Neoadjuvant treatment of ER-positive/HER2-negative breast cancer with aromatase inhibitors in sequence: Ki67 dynamics and biology shifts

Background. Neoadjuvant endocrine treatment (NET) is used for locally advanced hormone receptor (HR)-positive, HER-2-negative breast cancer in highly selected patients. Aromatase inhibitors (AI) are the preferred option in this setting. Although NET has been proposed to be similarly effective as neoadjuvant chemotherapy (NCT) for certain patients, there are no reliable markers currently available to guide post-surgery decisions. Changes in Ki67 levels and results provided by multi-gene arrays (MGA) before and after neoadjuvant treatment have been suggested to be promising and clinically relevant markers for further decision making. Methods. The NEOLETEXE trial was an open-label, intrapatient cross-over trial including patients with locally advanced HR-positive, HER2-negative breast cancer. Patients were randomized 1:1 to NET with either letrozole or exemestane for 3 months, followed by an intrapatient cross-over to the alternative treatment for another 3 months. Extensive biobanking was performed at multiple time points before and during neoadjuvant therapy. In a subset of patients, gene expression profiling was performed using the Prosigna® (PAM50) assay on both the diagnostic biopsy and the final surgical specimen. Immunohistochemical Ki67 expression was assessed in core biopsies obtained at baseline and in excision specimens collected at the time of surgery. Results. A total of 84 patients were included in the intention-to-treat analysis. The median age was 76 years. Pathological complete responses (pCR) occurred in 6% (n=5). The median follow-up time was 6,3 years. Only nine patients (10,7%) relapsed during the follow-up period. An analysis of Ki67 at baseline (Ki67b) and at the time of surgery (Ki67s) was performed, with levels of Ki67 10% or higher classified as Ki67high. Kaplan-Meier analysis showed that patients with low Ki67 levels at surgery (<10%) experienced significantly better recurrence-free survival (RFS) compared to the Ki67 high group (HR 0,07, CI 0.02-0.31, p < 0.001).Prosigna testing was performed at baseline and at the time of surgery in a cohort of 20 patients. At baseline, 35% (n = 7) of the Prosigna cohort were classified as Luminal A, 57,9% (n = 11) as Luminal B, and one patient was identified as HER2-enriched. At the time of surgery, most tumors were classified as Luminal A (80%, n = 16), two patients were categorized as HER2-enriched, and only one patient remained in the Luminal B category. Conclusions. Our trial strongly underlines that NET involving AI as monotherapy for locally advanced HR-positive breast cancer is a pragmatic and effective alternative to NCT in highly selected patients. Ki67 expression data and MGA data at surgery turned out to be promising markers to potentially guide post-neoadjuvant decision-making and should be tested in future clinical trials.