Abstract PS2-07-18: Prognostic effects of methylation-based HR and HER2 subtyping by liquid biopsy in ESR1 mutation-negative metastatic breast cancer

Abstract Background: Analysis of cell-free DNA (cfDNA) in blood has established clinical utility for the detection of somatic genomic alterations in cancer, but epigenetic signatures based on tumor-derived circulating methylated DNA is a less established approach. Here, we describe the application of a cfDNA methylation-based hormone receptor (HR) and HER2 breast cancer molecular subtype (BCMS) feature in clinically assigned HR+ HER2- MBC. Real word evidence (RWE) outcomes by epigenetic HR and HER2 status in ESR1-mutant cases will be reported at ESMO 2025. The aim of this study is to evaluate the prognostic properties of BCMS in patients with endocrine therapy (ET) resistance not associated with an ESR1 mutation. Methods: The GuardantINFORM™ RWE database includes de-identified administrative claims data from Guardant360® testing. Guardant360® integrates genomic and epigenomic components, including a methylation-based cfDNA BCMS feature to identify HR, HER2, and/or triple negative breast cancer (TNBC) status for patients with tumor fraction 0.5% with overall accuracy ranging from 85.73% to 91.88% (Tolkunov et al., AACR 2025). Included here were adult patients with aromatase inhibitor (AI)-treated MBC who received treatment after an ESR1 mutation (ESR1mut) was not detected. Clinical outcomes via "real-world time to treatment discontinuation” (rwTTD) and “real-world overall survival” (rwOS) were compared between patients assigned HR+/HER2- versus other categories (TNBC, HR-/HER2+, or HR+/HER2+). Propensity score matching was applied to balance baseline characteristics between study groups. Results: A total of 769 AI-exposed ESR1mut-negative clinically HR+/HER2- MBC cases had BCMS results available and no prior anti-HER2 therapy. Of these, 620 (80%) were HR+/HER2- by BCMS, while the remaining 20% comprised other subtypes (HR+/HER2+: n = 67, HR-/HER2+: n = 24, and HR-/HER2-: n = 63). Non-HR+/HER2- cases were grouped as there was a different subtype to that assigned clinically (HR+/HER2-). For patients who underwent ET only, rwTTD and rwOS were worse for cases not assigned HR+/HER2- (adjusted HR = 0.37, 95% CI 0.17-0.78, P = 0.009; adjusted HR = 0.18, 95% CI 0.42-0.77, P = 0.02). For patients who received ET plus CDK4/6 inhibition, rwTTD was worse (adjusted HR = 0.44, 95% CI 0.22-0.9, P = 0.024), and for those who received PIK3CA/AKT1 inhibition, there was no difference. Conclusions: cfDNA-based BCMS predicted RWE outcomes in previously AI treated patients receiving second-line therapy without ESR1 mutations. The subset categorized as not HR+/HER2- had significantly worse outcomes if assigned ET monotherapy. Additional RWE will be presented as well as a liquid biopsy-based genomic analysis of BCMS HER2+ cases that were classified as HER2-negative clinically. Epigenetic-based molecular breast subtyping may be a feasible, non-invasive clinical tool to reassess ER and HER2 status in MBC without requiring serial tissue biopsies. Citation Format: P. Razavi, N. Zhang, A. Hardin, A. Das, M. J. Ellis. Prognostic effects of methylation-based HR and HER2 subtyping by liquid biopsy in ESR1 mutation-negative metastatic breast cancer abstract. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS2-07-18.