Subungual melanoma: management considerations

The American Journal of Surgery 195 (2008) 244 –248

Clinical surgery—American

Subungual melanoma: management considerations Tzeela Cohen, M.D.a, Klaus J. Busam, M.D.b, Ami Patel, B.S.a, Mary S. Brady, M.D., F.A.C.S.a,* b

a Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA

Manuscript received August 25, 2006; revised manuscript March 20, 2007

Abstract Background: Patients with subungual melanoma (SM) often experience delayed diagnosis and present with deep primary lesions. Breslow depth of the primary lesion is often unknown before definitive resection, thus complicating treatment planning. Methods: Patients with SM treated at our institution from 1992 to 2004 were identified from our prospective melanoma database. Clinical and pathologic factors were reviewed; Student t test and Kaplan-Meier method were used for statistical analysis. Results: Forty-nine patients were identified; most were female (63%). The median age was 66 years (range 24 to 83). The most common site was the great toe (n ⫽ 21), followed by the thumb (n ⫽ 15). Eight patients had in situ disease; 6 were treated initially with wide local excision, and 4 of these eventually required amputation. The median Breslow depth of invasive lesions was 2.1 mm (range .2 to 11). Toe lesions were thicker than finger lesions (mean 3.5 vs 2.5 mm, P ⫽ .005). Patients with invasive SM of the toe had a less favorable outcome than those with finger lesions (5-year overall survival 40% vs 72%, respectively; P ⫽ .05). Sentinel lymph node (SLN) biopsy was performed in 30 patients and was positive in 5 (17%); all underwent completion lymphadenectomy. Median Breslow depth in patients with positive SLN was 4 mm (range 1.2 to 11). Four of 5 patients with positive SLN developed recurrence (median 16 months); 3 patients died of disease within 40 months. Conclusions: Patients with SM present distinct therapeutic challenges. They continue to present with deep primary melanoma, particularly on the toe. Undertreatment of early disease is associated with local recurrence. © 2008 Excerpta Medica Inc. All rights reserved. Keywords: Amputation; Management; Melanoma; Recurrence; Sentinel lymph node; Subungual

Subungual melanoma (SM) arises from the nail bed and accounts for 1% to 3% of all cutaneous melanoma in primarily white populations [1–3]. Among populations at low risk for cutaneous melanoma, SM accounts ⱕ20% of all melanoma diagnoses [4,5]. Most previous series antedate the introduction of SLN mapping in the management of patients with cutaneous melanoma [6,7]. We review our recent experience with patients with SM in an era of modern surgical management. Methods Our Institutional Review Board approved this study. An institutional prospective melanoma database was searched for records of melanoma patients with a subungual primary site. One hundred eight patients were identified who were seen at * Corresponding author. Tel.: ⫹1-212-639-8347; fax: ⫹1-212-794-5847. E-mail address: [email protected]

Memorial Sloan-Kettering Cancer Center (MSKCC) between 1992 and 2004 with the diagnosis of SM. Only patients who underwent surgical treatment for localized primary disease at MKSCC were included; patients who were seen for an opinion only or who presented with recurrent disease were not included. This left a final study group of 49 patients. Medical records were reviewed for demographic, clinical, and pathologic information. Statistical analysis was performed using SPSS software (SPSS, Chicago, Illinois), Student’s t test was used to compare groups; and survival distributions were estimated using the Kaplan-Meier method and compared using the log-rank test. Results Demographics There were 49 patients treated for clinically localized primary SM at MSKCC from 1992 to 2004. Most were female (n ⫽ 31 or 63%), and the median age was 66 years

0002-9610/08/$ – see front matter © 2008 Excerpta Medica Inc. All rights reserved. doi:10.1016/j.amjsurg.2007.03.010

T. Cohen et al. / The American Journal of Surgery 195 (2008) 244 –248 Table 1 Clinical and pathologic characteristics of 49 patients with SM treated at MSKCC from 1992 to 2004 Characteristics Sex Female Male Age (y) Median Range Tumor site Toe Finger Stage (AJCC) In situ Stage I Stage II Stage III Thickness (mm)* Median Range Ulceration* Present Absent Unknown Clark’s level* Level II Level III Level IV Level V Unknown

n ⫽ 49 31 18

245

sion of the nail bed for 14, 30, and 40 years, respectively, before diagnosis. Change in the appearance of the lesion lead to diagnostic biopsy results in all 3 patients.

% 63 37

66 24–83 26 23

53 47

8 13 22 6

16 27 45 12

2.1 0.2–11 20 17 4

49 41 10

3 1 23 10 4

6 2 47 20 8

AJCC ⫽ American Joint Committee on Cancer melanoma staging system [19]. * Among 41 patients with invasive SM.

(range 24 to 83). Most of the patients were white (n ⫽ 39), 4 were African American, and 6 were Asian. Forty-one patients were diagnosed with invasive melanoma, and 8 patients had in situ lesions. Approximately half of the patients had SM of the toe (n ⫽ 26 or 53%), and the remaining had SM of the finger. The most commonly affected digit was the great toe (n ⫽ 21 or 43%), followed by the thumb (n ⫽ 15 or 31%), and fifth finger (n ⫽ 5 or 10%). Symptoms and diagnosis The most common symptom was discoloration of the nail, which occurred in 24 patients (49%). Other signs and symptoms included a nonhealing wound, a mass, a split in the nail, and bleeding. Five patients had a history of trauma to the nail or digit. Delayed diagnosis was common. “Delay” was defined as ⬎3 months from the time a patient first noticed an abnormal lesion on the digit (ie, discoloration, split of nail, ulcer, or change in a longstanding lesion) to diagnostic biopsy. Among the 49 study patients, information regarding this interval was available for 43 patients (43 of 49 or 88%). Of these, 38 patients (88%) had delayed diagnosis. The median delay time was 24 months (range 4 to 132). Many patients were treated during this period for presumed onychomycosis. Three patients reported a long-standing pigmented le-

Primary tumor characteristics Clinical and pathologic characteristics of all patients (with invasive and in situ lesions) are listed in Table 1. SM of the toe was significantly thicker than that of the finger (mean 3.5 vs 2.5 mm, P ⫽ .005). Median difference between initial thickness as seen on the biopsy specimen and final thickness (after definitive resection) was .6 mm, and the difference was significantly higher in toe compared with finger lesions (mean 2.08 vs .89 mm; P ⫽ .02). Four patients, all with toe lesions, had intermediate thickness melanoma on initial biopsy specimen but were found to have thick melanoma after definitive amputation (range 5.8 to 11mm). Operative management of the phalanx In situ SM. Of the 6 patients with in situ melanoma who underwent WLE of the lesion, radial margins were 5 mm in 2 patients, “wide” or “generous” margins (but not precisely stated) in 3 patients, and 3 mm in 1 patient. Four patients subsequently required amputation, 3 secondary to persistent positive margins and 1 because of local recurrence 18 months from the initial procedure. Two patients with in situ disease were treated with distal amputation at initial presentation; 1 had recurrence locally despite negative margins of resection and was treated with WLE and skin grafting of the dorsal aspect of the remaining digit. In both patients with recurrent disease (1 with recurrence after amputation and 1 with recurrence after WLE), the recurrent disease was in situ (ie, there was no invasive component). Invasive SM. All but 3 patients with invasive melanoma were treated with amputation. Amputation was performed through the metatarsophalangeal joint for most toe lesions and through the proximal interphalangeal joint or the midphalanx for most digit lesions. Two patients with thin invasive melanoma (.2 and .6 mm) underwent WLE with 1-cm radial margins. Both patients had no evidence of disease at 29 and 24 months follow-up, respectively. One patient underwent biopsy only because of other comorbidities. Sentinel lymph node mapping Thirty patients with invasive melanoma underwent SLN mapping and biopsy at the time of initial surgery (Table 2). Table 2 SLN mapping in patients with SM (N ⫽ 30) treated at MSKCC from 1992 to 2004 Treatment

No. patients (%)

LN mapping and SLNB Identification of SLN Positive SLNB True-negative SLNB False-negative SLNB Completion LND after positive SLNB Positive CLND

30 (100) 30 (100) 5 (17) 23 (92) 2 (8) 5 (100) 1 (20)

LND ⫽ lymph node dissection; SLNB ⫽ sentinel lymph node biopsy.

246

T. Cohen et al. / The American Journal of Surgery 195 (2008) 244 –248

of the finger (40% vs 72%; P ⫽ .05). As with other cutaneous melanoma, patients with ulcerated lesions experienced a less favorable disease-free survival (Fig. 1). Patients with thick SM (Breslow depth ⱖ4 mm) had an extremely poor prognosis, with 25% 2-year disease-free survival and ⬍40% 4-year overall survival (Fig. 2).

Fig. 1. Kaplan-Meier disease-free survival distribution for patients based on primary tumor ulceration. The median 5-year disease-free survival was 33% versus 64% in patients with ulcerated compared with nonulcerated primary SM (P ⫽ .012).

Median Breslow thickness was 2.3 mm (range .6 to 11). The SLN was positive in 5 (17%). Median Breslow thickness in patients with positive SLN was 4 mm (range 1.2 to 11). All patients with a positive SLN underwent completion lymphadenectomy (CLND), and 1 of these patients had additional positive nodes. Clinical outcome With a median follow-up of 38 months, 13 of 49 (27%) patients developed disease recurrence. Two of 8 patients with in situ melanoma (25%) developed local recurrence; 1 patient had recurrence after previous WLE and was treated with distal amputation. A second patient, who had recurrence after previous distal amputation, was treated with WLE and skin grafting of the stump. Eleven of 41 patients with invasive melanoma developed recurrent disease (27%). Clinical and pathologic data for patients with recurrent invasive SM are listed in Table 3. Median Breslow thickness in patients experiencing recurrence was 5 mm (range 1.2 to 11 mm), and 8 lesions were ulcerated. Three patients developed in-transit recurrence; 2 of them had SM of the toe and were treated initially with amputation, SLN biopsy, and subsequent CLND. In these 2 patients, there was a significant discrepancy between preoperative Breslow thickness and final thickness (1.8 mm initial vs 11 mm final and 1.5 mm initial vs 7.5 mm final, respectively). Four of the 5 patients with positive SLN had recurrence; the median time to recurrence in these patients was 16 months, and 3 of them died of disease within 40 months. Median follow-up for 41 patients with invasive subungual melanoma was 38 months (range 3 to 142). The overall 5-year survival for the 41 patients with invasive SM was 58%. Patients with invasive SM of the toe had a significantly worse 5-year overall survival than patients with SM

Discussion SM has been considered a deadly form of melanoma for many years, with reported 5-year overall survival of only 27% [7]. More recent series, however, have reported 5-year overall survival ranging from 40% to 59% [1,3,6], which is consistent with our recent experience. The first clinical description of SM has been attributed to Boyer [8]. In 1834, he described a 58-year-old man with SM of the fifth digit. In 1886, Hutchinson, a British surgeon and academic, published his experience with this disease [9]. He used the term “melanotic whitlow” to describe SM, and made the important observation, known as “Hutchinson’s sign,” that SM could be distinguished from a benign lesion of the nail when pigment changes extend onto the soft tissues surrounding the nail, the proximal nail fold, and the periungual skin (Fig. 3). The poor prognosis observed in patients with invasive SM, especially of the toe, has been reported by others [10,11]. Factors that contribute to this include delayed diagnosis and consequent advanced Breslow depth as well as the high incidence of ulceration in the primary lesion. In many patients, a presumed diagnosis of subungual hematoma, trauma, and or onychomycosis leads to the long delay

Fig. 2. Kaplan-Meier overall survival distribution based on Breslow depth for patients with primary melanoma Breslow depth ⱕ1, ⬎1 to ⬍4, and ⱖ4 mm. Five-year survival was 100% in patients with primary melanoma ⱕ1 mm, 67% for those with primary melanoma ⬎1 to ⬍4 mm, and 19% for those with primary melanoma ⱖ4 mm Breslow depth (P ⫽ .05).

T. Cohen et al. / The American Journal of Surgery 195 (2008) 244 –248

Fig. 3. In situ subungual melanoma arising on the thumb of a 65-year-old woman. The patient had a 2-year history of an enlarging area of abnormal pigmentation involving her dominant right thumb, beginning proximal to the nail matrix.

in diagnosis [12]. A history of nail trauma is common and has been reported by others [13]. This may be caused by dystrophy of the nail due to the underlying tumor, making it more vulnerable to injury. A history of trauma was uncommon in our recent experience, occurring in only 10% of patients. Prognostic features important in patients with non-SM are also determinants of outcome in patients with SM. These include Breslow thickness, ulceration, and stage at diagnosis. Although Clark’s level is an important determinant of prognosis in thin non-SM, the small number of patients in our study precluded a meaningful assessment of its importance in patients with SM. Appropriate operative management of patients with SM involves WLE of the primary lesion, with radial margins of excision determined by the Breslow depth of the lesion. Patients with invasive SM should undergo resection with 1to 2-cm margins. A margin ⱖ2 cm is preferred in patients with melanoma ⬎2 mm in depth [14]. Wide excision for SM usually requires amputation because of the paucity of soft tissue between the tumor and the bone beneath the nail, even in patients with in situ or thin melanoma. Previous reports have suggested that distal compared with complete amputation of the digit is associated with high risk of local recurrence and poor survival. For this reason, amputation through the tarsometatarsal or carpo-

247

metacarpal joint has been proposed as the most appropriate approach [7,15]. More recent experience suggests that distal amputation is sufficient if appropriate cutaneous margins can be obtained and is associated with a low risk of recurrence and no diminution in survival [6,16]. Moehrle et al [2] reported a recent experience with “functional” surgery, avoiding amputation, and found a similar incidence of local recurrence and improved survival in the group of patients undergoing limited excision compared with patients undergoing amputation. Although the 2 groups shared similar clinical and pathologic characteristics, there was certainly selection bias because assignment to the more conservative treatment arm was done after obtaining the pathology results from the initial excision. In our experience, WLE for in situ melanoma was commonly associated with the need for eventual amputation because of local recurrence or persistent positive margins. Despite this, WLE for in situ disease on a functionally important digit, such as the finger, is not an unreasonable strategy as long as close clinical follow-up of the site is performed. Among patients with invasive melanoma undergoing amputation, only 1 patient developed local recurrence. In the current series, 17% of the patients undergoing SLNB had positive SLN biopsy results, and this number is consistent with results for elective lymphadenectomy in patients with SM [6,10]. In addition, the risk of a positive SLN biopsy result is similar to that seen in all cutaneous melanoma patients treated in our institution [17]. Nodal basin recurrence was uncommon after SLN staging, consistent with our institutional experience [17]. SLN status is the most accurate predictor of outcome in patients with intermediate-thickness melanoma (1 to 4 mm) [18]. Despite this, a recent prospective, randomized trial reported no impact on overall survival when patients undergoing WLE and SLN biopsy were compared with those undergoing WLE alone [20]. It is difficult to argue, therefore, that SLN biopsy is an important component of treatment in patients in whom the primary lesion warrants consideration of adjuvant therapy and/or intensive posttreatment surveillance because of advanced Breslow depth. Four of the 5 patients with positive SLN biopsy results developed recurrent disease at a median

Table 3 Patterns of recurrence in 11 of 41 (27%) patients who had recurrence after treatment for invasive SM No.

Site

Breslow

Clark

Ulcer

AJCC stage

SNB

LND

DFS

Recurrence

Status

1 2 3 4 5 6 7 8 9 10 11

Finger Toe Toe Toe Toe Toe Finger Toe Finger Finger Toe

1.2 11 7.5 4 6.9 5 3 1.8 8 4 7

IV IV V IV V V IV V V III V

U/N ⫹ ⫺ ⫹ ⫹ ⫹ ⫹ ⫺ ⫹ ⫹ ⫹

III III III III IIC IIC IIB IB IIC IIB II

Positive Positive Positive Positive Negative Negative Negative Negative N/A N/A N/A

Negative Positive Negative Negative N/A N/A N/A Negative N/A N/A Negative

18 9 17 16 27 30 14 44 7 26 17

Systemic Local and in-transit In-transit Systemic Nodal Systemic Nodal In-transit Systemic Nodal Nodal

DOD DOD AWD DOD AWD DOD DOD DOD AWD DOD DOD

AWD ⫽ alive with disease; DFS ⫽ disease-free survival in months; DOD ⫽ dead of disease; LND ⫽ lymph node dissection; N/A ⫽ not applicable; Nodal ⫽ regional lymph node basin; SNB⫽ sentinel-node biopsy; U/N ⫽ unknown.

248

T. Cohen et al. / The American Journal of Surgery 195 (2008) 244 –248

follow-up of 16 months; of those, 3 patients had advanced Breslow depth of the primary lesion (4 to 11 mm). The management of patients with SM is further complicated by the fact that the depth of the primary lesion is frequently unknown until definite resection, which commonly occurs at the time of SLN mapping and biopsy. We found significant discrepancies between the initial Breslow thicknesses, as seen on biopsy specimen, and thickness according to the final pathology report. The difference was most pronounced for patients with toe lesions, many of them being much thicker then estimated initially. Because distal amputation was the treatment of choice in all except 2 patients with invasive melanoma, we did not face the problem of positive or inadequate margins and the subsequent necessity for more proximal amputation in these patients. It may be prudent to repeat the biopsy before surgery if a deeper lesion is suspected or if the consequences of a more proximal amputation from a functional or cosmetic point of view are significant (ie, patients with subungual melanoma of the thumb or fingers). Patients with thick lesions (ⱖ4 mm) had extremely poor outcome, with only 25% of patients alive without disease at 2 years. Our series suggests that SLNB and subsequent CLND in node-positive patients with deep primary SM are unlikely to impart a survival benefit. Lymphedema occurring after CLND may also complicate further treatment in the setting of in-transit recurrence. Because accurate SLN mapping is not precluded by amputation, one alternative is to delay SLN mapping until determination of Breslow depth has been made. For patients with intermediate-thickness melanoma, injection of radiocolloid or blue dye just proximal to the amputated digit should result in accurate localization of the SLN. In this way, patients with primary lesions with very poor prognosis can be excluded from regional nodal basin staging. The role of SLN biopsy in patients with SM cannot be adequately addressed, however, in this small recent series and should be further clarified by larger studies. Comments Special consideration should be given to the management of patients with SM. Amputation of the distal phalanx with an appropriate margin of normal skin must be considered the standard approach to management of the primary lesion, and margins of excision used for patients with non-SM should be employed. Patients with early lesions managed with treatment less than amputation are at high risk for treatment failure. Close clinical follow-up of these patients is highly recommended.

Evaluation of the utility of SLN mapping and careful selection of patients for this procedure may decrease the likelihood of subjecting patients to the morbidity of CLND when it is unlikely to be beneficial and may complicate their subsequent management. Table 3. Acknowledgments The authors thank Liza Marsh for editorial expertise. References [1] O’Leary JA, Berend KR, Johnson JL, et al. Subungual melanoma. A review of 93 cases with identification of prognostic variables. Clin Orthop Relat Res 2000;206 –12. [2] Moehrle M, Metzger S, Schippert W, et al. “Functional” surgery in subungual melanoma. Dermatol Surg 2003;29:366 –74. [3] Blessing K, Kernohan NM, Park KG. Subungual malignant melanoma: clinicopathological features of 100 cases. Histopathology 1991;19:425–9. [4] Kato T, Suetake T, Sugiyama Y, et al. Epidemiology and prognosis of subungual melanoma in 34 Japanese patients. Br J Dermatol 1996; 134:383–7. [5] Hemmings DE, Johnson DS, Tominaga GT, et al. Cutaneous melanoma in a multiethnic population: is this a different disease? Arch Surg 2004;139:968 –72. [6] Heaton KM, el-Naggar A, Ensign LG, et al. Surgical management and prognostic factors in patients with subungual melanoma. Ann Surg 1994;219:197–204. [7] Patterson RH, Helwig EB. Subungual malignant melanoma: a clinical-pathologic study. Cancer 1980;46:2074 – 87. [8] Boyer A. Fungus hematode du petit doigt. Gaz Med Paris 1834;212. [9] Hutchinson J. Melanosis often not black: melanotic whitlow. Br Med J 1886;1:491– 4. [10] Daly JM, Berlin R, Urmacher C. Subungual melanoma: a 25-year review of cases. J Surg Oncol 1987;35:107–12. [11] Hayes IM, Thompson JF, Quinn MJ. Malignant melanoma of the toenail apparatus. J Am Coll Surg 1995;180:583– 8. [12] Metzger S, Ellwanger U, Stroebel W, et al. Extent and consequences of physician delay in the diagnosis of acral melanoma. Melanoma Res 1998;8:181– 6. [13] Mohrle M, Hafner HM. Is subungual melanoma related to trauma? Dermatology 2002;204:259 – 61. [14] Thomas JM, Newton-Bishop J, A’Hern R, et al. Excision margins in high-risk malignant melanoma. N Engl J Med 2004;350:757– 66. [15] Das Gupta T, Brasfield R. Subungual melanoma: 25-year experience. Ann Surg 1965;161:545–52. [16] Quinn MJ, Thompson JE, Crotty K, et al. Subungual melanoma of the hand. J Hand Surg [Am] 1996;21:506 –11. [17] Clary BM, Brady MS, Lewis JJ, et al. Sentinel lymph node biopsy in the management of patients with primary cutaneous melanoma: review of a large single-institutional experience with an emphasis on recurrence. Ann Surg 2001;233:250 – 8. [18] Balch CM, Buzaid AC, Soong SJ, et al. New TNM melanoma staging system: linking biology and natural history to clinical outcomes. Semin Surg Oncol 2003;21:43–52. [19] Balch CM, Buzaid AC, Soong SJ, et al. Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol 2001;19:3635– 48. [20] Morton DL, Thompson JF, Cochran AJ, et al. Sentinel-node biopsy or nodal observation in melanoma. N Engl J Med 2006;355(13):1307–17.