The sensitivity and specificity of IIC are similar to that of intraoperative frozen section evaluation. Therefore, IIC is a viable alternative to frozen sectioning when intraoperative evaluation is required. If SLN micrometastasis is used to determine the need for further lymphadenectomy, more sensitive intraoperative methods will be needed to avoid a second operation.

Background: The increasing utilization of lymphatic mapping techniques for breast carcinoma has made intraoperative evaluation of sentinel lymph nodes attractive. Axillary lymph node dissection can be performed during the initial surgery if the sentinel lymph node is positive, potentially avoiding a second operative procedure. At present the optimal technique for rapid sentinel lymph node assessment has not been determined. Both frozen sectioning and intraoperative imprint cytology are used for rapid intraoperative sentinel lymph node evaluation at many institutions. The purpose of this study is to evaluate experience with imprint cytology for intraoperative evaluation of sentinel lymph nodes in patients with breast cancer. Methods: A retrospective review of the intraoperative imprint cytology results of 678 sentinel lymph node mappings for breast carcinoma was performed. Sentinel nodes were evaluated intraoperatively by either bisecting or slicing the sentinel node into 4 mm sections. Imprints were made of each cut surface and stained with H&E and/or Diff-Quik. Permanent sections were evaluated with up to four H&E stained levels and cytokeratin immunohistochemistry. Intraoperative imprint cytology results were compared with final histologic results. Results: The sensitivity of imprint cytology was 53%, specificity was 98%, positive predictive value was 94%, negative predictive value was 82% and accuracy was 84%. The sensitivity for detecting macrometastases ( more than 2mm) was significantly better than for detecting micrometastases (<2 mm), 81 versus 21%, respectively (P < 00001). Conclusions: The sensitivity and

Islet cell neoplasms (ICNs) are uncommon tumors that may present with bizarre endocrine manifestations. Only rarely have their fine-needle aspiration (FNA) cytomorphologic characteristics been reported. The authors have studied FNAs of ICNs in ten patients, including immunocytochemistry (ICC) directly on the aspirates. FNA yielded moderately to markedly cellular specimens with numerous individual cells and large aggregates. Although moderate pleomorphism was present in three cases, striking uniformity in nuclear size and contour and in delicate nuclear membranes was evident. Typically the round nuclei were eccentric, imparting a plasmacytoid appearance. Cytoplasmic granularity was noted in only a minority of tumor cells. Five of the ICNs were positive for chromogranin (CG). In comparison, seven ICNs were positive for CG in tissue sections. None of the cytologic material showed immunoreactivity with insulin, glucagon, somatostatin, or gastrin. Histologic material showed positivity for these hormones in 33% (three of nine), 22% (two of nine), 22% (two of nine), and 11% (one of nine), respectively. Although CG is the most useful immunocytochemical marker for ICNs, and thus is helpful in confirming a diagnosis of ICN in FNA material, it is negative in half of the cases and may be negative when the histologic material is positive.

Sclerosing stromal tumors are rare, benign ovarian neoplasms of unknown etiology and histogenesis. Three sclerosing stromal tumors were evaluated by immunohistochemistry and electron microscopy and were compared to two thecomas and nonneoplastic ovarian mesenchymal tissue. The sclerosing stromal tumors and thecomas were positive for muscle-specific actin; immunoreactivity was intense in the cellular areas of the sclerosing stromal tumors and focal in the thecomas. This antigen was expressed in nonneoplastic stroma predominantly in a perifollicular (theca externa) distribution. Two sclerosing stromal tumors and both thecomas were vimentin positive. Desmin was present in nonvascular cells in one of each tumor type. Expression of vimentin diffusely and of desmin focally was present in nonneoplastic cortical stroma and surrounding follicles. All specimens were nonreactive for cytokeratin. Electron microscopy supported differentiation toward smooth muscle in the sclerosing stromal tumors but not in the thecomas. Such differentiation included aggregates of cytoplasmic filaments with interspersed dense bodies, pinocytotic vesicles, and basal lamina. Delicate, long processes interconnected cells, often with primitive junctions, in the hypocellular foci. Cytoplasmic lipid, which was present in the thecomas, was not well developed in the sclerosing stromal tumors. It is proposed that a population of muscle-specific actin-positive elements exists in the theca externa--the perifollicular myoid stromal cell--and that sclerosing stromal tumors may originate from them. Sclerosing stromal tumors and thecomas share many antigenic determinants and morphologic features and thus are probably closely related entities.