COX-2 expression in lobular in situ neoplasia of the breast: correlation with histopathological grading system according to the Tavassoli classification.

AIMS: There is considerable evidence to link cyclooxygenase (COX)-2 to the development of cancer. The aim of this study was to assess COX-2 expression and its subcellular localization in lobular in situ neoplasia (LIN) of the breast and to verify differences in COX-2 expression between different grades of lesions according to the Tavassoli classification. METHODS AND RESULTS: We analysed the expression of COX-2 protein by immunohistochemistry in tissue samples of 51 LIN lesions classified into three grades according to the Tavassoli classification. COX-2 immunostaining was observed in 78.4% of LIN samples and showed a prevalent membranous rather than cytoplasmic pattern. COX-2 was expressed in 16/17 (94.1%) LIN1, 22/25 (88%) LIN2 and 2/9 (22.2%) LIN3. As regards COX-2 expression, a statistically significant difference was found between LIN1 and LIN3 (P = 0.001) and between LIN2 and LIN3 (P =0.001). No difference was found between LIN1 and LIN2. Moreover, a significant negative correlation was found between LIN grade and COX-2 expression (P < 0.0001). CONCLUSIONS: COX-2 is highly expressed in LIN, supporting a role for this protein in the early stage of breast carcinogenesis, representing the rationale for using COX-2 selective inhibitors in the earliest stages of breast cancer.

Cell blocks of breast FNAs frequently allow diagnosis of invasion or histological classification of proliferative changes.

Two major limitations of breast fine needle aspiration (FNA) compared with core needle biopsies (CNB) are the inability to determine whether a cancer is invasive and to classify proliferative lesions. We studied 40 consecutive "rapid cell blocks" from breast FNAs with surgical pathology follow-up to test whether cell blocks can overcome these limitations. Of 25 carcinomas, invasion could be identified in the cell block sections in 11 (44%). One cystosarcoma phyllodes was suspected based on the cell block sections. Cell blocks from 12 of 14 benign breast FNAs showed sufficient cells to assign a histologic diagnosis of no hyperplasia (1 case, confirmed on follow-up) and usual hyperplasia (11 cases; confirmed in eight of 11 on follow-up). Specific histologic diagnoses included intraductal papilloma (2 cases), and in situ lobular neoplasia (2 cases). Cell blocks complement smears and monolayers and appear to overcome major limitations of breast FNA.

Influence of serum from healthy or breast tumor-bearing women on the growth of MCF-7 human breast cancer cells.

Sera from women healthy (HW) or with breast (BCW), ovarian or endometrial cancer, were added (10%) to the culture media of MCF-7 cells and cell proliferation assessed 4 days later to verify: a) whether sera from BCW, obtained before or 8 days after tumor ablaction, influence the proliferation of these cells, b) whether the effects of serum from BCW are specific for mammary tumor cells. Sera from BCW, but not sera from women with ovarian or endometrial cancer, increased MCF-7 cell proliferation in comparison with sera from HW. After surgical ablation of the breast tumors, serum's ability to increase MCF-7 cell proliferation decreased significantly. These effects cannot be explained by differences on serum levels of estradiol or melatonin. These results suggest the presence of growth-promoting substances of possible tumoral origin in serum of BCW, a fact that should be considered as support for the surgical treatment of tumor masses.

Contemporary approach to the classification of renal epithelial tumors.

Our understanding of the morphologic and genetic features of renal epithelial neoplasia has brought about profound changes in the classification of these tumors. It is clear that they represent a heterogeneous group of tumors with distinct histopathologic, genetic, and clinical features ranging from benign to high-grade malignant. "Granular" and "sarcomatoid" carcinomas are not distinct entities, since tumors with granular or spindle cell features may be seen in many tumor-types. Using conventional pathologic tools such as hematoxylin and eosin staining, histochemistry, immunohistochemistry, and electron microscopy, we are able to properly classify the majority of these tumors. Nevertheless, approximately 6% to 7% of cases are impossible to classify in this fashion, thus requiring molecular genetic studies for proper characterization.