Background. More than half of breast carcinomas express estrogen and progesterone receptors (ER and PR) and remain estrogen-dependent. Endocrine therapy with inhibitors of the estrogen cascade allows long-term and effective control of the disease in a significant number of patients, however, some patients demonstrate primary, and the majority - secondary resistance to such treatment.
 The aim of the study was to evaluate the predictive significance of the CCND1 and FGFR1 gene amplifications and PIK3CA mutations for endocrine therapy.
 Materials and methods. The study included 138 women with ER-positive primary metastatic breast cancer (BC) who received hormone therapy with aromatase inhibitors (AIs) (n = 69), tamoxifen (n = 65), goserelin (n = 2), or a combination of goserelin and tamoxifen (n = 2) as the first line treatment. CCND1 and FGFR1 gene amplifications were tested by digital droplet PCR, while mutations in exons 7, 9, and 20 of the PIK3CA gene were determined using high-resolution melting analysis and allele-specific PCR. We analyzed the associations between the presence of the mentioned genetic lesions, progression-free-survival (PFS) and response to treatment.
 Results. Amplifications of CCND1 and FGFR1 genes were identified in 24 (17.9%) and 28 (20.9%) of 134 successfully tested cases, respectively; 9 tumors were positive for both alterations. Amplifications were more prevalent in less differentiated tumors (p = 0.018). CCND1 amplification was associated with shorter PFS in patients receiving aromatase inhibitors (16.0 months vs. 32.4 months, HR = 3.16 [95% CI: 1.26-7.93], p = 0.014). FGFR1 status did not significantly affect PFS of AI-treated women, however, partial regress as a result of AI implementation was less frequent in FGFR1-amplified BC as compared to cases with normal FGFR1 copy number (p = 0.031).
 The frequency of PIK3CA mutations was 40.2% (49/122). They were more often observed in smaller tumors (p = 0.034), in PR-positive carcinomas (p = 0.012), and in cases with more extensive metastatic involvement (p = 0.029). The presence of PIK3CA mutations did not affect the results of treatment with AI or tamoxifen.
 Conclusion The presence of CCND1 and/or FGFR1 amplification is associated with worse results of AI therapy of metastatic breast cancer.