Comparison of Primary and Metastatic Malignant Melanoma of the Esophagus: Clinicopathologic Review of 10 Cases

It has been postulated that during early embryogenesis, melanoblasts migrate from the neural crest to various sites, including the epidermis, hair follicles, oral cavity, uvea, leptomeninges, and inner ear, which accounts for the development of primary malignant melanoma in these sites.1 The presence of esophageal melanocytes may be attributed to aberrant migration because the esophagus normally lacks melanoblasts.2-4 Primary esophageal melanoma (PEM), associated with melanocytosis of the esophagus, has been reported.5

Primary esophageal melanoma is an extremely rare tumor and accounts for only 0.1% to 0.2% of all primary esophageal malignancies.6-8 Similarly, esophageal metastasis of malignant melanoma is also uncommon. In 1964, Dasgupta and Brasfield9 reviewed 100 autopsy cases of metastatic melanoma to gastrointestinal tract and found only 4 patients who had esophageal metastases. Small bowel is the most common gastrointestinal site for metastatic melanoma, comprising 70% of all gastrointestinal metastasis.9 Because of the rarity of this disease, the literature on malignant melanoma of esophagus is limited, to our knowledge, to only a few small series and case reports that focused primarily on the clinicopathologic features of primary esophageal melanoma.10-13 There is no report in the literature that compares the distinguishing clinical and histopathologic features between PEM and metastatic melanoma of the esophagus. Thus, despite some of the established criteria based on the previous studies, the diagnosis of a primary esophageal melanoma remains challenging, and it is very difficult to definitively rule out a metastasis based on histology alone. For these reasons, we decided to study 10 cases of malignant melanoma of the esophagus, including 5 PEMs and 5 metastatic melanomas at our institution and to compare the clinicopathologic features between these 2 groups. The results from our study will help to further define the diagnostic criteria for PEM.

MATERIALS AND METHODS

Ten cases of esophageal melanomas, including 4 surgically resected specimens (3 primary and 1 metastatic), 2 autopsy cases (1 primary and 1 metastatic), and 4 cases reported on mucosal biopsies (1 primary and 3 metastatic), were identified in the files at the University of Texas, M. D. Anderson Cancer Center (Houston), covering a period of 49 years from 1957 to 2006. All 5 cases of PEM were reported based on clinicopathologic correlation with no history of melanoma. Formalin-fixed, paraffin-embedded tissue blocks were available from all cases except for 1 autopsy case of PEM (case 4). The institutional review board at the University of Texas M. D. Anderson Cancer Center approved this study.

Immunohistochemical studies for pancytokeratin, S100, and HMB-45 were performed on all cases except case 4. For case 4, melanin stain had been performed, and the tumor was positive for melanin pigment. The antibodies and dilutions used were S100 (1:40; 15E2E2, BioGenex, San Ramon, Calif) and HMB-45 (1: 50; HMB-45, Dako, Carpinteria, Calif). The pancytokeratin cocktail consisted of a mix of 4 different antibodies against cytokeratins: AE1/AE3 (1:50), CAM 5.2 (1:50), MNF116 (1:50), and 8/18 (SPM141, 1:25). Standard immunohistochemical staining techniques were used as described by Hsu et al.14 Appropriate positive and negative controls were used. The immunohistochemical staining results were categorized as positive or negative. All slides, including the hematoxylin-eosin and immunohistochemical-stained slides were reviewed jointly by a dermatopathologist (V.G.P.) and a gastrointestinal pathologist (H.W.). The gross features were obtained from pathology or autopsy reports. The following features were recorded: anatomic location, gross configuration, tumor size (centimeters), depth of invasion, predominant types of cytology (epithelioid, spindle, or mixed), modified Breslow thickness (millimeters), the presence or absence of an in situ component, radial growth phase, lymphovascular invasion, regional lymph node metastasis, and satellitosis. The radial growth phase was defined as the presence of intraepithelial proliferation of atypical melanocytes (in situ melanoma) beyond 3 epithelial ridges from the invasive component. The modified Breslow thickness was measured from the top of the mucosal surface to the deepest tumor cells.

Clinical information and follow-up data were obtained from review of patients' medical records and from the US Social Security Death Index (SSDI; http://ssdi.rootsweb.ancestry.com). Overall, survival was calculated from the time of diagnosis of esophageal melanoma to the time of death or to the time of last follow-up, at which point, the data were censored. Statistical analysis was performed using SPSS software (version 12 for Windows, SPSS, Chicago, Ill). The survival curves were constructed using the Kaplan-Meier method, and the log-rank test was used to evaluate the statistical significance of differences.

RESULTS

The clinical features of 5 patients with PEM and 5 patients with metastatic melanomas to esophagus are shown in Table 1. Among the 5 patients with PEM, 2 (40%) were men and 3 (60%) were women, with a median age at diagnosis of 62 years (range, 45-89 years). In comparison, 4 (80%) of 5 of the patients with metastatic melanoma were men, with a median age of 50 years (range, 30-67 years). All patients presented with dysphagia, except 1 patient, who had metastatic melanoma and presented with back pain from bone metastasis. The tumor was located in the distal third of the esophagus in 3 (60%) of 5 patients with PEM and in 4 (80%) of 5 patients with metastatic melanoma, in the middle third of the esophagus in 1 patient (20%) in each group, and in the proximal esophagus in 1 patient (10%) with PEM (Table 1). None (0%) of the 5 patients who had PEM had a history of melanoma or distant metastasis at the time of diagnosis other than regional nodal disease. In contrast, all (100%) 5 patients with metastatic melanoma to esophagus had a history of cutaneous malignant melanoma. The interval between the primary cutaneous melanoma and the metastatic tumor to the esophagus ranged from 11 months to 62 months (median interval, 50 months). All (100%) 4 patients with metastatic melanoma, whose follow-up information was available, had metastasis to other organs at the time of the diagnosis of esophageal metastasis (Table 1). One patient with PEM (case 4) and 1 patient with metastatic melanoma (case 9) were misdiagnosed as poorly differentiated squamous carcinoma on initial biopsies. The diagnosis of melanoma was confirmed at the time of autopsy (case 4) or on a subsequent biopsy (case 9).

Among the 10 patients, 4 (40%) underwent esophagogastrectomy (3 with PEM and 1 with metastatic melanoma), and autopsies were performed on 2 additional patients (1 with PEM and 1 with metastatic melanoma). The histopathologic features of these 6 cases are summarized in Table 2. Four (67%) of 6 tumors had a polypoid or exophytic configuration (Figure 1, A). In situ melanoma and radial growth phase was present in all 4 PEMs (100%; Figures 1, B, and 2, A and B). For case 2, in situ melanoma was present at the proximal esophageal resection margin, which was, grossly, 6.0 cm away from the tumor (Figure 1, B, inset). In addition, in situ melanoma was also present in the case of PEM reported on mucosal biopsy (case 5; Figure 3) and in the biopsy of case 4 but not in biopsies from the remaining 2 cases of PEM. In situ melanoma was, however, not present in any (0%) of the 5 patients with metastatic melanoma to esophagus. Of 4 patients with PEM, 2 (50%) had mixed epithelioid and spindle cell morphology (Figure 2, C and D), 1 (25%) had spindle cell morphology, and 1 (25%) had epithelioid morphology. In contrast, all (100%) 5 metastatic melanomas had only epithelioid morphology (Table 2), including 1 (20%) resected case, 1 (20%) autopsy case, and 3 (60%) reported on biopsies (cases 8, 9, and 10). It was interesting that 2 PEMs that had polypoid configuration (cases 1 and 2) showed invasion only into submucosa, and yet 1 of those cases (case 2) had metastasis in 3 (12%) of 25 regional lymph nodes. Satellitosis was present in only 1 PEM (case 1). Focal melanocytosis was present in 2 (50%) of 4 cases of PEM but not in the resected or the autopsy cases of metastatic melanoma to esophagus (Table 2). Lymphovascular invasion was present in 1 case of PEM evaluated but not in the other 3 PEMs or the 2 metastatic melanomas evaluated. Lymph node metastases were present in 3 (75%) of 4 PEMs documented and in 1 (50%) of 2 metastatic melanoma cases evaluated. Melanin pigments were present in all 10 cases.

Immunohistochemical studies were performed on 9 cases to confirm the diagnosis. S100 and HMB-45 were positive in 9 (100%) of 9 cases, and pancytokeratin was negative in all 9 cases (Figures 2 and 3). The in situ melanoma cells were also positive for S100 and HMB-45 and were negative for cytokeratin (Figures 2 and 3).

Patients with PEM had better survival rates (median survival, 24 months) compared with the patients diagnosed with metastatic melanoma to esophagus (median survival, 11 months; P = .03, log-rank test; Figure 4). All patients with metastatic melanoma to esophagus died within 1 year from the time of diagnosis of esophageal metastasis (Table 2). In contrast, 4 (80%) of 5 patients with PEM were alive at least 1 year after diagnosis (Table 2).

COMMENT

In this study, we compared the clinical and pathologic features of 5 cases of PEM and 5 cases of metastatic melanoma to esophagus to further define the criteria that could be used for distinguishing between the primary and metastatic melanoma of the esophagus. The clinicopathologic features that distinguish PEM from metastatic melanoma identified in this study were compared with those reported in the previously published series on esophageal melanoma and are shown in Table 3. The median age for patients with PEM was older than for those with metastatic melanoma to esophagus in this study. This was corroborated by the previous reports.10-13 Similar to previous reports (Table 3), we found that no history of melanoma was present in any of the 5 patients who were diagnosed with PEM.10,11,13 In contrast, a history of cutaneous melanoma was present in all 5 patients who had metastatic melanoma to esophagus. However, the difference in a history of melanoma between the PEM and metastatic melanoma groups in this study could be the result of the selection criteria used for PEM. All 5 patients with metastatic melanoma to esophagus had metastasis to at least one other organ at the time of diagnosis of metastatic melanoma to esophagus. Therefore, our data suggest that metastasis to esophagus is a late event during the disease progression of cutaneous melanoma and is often associated with metastasis to other organs at the time of the diagnosis. No distant metastasis was detected at the time of diagnosis in the 5 patients with PEM in this study or in the 6 patients in the Li et al study.13 However, distant metastasis was present in 4 (40%) of the 10 cases of PEM reported by Lohmann et al11 and 2 (33%) of the 6 cases reported by DiCostanzo and Urmacher.10 Thus, a history of malignant melanoma and metastasis to other anatomic sites or organs is important for distinguishing between PEM and metastatic melanoma to esophagus. In the absence of previous history, a careful physical examination of the skin, eyes, and other organ systems may be needed to rule out a metastatic melanoma from another anatomic site before making a final diagnosis of PEM.

The histopathologic criteria for diagnosing PEM have not been clearly established. The most challenging question is how to differentiate a primary esophageal melanoma from a metastatic lesion to the esophagus. In 1953, Allen and Spitz15 set forth the main diagnostic criteria for primary melanoma of skin and mucous membranes as the presence of a junctional or in situ melanoma component in the intact epithelium that overlays the invasive melanoma. As long as there is no ulceration of the overlying epithelium, the absence of junctional change (in situ melanoma) in the overlying epithelium may be interpreted as very strong evidence for metastasis or recurrence.15 Consistent with these criteria, we found that in situ melanoma was present in all 5 cases of PEM, and radial growth phase was present in all 3 resected cases and 1 autopsy case of primary lesions. Similarly, in situ melanoma was detected in 9 (90%) of the 10 cases reported by Lohmann et al11 and 5 (83%) of the 6 cases reported by DiCostanzo and Urmacher (Table 3).10 The in situ melanoma or radial growth phase were not present in any of the metastatic melanomas to esophagus in this study. It may, however, be difficult to assess this criterion from a small mucosal biopsy or from a tumor that is severely ulcerated. In fact, in situ melanoma was present in only 2 (40%) of the 5 primary cases in the initial mucosal biopsies in our study. Therefore, other histologic features and clinical information should be considered. In 1989, Sabanathan et al,12 proposed 2 additional criteria for primary esophageal melanoma: presence of esophageal melanocytosis and a diagnosis based on exclusion. Esophageal melanocytosis is characterized by the presence of increased melanocytes/pigmentation of the keratinocytes in the basal layer of esophageal squamous mucosa, and the melanocytes do not show nuclear or cellular atypia.16 Focal melanocytosis was present in 2 (66.7%) of the 3 resected PEM cases in this study and 2 (33%) of the 6 cases reported by DiCostanzo and Urmacher,10 but not in either the resected case or the autopsy case of metastatic melanoma in this study. Therefore, the presence of melanocytosis would be helpful in the differential diagnosis between PEM and metastatic lesions, particularly in the resection specimen. However, because of the focal nature of melanocytosis, as noted in this study and previous studies, adequate sampling of the uninvolved esophageal mucosa may be needed to document the presence or absence of esophageal melanocytosis. Furthermore, we found that mixed epithelioid and spindle cell morphology was present in 2 (50%) of the 4 primary esophageal melanomas but not present in the 2 metastatic melanomas. The mixed epithelioid and spindle cell morphology may be a helpful feature in the differential diagnosis between primary and metastatic melanoma. Given the limited number of cases examined in this study, additional studies may be needed to further confirm these findings.

Malignant melanoma, either esophageal primary or metastatic to esophagus, was also difficult to distinguish from other more common esophageal malignancies clinically and histologically. Grossly, melanoma can be pigmented or amelanotic and covered by squamous epithelium or show surface ulceration.17-20 Given the rarity of melanoma in esophagus and the deceptive histologic nature of melanoma, which often shows either epithelioid or sarcomatoid features, it is important to differentiate esophageal melanoma from poorly differentiated squamous carcinoma or adenocarcinoma, lymphoma, or sarcoma. In fact, 2 (20%) of the 10 patients in this study had been misdiagnosed as having poorly differentiated squamous carcinoma on initial biopsies. Similarly, melanoma of the esophagus was reported in only 1 (17%) of the 6 cases reported by Li et al13 and 3 (50%) of the 6 cases reported by DiCostanzo and Urmacher.10 Therefore, malignant melanoma should be considered when a poorly differentiated neoplasm is encountered on esophageal biopsies and confirmed by immunohistochemistry.

Most of the patients with PEM had metastasis either to regional lymph nodes or to other distant sites at the time of surgery or diagnosis. Four (80%) of 5 patients had regional lymph node metastasis, which was consistent with the previously reported10,11,13 70% to 83% metastasis in either regional lymph nodes or distant sites. The prognosis of patients with PEM is poor. However, the prognosis for patients with metastatic melanoma is worse, with a median survival of 11 months compared with 24 months for patients with PEM in this study. Our data are consistent with the previous reports17,18 that none of 11 patients with metastatic melanoma to esophagus survived for 1 year. The treatments for PEM include surgery, radiotherapy, and chemotherapy. Most patients undergo surgical resection with a combination of the other 2 modalities.21 Many patients with esophageal metastases benefit from palliative surgical treatment because of the obstructive nature of these lesions, often causing severe pain and life-threatening problems, such as hemorrhage and difficulty in swallowing. Decisions must be made regarding the probability of benefit versus the expectancy of survival because significant mortality and morbidity exist for this surgical procedure.22

In summary, melanoma of the esophagus is a rare and aggressive disease. The presence of in situ melanoma, radial growth phase, melanocytosis, and mixed epithelioid and spindle cell morphology, in the context of no prior history of melanoma, may help establish the diagnosis of PEM. This distinction is important because patients with PEM have a better prognosis than patients with metastatic malignant melanoma. Immunohistochemical study for melanoma markers and cytokeratin, careful searching for an in situ component, and review of clinical history are critical to the diagnosis.

© 2008 College of American Pathologists Provided by ProQuest LLC. All Rights Reserved.