HER2-enriched subtype and novel molecular subgroups drive aromatase inhibitor resistance and an increased risk of relapse in early ER+/HER2+ breast cancer

Background Oestrogen receptor positive/human epidermal growth factor receptor positive (ER+/HER2+) breast cancers (BCs) are less responsive to endocrine therapy than ER+/HER2- tumours. Mechanisms underpinning the differential behaviour of ER+HER2+ tumours are poorly characterised. Our aim was to identify biomarkers of response to 2 weeks' presurgical AI treatment in ER+/HER2+ BCs. Methods All available ER+/HER2+ BC baseline tumours (n=342) in the POETIC trial were gene expression profiled using BC360 (TM) (NanoString) covering intrinsic subtypes and 46 key biological signatures. Early response to AI was assessed by changes in Ki67 expression and residual Ki67 at 2 weeks (Ki67(2wk)). Time-To-Recurrence (TTR) was estimated using Kaplan-Meier methods and Cox models adjusted for standard clinicopathological variables. New molecular subgroups (MS) were identified using consensus clustering. Findings HER2-enriched (HER2-E) subtype BCs (44.7% of the total) showed poorer Ki67 response and higher Ki67(2wk) (p<0.0001) than non-HER2-E BCs. High expression of ERBB2 expression, homologous recombination deficiency (HRD) and TP53 mutational score were associated with poor response and immune-related signatures with High Ki67(2wk). Five new MS that were associated with differential response to AI were identified. HER2-E had significantly poorer TTR compared to Luminal BCs (HR 2.55, 95% CI 1.14-5.69; p=0.0222). The new MS were independent predictors of TTR, adding significant value beyond intrinsic subtypes. Interpretation Our results show HER2-E as a standardised biomarker associated with poor response to AI and worse outcome in ER+/HER2+. HRD, TP53 mutational score and immune-tumour tolerance are predictive biomarkers for poor response to AI. Lastly, novel MS identify additional non-HER2-E tumours not responding to AI with an increased risk of relapse.