In early-stage breast cancer patients, overexpression of c-erbB2 is a marker of lack of efficacy of adjuvant tamoxifen.

We report 62 cases of invasive micropapillary carcinoma of the breast characterized by delicate pseudopapillary structures lacking a fibrovascular core and by tubuloalveolar structures freely floating in clear, empty spaces. All patients but 1 were women (median age, 57 years; range, 25-89 years). Tumor size ranged from 0.7 to 10 cm (median, 2.8 cm); 54 (87%) were grade 3. Psammoma bodies were identified in 29 (47%). Focal to massive lymphatic permeation was present in 39 (63%). Architectural features were retained in the node metastases, dermal lymphatics, and recurrences. Fifty-six patients (90%) had metastatic axillary nodes: 18 tumors were estrogen receptor-positive (32%); 11 were progesterone receptor-positive (20%); HER2/neu was overexpressed in 53 (95%) and p53 in 39 (70%). A peculiar immunoreactivity for MUC1 limited to the cytoplasmic membrane oriented toward the stroma and an absence of immunoreactivity for E-cadherin in the same side of the cytoplasmic membrane indicated inversion of cell polarization and a disturbance in the cell adhesion molecules. Of 41 patients with available follow-up, 29 (71%) had local recurrence (mean, 30 months) and 20 (49%) died of disease. These results underscore the aggressive behavior and poor prognosis of this breast carcinoma variant. Aggressive preoperative neoadjuvant chemotherapy should be considered.

Background. This study assessed the prognostic effect of lymphatic and blood vessel invasion (LVI and BVI) on survival in a retrospective sample of 1408 patients with breast cancer. Methods. Survival analysis was evaluated by univariate (Kaplan‐Meier product limit method and log rank test) and multivariate (Cox model) analysis. Correlations among variables were studied by contingency tables and statistical significance was evaluated by chi‐square test. Results. Lymphatic vessel invasion was present in 34.2% of cases, and BVI in 4.2%. Lymphatic vessel invasion correlated with BVI (P < 0.0001), and both were correlated with metastatic axillary lymph nodes and increasing tumor size and grade; BVI was sporadic (only 10 cases) among lymph node negative patients. Although LVI was more frequent among premenopausal patients and those with ductal carcinomas, BVI was unrelated to menopausal status and tumor type. Lymphatic vessel invasion and BVI were associated with poor survival by univariate analysis (P < 0.0001). By multivariate analysis, relative risk of death was significantly increased when LVI was present both in the whole series as well as in the lymph node negative and lymph node positive subgroups; the prognostic role of LVI was independent of menopausal and lymph node status, tumor size, tumor grade, or adjuvant treatment. In the lymph node negative sample, LVI had strong prognostic power. In the lymph node positive sample, the prognostic role of LVI was also independent of the number of lymph nodes with metastases. Blood vessel invasion had no prognostic role in any subgroup. Conclusion. The prevalence of BVI was particularly low in this study, and the question of its possible prognostic role for patients with breast cancer should be assessed with methods that amplify its detection. LVI is a strong prognostic factor for operable patients with breast cancer. In lymph node negative patients, LVI is a predictor of poor prognosis for those who are consequently candidates for adjuvant therapy. Similarly, in lymph node positive patients, LVI is a predictor for a high risk of death for those who are candidates for highly intensive adjuvant strategies. Cancer 1995; 76:1772–8.

The expression of growth factors, such as transforming growth factor a (TGFa), amphiregulin (AR) and CRIPTO, a type-I tyrosine-kinase growth factor receptor (erbB-2), and a tumor-suppressor gene (p53). that have been implicated in the development and/or the progression of breast cancer, was evaluated by immunohistochemistry in I00 human primary infiltrating breast carcinomas (IBC). AR and CRIPTO imrnunoreactivity was also assessed in 55 human breast ductal carcinomas in situ (DCIS). Within the I00 IBC, 80,50,73, 17, and 34 tumors expressed moderate to high levels of TGFol, AR, CRIPTO, erb6-2, and p53 respectively. In addition, AR and CRIPTO immunoreactivity were found in I I and in 26 out of 55 DCIS respectively. In contrast, only 4, 3, and 2 out of 10 normal mammary-gland samples were weakly positive for TGFol, AR, and CRIPTO expression, respectively, whereas none was positive for erbB-2 or p53. Within the I00 IBC, expression of erb6-2 significantly correlated with high histologic and nuclear grading, with high growth fraction, and with estrogen-receptor(ER)-and progesterone-receptor(PgR)-negative tumors. A statistically rignificant correlation was also observed between p53 expression and high histologic grading, high growth fraction, and PgRnegative tumors. In contrast, no significant correlations were found between TGFol, AR, and CRIPTO imrnunoreactivity and various clinicopathological parameters, with the exception of a positive correlation between TGFa and ER expression. These data demonstrate that TGFol, AR, and CRIPTO expression are significantly increased in malignant mammary epithelium relative to normal epithelium. In particular, the differential expression of CRIPTO may serve as a potential tumor marker for breast carcinogenesis.o 1996 Wilq-Liss, Inc.

We analyzed retrospectively the relationships and the prognostic significance of four anatomopathological features (elastosis, fibrosis, necrosis, inflammatory cell reaction) of the primary tumor in a series of 1,457 cases of infiltrating ductal carcinoma observed at our institution from January 1978 to December 1991. Necrosis, elastosis, fibrosis and inflammatory cell reaction were strongly associated among themselves (all p < 0.0001), the only exception being necrosis and elastosis. Necrosis was significantly related to tumor size (odds ratio [OR] = 5.40, p < 0.0001) and tumor grade (OR = 2.22, p < 0.0001). Univariate analysis showed that the presence of necrosis and cell reaction were significantly related to worse survival (p < 0.0001 and p = 0.03, respectively). Multivariate analysis, including the four variables plus nodal status, tumor size, grading, adjuvant therapy, age and first order interactions, revealed that greater tumor size (p < 0.0001), positive nodal status (p < 0.0001), higher histologic grade (p < 0.0001) and presence of inflammatory cell reaction (p = 0.0007) independently worsened survival. On the other hand, adjuvant therapy had a significant independent role in preventing deaths (p = 0.03). The only first-order interaction retained in the model was that between grading and cell reaction (p = 0.002). Cell reaction had a different prognostic behaviour in the groups G1-G2 and G3: in the former group, survival was worse (p = 0.0001) when the inflammatory cell reaction was present. In conclusion, we demonstrate that cell reaction is an independent prognostic factor in the G1-G2 subgroup of patients, and propose a hypothesis as to the role of cell reaction in primary breast cancer.