OBJECTIVE: Confirmation of a breast origin for triple-negative breast cancer (TNBC) is sometimes problematic. The traditional breast markers GATA-binding protein 3 (GATA3), mammaglobin (MGB) and gross cystic disease fluid protein 15 (GCDFP15) have shown limitations in identifying TNBC. Here, we aimed to examine the diagnostic potential of the newly proposed TNBC marker, Sry-related high-mobility-group/HMG box 10 (SOX10).
RESULTS: We analysed and compared SOX10 expression with GATA3, MGB and GCDFP15 expression in a test cohort of 1838 invasive breast cancers (IBCs) by using tissue microarrays. The findings from the test cohort were further examined with a validation cohort of 42 TNBCs in whole sections. The overall expression rates of SOX10, GATA3, MGB and GCDFP15 were 6.9%, 83.1%, 47.0%, and 34.8%, respectively. Among the TNBCs within this cohort, the expression rates of SOX10, GATA3, MGB and GCDFP15 were 31.3%, 34.5%, 27.9%, and 25.2%, respectively. SOX10 was strongly associated with TNBC (P < 0.001), whereas all other traditional markers were associated with non-TNBC (P < 0.001 for all). In addition, SOX10 was more correlated to basal-like breast cancer (BLBC) (P = 0.001) than five-marker-negative subtype among the TNBCs. A high expression rate of SOX10 (81%) was confirmed in the validation cohort. Additionally, SOX10 expression was inversely correlated with GATA3 and GCDFP15 expression, so they may complement each other in TNBC detection. The SOX10-GATA3 combination yielded a sensitivity of 60.3% for TNBC detection in the test cohort.
CONCLUSIONS: SOX10 is a reliable marker for identifying TNBC, and complements GATA3. The SOX10-GATA3 combination may be used as a sensitive TNBC marker.

Mammary analogue secretory carcinoma of salivary gland (MASC) is a tumor with histopathologic and immunophenotypic features mimicking secretory carcinoma of the breast harboring the ETV6 split. The expression of mammaglobin, S-100, Ki-67, P63 and ETV6 split were detected in twelve cases of acinar cell carcinoma and fourteen cases of mammary analogue secretory carcinoma of salivary gland by immunohistochemistry and fluorescence in situ hybridization respectively. The expression of ETV6 gene split was detected in fourteen mammary analogue secretory carcinomas of salivary gland with positive expression of mammaglobin. Eight of mammary analogue secretory carcinomas of salivary gland also tested positive for the ETV6 gene split via fluorescence in-situ hybridization (FISH). The concordance rate of the immunohistochemistry and FISH was 72.3%. Mammaglobin and ETV6 gene split detection could help to distinguish mammary analogue secretory carcinoma of salivary gland. The mammary analogue secretory carcinoma of salivary gland specimens were also examined under transmission electron microscope. And apical junctional complexes were observed in the loosely connected tumor cells.

OBJECTIVE: Despite numerous studies on the utility of GATA-3 as breast cancer marker, its comparison with other breast markers, its concordance between primary and metastatic tumors and its expression in primary cancers from sites with frequent breast metastases remains unclear.
METHODS: To address these questions, totally 993 invasive breast cancers (IBC), 254 paired nodal metastases, 23 distant metastases, and 208 lung carcinomas were included. GATA-3 expression was analyzed by immunohistochemistry and compared to other breast markers [gross cystic disease fluid protein 15 (GCDFP-15) and mammaglobin (MGB)].
RESULTS: GATA-3 was expressed in 82.5% of IBC, predominantly in luminal (93.9%), and lower in non-luminal cancers [59.6% of HER2 overexpressing (HER2-OE) and 38.1% of triple negative breast cancer (TNBC) subtypes]. GATA-3 identified more IBC than GCDFP-15 (23.9%) and MGB (46.6%). However, MGB showed a comparable sensitivity for non-luminal cancers to GATA-3. Combining MGB and GATA-3 improved sensitivity for both HER2-OE (80.8%) and TNBC cases (55.4%). GATA-3 showed a high sensitivity for nodal metastases and distant metastases, with good concordance with primary tumors. GATA-3 was expressed in 1.0% of lung carcinomas, with sensitivity and specificity of 82.5 and 99.0% in differentiating IBC and lung carcinoma.
CONCLUSIONS: GATA-3 expression was the highest in luminal breast carcinomas, and showed higher sensitivity than GCDFP-15 and MGB. However, in the poorly differentiated IBC, its utility was still limited. One should be aware of the possible GATA-3 expression in lung carcinomas.

GATA3 has been recognized as the novel marker for identifying primary and metastatic breast carcinomas, consistently showing that GATA3 was significantly more sensitive than traditional markers gross cystic disease fluid protein 15 (GCDFP15) and mammaglobin (MGB). However, clinically useful groups of breast carcinomas status were not identified, which were determining appropriate treatment strategy, affecting the prognosis. In this study, we undertook a comparative study of the marker GATA3 and GCDFP15 and MGB in clinically useful groups of paired primary and metastatic breast cancer. We retrieved 64 cases of matched primary and metastatic breast cancer from the surgical pathology archive at our institution. According to the emerging 2015 St. Gallen Consensus, the clinically useful groups were divided into ER and/or PR (+), HER2 (-), abbreviated as A; ER and/or PR (+), HER2 (+), abbreviated as B; ER and PR (-), HER2 (+), abbreviated as C; ER, PR and HER2 (-), abbreviated as D; each group contained 16 cases (n=16). Tissue microarrays were created, with three 1-mm punch specimens from each case. The tissue microarrays were cut at 4-μm thickness and stained with monoclonal antibodies to GATA3, GCDFP15, and MGB. Staining intensity (0-3+) and extent (0%-100%) were scored with an H-score calculated (range, 0-300). Sensitivities by varying H-score cutoffs (any; ≥50; ≥150) for a positive result in the clinically useful groups of matched primary or metastatic breast cancer among GATA3, GCDFP15, and MGB. GATA3 was significantly more sensitive than GCDFP15 and MGB A and B groups (P<.05) rather than C and D groups (P>.05). However, GATA3 in conjunction with GCDFP15 and MGB detection could improve the sensitivity of C group (P<.05) rather than D group (P>.05). Significantly, good coincidence was observed between primary and metastatic tumor GATA3 expression (κ value = 0.826 >0.75) as compared with the coincidence of GCDFP15 (κ value =0.492 <0.75) and MGB (κ value =0.593 <0.75) (both P<.05). In conclusion, GATA3 expression did not show the same sensitivity for the clinically useful groups of breast cancer. GATA3 expression is positively correlated with ER-positive, PR-positive, and HER2-positive carcinomas. In addition, the matched primary and metastatic tumor expression of GATA3 shows good coincidence. We propose the careful selection of GATA3 for identifying hormone receptor negativity of breast cancer, especially in the case of triple-negative breast cancer.