GATA3 has been recognized as a promising marker for primary urothelial carcinoma (UC), consistently showing higher expression levels than urothelial markers thrombomodulin and uroplakin III. However, expression of GATA3 in comparison with UC-associated markers CK7 and p63 has not been systematically studied. Moreover, no studies have been conducted to establish GATA3 sensitivity in regional metastases. In this study, high-density tissue microarrays were constructed from 69 matched paired primary and metastatic bladder tumors including pure urothelial UCs with papillary (n=48) or flat phenotype (n=9), mixed tumors with micropapillary, glandular, small cell, squamous, giant cell, and plasmacytoid features (n=9), and 3 adenocarcinomas. GATA3 was expressed in 62/69 (90%) primary UC and 64/69 (93%) metastases, with significantly higher staining intensity in nodal metastases (P=0.03). In primary tumors, GATA3 was positive in 44/48 (92%) papillary UCs, 9/9 (100%) flat UCs, 8/9 (89%) mixed UCs, and 1/3 (33%) adenocarcinomas, whereas in metastases these numbers were 45/48 (94%), 9/9 (100%), 8/9 (89%), and 2/3 (67%), respectively. The majority of positive cases showed strong diffuse nuclear reactivity: 75% of primary UCs and 79% of metastases. GATA3 sensitivity in primary and metastatic UCs was comparable to that of CK7 and superior to that of p63 (P<0.05). GATA3 specificity was computed in comparison with its morphologic mimics expressing CK7 and p63, including 208 primary and 24 metastatic tumors from the lung, cervix, and head and neck regions. Strong GATA3 expression was present in 2/51 (4%) cervical carcinomas, whereas weak GATA3 expression was present in 7/51 (14%) cervical, 6/74 (8%) head and neck cancers, and 2/83 (3%) lung carcinomas. Remaining 191 primary and 24 metastatic tumors were GATA3 negative. Therefore, specificity of GATA3 calculated on the basis of morphologic and immunophenotypic UC mimics from lung, cervix, head and neck was 92%. Our findings demonstrate high sensitivity and specificity of the GATA3 diagnostic marker, with not only maintained but increased expression in regional metastases.
*Department of Pathology, University of Chicago, Chicago, IL
Departments of †Biostatistics
‡Pathology, University of Washington, Seattle, WA
Presented in part at the United States and Canadian Academy of Pathology (USCAP) 2013 Annual Meeting, Baltimore, MD, March 4-5, 2013.
Conflicts of Interest and Source of Funding: The authors have disclosed that they have no significant relationships with, or financial interest in, any commercial companies pertaining to this article.
Correspondence: Maria S. Tretiakova, MD, PhD, Department of Anatomic Pathology, University of Washington, NJB220, Box 359791, 325 9th Ave, Seattle, WA 98104 (e-mail: firstname.lastname@example.org).