12 which accounted for the majority of patients in our experience. Achieving 2-cm margins often requires resection of the distal urethra or anus.
To completely resect the primary lesion with adequate margins, some gynecologic oncologists maintain that for women with melanoma seemingly isolated to the vagina, pelvic exenteration is the best therapeutic approach.13 However, our study found an improvement in overall survival of only 5 months for women who underwent pelvic exenteration compared with women who underwent wide local or radical excision and adjuvant radiation therapy. Without a clear improvement in survival, surgeons might consider wide excision and adjuvant radiation therapy for women with vaginal melanoma instead of the more radical, more morbid pelvic exenteration. However, the number of patients who underwent pelvic exenteration in our study was small (n=4), so definitive recommendations are obviously difficult to make. We do, however, perform positron emission tomography imaging in all patients before consideration for exenteration to rule out nodal spread or distant metastases.
Although melanoma has historically been considered “radioresistant,” our data support wide local or radical excision followed by adjuvant radiation therapy as a reasonable approach to treating vaginal melanoma. Although the 13-month improvement in survival for women who received adjuvant radiation therapy after wide local or radical excision compared with women who underwent wide local or radical excision alone was not statistically significant, we believe this is likely due to small numbers of patients in each of those cohorts. Supporting this hypothesis is the fact that adjuvant radiation therapy greatly reduced local recurrence of melanoma at the primary resection site. A similar finding was reported by Irvin et al, who reviewed six patients with malignant melanoma of the vagina treated with wide excision, radiation therapy, or both.6 Although all six patients had recurrence, the four women treated with single-modality therapy (ie, surgery or radiation) had recurrence in the vagina, whereas the two women treated with combination therapy (ie, surgery and radiation) had recurrence in the lung.
Similar to our findings in vaginal melanoma, postoperative adjuvant radiation therapy in mucosal melanomas of the head and neck reduces the risk of local recurrence. Temam et al found a relative risk of recurrence of 0.4 (95% CI 0.2–0.9) in patients with mucosal melanoma who received adjuvant radiation therapy compared with those treated with surgery alone.14 Other investigators in head and neck mucosal melanomas have found similar reduction in local recurrence risk with adjuvant radiation therapy, albeit in studies with smaller sample sizes.15,16 Although radiation therapy may reduce the risk of local recurrences, we did not find a difference in overall recurrence rates or survival between patients who received adjuvant radiation therapy and those who did not. Temam et al likewise found that the observed reduction in the risk of local recurrence of head and neck melanomas with adjuvant radiation therapy did not translate into a survival advantage as the risk of distant metastases was not reduced.14
Like vaginal melanoma, anal–rectal melanoma is a mucosal lesion with a very poor prognosis. The controversy of whether superior treatment consists of exenteration or wide radical excision with postoperative radiation therapy for vaginal melanoma has undergone an analogous debate among those surgical oncologists who treat that disease. Analogous to exenteration, the “classic” surgical approach for treating patients with anal–rectal melanoma had been abdominoperineal resection, a debilitating surgery that typically commits the patient to a permanent colostomy. This morbid procedure, however, resulted in 5-year survival rates of only 0–25%.17–20 More conservative therapy of sphincter-sparing surgical resection and adjuvant radiation, however, resulted in a 5-year survival rate of 31%.21 Although no randomized study directly comparing abdominoperineal resection with sphincter-sparing surgery for anal–rectal melanoma exists, it appears as although the two approaches result in similar overall survival. Likewise, we found that morbid exenterative surgery and conservative radical excision with radiation therapy have seemingly similar overall survival.
Because distant metastases were a component in 78% of cases of recurrence, one must wonder whether adjuvant systemic therapy might be of worth in these high-risk tumors. Multiple traditional cytotoxic agents, including dacarbazine, temozolomide, and platinum compounds, both as single agents and in combination, have been evaluated in the treatment of melanoma, with limited or no success. The addition of interleukin-2 to traditional cytotoxic agents has failed to show an improvement in overall survival but has considerably increased toxicity.22 Currently, the only Food and Drug Administration–approved agent for adjuvant therapy in cutaneous melanoma is interferon-α (IFN-α). Multiple phase III studies have shown a significant improvement in disease-free survival in patients who received IFN-α as adjuvant therapy compared with patients who were observed.23 However, the effect on overall survival in these studies has been mixed, leading some to voice concern over whether routine administration of IFN-α is cost-effective. Most oncologists have decided that the benefits do outweigh the costs and the risks, and prescription of IFN-α has, for the most part, become standard for patients with high-risk cutaneous melanoma after surgical resection.23
Because of the rarity of vaginal melanoma, the utility of regional lymph node dissection or even sentinel lymph node biopsy in women with this disease remains unknown. Our study found that lymph node status was of great prognostic significance, but the therapeutic benefits are unclear. One of the difficulties in performing lymphadenectomies in women with vaginal cancer is deciding which nodal basins to resect. Traditional wisdom argued that the lower third of the vagina had lymph drainage to nodes in the groin and that the apex of the vagina had lymph channels that led to the pelvis. However, lymphatic mapping studies of the vagina show that these predictions of drainage based on anatomic considerations do not always hold.24 One approach, then, would be to consider performing lymphatic mapping in women with vaginal melanoma to delineate the lymphatic drainage of the primary tumor. For years, this was the approach in treating cutaneous melanoma, but now there is debate among surgeons as to the utility of sentinel node biopsy in patients with that disease. A large prospective, phase III study failed to show a survival advantage for patients with cutaneous melanoma who underwent sentinel node biopsy with immediate lymphadenectomy for metastatic disease compared with patients who were simply observed after excision of the primary lesion, with nodal dissection performed only if there was clinical recurrence.25 The authors found the sentinel node status to be of prognostic value only. Likewise, we believe nodal status in women with vaginal cancer is likely of prognostic value only, and therefore a procedure with low morbidity such as sentinel node biopsy should be considered, whereas complete inguinofemoral or pelvic lymphadenectomies should be approached with caution.
This study certainly suffers from the limitations of all retrospective chart reviews in a rare disease that relies heavily on the quality of documented data in the medical record. Fortunately, owing to the single-institution aspect of this study, we had limited missing data and a consistency of standardized follow-up intervals and times among the faculty. Other limitations of retrospective studies include selection, referral, and treatment biases. This retrospective cohort included every patient with stage I vaginal melanoma who presented to our institution between 1980 and 2009 so selection bias likely was not an issue. However, as MD Anderson is a large referral center, our study population may not reflect the general population with this disease. One example of how referral bias may have influenced this study population is the fact that 84% of patients were white whereas only 3% were African American. Unlike cutaneous and vulvar melanomas, where whites are significantly more at risk than African Americans for developing disease, vaginal melanoma, a mucosal lesion, has not been shown to occur more frequently in whites compared with African Americans.3 In addition, the long observational period could add treatment bias as management of vaginal melanoma might change over time. Although the numbers were too small to perform statistical analysis, it did not appear that initial management of vaginal melanoma changed notably in each decade examined.
In summary, even when malignant melanoma is clinically limited to the vagina, this cancer carries an exceedingly poor prognosis. Surgical resection (wide excision or pelvic exenteration) improves overall survival compared with radiation therapy or chemotherapy or both. Few patients in this study underwent pelvic exenteration, but those patients did not have a significantly longer overall survival than the patients treated with more conservative surgery. Wide excision with adjuvant pelvic radiation therapy reduces local recurrence and may increase survival, although the survival difference between patients who underwent irradiation and those who did not was not statistically significant. As 93% of patients who had recurrence after wide excision and adjuvant radiation therapy had recurrence at least at one distant site, systemic therapy with IFN-α might make clinical sense. Ideally, a phase II study of wide excision followed by pelvic radiation therapy and IFN-α could be undertaken to evaluate this regimen. Because of the extreme rarity of vaginal melanoma, however, a clinical trial in this specific disease is unfortunately unlikely.
1. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics, 2009. CA Cancer J Clin 2009;59:225–49.
2. Creasman WT, Phillips JL, Menck HR. The National Cancer Data Base report on cancer of the vagina. Cancer 1998;83:1033–40.
3. Hu DN, Yu GP, McCormick SA. Population-based incidence of vulvar and vaginal melanoma in various races and ethnic groups with comparisons to other site-specific melanomas. Melanoma Res 2010;20:153–8.
4. Tjalma WA, Monaghan JM, de Barros Lopes A, Naik R, Nordin A. Primary vaginal melanoma and long-term survivors. Eur J Gynaecol Oncol 2001;22:20–2.
5. Signorelli M, Lissoni AA, Garbi A, Perego P, Mangioni C. Primary malignant vaginal melanoma treated with adriamycin and ifosfamide: a case report and literature review. Gynecol Oncol 2005;97:700–3.
6. Irvin WP Jr, Bliss SA, Rice LW, Taylor PT Jr, Andersen WA. Malignant melanoma of the vagina and locoregional control: radical surgery revisited. Gynecol Oncol 1998;71:476–80.
7. Reid GC, Schmidt RW, Roberts JA, Hopkins MP, Barrett RJ, Morley GW. Primary melanoma of the vagina: a clinicopathologic analysis. Obstet Gynecol 1989;74:190–9.
8. McLaughlin CC, Wu XC, Jemal A, Martin HJ, Roche LM, Chen VW. Incidence of noncutaneous melanomas in the U.S. Cancer 2005;103:1000–7.
9. Shah CA, Goff BA, Lowe K, Peters WA 3rd, Li CI. Factors affecting risk of mortality in women with vaginal cancer. Obstet Gynecol 2009;113:1038–45.
10. Ragnarsson-Olding B, Johansson H, Rutqvist LE, 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.
11. Buchanan DJ, Schlaerth J, Kurosaki T. Primary vaginal melanoma: thirteen-year disease-free survival after wide local excision and review of recent literature. Am J Obstet Gynecol 1998;178:1177–84.
12. Testori A, Rutkowski P, Marsden J, Bastholt L, Chiarion-Sileni V, Hauschild A, et al. Surgery and radiotherapy in the treatment of cutaneous melanoma. Ann Oncol 2009;20 suppl 6:vi22–9.
13. Geisler JP, Look KY, Moore DA, Sutton GP. Pelvic exenteration for malignant melanomas of the vagina or urethra with over 3 mm of invasion. Gynecol Oncol 1995;59:338–41.
14. Temam S, Mamelle G, Marandas P, Wibault P, Avril MF, Janot F, et al. Postoperative radiotherapy for primary mucosal melanoma of the head and neck. Cancer 2005;103:313–9.
15. Vongtama R, Safa A, Gallardo D, Calcaterra T, Juillard G. Efficacy of radiation therapy in the local control of desmoplastic malignant melanoma. Head Neck 2003;25:423–8.
16. Newlin HE, Morris CG, Amdur RJ, Mendenhall WM. Neurotropic melanoma of the head and neck with clinical perineural invasion. Am J Clin Oncol 2005;28:399–402.
17. Konstadoulakis MM, Ricaniadis N, Walsh D, Karakousis CP. Malignant melanoma of the anorectal region. J Surg Oncol 1995;58:118–20.
18. Ross M, Pezzi C, Pezzi T, Meurer D, Hickey R, Balch C. Patterns of failure in anorectal melanoma: a guide to surgical therapy. Arch Surg 1990;125:313–6.
19. Luna-Perez P, Rodriguez DF, Macouzet JG, Labastida S. Anorectal malignant melanoma. Surg Oncol 1996;5:165–8.
20. Roumen RM. Anorectal melanoma in The Netherlands: a report of 63 patients. Eur J Surg Oncol 1996;22:598–601.
21. Ballo MT, Gershenwald JE, Zagars GK, Lee JE, Mansfield PF, Strom EA, et al. Sphincter-sparing local excision and adjuvant radiation for anal-rectal melanoma. J Clin Oncol 2002;20:4555–8.
22. Rosenberg SA, Yang JC, Schwartzentruber DJ, Hwu P, Marincola FM, Topalian SL, et al. Prospective randomized trial of the treatment of patients with metastatic melanoma using chemotherapy with cisplatin, dacarbazine, and tamoxifen alone or in combination with interleukin-2 and interferon alfa-2b. J Clin Oncol 1999;17:968–75.
23. Sabel MS, Sondak VK. Point: interferon-alpha for adjuvant therapy for melanoma patients. J Natl Compr Canc Netw 2004;2:61–8.
24. Frumovitz M, Gayed IW, Jhingran A, Euscher ED, Coleman RL, Ramirez PT, et al. Lymphatic mapping and sentinel lymph node detection in women with vaginal cancer. Gynecol Oncol 2008;108:478–81.
25. Morton DL, Thompson JF, Cochran AJ, Mozzillo N, Elashoff R, Essner R, et al. Sentinel-node biopsy or nodal observation in melanoma. N Engl J Med 2006;355:1307–17.
Figure. No caption available.