Attenuated podoplanin staining in breast myoepithelial cells: A potential caveat in the diagnosis of lymphatic invasion

Although podoplanin expression in breast myoepithelial (ME) cells has been reported to be readily distinguishable from that of lymphatic endothelium (LE), we recently encountered a case in which attenuated expression led to an incorrect interpretation of lymphatic invasion. This prompted us to further analyze the pattern and intensity of podoplanin expression in LE and in ME cells surrounding both non-neoplastic and neoplastic breast epithelium in 42 cases of breast carcinoma. In addition, cases with presumed lymphatic invasion (tumor cells within podoplanin-positive lymphatic-like structures) were further characterized on the basis of histologic review and results of additional endothelial (CD31 and/or CD34) and ME (p63±calponin) immunostains. Normal LE always displayed a strong (3+) linear podoplanin expression pattern, whereas ME cells surrounding non-neoplastic breast epithelium had a granular, branching membranous staining pattern that was either moderate (2+) or strong (3+) in 13 (34%) and 25 (66%) foci, respectively. ME cells surrounding ductal carcinoma in situ (DCIS) displayed weaker (1+ to 2+) podoplanin expression with 26 (72%) foci showing only a residual thin/discontinuous pattern of expression, whereas the other 10 (28%) foci showed a pattern similar to that around non-neoplastic epithelium. Further evaluation of 10 foci of presumed lymphatic invasion confirmed the presence of lymphatic invasion in 6 (60%) cases, whereas 4 displayed surrounding ME cells and were thus interpreted as DCIS. One case, in which both DCIS and lymphatic invasion were unequivocally present, displayed a few nests of tumor cells surrounded by both podoplanin-positive and p63-positive cells in different areas of the periphery and could not be readily classified. These findings represent an important caveat in the diagnosis of lymphatic invasion and, accordingly, one should always interpret results of podoplanin staining in the context of the histologic appearance and/or consider evaluation of additional endothelial or ME immunostains, especially when the characteristically strong linear pattern of expression of LE is not evident.