Genomic predictors of neoadjuvant chemotherapy (NACT) response in breast cancer (BC).

Abstract

e12122 Background: We assessed effects of NACT on BC mutational landscape. Methods: Baseline (BL) and post-NACT tumor / matched normal DNA from 12 newly diagnosed BC patients on NACT (4 x doxorubicin/cyclophosphamide + low dose sunitinib; NCT01176799) were subject to whole exome sequencing. Nonsynonymous somatic single nucleotide variants from 34 genes in known BC signaling pathways were evaluated for changes in mutant variant allele frequency (VAF) according to clinical outcome. Poor outcome was defined as <50% target lesion reduction after NACT or BC relapse / progression (PD) within 2 years; significant change was defined as > 0.2 difference in BL vs post-NACT mutant VAF. Results: Mean tumor size was 6.4 + 2.9cm; 50% were N+; 8% were M1; 7/12 patients had poor outcomes. Tumors harbored mutations in PI3K (58%), NOTCH (42%), Wnt (42%), TP53 (33%) and FOXA (17%) pathways. Change in no. of somatic mutations post-NACT correlated with outcome (mean percent change +14% vs -30% in patients with poor vs good outcome, p=0.04). 11 patients had >1 of 23 putative driver mutations identified ( Table 1). Mutant VAF declined significantly in those with good outcomes, except for a new NOTCH2 mutation in A2 and rise in mutant VAF in A4. In patients with poor outcomes, mutant VAF persisted or rose, and emergent mutations (AKT1, PIK3CA) occurred in 2 patients. Conclusions: Chemoresistance and emergent mutations were revealed by tracking mutant VAF in BC patients on NACT. Clinical trial information: NCT01176799. [Table: see text]