Actionable Pharmacogenomic Biomarkers in Breast Cancer Therapy: From Drug Response to Clinical Implementation

Abstract

Breast cancer therapy increasingly depends on biomarkers that identify drug sensitivity, acquired resistance, or treatment-related toxicity. Pharmacogenomic information becomes clinically relevant when a molecular result changes drug selection, dose, treatment sequence, or monitoring; consequently, germline and somatic biomarkers are most informative when interpreted within the clinical course of disease, from baseline pathological classification to molecular reassessment after treatment pressure and clonal evolution. This review examines the clinical actionability of germline and somatic biomarkers across this care pathway. PubMed/MEDLINE was searched for English-language literature published from January 2015 to August 2026, while reference lists of major guidelines and pivotal trials were screened for landmark evidence published before 2015. Targeted searches of ASCO, ESMO, CPIC, U.S. FDA, and publisher sources were additionally used to verify contemporary guidelines, regulatory decisions, companion-diagnostic requirements, approved dosing information, and pivotal clinical trials. Eligible evidence included clinical guidelines, regulatory documents, randomized or prospective studies, large translational cohorts, and highquality reviews directly addressing biomarker-guided treatment, dose modification, toxicity, resistance, or implementation. The evidence was synthesized as a structured narrative review; systematic-review and meta-analytic procedures were not applied. Clinically established applications span ER/PR/HER2-directed treatment selection, germline BRCA1/2-directed PARP inhibition, PIK3CA/AKT1/PTEN-directed pathway inhibition, ESR1-directed endocrine sequencing, and DPYD-guided fluoropyrimidine dose modification. CYP2D6 can inform tamoxifen interpretation when genotype is available, although outcome evidence and clinical uptake remain heterogeneous. In addition, PALB2 has prospective evidence of PARP-inhibitor sensitivity, whereas ATM and CHEK2 do not currently show equivalent predictive value. Liquid biopsy can identify acquired resistance alterations and support longitudinal testing; in contrast, integrated multi-omics, artificial intelligence/ machine learning, and quantitative systems pharmacology remain predominantly investigational. These findings support a context-specific approach to biomarker use, with routine implementation constrained by uneven evidence maturity, assay limitations, and access to matched therapies. Investigational approaches require prospective validation before informing prescribing. Clinical pharmacogenomic interpretation depends on the therapeutic question, specimen type, analytical method, strength of evidence, treatment consequence, and timing of reassessment. When these elements are integrated across the disease course, biomarkers with validated prescribing implications can be distinguished from findings that remain context-dependent or exploratory, thereby supporting treatment individualization without overstating molecular associations.