Prognostic and predictive impact of baseline gene expression (exp) in the NATALEE trial of adjuvant (adj) ribociclib (RIB) + nonsteroidal aromatase inhibitor (NSAI) in HR+/HER2− early breast cancer (EBC).

Abstract

501 Background: In the NATALEE clinical trial, RIB + NSAI significantly improved invasive disease–free survival (iDFS) vs NSAI alone in patients (pts) with stage II/III HR+/HER2− EBC. We report on the prognostic and predictive value of baseline gene exp in NATALEE. Methods: Pts were randomized 1:1 to RIB + NSAI or NSAI alone. Men and premenopausal women received goserelin. Eligible pts had anatomical stage IIA (if N0 with additional risk factors [G3, or G2 with Ki67 ≥20% or high genomic risk] or N1 [1-3 axillary lymph nodes]), IIB, or III disease per AJCC (8th ed). Gene exp in surgical samples was profiled by NanoString BC360 panel. PAM50-based intrinsic subtype × treatment (tx) interaction was estimated by likelihood ratio test. Prognostic and predictive effects of PAM50 subtypes and genomic risk/proliferation signature scores created from gene exp data were assessed with Cox proportional hazards models. Differences in genomic risk and proliferation scores were analyzed by Wilcoxon test. Associations between gene exp and RIB benefit were estimated with Cox models. Results: In all, 3022 baseline surgical samples were analyzed. iDFS benefit with RIB + NSAI was consistent in the biomarker (hazard ratio [HR], 0.72) and intent-to-treat populations (HR, 0.71). PAM50 subtype distribution was comparable across tx arms and differed between N0 and N1-N3 in pts with luminal A (LumA; 50% vs 68%), luminal B (lumB; 41% vs 26%), and basal-like (BSL; 6.4% vs 2.5%) disease, with similar percentages for HER2-enriched (HER2E; 2.5% vs 3.0%). PAM50 subtypes were strongly prognostic (HRs vs LumA: LumB, 1.39; HER2E, 2.62; BSL, 3.92). RIB had benefit across all PAM50 subtypes (HRs: LumA, 0.77; LumB, 0.71; HER2E, 0.50; BSL, 0.42), with no significant subtype × tx interaction ( P = .34). Higher genomic risk signature or proliferation signature scores showed a trend for increased RIB benefit, again with no significant interaction. Pts with N0 vs N1-N3 had significantly higher genomic risk scores and proliferation scores. Exploratory analysis of the predictive effect of gene exp on RIB benefit identified several genes for which higher (eg, CEACAM6 , NOD2 , and GPX3) or lower exp (eg, GATA3 , SLC39A6 , and MAPT ) was associated with increased RIB benefit. Conclusions: In this analysis of NATALEE, which examined the largest dataset of surgical tumor samples from any adj CDK4/6i trial in HR+/HER2− EBC, RIB had benefit across all PAM50 subtypes, with a trend for increased benefit in pts with higher genomic risk signature or proliferation signature scores. Baseline exp levels of several genes were associated with differential RIB benefit in EBC, showing potential predictive and prognostic value. The findings reinforce the therapeutic benefit of RIB in combination with ET across HR+/HER2− EBC populations. Clinical trial information: NCT03701334 .