Few data are available on survival of breast cancer (BC) patients (pts) treated with neoadjuvant endocrine therapy (NET). Understanding how tumor response to NET is related to relapse risk would help clinicians make decisions about additional treatments for these pts.

The primary objective was to assess 5-years relapse-free survival (RFS) in previously published two sister phase II studies evaluating anastrozole and fulvestrant as NET in post-menopausal HR-positive/HER2-negative BC pts (CARMINA02 NCT00629616 ; HORGEN NCT00871858). Prognostic factors of RFS were explored by pooling the two databases : age (<70 years versus>70), tumor and node staging, histological type and grade, clinical response, pathological response according to the Sataloff classification, baseline and surgical ER Allred score and Ki-67 expression, treatment arm, preoperative endocrine prognostic index (PEPI) score, adjuvant chemotherapy. The factor study (Horgen vs. Carmina02) was also introduced in the model as covariate. Univariate and multivariate Cox regression analysis (covariates from univariate model with P-value<0.15) were performed.

Intention-to-treat analysis is available for 217 of 236 pts included in the two trials. Post-menopausal pts with T2–T4, N0–N1, M0, HR-positive/HER2-negative BC were randomized to receive anastrozole (PO, 1mg daily) or fulvestrant (IM, 500mg on D1, D15 and D29 and then every 4 weeks) for 4 to 6 months before surgery. Mean age at diagnosis was 71.2 years. Baseline tumor characteristics were : 73.7 % T2, mean tumor size 44.3mm, 66.4 % N0, 19.4 % SBR grade 1, 98.6 % ER Allred score 6-8, mean Ki67 score 17.5 % with Ki67>20 % in 29.5 % of cases. A clinical response (complete or partial) was observed in 50.2 % of pts and pathological response rate defined as TA or TB and NA or NB was 15.7 %. Mean Ki67 at surgery was 8.8 % (11.4 % of cases>20 %). Among the pts, 12.4 % were classified in PEPI risk group I (risk score 0), 35.5 % in group II (score 1–3) and 30.4 % in group III (score>4) (PEPI not available for 21.7 % of pts) ; 21.7 % of pts received adjuvant chemotherapy. Median follow-up is 65.2 months. RFS and overall survival at 5 years are 83.7 % [95 % CI : 77.9–88 %] and 92.7 [95 % CI : 88.2–95.6 %] respectively. On univariate analysis, tumor staging (T3–4 versus T2, P=0.0001) and PEPI score (I+II versus III, P=0.0004) are associated with RFS. So far, no relapses/death have been observed in pts with pathological response (n=34, 15.7 %). Multivariate analysis (T, baseline Ki67, ER Allred score at surgery, PEPI and study) shows that PEPI group III is associated with significantly worse RFS (P=0.0014, HR : 3.9 [95 % CI : 1.7–9.0]).

Post-menopausal HR-positive/HER2-negative BC pts receiving NET have a very favorable outcome. Within this population eligible for NET, the PEPI score identifies a subset of pts of poorer prognosis who are candidate for further additional treatment like chemo/or targeted therapies. The PEPI may be a tool routinely used for individualization of adjuvant treatment in NET setting. According to recent published studies, further analysis will be performed with Ki67>10 % and PEPI risk (group I versus groups II–III).