Combined use of SOX10 and GATA3 in mammary carcinoma

The complementary coverage of different subsets of breast cancer by GATA3 and SOX10 makes their use in combination appealing for routine clinical use, but study of these markers has been largely limited to cases with high or absent ER expression. Here we report SOX10 and GATA3 immunostaining in parallel using a tissue microarray containing 246 invasive breast carcinoma cases with a range of ER expression. GATA3 and SOX10 were positive in 93 % (229/246) and 15 % (38/246) of cases overall and in 63 % (24/38) and 74 % (28/38) of triple negative breast carcinomas (TNBC), respectively; SOX10 was positive in 15 of the 17 cases that lacked GATA3 expression (88 %). SOX10 was also positive in 3 % (6/196) of ER + cases, including 50 % of cases with low ER (3/6), 20 % with intermediate ER (3/15), and 0 % with high ER (n = 175), so that ER-low cases more strongly resembled TNBC than those with high ER expression. GATA3 expression was lower in cases that co-expressed SOX10 in comparison to those that were positive for GATA3 alone. Less than 1 % (2/246) of cases were negative for both GATA3 and SOX10. Therefore, SOX10 is a useful adjunct to GATA3 in the detection of TNBC and cases with low ER expression and/or reduced GATA3 intensity relative to that typical of breast cancers with higher ER expression. Moreover, given such high sensitivity, metastatic tumors lacking either GATA3 or SOX10 are unlikely to be of breast origin. Additional study is necessary to determine the extent to which SOX10 may also improve specificity and to characterize its biologic significance in breast cancers with low ER expression.