ESR1 gene amplification and protein expression in 946 patients with resected breast cancer: A Hellenic Cooperative Oncology Group (HeCOG) translational research study.
592 Background: Discrepant data have been reported on the incidence of ESR1 gene amplification in breast cancer, its correlation to clinicopathologic characteristics and its impact on prognosis. Methods: Formalin-fixed paraffin-embedded (FFPE) tumor tissue samples from 946 patients participating in two adjuvant chemotherapy phase III trials (HE10/97 and HE10/00) were centrally assessed in tissue microarrays by immunohistochemistry (IHC) for estrogen receptor (ER), progesterone receptor (PgR) and human epidermal growth factor receptor 2 (HER2). FISH was also performed for HER2, TOP2A and ESR1 using commercially available dual (for ESR1) and triple (for HER2 and TOP2A) hybridization probes. Results: The majority of patients had >T2 (69%), node-positive, ER-positive (>1% stained cells, 73%) tumors managed with resection, chemotherapy and hormonotherapy (76%). Only 38 tumors (4.0%) had ESR1 gene amplification (ESR1/CEN6 ratio >2) and 514 (54.3%) ESR1 gene gain (1<ratio<2). The number of ESR1 gene copies was 3-4 in 251 (26.5%) and 5-10 in 42 (4.4%) of cases. HER2 and TOP2A gene amplification was seen in 234 (25.3%) and 101 (10.9%) of tumors, respectively. We studied the immunohistochemical expression of ER protein by evaluating the percentage of stained cells, the Allred score (0-2, 26.8%; 3-6, 62.5%; 7-8, 10.7% of tumors) and the semiquantitative H-Score (50-100, 13.8%; 101-200, 36.8%; 201-300, 15% of tumors). ESR1 gene amplification was significantly associated with taxane therapy, age >50, postmenopausal status, grade III-IV, absence of HER2 amplification and ER protein expression (p<0.05). At a median follow-up of 92 months, univariate Cox regression analysis showed that ER protein expression, but not ESR1 gene status, was a predictor of favorable outcome. In multivariate analysis, tumor size >5 cm, >4 involved nodes and negative/low ER protein expression by Allred score were independent adverse prognostic factors. Conclusions: Our data showed a rather low incidence of ESR1 gene amplification and failed to confirm its prognostic/predictive utility. ESR1 mRNA expression data will be presented at the meeting.