Diagnostic utility of WT1 immunostaining in ovarian sertoli cell tumor.

WT1, the Wilms tumor gene product, can be expressed in various tumors from different anatomic sites, including some types of ovarian tumors. Regarding the latter, most studies have focused on surface epithelial-stromal tumors in which serous carcinomas are usually positive and endometrioid carcinomas are negative. Very few studies have specifically investigated this marker in ovarian sex cord-stromal tumors; however, limited data in the literature suggest that WT1 may be frequently expressed in sex cord-stromal tumors. As pure Sertoli cell tumor can be in the histologic differential diagnosis of endometrioid tumors (particularly borderline tumor and carcinoma) and carcinoid, immunostaining for WT1 might be of diagnostic value. Immunohistochemical staining for WT1 was performed in 108 ovarian tumors: pure Sertoli cell tumor (n=26), endometrioid borderline tumor (n=25), classic well-differentiated endometrioid carcinoma (n=23), sertoliform endometrioid carcinoma (n=12), and carcinoid (n=22). Additionally, inhibin and calretinin immunostaining were performed in all cases of Sertoli cell tumor for purposes of comparing expression with WT1. Extent of immunostaining was scored on a 0 to 4+ semiquantitative scale, and immunohistochemical composite scores based on a combination of extent and intensity of immunostaining were calculated in positive cases (possible range, 1 to 12). Nuclear expression of WT1 was present in 96% of Sertoli cell tumors, 16% of endometrioid borderline tumors, 13% of classic well-differentiated endometrioid carcinomas, 25% of sertoliform endometrioid carcinomas, and 0% of carcinoids. In Sertoli cell tumors, expression was diffuse (>50% of positive cells) in all positive cases. When positive in the non-Sertoli cell tumors, the extent of expression tended to be focal to patchy (50% or less positive cells). In Sertoli cell tumors, inhibin and calretinin were expressed in 96% and 54% of cases, respectively. The extent of expression of inhibin tended to be diffuse, similar to WT1; however, the extent of immunostaining for calretinin tended to be focal to patchy. The immunohistochemical composite scores for WT1, inhibin, and calretinin were 11.2, 7.6, and 4.8, respectively. Coordinate patterns for the extent of expression of WT1, inhibin, and calretinin in pure Sertoli cell tumor showed that all 3 markers were positive in 54% of cases; however, 42% were positive for WT1 and inhibin but negative for calretinin. In cases positive for both WT1 and inhibin, expression of both markers was diffuse in 84% of cases, but WT1 was diffuse while inhibin was focal to patchy in 16% of cases. We conclude that ovarian Sertoli cell tumor should be added to the growing list of WT1-positive tumors. This marker is useful for the distinction of Sertoli cell tumor from endometrioid tumors and carcinoid. The diagnostic utility of WT1 in Sertoli cell tumor is similar to inhibin but better than that of calretinin.