Melanoma Arising in Association With Blue Nevus

A Clinical and Pathologic Study of 24 Cases and Comprehensive Review of the Literature

Sanam Loghavi; Jonathan L Curry; Carlos A Torres-Cabala; Doina Ivan; Keyur P Patel; Meenakshi Mehrotra; Roland Bassett; Victor G Prieto; Michael T Tetzlaff


Mod Pathol. 2014;27(11):1468-1478. 

In This Article


Clinical Features

Clinical features of the patients are summarized in Table 1. Overall, there were 9 male and 15 female patients with a mean age of 49 years (range: 20–85 years; median: 50 years). The majority of patients were Caucasian (22/24; 92%). One patient was Black (4%) and one was Hispanic (4%). The tumors were located on the head and neck region in 12 patients (50%), the trunk in 5 patients (21%), and the buttock/sacrococcygeal region in 4 patients (17%); the perianal area (1 patient; 4%), the penis (1 patient; 4%), and the spinal cord (1 patient; 4%) were less common sites of involvement. No case arose on the extremities in our series.

Clinically, most patients (10/24; 42%) endorsed a history of a recently enlarging and/or changing pigmented lesion, which had been present for many years prior to the diagnosis. In patient 1, the lesion was identified during a routine dermatologic exam. Patient 2 presented with metastatic tumor. Patient 4 presented with blurry vision, and patient 14 presented with exophthalmos. Patient 16 was receiving chemotherapy (intraperitoneal and intravenous cisplatin and paclitaxel) for ovarian cancer and noticed a darkly pigmented lesion on her scalp after chemotherapy-induced alopecia. Patient 21 presented with sciatic nerve pain. Initial presenting clinical features were not available for the remaining eight patients.

Histopathologic Findings

Histopathologic findings are summarized in Table 2. An example of melanoma blue nevus type situated in close proximity to the epidermis is shown in Figure 1. The lesions lacked a prominent intraepidermal component in all cases, except one (case 1). In addition, many cases arose in dermal and subcutaneous sites and were situated a significant distance from the epidermis. Figure 2 illustrates an example of a case arising in the perianal soft tissue where the largest tumor dimension was measured to represent tumor thickness due to lack of epidermis. Furthermore, in eight cases, the epidermis was not available for review; therefore, a calculation of Breslow thickness was not always feasible. The mean tumor thickness in our series was 20.9 mm (range: 0.6–130 mm; median 12 mm). Mitotic figures were identified in all cases; the mean mitotic figure count was 6.5/mm2 (range: 1–30/mm2; median 3). Perineural invasion was present in nine cases (38%) and satellitosis was identified in four cases (17%). In contrast, despite the overall relatively large size of the lesions, lymphovascular invasion (three cases; 13%), regression (two cases; 8%), and ulceration (one case; 4%) were comparatively rare phenomena in this series of melanomas.

Figure 1.

Melanoma arising in association with common and cellular blue nevus on the scalp of a 49-year-old woman (case 16). (a) Shave biopsy specimen of a pigmented lesion reported to have been present for years on the scalp of a 49-year-old female reveals epidermal ulceration and a proliferation of cytologically benign spindled and dendritic melanocytes throughout the subjacent dermis (H&E, × 40). (b) Higher power examination of the lesion reveals attributes of a proliferation of benign spindled melanocytes (H&E, × 400). (c) Excision of the lesion site 2 months later reveals an asymmetric, distinctly biphenotypic proliferation of markedly atypical melanocytes forming an expansile and infiltrative nodule in the dermis with extension into the subcutaneous adipose tissue to a depth of ~10.8 mm arising in association with a benign blue nevus component (H&E × 20). (d) Common and cellular blue nevus components are associated with the melanoma and exhibit a benign spindled morphology similar to that seen in the original shave biopsy specimen with a plate-like and dumbbell-shaped architecture (H&E, × 100; inset; H&E × 200). (e) Cytologically malignant melanocytes proliferate adjacent to the blue nevus component and form an expansile nodule with central necrosis. These cells exhibit an epithelioid morphology with increased pale amphophilic cytoplasm and enlarged oval nuclei with prominent nucleoli. Mitotic figures are present (H&E, × 100; inset: × 400).

Figure 2.

Melanoma arising in association with a common blue nevus in the perianal soft tissue of a 67–year-old man (case 24). (a) Excisional biopsy specimen of a pigmented lesion arising in the perianal soft tissue of a 67–year-old male reveals a distinctly biphenotypic melanocytic proliferation comprising a centrally located nodular proliferation of atypical melanocytes with a surrounding less cellular spindled melanocytic proliferation. (H&E, × 20). (b) Higher power examination of the lesion at the border of the biphenotypic proliferation reveals benign spindled melanocytes abutting a more cellular proliferation of atypical epithelioid melanocytes (H&E, × 100). (c) Higher power examination of the spindled area reveals a proliferation of benign spindled melanocytes associated with dense fibrosis and pigmented macrophages (H&E, × 200). (d) Cytologically malignant melanocytes forming an expansile nodule. These cells exhibit an epithelioid morphology with increased pale amphophilic cytoplasm and enlarged oval nuclei with prominent nucleoli. Mitotic figures are present (H&E, × 100; inset: × 400).

In the cases for which we were able to assess cytomorphologic features, the tumor cells ranged from purely malignant epithelioid melanocytes in eight cases (33%) to a combination of malignant epithelioid and spindled melanocytes in 12 cases (50%). Twenty-three cases contained an unequivocal benign nevus component: six cases with common blue nevus (25%), seven cases with cellular blue nevus (29%), nine cases with both common and cellular blue nevus (38%), and one case with blue nevus/neurocristic hamartoma (4%). In one case (4%), the melanoma was present in association with an atypical cellular blue nevus. The benign component ranged from focal (≤20% of the lesion) and present only at the periphery of the lesion (14 cases), to comprising ~20–50% of the lesion (seven cases), to comprising ≥50% of the proliferation with only a focal malignant component (in three cases; most prominent in case 21). The transition between the benign and malignant components was typically abrupt. The presence of the benign blue nevus areas in our case series was critical in the distinction of primary tumors from metastases.

Molecular Data

Tumors from four patients were analyzed for mutations in the BRAF gene; all were wild type for BRAF. An additional five samples were tested for GNAQ mutation; all cases were wild type for GNAQ. One sample was tested for both BRAF and GNAQ mutations and was wild type for both genes.

Clinical Follow-up and Outcome

Clinical follow-up was available for 21 patients. Overall, follow-up times ranged from 0.04 to 11.6 years with a median of 2.1 years.

Fifteen patients (15/24: 63%) underwent sentinel lymph node biopsy. Among these, four (4/15: 27%) had negative sentinel nodes and did not develop subsequent regional or distant metastasis (mean 1.62 years; range: 0.04–5.2 years). Six patients (6/15: 40%) initially had negative sentinel lymph nodes, but later developed either regional metastasis (one patient (1/15: 7%) with microscopic satellitosis concurrent with negative sentinel lymph node biopsy) or subsequent distant metastases, five patients (5/15: 33%): involving the lung, liver, mediastinal lymph nodes, and subcutaneous soft tissue in order of frequency; mean 3.2 years; range: 0.1–10.0 years). Finally, sentinel lymph node involvement was detected for five patients (5/15: 33%); among these, two (2/15:13%) subsequently developed additional regional metastases (involving the regional lymph nodes; 0.1 years later each), and three patients (3/15: 20%) developed both regional lymph node and distant metastases (involving the liver, renal parenchyma, and subcutaneous soft tissue; mean 0.3 years to first event; range: 0–1.0 years after diagnosis).

An additional six patients (6/24: 25%) did not undergo sentinel lymph node sampling but had follow-up information available. Among these, two patients (2/6: 33%) developed neither local recurrence nor regional or distant metastasis; three patients (3/6: 50%) developed only distant metastases (mean 2.9 years; range: 0–8.5 years); and one patient (1/6: 17%) developed both regional and distant metastases (0.7 years from diagnosis).

Distant metastatic sites most commonly included the liver (8/12: 67%); the lung (4/12: 33%); subcutaneous tissue (3/12: 25%); bone (2/12: 17%); kidney (2/12: 17%); distant lymph nodes (2/12: 17%); abdomen (1/12: 8%); spleen (1/12: 8%); adrenal glands (1/12: 8%) and retroperitoneum (1/12: 8%). Seven patients (29%) had died at the time of our study; among these 6 (25%) died of disease with documented distant metastasis. The cause of death for one patient was unknown.

Statistical Analysis and Risk Stratification

Detailed results from univariate Cox proportional hazards regression models for time to distant metastasis and recurrence-free survival are presented in Table 3 and Table 4, respectively. These demonstrate a significant association between tumor thickness and recurrence-free survival (hazard ratio=1.02 per mm; P=0.04) and reduced time to distant metastasis (hazard ratio=1.03 per mm; P=0.02) with a similar trend toward reduced time to local recurrence (hazard ratio=1.02 per mm; P=0.07). No other parameter (including age, anatomic location, mitotic figures, lymphovascular or perineural invasion, necrosis, or type of associated blue nevus) emerged as significant. None of the covariates of interest were significantly associated with overall survival (data not shown).