Prognostic impact of estrogen receptor (ER) level changes during progression for patients with both ER-positive (ER+) primary breast cancer and paired recurrence

Background: We have previously reported that ER+ breast cancer (BC) patients (pts) who become ER-negative at relapse have a poorer overall survival (OS) as compared to those still ER+ at relapse [Dieci et al., Ann Oncol 2013]. Our aim is to evaluate whether, among the group of patients with an ER+ status on both primary and recurrence, changes in the level of ER expression may be of prognostic value. Methods: A total of 81 pts with ER+ primary BC and ER+ paired recurrence who underwent relapse biopsy at Modena University Hospital were studied. ER status was assessed by IHC and the cutoff for ER-positivity was >=10%. Samples were defined as ER-high (>50%) and ER-low (>=10% and <=50%). HER2-status was defined according to IHC and/or FISH results. OS was calculated as the time interval between primary BC diagnosis and death or last follow up. Results: Biopsied recurrences were: distant (86%) and local relapses (14%). Fifteen percent of primary and 21% of recurrent tumors were HER2-positive. Sixty-two pts maintained the same ER level (i.e. high or low) on both primary and relapse (ER-level concordant), whereas 19 changed from ER-high to ER-low or viceversa (ER-level discordant). No difference in OS was observed between the ER-level concordant and the ER-level discordant groups (p=0.3). However, we identified those pts whose ER-high primary BC turned into ER-low as having a particularly poor outcome. Indeed, 10yrs-OS rates were 51% for the ER-level concordant group, 50% for pts changing from ER-low to ER-high and 14% for pts changing from ER-high to ER-low (p=0.0019). Finally, we focused on the subset of pts starting from an ER-high primary BC and showing an ER+/HER2-negative phenotype on both primary and relapse (n=51). The drop of ER-level expression below the 50% cut-off at relapse was confirmed as a poor prognostic factor, as compared to pts maintaining an ER-high level (10yrs-OS 53% vs 17%, p=0.0063). Conclusions: We demonstrated that, even in the case of maintenance of the same single-receptor status (ER+) and/or tumor phenotype (ER+/HER2-negative) between primary BC and recurrence, relapse biopsy may provide relevant prognostic information.