Background: Anorectal malignant melanoma is a rare tumor with an extremely poor prognosis. DNA flow cytometric study as well as detailed immunohistochemical study have not been reported previously.
Methods: Eighteen cases of anorectal melanoma were studied, including immunohistology for melanoma markers and epithelial markers and DNA flow cytometric study of paraffin blocks.
Results: Most patients were Ashkenazi Jews, compared with Sephardi Jews and Arabs. Of the 17 patients followed, 14 died of disease at 4-39 months from presentation. Three patients were alive with disease at 12, 53, and 72 months of follow-up. Tumor thickness ranged from 3-35 mm (mean, 12.8 mm). The 2 long term survivors had tumor thickness < or = 7 mm. No correlation was found between the mode of primary surgical treatment (8 patients: abdominoperineal resection; 10 patients: local excision) and outcome. Vimentin, HMB-45, and S-100 protein stainings were positive in 18, 17, and 15 tumors, respectively. Polyclonal carcinoembryonic antigen (CEA), broad-spectrum cytokeratin, epithelial membrane antigen, monoclonal CEA, and TAG-72 (B72.3) stainings were positive in 13, 3 (only focal and rare staining), 2, 0, and 0 tumors, respectively. Thirteen tumors had adequate material for DNA analysis, and all were DNA aneuploid. S-phase fraction could be assessed in 11 tumors and ranged from 7.7-24% (mean, 14%). An S-phase fraction of < 10% was observed in the 2 long term survivors.
Conclusions: Anorectal melanoma in this study carried a grave prognosis. The frequent staining for polyclonal CEA (with negative monoclonal CEA staining) was probably due to nonspecific cross-reacting antigens. The occasional staining for epithelial markers warrants a comprehensive immunohistochemical study to ensure a correct diagnosis, especially in small biopsies of amelanotic undifferentiated tumors that lack junctional changes. The aneuploidy of all tested tumors reflected their highly malignant behavior. A trend toward longer survival was observed in patients with thin tumors and an S-phase fraction of < 10%. However, due to the small number of survivors, the latter observation should be further tested in a larger scale series.