Prognostic significance of Ki-67 and topoisomerase II[alpha] expression in infiltrating ductal carcinoma of the breast: A multivariate analysis of 863 cases

• Published in: Breast cancer research and treatment 1999-05, Vol.55 (1), p.61
• Main Authors: Rudolph, Pierre, MacGrogan, Gaetan, Bonichon, Francoise, Frahm, Sven-Olaf
• Format: Article
• Language: English
• Subjects: Breast cancer; Cancer research; Cancer therapies
• ISSN: 0167-6806; 1573-7217

To evaluate the prognostic relevance of Ki-67 and topoisomerase IIalpha expression in relation to tumor stage, grade, and hormone receptor content, 942 ductal infiltrating carcinomas of the breast were examined by means of the monoclonal antibodies Ki-S11 (Ki-67) and Ki-S4 (topoisomerase IIalpha). pS2, c-erbB2, and p53 were additionally considered as prognostic variables. The median follow-up time was 149 months. Eight-hundred-and-sixty-three tumors reacted with Ki-S11 and Ki-S4; the labeling indices of the two antigens were closely associated (r = 0.93). Both correlated positively with the tumor size, c-erbB2, and p53 expression, and negatively with patient age, hormone receptor content, and pS2 immunostaining. In the univariate analysis, Ki-S11 and Ki-S4 scores, nodal status, tumor size, tumor grade, and progesterone receptor content strongly predicted both overall and metastasis-free survival (p < 0.00001). Estrogen receptor status, p53, and c-erbB2 were of minor significance. Concerning overall survival, multivariate Cox regression analysis selected a Ki-S4 score >25% (p < 0.00001) next to the nodal status, and before tumor size, progesterone receptor content, and patient age. Independent predictors of the occurrence of distant metastases were nodal status, Ki-S4, tumor size, grade 1, and progesterone receptor negativity, in that order. The Ki-S11 score was of independent prognostic significance only if examined as a continuous variable. We conclude that topoisomerase IIalpha expression as assessed by monoclonal antibody Ki-S4 may add valuable information to current prognostic models for breast cancer. Its predictive value appears to be essentially related to the proliferative activity of tumor cells.