Abstract 2671: Divergent evolutionary trajectories of the molecular subtypes of breast cancer are driven by differential selection and selective epistasis

Abstract The molecular subtypes of breast cancer—defined by estrogen-receptor, progesterone-receptor, and human epidermal growth factor receptor status—have established differences in mutational profiles, prognoses, and response to standard treatments. However, the evolutionary trajectories underlying these differences are not well understood. Knowledge of these differential trajectories would empower predictive modeling of treatment benefit. To characterize these trajectories, we first aggregated a multi-institutional cohort of 23, 247 breast cancer patients with whole exome or targeted sequencing data, including from AACR Project GENIE. We then calculated variant-specific background mutation rates, and subsequently quantified the selective benefit of known and potential driver mutations across breast cancers and within each subtype. We then assessed how the selective benefit of such mutations changed throughout cancer progression and due to the epistatic influence of prior mutations. Finally, we confirmed the ordering of mutational acquisition determined by epistasis by performing phylogenetic reconstruction of the evolution of ten breast cancers. In a subtype-agnostic analysis, we found that PIK3CA H1047R and H1047L, SF3B1 K700E, and multiple TP53 variants experienced the strongest selection. Interestingly, the selective benefit of mutations to oncogenesis varied between subtypes: the PIK3CA H1047R variant experienced maximal selection in the Luminal A subtype, while the TP53 I195T variant experienced maximal selection in TNBC. Additionally, we found that PIK3CA E542K experienced significantly higher selection in the Luminal A and B subtypes than the TNBC subtype. These mutations also experienced differential selection throughout cancer progression: PIK3CA and CDH1 variants were under stronger selection during early carcinogenesis, whereas TP53, ERBB2, and multiple other variants experienced stronger selection in late stage cancer. These differences in subtype- and stage-specific selection are partially explained by asymmetric synergistic and antagonistic epistatic interactions between mutations in PIK3CA, TP53, AKT1, CDH1, and other driver genes, which create preferential mutation orderings and combinations. These analyses enable inference of subtype-specific tumor evolutionary trajectories with quantitative measures of the selective benefit of mutations across molecular contexts and stages of cancer progression. Such measures are useful for predicting response to targeted therapeutics and therefore facilitate the development and testing of targeted therapies—and the optimization of current regimens—for the molecular subtypes of breast cancer. Citation Format: Jessica Liu, Krishna Dasari, Nicholas Fisk, Carolyn Kravitz, Lajos Pusztai, Jeffrey Townsend. Divergent evolutionary trajectories of the molecular subtypes of breast cancer are driven by differential selection and selective epistasis [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 2671.