Historical early treatment effects of adjuvant endocrine therapy for breast cancer in high-risk subgroups: Reanalysis of BIG 1-98, SOFT and TEXT.

508 Background: Clinical trials testing adjuvant endocrine therapy (ET) have not historically limited enrollment to patients with clinically high-risk breast cancer (BC), nor estimated treatment (trt) effects prior to 5 yrs. In light of recent trial results for adjuvant CDK4/6 inhibitors, we estimated early relative and absolute trt effects of aromatase inhibitor (AI) vs tamoxifen (T), with ovarian suppression (OFS) if premenopausal, in high risk subgroups. Methods: From pts enrolled in adjuvant phase 3 randomized clinical trials BIG 1-98 (postmenopausal, 5yr AI v T), TEXT (5yr AI+OFS v T+OFS) and SOFT (5yr AI+OFS v T+OFS v T), we identified subgroups with HR+/HER2- BC and high risk features (≥4 pLN; or 1-3 pLN with grade 3, pT≥5 cm and/or Ki-67≥20% [retrospectively centrally assessed]). TEXT randomized at start of adjuvant trt; BIG 1-98 and SOFT after chemotherapy. Disease-free survival (DFS) was defined from randomization to invasive recurrence at local, regional, distant or contralateral breast, second non-breast malignancy or death. We estimated trt effects as differences in 2, 3 and 5yr DFS Kaplan-Meier estimates (KM diffs at t yrs) and hazard ratios over time intervals of 0-2, 0-3 and 0-5 yrs (HRs over 0 to t yrs) to approximate the maturity of trial results with increasing follow-up. Results: The high-risk HR+/HER2- subgroups were 695/4922, 707/2660 and 526/3047 pts in BIG 1-98, TEXT and SOFT with 202, 122, 142 DFS events observed by 5 yrs since randomization. 60%, 46% and 44% of pts had ≥4 pLNs; 42%, 90% and 92% had chemotherapy. The table summarizes results. Conclusions: In re-analysis of clinical high risk HR+/HER2- subgroups of 3 trials each ̃700 pts, 5 yrs AI vs T had similar magnitude of early trt effects after 2-3 yrs follow-up of all pts vs trt effects observed after 27 mos median follow-up in monarchE trial (HR=0.70; KM diffs 2.7% at 2 yrs, 5.4% at 3 yrs; n=5637). Relative and absolute trt effects sometimes diminished when estimated over 5 yrs rather than over 2 yrs, but meaningful absolute differences remained at 5 yrs in contrast to Penelope-B trial results. Design and interpretation of high risk HR+/HER2- early BC trials may depend on timing of randomization and selection of backbone adjuvant trts; follow-up >5 yrs must remain standard. [Table: see text]