Comparison of mitotic index, in vitro bromodeoxyuridine labeling, and MIB-1 assays to quantitate proliferation in breast cancer.

To investigate the hypothesis that in vitro bromodeoxyuridine (BrDu) labeling might be superior to MIB-1 immunostaining for prognostic value, because it more selectively labels cells during the S phase. Four hundred eighty-six patients with breast cancers (59% lymph node-negative, 41% lymph node-positive) surgically excised between 1988 and 1993 (median follow-up, 62 months) were evaluated for cellular proliferation using prospective in vitro BrDu uptake assays, retrospective mitotic indices, and MIB-1 labeling. MIB-1, BrDu labeling, and mitotic index-derived proliferation data were highly correlated. Each was similarly associated with most other markers of prognosis, although these relationships were not identical. By univariate analysis, nodal status was the most significant prognostic variable for all patients. Higher BrDu labeling index, MIB-1 immunolabeling, and mitotic index were also associated with shortened disease-free survival (DFS) and disease-specific survival for the entire patient group, as well as for node-negative patients. The association between cellular proliferation and survival was much weaker for node-positive patients. Multivariate models confirmed that nodal status, tumor size, and proliferation data predicted survival in all patients as well as those with node-negative disease, although MIB-1 was somewhat more closely associated with outcome than mitotic index or in vitro BrDu data. For patients with T1NOMO disease (n = 172), the only significant predictors of DFS were proliferation rate (mitotic index or MIB-1) and tumor grade. Proliferation rate predicts recurrence and survival in breast cancer. This effect is more pronounced in node-negative patients. In vitro BrDu data are not superior to MIB-1 and mitotic counting.