Immunohistochemistry as a diagnostic aid in the evaluation of ovarian tumors

Aspects of immunohistochemistry (IHC), which are useful in the diagnosis of ovarian tumors (mostly neoplasms but also a few tumor-like lesions), are discussed. The topic is first approached by considering the different growth patterns and cell types that may be encountered. Then a few other specific situations in which IHC may assist are reviewed. Selected findings largely, or only, of academic interest are also mentioned. One of the most common situations in which IHC may aid is in the evaluation of tumors with follicles or other patterns which bring a sex cord-stromal tumor into the differential. The distinction between a sex cord tumor and an endometrioid carcinoma with sex-cord-like patterns may be greatly aided by the triad of epithelial membrane antigen (EMA), inhibin, and calretinin, the latter two being typically positive and EMA negative in sex cord tumors, the converse being typical of endometrioid carcinoma. It should be emphasized that granulosa cell tumors may be inhibin negative and, albeit less specific, calretinin is more reliable in evaluating this particular issue. Lack of staining for inhibin and calretinin may also be supportive in leading to consideration of diverse other neoplasms that may form follicles, including metastatic tumors as varied as carcinoid and melanoma. The well-known staining of the latter neoplasm for S-100 protein and HMB-45 may be very helpful in evaluating melanomas with follicular or other unusual patterns, a challenging aspect of ovarian tumor interpretation. The most common monodermal teratoma, struma ovarii, usually has an overt follicular pattern and is easily recognized, but recognition of unusual appearances ranging from oxyphilic to clear cell to various patterns of malignant struma may be greatly aided by a thyroglobulin or TTF 1 stain. IHC for neuroendocrine markers may assist in the diagnosis of primary and metastatic carcinoid tumor. The broad differential diagnosis of glandular neoplasms with an endometrioid-pseudoendometrioid morphology, or mucinous cell type, has been the subject of much exploration in recent years, particularly the distinction between primary and metastatic neoplasms. The well-known CK7 positive, CK20 negative phenotype of primary endometrioid carcinoma, and the converse profile in most metastatic large intestinal adenocarcinomas with a pseudoendometrioid morphology, has been much publicized but albeit an appropriate supportive adjunct in many cases, exceptions from the typical staining pattern may be encountered. It is even less helpful in the case of primary versus metastatic mucinous neoplasia. Evaluation of the expression of mucin gene products has shown mixed, essentially unreliable, results. Experience with other new markers, such as CDX-2, villin, beta catenin, and P504S (racemase), is limited but is in aggregate promising with regard to providing some aid in this area. The rare differential of metastatic cervical adenocarcinoma versus primary ovarian mucinous or endometrioid carcinoma may be aided by strong p16 staining of the former. Staining for alpha-fetoprotein may aid in confirming the diagnosis of endometrioid-like (and hepatoid) variants of yolk sac tumor. Ependymoma of the ovary may also have an endometrioid-like glandular pattern, but positive stains for glial fibrillary acidic protein contrast with the negative results in others neoplasms with a similar pattern. Immunostains may be very helpful in the evaluation of oxyphilic tumors and tumor-like lesions and in some unusual forms of clear cell neoplasia, such as clear cell struma, both subjects being reviewed herein. Immunostains may highlight both the presence and extent of epithelial cells in a variety of circumstances, including microinvasive foci in cases of serous borderline tumors and mucinous carcinomas, and in determining the extent of carcinoma cells and reactive cells within mural nodules of mucinous neoplasms. As in tumor pathology in general, various markers may be crucial in the diagnosis of small round cell tumors of the ovary, and familiar markers of epithelial, lymphoid, leukemic, and melanocytic neoplasms may assist in the analysis of high grade tumors with a poorly differentiated carcinoma, lymphoma-granulocytic sarcoma, malignant melanoma differential. The evaluation of ovarian cystic lesions may be aided by thyroglobulin or TTF 1 (cystic struma), glial fibrillary acid protein (ependymal cysts), and inhibin-calretinin (follicle cysts and unilocular granulosa cell tumors). Stains for trophoblast markers may occasionally aid in the evaluation of germ cell tumors, although routine stains should usually suffice; they may be of academic interest in confirming trophoblastic differentiation in some high grade surface epithelial carcinomas.