Predictive value of different prognostic factors in breast cancer recurrences: Multivariate analysis using a logistic regression model

Background: The aim of this study was to compare the sensitivity of different pre-operative parameters in patients with breast cancer (BC) recurrence using univariate and multivariate analysis. Materials and Methods: We retrospectively analyzed a series of 387 women (median age 60 years, range 3583 years) who underwent curative surgery for pT1-2 BC. The patients were divided into two groups: Group 1: 325 (84.0%) patients with no evidence of disease during a median follow-up of 53 months (range 25-149 months) and Group 2: 62 (16.0%) patients who developed local or distant recurrences. Results: Univariate analysis showed significant (p < 0.01) differences between the two Groups in age, size and grading of the tumor and hormone receptor rate. MIB1 proliferation rate, serum markers CEA and CA 15-3, and lymph node status were not useful in predicting relapse. Multivariate analysis using a logistic regression model showed that only age, size of the tumor and hormone receptor rate independently correlate with the onset of recurrences. Conclusion: There is no clear correlation between BC recurrence and the majority of the prognostic factors available. Multivariate analysis of several pre-operative parameters may help to correctly select the high risk population.