Prognostic role of estrogen receptor (ER) expression in breast cancer (BC) metastases and its dynamics from primary to metastatic disease: Results from a large multicentric cohort of patients with phenotypically stable ER+(≥10%)/HER2- BC.

Abstract

1067 Background: ER expression is one of the main determinants of prognosis in patients (pts) with BC. Phenotypic conversion from ER+ (ER>=10%)/HER2- primary BC towards ER<10%/HER2- advanced BC has a well-known negative impact on outcome. However, in the specific context of phenotypically stable ER+/HER2- BC, the prognostic impact of ER expression in metastases or ER dynamics during disease evolution, remains largely understudied. Methods: We enrolled pts with advanced BC undergoing biopsy of a metastatic site. ER+ was defined as ER>=10%. ER expression was evaluated both as continuous and categorical variable (categories: 10-30%, 30-50%, 50-100%). Overall survival (OS) was the primary study endpoint. Cox multivariable models included covariates associated with OS in univariate analysis. Results: Among 1114 pts, 410 had ER+/HER2- phenotype in both primary and metastatic tumor specimens. In this subgroup, ER expression (both continuous and categoric) in metastases had significant prognostic value: for each 10% lower ER expression, the risk of death increased by 6.7% (p=0.011). Pts with ER 10-30% had significantly worse OS than those with ER 50-100% (HR 0.62, p=0.023), and numerically shorter than ER 30-50% (HR 0.51, p=0.063). The table shows the evolution of ER categories from primary BC to metastases. ER expression dynamics also had prognostic impact. Regarding continuous ER expression changes in paired primary vs. metastatic BC, each 10% ER increase corresponded to 5.9% decrease in the risk of death (p=0.008). Pts whose tumors showed an increased ER expression from 10-30% to 50-100% had the most favorable outcome overall, with better OS compared to pts with persistently low ER levels (HR 6.73, p=0.002), or to pts with decreased ER expression in metastases - particularly pts whose ER levels dropped from 50-100% to 10-30% (HR 2.89, p=014). These pts also had superior OS when compared to pts with persistently high ER levels (HR 2.08, p=0.043). The prognostic impact was preserved at the multivariate analysis (including age, grade, visceral/non-visceral disease, biopsy site). Conclusions: Intratumor ER expression and dynamics may in part explain the prognostic heterogeneity of pts with ER+/HER2- stable phenotype from primary to advanced BC. Pts with lower ER levels in metastases are prognostically disadvantaged. Dynamic changes in ER expression provide additional insights beyond those captured by single-point assessment. Interestingly, pts whose tumors shifted from ER 10–30% to ER 50–100% showed the most favorable prognosis, even outperforming those with consistently high ER levels. Metastases, ER 10-30% 30-50% 50-100% Total Primary BC, ER n % n % n % n % 10-30% 5 1.2 0 0 18 4.4 23 5.6 30-50% 7 1.7 4 1.0 11 2.7 22 5.4 50-100% 24 5.9 21 5.1 320 78.0 364 89 Total 36 8.8 25 6.1 349 85.1 410 100