Abstract Background: Eribulin is a non-taxane, microtubule dynamics inhibitor approved for the treatment of patients with locally advanced or metastatic breast cancer. To date, no biomarker exists to prospectively select patients who will derive the maximum benefit from this chemotherapeutic. In this study, we explored, in a prospective clinical trial, the efficacy and the association of pre-treatment expression of mRNA with response in patients with HER2-negative breast cancer treated with neoadjuvant eribulin. Results of the HR+ cohort are presented here. Methods: SOLTI1007 is a phase II, open-label, single-arm, exploratory pharmacogenomic study of single agent eribulin as neoadjuvant treatment for stage I-II HER2-negative breast cancer (n=100 HR+ and n=100 HR-negative). Patients received 1.4 mg/m2 of eribulin mesilate intravenously on Days 1 and 8 every 21-day cycle, for 4 cycles. Baseline, C2D1, and post-treatment (surgical) formalin-fixed, paraffin-embedded tissue samples were collected and gene expression profiled. The PAM50 genes and 492 additional breast cancer-related genes were evaluated using the nCounter platform. Research-based intrinsic subtype and the Risk of Relapse based on subtype and proliferation (ROR-P) were evaluated using the Parker et al. JCO 2009 algorithm. The association of each signature and pathological complete response in the breast (pCRb) was evaluated using univariate logistic regression models. Genes found differentially expressed between screening and surgical specimens were identified using a paired two-class Significance Analysis of Microarrays. Results: 92 evaluable patients were included. Mean age: 55.7. Mean tumor size: 3.7 cm. Mean Ki-67 %: 30.3. The overall pCRb rate was 5.4%. Distribution of the intrinsic subtypes was as follows: Luminal A (n=36, 39.1%), Luminal B (n=34, 37.0%), Basal-like (n=10, 10.9%), Normal-like (n=10, 10.9%) and HER2-enriched (n=2, 2.2%). pCRb rates differed significantly by subtype (p=0.009): Luminal B (11.7%, 4/34), HER2-enriched (50%, 1/2), Luminal A (0%, 0/36), Basal-like (0%, 0/10), Normal-like (0%, 0/10). pCRb rate differed significantly by ROR-P (p=0.018): ROR-P high (16.7%, 4/24), ROR-P med (1.9%, 1/52), ROR-P low (0%, 0/16). Ki67 % did not predict pCRb (p=0.406). Subtype change at surgery occurred in 48.3% (14/29) of Luminal Bs, 13.8% (4/29) of Luminal As, 100% (2/2) of HER2-enriched and 12.5% (1/8) of Basal-like tumors. 78.6% (11/14) of subtype changes in Luminal B disease were to Luminal A. Compared to screening biopsies, 136- and 141- genes were found up- and down-regulated in surgical specimens (False Discovery Rate <5%). The up-regulated gene list was enriched for response to hormone stimulus (e.g. ESR1, BCL2 and ERBB4), negative regulation of apoptosis (e.g. IL6, EGFR and PTEN) and angiogenesis (e.g. ANGPTL4, HIF1A and TGFBR2). The down-regulated gene list was enriched for cell cycle (e.g. CCNB1, RAD17 and MKI67), DNA-repair (e.g. BRCA1, BRCA2 and ATR) and microtubule cytoskeleton organization (e.g. AURKA, CENPA and KIF23). Conclusions: From a response and biological perspective (i.e. induction of a Luminal A phenotype), patients with Luminal B disease might benefit the most from eribulin therapy. Strategies combining eribulin with endocrine therapy seem warranted in Luminal B breast cancer. Citation Format: Prat A, Ortega V, Paré L, Galván P, Oliveira M, Nucíforo P, Lluch A, Morales S, Amillano K, Lopez R, González R, Manso L, Martínez J, Llombart A, de la Peña L, di Cosimo S, Rubio IT, Harbeck N, Baselga J, Cortés J. Efficacy and gene expression results from eribulin SOLTI1007 neoadjuvant study. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P3-07-66.