Characterization of Functional Estrogen Receptor (ER) Dependence Via Comprehensive Epigenomic Liquid Biopsy Stratifies Endocrine Therapy (ET) Responders with Metastatic Breast Cancer (MBC)

Abstract Background: Endocrine therapy (ET) is standard of care for ER+ MBC, but virtually all patients (pts) develop ET resistance over their disease course. Biomarkers predictive of loss of ER dependence to guide therapy (tx) are lacking. Methods: We identified 122 pts with ER+/HER2- MBC enrolled in the EMBRACE study at Dana Farber Cancer Institute who received 1-2 lines (L) of ET and had a blood drawn within +/- 30 days of switch from 1L to 2L (n=77) or from 2L to 3L (n=53). Using 1mL plasma, we profiled genome-wide enhancers, promoters and DNA hypermethylation to infer pathway activation and gene expression levels. Samples that did not meet quality thresholds (n=16) or had tumor fraction <0.5% (n=48) were excluded. We applied the Precede ER dependence index (PERDI) that quantifies relative activities of ER-driven vs estrogen-starvation-induced enhancers and classified samples as PERDI-high or -low using a predefined threshold. We investigated the association between PERDI and time-to-next-treatment (TTNT). A conditional landmark approach defined TTNT beginning from 30 days after the start of tx of interest (ET or non-ET) until the start of the next tx line, considering only tx switch for tumor progression as events. Time-dependent ROC curves were used in the prespecified primary analysis investigating the association between PERDI and early progression on ET. Kaplan Meier methods and Cox proportional hazard models were used in additional analyses. Results: A total of 71 pts (79 samples) were included: 57 pts had draws collected before ET (40 switch from 1L to 2L ET, 17 from 2L to 3L ET); 22 pts had samples collected from 2L ET to 3L non-ET. Most common next-line ET regimens were SERD (n=17), SERD + CDK4/6 inhibitors (i) (n=14), aromatase inhibitors (AI) + mTORi (n=9), SERD + other targeted therapy (n=6). Among 57 pts starting next-line ET, PERDI was associated with landmark TTNT (median TTNT 2.4 months (mo) with low vs 4.1 mo with high PERDI, HR 2.65, 95% confidence interval [CI] 1.24-5.66, p=0.012). However, PERDI did not reliably identify early progressors to next-line ET (time-dependent AUC at 2 mo c = 0.51). Among 22 pts who switched from ET to non-ET (primarily chemotherapy), PERDI was not associated with TTNT (HR 0.66, 95% CI 0.23-1.85, p=0.426), suggesting a potential predictive role specifically for ET. Among pts with high PERDI but poor response to next-line ET, preliminary analyses identified 6 outliers for activity of known resistance pathways (e.g. FGFR1, ERBB2). Conclusions: We observed a significant association between PERDI and benefit from ET. Our exploratory analysis integrating resistance pathways beyond ER further identified pts with high PERDI but ET resistance. Upon further clinical validation, this blood based comprehensive epigenomic assay could become a valuable tool to guide tx for pts with ER+/HER2- MBC. Citation Format: Stefania Morganti, Jonathan Beagan, Ningxuan (Shirley) Zhou, Khoi Nguyen, Kalie Smith, Ashka Patel, Catherine Stever, Katheryn Santos, Molly Skeffington, Olivia D'Amico, Travis Clark, Justin Finkle, Anthony D'Ippolito, Mike Zhong, Jamey Guess, Kristian Cibulskis, Aparna Gorthi, Tyrone Tamakloe, Charlene O'Brien, Baovy Tran, Mary McGillicuddy, Nabihah Tayob, Sara M. Tolaney, Nancy U. Lin, Hillary Heiling, Corrie A. Painter, Matthew L. Eaton, J. Carl Barrett, Heather A. Parsons. Characterization of functional estrogen receptor (ER) dependence via comprehensive epigenomic liquid biopsy stratifies endocrine therapy (ET) responders with metastatic breast cancer (MBC) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 1020.