Background: Expression of aromatase by malignant breast epithelial cells and/or the surrounding stroma implies local estrogen production that could influence the outcome of endocrine therapy for breast cancer.
Methods: A validated immunohistochemical assay for aromatase was applied to samples from the P024 neoadjuvant endocrine therapy trial that compared tamoxifen and letrozole. The presence of aromatase expression by tumor or stromal cells was correlated with tumor response, treatment induced changes in proliferation index (Ki67), relapse-free survival (RFS) and breast cancer-specific survival (BCSS).
Results: Tumor and stromal aromatase expression were highly correlated (P = 0.0001). Tumor cell aromatase, as a semi-continuous score, also correlated with smaller tumor size at presentation (P = 0.01) higher baseline ER Allred score (P = 0.006) and lower Ki67 levels (P = 0.003). There was no significant relationship with clinical response or treatment-induced changes in Ki67. However, in a Cox multivariable model that incorporated a post-treatment tumor profile (pathological T stage, N stage, Ki67 and ER status of the surgical specimen), the presence of tumor aromatase expression at baseline sample remained a favorable independent prognostic biomarker for both RFS (P = 0.01, HR 2.3, 95% CI 1.2-4.6 for absent expression) and BCSS (P = 0.008, HR 3.76, 95% CI 1.4-10.0).
Conclusions: Autocrine estrogen synthesis may be most characteristic of smaller, more indolent and ER-rich breast cancers with lower baseline growth rates. However, response to endocrine treatment may not depend on whether the estrogenic stimulus has a local versus systemic source.