Management of vulvar melanoma

GYNECOLOGIC

ONCOLOGY

45, 2.54-258 (1992)

Management of Vulvar Melanoma EDWARD L. TRIMBLE, JOHN

L. LEWIS, JR., LAURA L. WILLIAMS, JOHN P. CURTIN, DOUGLASS CHAPMAN, JAMES M. WOODRUFF, STEPHEN C. RUBIN, AND WILLIAM J. HOSKINS Memorial

Sloan-Kettering

Cancer Center, New York, New York 10021

Received July 24, 1991

Considerable debate centers on the optimal treatment for vulvar melanoma, as well as those clinicopathological factors influencing prognosis. We reviewed 80 patients with vulvar melanoma seen between 1949 and 1990. Primary tumors were assessedaccording to Chung (47 patients) and Breslow (65 patients) microstaging systems.Fifty-nine patients (76%) underwent radical vulvectomy, ten patients (13%) had a partial vulvectomy, and nine patients (12%) had a wide local excision. Fifty-six also underwent inguinal node dissection. Median follow-up was 193 months. Median survival was 63 months. Ten-year survival by Chung level was as follows: I 100%; II, 81%; III, 87%; IV, 11%; V, 33%. Ten-year survival by tumor thickness was as follows: 0.75 mm, 48%; 0.751.5 mm, 68%; 1.51-3.0 mm, 44%; >3.0 mm, 22%. Increased depth of invasion was associated with increased incidence of inguinal node metastasis. Cox regression analysis demonstrated prognostic significance for tumor thickness (P < O.OOl), inguinal node metastasis (P < O.OOl), and older age at diagnosis (P < 0.001). Radical vulvectomy did not seemto improve survival over less radical procedures. Based on this experience, we recommend radical local excision for patients with malignant melanoma of the vulva. Patients who have more than a superficially invasive melanoma should also have inguinal lymph node dissection. Q 1992 Academic

Press, Inc.

INTRODUCTION melanomas compose less than 10% of melanomas in women and of vulvar malignancies in general. Their rarity has made evaluation of staging methods, treatment, and survival difficult. Strategies for management have been patterned primarily on the treatment of epidermoid carcinoma of the vulva. More recently, based on experience in cutaneous melanoma, less radical surgery has been recommended. To address the issues of survival, we elected to study outcomes in patients previously reported from this institution, as well as those treated since that time [l]. As surgical management of this entity has grown more conservative over this period, we also chose to compare the Vulvar

254 0090~8258/92$4.00 Copyright 0 1992 by Academic Press, Inc. All right5 of reproduction in any form reserved.

effect on survival of radical vulvectomy versus local excision. In addition, we sought to define clinicopathological factors affecting regional lymph node metastasis and survival. MATERIALS AND METHODS Through a review of discharge diagnoses, pathology records, and previously reported cases, we identified 84 patients with vulvar melanoma seen at Memorial SloanKettering Cancer Center between 1949 and 1990. Fortyfour cases were previously reported. Additional followup is provided on 15 of them. Forty are new cases. Slides were reviewed by one author (J.M.W.). Two caseswere excluded, as a review of their slides showed atypical melanocytic nevi. Slides of 2 patients’ lesions were not available for review, leaving 80 patients in the series. Microstaging was assigned by the Chung level (47 patients) and the Breslow tumor thickness (65 patients) of the original lesion [1,2]. As many of the original slides from the diagnostic biopsies had been returned to outside institutions, it was not possible to assign a Chung level or a Breslow tumor thickness to all 80 patients. Survival was calculated by the method of Kaplan and Meier [3]. Differences between survival curves were evaluated by the log-rank test [4]. Cox regression was used for multivariate analysis of the effects of age and tumor thickness on survival [5]. RESULTS All patients were Caucasian. Mean age at diagnosis was 58.5 years (range, 14 to 88). The most common presenting symptoms were bleeding (17%), itching’ or burning (16%), and pain (7%). Thirteen (16%) reported a preexistent nevus. Eleven had initial surgical treatment elsewhere. Sixty-seven patients had primary surgical treat-

MANAGEMENT

OF VULVAR

TABLE 2 Cox Regression Analysis of Histopathologic Factors Affecting Survival

TABLE 1 Initial Surgical Treatment Radical vulvectomy Without inguinal node dissection With bilateral inguinal node dissection With bilateral inguinal node dissection and pelvic lymph node biopsy With ipsilateral inguinal node dissection

12 20 25

Hemivulvectomy Without inguinal node dissection With bilateral inguinal node dissection With ipsilateral inguinal node dissection With anterior exenteration and partial vaginectomy

6 2 1 1 10

Subtotal Wide local excision Without inguinal node dissection With bilateral inguinal node dissection With ipsilateral inguinal node dissection and pelvic lymph node biopsy

6 2 1

Subtotal

9

Not treated

2 80

Total

’ One patient, with Chung level V disease and Breslow tumor thicknessof 15 mm, declined radical vulvectomy and inguinal node dissection.

ment at Memorial Sloan-Kettering Cancer Center. Two patients not treated, one because of advanced metastatic disease at time of diagnosis and the other due to a syn-

A

( ( ( ( (

age >- 50, < 60 og* >- 60, < 70 ag, b- 70 TICK

MARK(l)

INDICATES

LAST

6 13 14 23 22

PTS. PTS. PTS. PTS. PTS.

6 4 7 6 2

P value P = 0.001 P = 0.001

Age

2 59

opa s 40 age B- 40, s 50

Histopathologic factors (n = 65) Tumor thickness

Subtotal

0 0

255

MELANOMA

CENSORED) CENSORED) CENSORED) CENSORED) CENSORED)

FOLLOW-UP

chronous adenocarcinoma of the pancreas, were excluded from survival analysis. Table 1 documents the intial surgical treatment. Fifty-nine patients underwent radical vulvectomy, ten patients underwent hemivulvectomy, and nine patients underwent wide local excision. Fifty-three patients underwent inguinal lymphadenectomy. Nine patients were lost to follow-up; at time of last contact, four of these had no evidence of disease, and five were alive with disease. Median follow-up among all patients was 193 months. Median survival was 63 months. Younger age at diagnosis bore a better prognosis than did older age (Fig. 1). To explore a confounding factor, that is, that younger patients might present with earlier stage disease, we used stepwise Cox regression with two eligible terms to see if one might be eliminated. Even after tumor thickness entered the Cox model, age at diagnosis retained independent prognostic information. Table 2 shows the factors evaluated using Cox regression and their associated P values. Microstaging according to Chung levels was a more accurate predictor of both survival and risk of nodal disease than the Breslow tumor thickness (Tables 3 and 4). Survival curves by Chung level and Breslow tumor thickness are shown in Figs. 2 and 3. Similar curves were noted for patients with Chung levels II and III, while Chung levels IV and V also clustered together. The 12 patients with tumors less than 0.75 mm thick had only 24% survival at 15 years. Although the number of patients TABLE 3 Survival by Chung and Breslow Microstaging Systems No. of pts

5 Years (%I

10 Years (%I

15 Years (%I

20 Years (%I

1 12 8 20 6

100 81 87 45 33

100 81 87 11 33

58 65 6 -

58 65 -

Breslow tumor thickness (mm) co.75 12 0.75-1.50 10 1.51-3.0 9 >3.0 34

48 79 56 44

48 68 44 22

24 54 44 11

54 -

Chung level I II III IV V

8

.oo

1 51.00

I 102.00 months

I 1, 153.00 204.00 from diagnosis

I 255.00

FIG. 1. Survival by age at diagnosis.

b, 306.00

256

TRIMBLE ET AL.

TABLE 4 Incidence of Metastatic Disease in Inguinal Lymph Nodes by Chung and Breslow Microstaging Systems Chung level I II III

Breslow tumor thickness (mm) CO.76 0.76-1.50 1.51-3.0 >3.0

14% 0% 33% 46%

z

.oo

TICK

MARK(l)

II,Ill IV,V

( 20 PTS. ( 26 PTS. INDICATES

LAST

13 CENSORED) 3 CENSORED)

FOLLOW-UP

(l/7) (O/5) (3/9) (11/24)

in this group is small, the finding suggests that even shallow lesions may prove aggressive. We are at a loss to explain why this group should have a survival at 15 years lower than that of patients with deeper lesions (0.75-3.0 mm). Among those patients with melanomas classified in Chung levels II and III, there appears to be a steep dropoff in survival between 10 and 15 years. The etiology for this is unclear, but it underscores the need for close and prolonged follow-up in these patients. Metastatic disease in the inguinal lymph nodes presaged decreased survival (Fig. 4). The incidence of metastatic groin disease by Chung and Breslow levels is shown in Table 4. Only 2 patients had metastatic disease in the pelvic lymph nodes. Both had inguinal metastases as well. One died of her disease; the other was alive with disease when lost to

0

Chung 1.wI1 Chung I.Y.II

N/A 0% (O/7) 29% (217) 50% (8/16) 67% (2/3)

IV V

0 0 A

0 0

twr tumr tumr turn,

thicknet thickneaa thicknrrr thicknort

x- .75 .76 1.51 )- 3.1

TICK

MARK(I)

INDICATES

51I.oo

I lll2.00 months

mn 1.5 mn 3 mn “I”

( I2 PTS. ( IO PTS. ( 9 PTS. ( 34 PTS. LAST

4 5 4 6

FIG. 2. Survival by Breslow tumor thickness (log-rank test, P = 0.06).

I

102.00 months

I

I

153.00 204.00 from diagnosis

I

255.00

31

1.00

follow-up. Two patients who each had metastasis to a single inguinal lymph node are alive without evidence of disease at 6% and 17 years, respectively. The extent of vulvar surgery did not appear to affect survival (Fig. 5). DISCUSSION

Our patients’ race, age, presenting symptoms, and preexistent nevi are consistent with previous reports. Advanced age at diagnosis has been noted as a risk factor 0 0

FOLLOW-UP

diagnosis

I 5, .oo

FIG. 3. Survival: Chung microstaging levels II and III versus levels IV and V (log-rank test, P = 0.0001).

CENSORED) CENSORED) CENSCRED) CENSORED)

I a, I I 255.00 306.00 153.00 204.00 from

x ‘.oo

negative po,ltirr TICK

x-c.,

groin groin MARK(()

44.00

nodrr nodes

( 35 PTS. ( 16 PTS.

INDICATES

66.00 months

LAST

17 CENSMIED) 3 CENSDRED)

FOLLOW-UP

132.00 173.00 from diagnosis

FIG. 4. Survival: inguinal node involvement (log-rank test, P = 0.0092).

MANAGEMENT 0 0

ride local hmivulvectmy

A

rodisol

TlCK

x .oo

(

excision

vulvsctcq

MARK(l)

I 51.oo

INDICATES

LAST

9 PTS.

.? CENSORED)

( 10 PTS. ( 59 PTS.

2 CENSORED) 7.1 CENSORED)

OF VULVAR

FOLLOW-UP

a, I I I I 506.00 204.00 255.00 153.00 102.00 months

from

diagnosis

FIG. 5. Survival: initial surgical treatment (log-rank test, P = 0.53).

for recurrence and death in cutaneous melanoma. Similar findings have been noted in patients with vulvar melanoma by Rose et al. [6] and Bradgate et al. [7]. Our study confirms these observations. Current FIG0 recommendations for staging vulvar melanoma parallel those for vulvar carcinoma. Cutaneous melanoma is staged according to the recommendations of the American Joint Committee on Cancer (stage I, local disease; stage II, regional spread to lymph nodes; stage III, distant metastases). Clark et al. argued for microstaging of cutaneous melanoma based on the level of penetration into the dermis [8]. Breslow and Thickness proposed microstaging based on tumor thickness [2]. Chung et al., noting that the absence of a well-defined papillary dermis in part of the vulva made Clark levels III and IV unusable, put forward a modification of the Clark levels [l]. They suggested using depth of penetration beneath the epidermis to distinguish these intermediate levels (Fig. 6). In reviewing 3500 casesof cutaneous stage I melanoma, Balch et al. found survival to be more closely associated with Breslow levels than with Clark levels [9]. Podratz et al. reported 48 patients with vulvar melanoma, staged according to FIGO, with microstaging assigned by Clark and Breslow levels [lo]. They found survival in patients with Clark level V significantly decreased from that in level I-IV patients and tumor thickness greater than 3.00 mm as compared to thinner lesions. Our study suggests that Chung microstaging offers an even better correlation with prognosis than does Breslow microstaging. Rose et al. [6], Davidson et al. [ll], and Bradgate et

257

MELANOMA

al. [7] noted no difference in survival between those pa-

tients treated with radical vulvectomy and those treated with wide local excision. Podratz et al., on the other hand, noting the high incidence of local recurrences in women with deeply invasive vulvar melanoma, such as Clark’s level III-V, have argued for 4- to 5-cm margins, even exenterative procedures, to increase survival in these patients [lo]. For more superficial lesions, however, they recommend radical local excision. In stage I cutaneous disease Urist et al. have recommended a l- to 2-cm margin for thin melanomas (co.76 mm in thickness) and 3 cm for intermediate and thick melanomas (bO.76 mm) [12]. Veronesi et al. randomized 612 patients with primary cutaneous melanoma no thicker than 2 mm to narrow (lcm margins) or wide (margins of 3 cm or more) excision. They found no difference in survival between the two groups [13]. Our findings indicate that wide local excision produces survival as effectively as do more radical procedures. (Fig. 6). While our data do not allow us to make a definite statement about margin width, it would seem that the recommendations made by Urist et al. for cutaneous melanoma would be applicable to melanoma of the vulva. We thus advocate radical local excision, to obtain l- to 2-cm margins, for thin melanomas (Chung levels II), and 3 cm for thicker melanomas of the vulva (Chung levels III-IV). Chung level I lesions (melanoma in situ) may be treated with simple excision and close follow-up.

LEVELS

OF INVOLVEMENT

IN VULVAR

MELANOMA

SUBCUTANEOUS

*As

measured from the granular surface epithelium

layer of

FIG. 6. Chung levels of involvement in melanoma of the vulva. Level I (melanoma confined to the surface epithelium and pilar sheath) and level V (tumor extension into the underlying adipose tissue) are identical to Clark’s levels I and V. Levels II, III, and IV are determined by measurements from the granular layer of the vulvar skin or outermost layer of the squamous mucosa. (Reprinted with permission from The American College of Obstetricians and Gynecologists, Obstet. Gynecol. 45, 638-646, 1973).

258

TRIMBLE ET AL.

statistics arising in its consideration, Cancer Chemother. Rep. 50, In cutaneous melanoma occult regional lymph node 163-170 (1966). involvement has been calculated to increase with tumor 5. Cox, D. R. Regression models and life tables (with discussion) thickness. Our study details similar findings in vulvar melJ. R. Statist. Sot. B 34, 187-220 (1972). anoma (Table 4). Podratz et al. [lo] have documented 6. Rose, P. G., Piver, M. S., Ysukada, Y., and Lau, T. Conservative the adverse impact on survival of metastatic disease in therapy for melanoma of the vulva, Am. .I. Obstet. Gynecol, 159, the regional lymph nodes, which our study confirms. Met52-55 (1988). astatic disease in the pelvic lymph nodes was not found 7. Bradgate, M. G., Rollason, T. P.,‘McConkey, C. C., and Powell, J. Malignant melanoma of the vulva: A clinicopathological study in the absence of disease in the inguinal lymph nodes. of 50 women, Br. J. Obstet. Gynaecol. 97, 124-133 (1990). Thus, there is no justification at this time for the routine 8. Clark, W. H., From, L., Bernardino, E. A., and Mihm, M. C. The removal of pelvic lymph nodes. histogenesis and biologic behavior of primary human melanomas The role of elective regional lymphadenectomy in cuof the skin, Cancer Res. 29, 705-726 (1969). taneous melanoma remains controversial. Balch has pos9. Balch, C. M., Soong, S. J., Shaw, H. M., and Milton, G. W. An tulated a benefit to node dissection in cutaneous lesions analysis of prognostic factors in 4000 patients with cutaneous melof intermediate thickness, between 1 and 4 mm [ 141.With anoma, in Cutaneous melanoma (C. M. Balch and G. W. Milton, Eds.), Lippincott, Philadelphia, pp. 321-352 (1985). shallow lesions of the vulva (less than 1 mm in thickness), the low incidence of inguinal node metastasis and absence 10. Podratz, K. C., Gaffey, T. A., Symmonds, R. E., Johansen, K. L., and O’Brien, P. C. Melanoma of the vulva: An update, Gynecol. of benefit in cutaneous melanoma would argue against Oncol. 16, 153-168 (1983). elective inguinal lymph node dissection. Regional node 11. Davidson, T., Kissin, M., and Westbury, G. Vulvo-vaginal meldissection is a reasonable option for Chung level III-V anoma-should radical surgery be abandoned?, Br. J. Obstet. Gylesions, since nodal involvement is common, and longnaecol. 94, 473-476 (1987). term survival may be achieved even with positive nodes. 12. Urist, M. M., Balch, C. M., and Milton, G. W. Surgical manageEffective systemic therapy will be necessary before surment of the primary melanoma, in Cutaneous melanoma (C. M. Balch and G. W. Milton, Eds.), Lippincott, Philadelphia, pp. 71vival can be improved for patients with Chung level V 90 (1985). lesions.

REFERENCES Chung, A. F., Woodruff, J. M., and Lewis, J. L., Jr. Malignant melanoma of the vulva, Obstet. Gynecol. 45, 638-646 (1975). Breslow, A. Thickness, cross-sectional area, and depth of invasion in the prognosis of cutaneous melanoma, Ann. Surg. 172, 902-908 (1970). Kaplan, E. L., and Meier, P. Nonparametric estimation from incomplete observations, J. Am. Stat. Assoc. 53, 457-481 (1958). Mantel, N. Evaluation of survival data and two new rank order

13. Veronesi, U., Cascinelli, N., Adamus, J., Balch, C., Bandiera, D., Barchuk, A., Bufalino, R., Craig, P., De Marsillac, J., Durand, J. C., van Geel, A. N., Holmstrom, H., Hunter, J. A., Jorgensen, 0. G., Kiss, B., Kroon, B., Lacour, H., Lejeune, F., MacKie, R., Mechl, Z., Mitrov, G., Morabito, A., Nosek, H., Panizzon, R., Prade, M., Santi, P., Van Slooten, E., Tomin, R., Trapeznikov, N., Tsanov, T., Urist, M., and Wozniak, K. D. Thin stage I primary cutaneous malignant melanoma: Comparison of excision with margins of 1 or 3 ems, N. En@. J. Med. 318, 1159-1162 (1988). 14. Balch, C. M. The role of elective lymph node dissection in melanoma: Rationale, results, and controversies, J. Clin. Oncol. 6, 163172 (1988).