Molecular characterisation of the residual disease after neoadjuvant endocrine therapy in ER+/HER2- breast cancer uncovers biomarkers of tumour response

Abstract

Background: Neoadjuvant endocrine therapy (NET) in oestrogen receptor-positive /HER2-negative breast cancer (ER+/HER2- BC) allows real-time evaluation of treatment sensitivity by monitoring tumour response and offers the opportunity of personalised therapy. However, the lack of reproducible biomarkers to assess response and long-term prognosis after NET is a significant barrier to increase its indications.

Methods: In this study we searched for clinically relevant molecular reporters of response to NET in a multicentre population of ER+/HER2- BC patients (n = 87) by using: PAM50 gene expression panel and immunohistochemical evaluation of key proteins involved in tumorigenesis.

Results: Our PAM50 analyses show that tumours changing from luminal A to normal-like subtype after NET presented better radiological and pathological tumour responses, a significant larger decrease in Ki67 at surgery, lower preoperative endocrine prognostic index score (PEPI) and lower tumour cellularity size (TCS) than those with persistent luminal A status. Patients with the highest response to NET showed the largest decrease in PAM50-derived risk of recurrence (ROR) following NET. In addition, the percentage of p53 positive cells was associated with decreased response to NET.

Conclusions: Our findings highlight the change of intrinsic subtype from luminal A to normal-like after NET as a putative biomarker characterising the patient population that obtains the highest benefit from NET. Our study also suggests that changes in PAM50-derived ROR score and p53 evaluation could also help to identify those patients. Thus, this study uncovers potential biomarkers of response to NET and prognosis, which should be validated in independent cohorts, helping to the implementation of NET in the clinical practice.

Keywords: Aromatase inhibitors; ER+/HER2- breast cancer; Neoadjuvant endocrine therapy; Normal-like subtype; PAM50; ROR; p53.

Fig. 1

Association of changes in PAM50-intrinsic subtype after NET with response to therapy. (A) Alluvial diagram showing PAM50-associated intrinsic subtypes distribution in pre- and post-NET matched samples. (B) Pie charts representing the PAM50-intrinsic subtype distribution in pre-NET (upper panel) and post-NET (lower panel) samples. (C–H) Association between changes in intrinsic subtype (persistent luminal A (Persistent LumA) or change from luminal A to normal-like (LumA to Normal)) with radiological tumour response (rad-TR) (C), pathological tumour response (path-TR) (D), mPEPI (E), percentage (%) of Ki67 positive (+) cells at surgery (F), tumour cellularity size (TCS) (G) and in high or low TCS subgroups (H). Lum = Luminal, → = change, HER2-E = HER2-Enriched, CR= Complete response, PR = partial response, PD = Progressive disease, SD = Stable disease.

Fig. 2

Association of changes in ROR after NET with response to therapy. (A) ROR-S values in matched pre- and post-NET samples. (B) ΔROR-S values and subgroups definition considering cut-offs at: ΔROR-S = 0 (biological cut-off) and ΔROR-S = −2 (novel cut-off separating Q1 from Q2-Q4). (C–E) Association between change in ROR-S score after NET (ΔROR-S) with mPEPI (C), percentage of Ki67 positive (+) cells at surgery (D) and ΔKi67 (E).

Fig. 3

Association between p53 and response to NET. (A) Percentage (%) of p53 positive (+) cells in matched pre- and post-NET samples. The lines matching the pre- and post-treatment samples are colour-coded to indicate the magnitude and direction of the change: Blue lines indicate a decrease (n = 19), black lines indicate a decrease to zero (n = 23), red lines indicate an increase (n = 7) and green lines indicate no change (n = 2). (B) Representative pictures of 2 different patients showing the decrease of p53 (in pre- and post-NET samples) after NET. Scale bars indicate 100 µm. From each image, a region of interest has been zoomed to improve the visualisation of the staining. (C-D) Association of p53+ cells (%) pre-NET with mPEPI score (C) and percentage of Ki67 positive (+) cells at surgery (D). (E–G) Association of p53+ cells (%) post-NET with mPEPI (E), percentage of Ki67 positive (+) cells at surgery (F) and high or low TCS subgroups (G). (H–I) Association of the changes in p53 (∆p53) with mPEPI score (H) and high or low TCS subgroups (I).