LAUSR

To the Editor, I read with interest the paper by Panse et al recently published in the Journal of Cutaneous Pathology about frequent patchy immunoexpression of the “neuroendocrine marker” CD56 in basal cell carcinomas (BCCs). In that study, 18 (90%) of 20 BCCs had patchy CD56 positivity (less than 70% of the tumor cells), whereas only 2 (10%) had diffuse CD56 positivity (more than 70% but less than 100% of the tumor cells). The authors concluded that the CD56 immunoexpression pattern could distinguish BCCs (patchy CD56 immunoexpression) from Merkel cell carcinomas (diffuse CD56 immunoexpression). In addition to agreeing with this conclusion, I would like to present additional information about the interesting immunoexpression of “neuroendocrine markers” in BCCs, as well as that of “myoepithelial markers.” Frequent CD56 immunoexpression in BCCs has been observed by myself (9 [90%] of 10 consecutive excised BCCs were positive; clone 1B6, dilution: 1:5; Nichirei, Tokyo, Japan; unpublished data) and Terada (63 [95%] of 66 BCCs were positive). BCCs often express chromogranin A, another “neuroendocrine marker,” as well. I observed chromogranin A immunostaining (polyclonal, dilution: 1:4; Nichirei) in 7 (70%) of the 10 BCCs noted above, and reported rates are 27% (18/66) and 72.2% (24/33). In general, neither marker was expressed in all tumor cells; instead, expression was usually patchy (Figure 1A-C). I observed CD56 positivity in 20% (5-80%) of the tumor cells and chromogranin A positivity in 10% (5-50%). Unlike CD56 and chromogranin A, synaptophysin, a third “neuroendocrine marker,” is rarely expressed in BCCs. Immunoexpression of synaptophysin was negative (0%, 0/10; clone 27G12, prediluted; Nichirei) in my BCC panel and infrequent (9%, 3/33) in a previous study. Synaptophysin was immunoexpressed in BCCs at a relatively high rate (18%, 12/66) in another study; however, this result should be interpreted with caution because the figures showed synaptophysin immunostaining in stromal cells, as well as weak synaptophysin staining in tumor cells. As discussed above, frequent patchy immunoexpression of CD56 and chromogranin A is a characteristic of BCCs; however, in general, the immunoexpression of these markers is not diffuse, and synaptophysin immunoexpression is disproportionately negative. Furthermore, in a previous electron microscopy study, the tumor cells of BCCs did not contain any definite neuroendocrine granules. Therefore, it is doubtful that BCCs undergo true neuroendocrine differentiation. The few reports of BCCs with neuroendocrine differentiation would simply be describing conventional BCCs. However, these “neuroendocrine markers” could aid the pathological diagnosis of BCC. Panse et al reported that CD56 is a useful marker for distinguishing BCCs from Merkel cell carcinomas. Moreover, I believe that CD56 and chromogranin A would also be helpful in distinguishing BCC from various other tumors such as Bowen disease, actinic keratosis, and squamous cell carcinoma. Actually, I often use them as “BCC markers” in my dermatopathology practice and note that I have not observed CD56 or chromogranin A immunoexpression in the more than 10 cases of the aforementioned tumors (unpublished data). However, it is important to avoid mistaking BCCs for neuroendocrine tumors including Merkel cell carcinomas when CD56 and chromogranin A immunostaining is positive. According to Panse et al, this can be accomplished by assessing the staining pattern of cytokeratin 5/6, along with that of CD56; BCCs diffusely express cytokeratin 5/6, whereas Merkel cell carcinomas do not. Cytokeratin 20 immunostaining might also aid the differential diagnosis; Merkel cell carcinomas express cytokeratin 20 with a cytoplasmic dot-like pattern, whereas BCCs do not. Similar to “neuroendocrine marker” immunoexpression, “myoepithelial marker” immunoexpression in the tumor cells of BCCs is also under-recognized. In one study, two “myoepithelial markers,” alphasmooth muscle actin (α-SMA) and calponin, were immunoexpressed in 17 (53%) and 6 (19%) of 32 BCCs, respectively; immunoreactivity was usually focal. I also observed the immunoexpression of α-SMA (clone 1A4, prediluted; DAKO, Glostrup, Denmark) in 7 (70%) of 10 BCCs (Figure 2A,B). In addition, p63 and pan-cytokeratin are diffusely positive in all BCCs. Thus, more than half of conventional BCCs could satisfy the immunohistochemical criteria for myoepithelial differentiation. There are several reports of myoepithelial differentiation of BCCs, especially those with plasmacytoid or signet-ring cytomorphologic Received: 7 June 2017 Revised: 9 July 2017 Accepted: 18 July 2017