The distinction between breast and müllerian carcinomas from each other and from tumors with a similar cytokeratin profile can be difficult. We tested the usefulness of 2 new markers, NY-BR-1 and PAX8, by staining a variety of breast and gynecologic carcinomas, along with tumors of pancreas, bile ducts, stomach, and gastroesophageal junction. NY-BR-1 expression (ie, H score >10) was seen in 58.4% of breast carcinomas (111/190), 5.6% of müllerian carcinomas (8/142), 7% of pancreatic tumors (1/15), 0% of cholangiocarcinomas (0/22), 0% of gastric tumors (0/36), and 0% of gastroesophageal carcinomas (0/25). All 188 breast carcinomas were negative for PAX8. PAX8 expression was seen in 72.4% of müllerian tumors (105/145). All pancreatic tumors (n = 15), cholangiocarcinomas (n = 23), and gastric (n = 35) and gastroesophageal junction (n = 25) carcinomas were negative for PAX8. Addition of NY-BR-1 and PAX8 in a panel would be useful in distinguishing breast cancer, gynecologic tumors, and tumors of the upper gastrointestinal tract.

There are a few studies that have evaluated a panel of stains on a single large data set of breast cancers, which is required for direct comparison between antibodies. The immunohistochemical panel in this study was chosen to include breast-specific markers and markers that are expressed in tumors resembling breast cancer. The individual marker positivity in decreasing order was 95% (177/186) for GATA-3, 92% (172/186) for cytokeratin (CK)7, 80% (151/189) for AR, 80% for estrogen receptor (158/198), 69% for progesterone receptor (137/198), 55% (105/190) for NY-BR-1, 52% (99/189) for mammaglobin, 31% (59/191) for vimentin, 26% (51/195) for GCDFP-15, 0.5% (1/186) for CK20, and 0% (0/188) for PAX-8. When tumors were categorized based on estrogen receptor and HER2 status; a total of 45 profiles were identified. In addition, some tumors showed an unconventional profile-although the majority of breast carcinomas were CK7-positive/CK20-negative, a CK7-negative/CK20-negative profile was seen in ∼8% of the cases. Such a profile can create confusion in investigation of a carcinoma of unknown origin. The results define the individual sensitivity of each marker and establish a baseline diagnostic profile of breast cancer in a large data set. In addition, the results support the use of immunohistochemical panel for confirming or determining breast as the source of metastasis.

No abstract