Abstract P5-09-01: Comprehensive gene assessment of estrogen receptor positive breast cancer reveals that HER2 plays an important role in resistance to neoadjuvant letrozole

Abstract Background 10% of all breast cancers are HER2+/ER+ and these cancers exhibit both intrinsic and acquired resistance to endocrine therapy. They have a worse prognosis than HER2-/ER+ cancers. Aims 1. To investigate the role of HER2 in response to neoadjuvant Letrozole. 2. To predict which HER2+ cancers do not respond to Letrozole. Methods 23 postmenopausal women with large, operable, locally advanced HER2+/ER+ breast cancer treated with neoadjuvant Letrozole had response assessed by periodic 3D ultrasound. Core biopsies taken at 0, 14 days and 3 months of treatment. RNA were extracted, amplified, labelled and hybridised to Illumina HT-12 whole genome beadarrays. A group of patients with ER+/HER2- disease were identified to compare clinical and molecular response. Results 13 (57%) HER2+/ER+ patients responded (R) and 10 (43%) patients did not (NR). HER2 expression was significantly higher at baseline in the NR group (p = 0.005). There were differences in gene expression between HER2+/ER+ R and NR and between the HER2+/ER+ and HER2-/ER+ NR groups. In the HER2+/ER+ NR group, there was virtually no change in gene expression during treatment with Letrozole. Table 1 Up Regulated GenesDown Regulated Genes n0-14 days14 days-3 months0-3 months0-14 days14 days-3 months0-3 monthsHER2-ve Responders43122833505127213HER2-ve Non Responders1588361432340129HER2+ve Responders13541272703831125HER2+ve Non Responders10851117106 Considerable gene changes with overlap in the genes that changed most was evident in HER2+/ER+ and HER2-/ER+ responders. Analysis of 55 estrogen sensitive proliferation genes revealed significantly less reduction in the HER2+/ER+ NR group than in the HER2+/ER+ R group (AUC = 10.57 vs 27.93 respectively; p<0.0001). This difference was apparent by 14 days (AUC = 25.55 vs 17.94; p = 0.005). The HER2+/ER+ NR group had significantly less reduction in these 55 estrogen sensitive proliferation genes when compared to the HER2-/ER+ NR group at 14 days (AUC = 17.94 vs 18.05 respectively; p = 0.0007) and by 3 months (AUC = 10.57 vs 26.26; p<0.0001). In logistic regression analysis, HER2 status was predictive of disease progression (p = 0.048). A single gene classifier predicted response to endocrine therapy in the ER+/HER2+ group. This was accurate in 94% of patients in the test set (n = 23). This classifier was also predictive of progression free survival (p = 0.006) in HER2+/ER+ patients. Conclusions In this large cohort of patients treated with neoadjuvant letrozole: • HER2+/ER+ cancers have a low rate of response. • Changes in gene expression are similar in HER2+/ER+ and HER2-/ER+ responders. • HER2+/ER+ cancers that do not respond to letrozole have few gene changes and little reduction in estrogen signalling pathways. • In contrast HER2-/ER+ non responders show significant gene changes and reduced expression of estrogen sensitive proliferation genes. • A single gene classifier at diagnosis in ER+/HER2+ has been identified which predicts response to Letrozole with 94% accuracy. • Validation continues. This single gene classifier may provide a simple test to predict which HER2+/ER+ cancers are endocrine resistant. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P5-09-01.