Journal Logo

Letters to the Editor

Re-evaluation of “Polymorphous Sweat Gland Carcinoma”

Goto, Keisuke MD

Author Information
The American Journal of Dermatopathology: September 2019 - Volume 41 - Issue 9 - p 695-697
doi: 10.1097/DAD.0000000000001249
  • Free

To the Editor:

“Polymorphous sweat gland carcinoma” was originally proposed in 19941 as a low-grade malignant adnexal tumor. The original article revealed 9 cases of this tumor type occurring in 3 men and 6 women aged between 42 and 70 years. Interestingly, 8 of the 9 tumors developed on extremities while only one was on the patient's lower mid-abdomen. These tumors were well-demarcated, 3.8- to 5.0-cm–sized lesions with a variety of histopathologic growth patterns, including solid, trabecular, tubular, cribriform, pseudopapillary, and cylindromatous, and they also exhibited prominent stromal changes including hemorrhage, hyalinization, and cystic change. The cells of these tumors displayed moderate cytological atypia and mitoses. Two of 6 followed-up cases showed local recurrence and 1 regional lymph node metastasis. However, except for 3 cases in 2 English-language articles,2,3 few new cases had been reported until 2017. In addition, it has been indicated that the histopathologic features of these tumors are very similar to, and possibly in a similar or the same spectrum as, adenoid cystic carcinoma.4,5 Recently, a report presented 3 new cases named “polymorphous sweat gland carcinoma” in the American Journal of Dermatopathology.6 The authors showed the histopathologic, immunohistochemical, and genetic data of these 3 cases, thereby demonstrating that the tumors represent a distinct entity from adenoid cystic carcinoma. Moreover, the authors revealed a characteristic immunoprofile, including diffuse positivity of p16, strong positivity of CD56, and focal positivity of chromogranin A. The authors suggested that these immunohistochemical markers could be effective for distinguishing “polymorphous sweat gland carcinoma” from other tumor types. However, I believe these cases are actually adenoid-type (a.k.a. adenoid cystic type) basal cell carcinoma with focal carcinoid-like pattern rather than “polymorphous sweat gland carcinoma.”

Regarding their clinical aspects, all 3 cases occurred in the head and neck area (2 cases were on the neck and 1 was on the ear), whereas the extremities are known to be the most common site (10/12) of “polymorphous sweat gland carcinoma”; the cases of the head and neck have never been demonstrated among the previously reported 12 cases in English-language literature.1–3 However, head and neck sites are rather common for basal cell carcinoma.

Histopathologically, the features of case 1, presented in Figures 2 and 3 in the article,6 and case 2, in Figure 4, fit the diagnosis of adenoid-type basal cell carcinoma with focal carcinoid-like pattern. These tumors exhibit a definite palisade cell arrangement at the peripheries of the tumor nests, as shown in Figures 2, 3A, B, and 4F of the article,6 with prominent mucin deposition in and around the tumor nests, as shown in Figures 3A–E and 4A–C, F of the article.6 The cleft-like spaces between the tumor nests and the surrounding stroma because of stromal retraction can also be seen in Figures 3D and 4A, F of the article.6 The tumors also form mucin-filled pseudoglandular spaces in Figures 3A, B, E and 4A, C, F of the article,6 with “pseudoglands” confirmed by the negativity of epithelial membrane antigen (EMA) and carcinoembryonic antigen (CEA) immunohistochemical staining. These features do not fit any sweat gland tumors, but they exactly match basal cell carcinoma.

In immunohistochemistry, all cases showed the same characteristic profiles as that of basal cell carcinoma. The positivity for cytokeratin AE1/AE3, cytokeratin 5/6, p40, p63, and BerEP4, and the negativity for EMA, CEA, CD117, S100 protein, and DOG1 are compatible with the immunohistochemical profile of basal cell carcinoma, rather than being suggestive of sweat gland skin tumors.7–9 Strangely, in the original article,1 EMA, CEA, and S100 protein were positive in all the 9 tumors, supporting the diagnosis of sweat gland tumors instead. Moreover, basal cell carcinoma often immunoexpresses CD56 (9/10 cases, 90%) (Fig. 1) and chromogranin A (7/10 cases, 70%) but does not express synaptophysin (0/10, 0%),10 as shown in the 3 cases from the article.6 In addition, basal cell carcinoma often exhibits p16 positivity in various degrees (Fig. 1). The article emphasized the strong and diffuse positivity for p16 in their cases; however, Figure 5C actually shows a lack of nuclear p16 expression in many tumor cells, with prominent p16 expression in several stromal cells.6 This means that the p16 staining might be nonspecific and unreliable. Therefore, the immunohistochemical profile of the 3 cases is likely to be same as that of basal cell carcinoma.

An adenoid-type basal cell carcinoma shown at a magnification of ×2 (A–C) and ×20 (D–F). Hematoxylin and eosin stain (A and D) and stains for CD56 (B and E) and p16 (C and F).

Moreover, in molecular aspects, the study failed to detect human papilloma virus by DNA in situ hybridization or MYB-NFIB gene fusion by fluorescent in situ hybridization. However, because basal cell carcinoma exhibits neither MYB-NFIB fusion nor human papilloma virus infection, these analyses would be useless to rule out the possibility of basal cell carcinoma.

In conclusion, the cases in the article,6 at least case 1 and case 2 of the article, called “polymorphous sweat gland carcinoma,” are presumably adenoid-type basal cell carcinomas, although case 3 in the article did not show any histopathologic images but exhibited a similar immunohistochemical profile to the other 2 cases. The previously reported 12 cases have not been well evaluated yet; however, the tumors of this group might actually be classifiable as other tumor types including adenoid-type basal cell carcinomas or relatively circumscribed adenoid cystic carcinomas with predominant solid and/or trabecular components. Further cases are required to better evaluate the mysterious entity of “polymorphous sweat gland carcinoma.”


1. Suster S, Wong TY. Polymorphous sweat gland carcinoma. Histopathology. 1994;25:31–39.
2. Ronnen M, Ben-Dor D, Huszar M. Recurrent polymorphous sweat gland carcinoma of the skin. J Am Acad Dermatol. 2002;46:914–916.
3. Walker A, Mesinkovska NA, Emanuel PO, et al. Polymorphous sweat gland carcinoma: a report of two cases. J Cutan Pathol. 2016;43:594–601.
4. Kazakov DV, Michal M, Kacerovska D, et al. Lesions with prominent apocrine and eccrine differentiation. In: Kazakov DV, Michal M, Kacerovska D, et al, eds. Cutaneous Adnexal Tumors. Philadelphia, PA: Wolters Kluwer Health; 2012:1–171.
5. Goto K. Other rare malignant sweat gland tumors. In: Ansai SI, Goto K, eds. Atlas of Cutaneous Adnexal Tumors. Tokyo, Japan: Igaku-shoin; 2018:88–93 [in Japanese].
6. Ronen S, Aguilera-Barrantes I, Giorgadze T, et al. Polymorphous sweat gland carcinoma: an immunohistochemical and molecular study. Am J Dermatopathol. 2018;40:580–587.
7. Fan YS, Carr RA, Sanders DS, et al. Characteristic Ber-EP4 and EMA expression in sebaceoma is immunohistochemically distinct from basal cell carcinoma. Histopathology. 2007;51:80–86.
8. Goto K. Immunohistochemistry for CD117 (KIT) is effective in distinguishing cutaneous adnexal tumors with apocrine/eccrine or sebaceous differentiation from other epithelial tumors of the skin. J Cutan Pathol. 2015;42:480–488.
9. Goto K. The role of DOG1 immunohistochemistry in dermatopathology. J Cutan Pathol. 2016;43:974–983.
10. Goto K. Under-recognized immunoexpression of “neuroendocrine markers” and “myoepithelial markers” in basal cell carcinomas: does it indicate true neuroendocrine and myoepithelial differentiation? J Cutan Pathol. 2017;44:991–993.
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.