Abstract 3371: Gene deletions in estrogen receptor positive HER2-negative breast cancer are associated with worse outcome than predicted by diagnostic-grade gene expression assays

Abstract Introduction: Currently, risk of recurrence in ER+ HER2- breast cancer is determined by pathological stage and histological grade, in combination with a Gene Expression Assay (GEA) including Oncotype, Prosigna, Breast Cancer Index and EPClin. Higher risk disease justifies systemic therapy beyond endocrine therapy alone. GEAs need improving as these assays can misclassify risk, i.e., recurrences still occur in patients assigned low risk. Thus, approaches to identify high-risk cases more accurately would reduce under-treatment. Herein we explore whether somatic mutation and copy number data from an established 83 gene targeted sequencing panel can improve risk stratification based on clinical factors and GEA alone. Methods: Targeted gene-sequencing (TGS) was undertaken in the Anastrozole (A) or Tamoxifen (T) alone or Combined (C) (ATAC) adjuvant trial for estrogen receptor-positive (ER+) breast cancer to identify mutations and copy number variants that are prognostic variables independent of clinical factors and GEA-based risk assessments. Patients in the combination arm were excluded for this study. Forward Stepwise Cox Regression was applied to identify candidates, whose prognostic effects were validated in independent datasets. Results: Investigation of prognostic interactions of gene alterations with four GEA mentioned above identified three adverse gene deletions (del) including: 1) MAGI3del at chr1p.13.2 (HR=2.43-2.86, p-value< 0.005), a newly assigned tumor suppressor, that associated with other adverse deletion events on chr1p13.2-36.22 including ARID1A/JAK1/mTOR; 2) METdel on chr7q.31.2 (HR=1.86-2.15; p<0.05), adjacent to the tumor suppressor CAV1; and 3) RB1CC1del on chr8q.11.23 (HR=3.41-4.30; p<0.05), also a putative tumor suppressor. Potential clinical impact included high-relapse risk reassignment for Oncotype/Prosigna/Breast Cancer Index (BCI)-defined low and/or intermediate-risk cases if MAGI3del was present, and high-risk reassignments of BCI intermediate-risk cases if any of MAGI3/MET/RB1CC1 deletions were present, with lower-risk reassignment if absent. Conclusions: Copy number variation (CNV) analysis identified three high-risk deletion events (MAGI3, MET, RB1CC1) in ER+ HER2- breast cancer that predicted worse clinical course than predicted by diagnostic-grade gene expression assays. Thus, the incorporation of CNV-detection into prognostic algorithms has the potential to reduce under-treatment due to low/intermediate-risk mis-assignment. These data are proof-of-concept that prognostication in ER+ HER2- disease can be improved by TGS, while also identifying under-studied gene deletions underlying the highly variable relapse trajectories that typifies ER+ breast cancer. Citation Format: Jonathan T. Lei, Kim Chu, Brian Li, Eugene Schuster, Zachary L. Skidmore, Shyam M. Kavuri, Vikrant Mehta, Nicholas C. Spies, Jason Kunisaki, Catrina Fronick, Robert S. Fulton, Malachi Griffith, Richard Buus, Jack Cuzick, Mitch Dowsett, Meenakshi Anurag, Obi L. Griffith, Matthew J. Ellis. Gene deletions in estrogen receptor positive HER2-negative breast cancer are associated with worse outcome than predicted by diagnostic-grade gene expression assays [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 3371.