Prognostic value of estrogen and progesterone receptors in primary infiltrating ductal breast cancer. A sequential multivariate analysis of 1262 patients.

Nine prognostic variables were evaluated for their significance in predicting the overall survival (OS), the length of disease-free survival (DFS) and the length of metastasis-free survival (MFS) of 1262 patients with primary breast cancer. The variables studied were: UICC clinical stage; menopausal status; histologic grade; number of involved nodes; anatomic tumor size; estrogen and progesterone receptors; local and adjuvant therapies. Three sequential multivariate analyses, at 2, 5 and 10 years, using the Cox proportional hazard regression model, were carried out to identify those variables most highly related to the criteria studied (overall, disease-free, metastasis-free survivals) and especially to fully evaluate the effects of hormonal receptors on prognosis and their stability over time. Our results showed that number of involved nodes and histologic grade were the most significant prognostic factors for all periods of time and whatever the criterion studied; ER had no predictive value while PR was an independent prognostic factor for metastasis-free survival at 2 years (P = 0.01) and 5 years (P = 0.02) but lost its significance at 10 years (P = 0.06). In the subgroup of 261 patients who received prolonged post-operative adjuvant chemotherapies, PR was the main prognostic factor for MFS at 2 years (P = 0.03) and the second at 5 years (P = 0.05) just after number of involved nodes. In the 1001 patients who did not receive prolonged post-operative adjuvant chemotherapies ER was significant for MFS at 5 and 10 years. The present data urge the need for a periodic redefinition of prognostic factors in primary breast cancer.