This study aimed to assess the diagnostic value of vimentin expression in differentiating endometrioid adenocarcinoma of primary uterine corpus and ovarian origin. Immunohistochemical analyses for the expression of vimentin in tumoral epithelial cells were performed on 149 endometrioid adenocarcinomas wherein the primary sites were not in question, including whole tissue sections of 27 carcinomas of uterine corpus origin (and no synchronous ovarian tumor), 7 carcinomas of ovarian origin (and no synchronous uterine corpus tumor) and a tissue microarray (TMA) containing 91 primary uterine corpus and 24 primary ovarian carcinomas. We also assessed 15 cases that synchronously involved the uterine corpus and ovary, 15 cases of metastasis to organs/tissues other than uterine corpus or ovary as well as 7 lymph node metastases. Vimentin was negative in 97% (30/31) of primary ovarian carcinomas. In contrast, 82% (97/118) of primary uterine corpus carcinomas were vimentin-positive. Vimentin expression was discordant in 53% of synchronous tumors. The sensitivity and specificity of negative vimentin staining in predicting an ovarian primary were 97% and 82%, respectively, whereas parallel values for positive vimentin staining in predicting a primary uterine tumor were 82% and 97%, respectively. The pattern of vimentin expression in all cases was maintained in their respective regional lymph nodes and distant metastases. In conclusion, ovarian and uterine corpus endometrioid adenocarcinomas have different patterns of vimentin expression. If validated in larger and/or different data sets, these findings may have diagnostic value in distinguishing metastatic lesions from double primary tumors involving both sites.

Pathologic intraoperative consultation (IOC) is a common approach for segregating the subset of patients with endometrial cancer who likely require a lymphadenectomy. We evaluate factors related to the performance and value of IOC, including the accuracy of frozen sections, “gross-only examinations,” and obtaining random sections when a gross lesion is not apparent. IOC was performed by gross examination only in 17 (8%) of 250 cases, the specificity and negative predictive value of which in diagnosing cancer were 100% and 85%, respectively. Among the 64 cases wherein a gross lesion was not apparent and random sections were examined, a final diagnosis of carcinoma was rendered in 20, of which only three (15%) had a diagnosable malignancy on the random section. The frozen-section/final diagnosis concordance was 80% for tumor grade. Determining the depth of myometrial invasion was problematic, with 36% underestimation and 2.6% overestimation. Obtaining random sections in the absence of a gross lesion has no significant benefit, and a negative result is likely to provide inaccurate data to the surgeon. Frozen-section analyses are a generally reliable tool to determine “low-risk” pathologic parameters that were evaluated herein when a gross lesion is present.