We conclude that tumors with overexpression of the c-erbB-2 oncogene are less responsive to cyclophosphamide, methotrexate, and fluorouracil (CMF)-containing adjuvant therapy regimens than those with a normal amount of gene product.

Several small studies have reported that having a high percentage of breast tumor cells that express the proliferation antigen Ki-67 (ie, a high Ki-67 labeling index) predicts better response to neoadjuvant chemotherapy. However, the predictive value of a high Ki-67 labeling index for response to adjuvant chemotherapy is unclear. To investigate whether Ki-67 labeling index predicts response to adjuvant chemoendocrine therapy, we assessed Ki-67 expression in tumor tissue from 1924 (70%) of 2732 patients who were enrolled in two randomized International Breast Cancer Study Group trials of adjuvant chemoendocrine therapy vs endocrine therapy alone for node-negative breast cancer. A high Ki-67 labeling index was associated with other factors that predict poor prognosis. Among the 1521 patients with endocrine-responsive tumors, a high Ki-67 labeling index was associated with worse disease-free survival but the Ki-67 labeling index did not predict the relative efficacy of chemoendocrine therapy compared with endocrine therapy alone. Thus, Ki-67 labeling index was an independent prognostic factor but was not predictive of better response to adjuvant chemotherapy in these studies.

Our data provide a strong methodologic and clinical multicenter validation of the predefined prognostic 76-gene signature in LNN breast cancer patients.

CD10 antigen is a 100-kDa-cell surface zinc metalloendopeptidase expressed in a variety of normal and neoplastic lymphoid and nonlymphoid tissues including melanomas. It was recently shown that metastatic melanomas express more CD10 than primary tumors. We evaluated CD10 expression in tumor and stromal cells in 70 biopsies with primary and 28 with metastatic malignant melanomas. Ki-67, Bcl-2, and Bax were also examined to investigate whether CD10 expression is associated with tumor proliferation index or factors of apoptosis. Formalin-fixed/paraffin-embedded tissues were studied by immunohistochemistry. More advanced primary tumors had higher CD10 expression in the tumor cells (r ¼ 0.27, P ¼ 0.03 for Clark levels and r ¼ 0.29, P ¼ 0.02 for Breslow) and higher Ki-67 proliferation fraction (r ¼ 0.32, P ¼ 0.007 for Clark levels and r ¼ 0.32, P ¼ 0.001 for Breslow). Similarly, CD10 expression in the intratumoral stromal cells was also higher in primary tumors with higher Clark level (P ¼ 0.04, linear-by-linear association) and tumor thickness according to Breslow (r ¼ 0.33, P ¼ 0.01). The presence of CD10 þ peritumoral stromal cell cuffs was also positively associated with tumor thickness according to Breslow (r ¼ 0.27, P ¼ 0.05). Also, expression of CD10 and Ki-67 were significantly higher in metastatic than in primary tumors (P ¼ 0.01 and 0.02 respectively), but Bcl-2 expression was higher in primary melanomas (P ¼ 0.02). We conclude that CD10 expression in malignant melanoma is associated with tumor progression.

The 76-gene signature defines high-risk patients who benefit from adjuvant tamoxifen therapy. Although we did not study the value of chemotherapy in this study, low-risk patients identified by the 76-gene signature have a prognosis good enough that chemotherapy would be difficult to justify. The prognosis of these patients is sufficiently good, in fact, that a disease-free benefit for tamoxifen therapy is difficult to prove, though benefits in terms of loco-regional relapse and a reduction in risk for contralateral breast cancer might justify hormonal therapy in these patients.