Breast cancer is the most common cancer in women globally and shows significant heterogeneity. This heterogeneity is present at inter- and intratumoural levels and influences disease progression, therapeutic response along with clinical outcomes. Intertumoural heterogeneity arises from molecular and genomic differences in patients with distinct clinical subtypes and variable prognosis. Intratumoural heterogeneity comes from interactions between tumour microenvironmental factors and intrinsic cancer cell properties like genetic and epigenetic diversity, transcriptomic diversity, proliferative capacity, and stemness. Cancer cell plasticity allows dynamic reprogramming and helps in adapting to environmental pressure which promotes survival and evolution. Processes like clonal evolution, epithelial- mesenchymal transition (EMT), and interclonal cooperation further stimulate tumour progression, metastasis, and resistance to therapy. Heterogeneity is a major hurdle in designing treatment modality, mainly for metastatic breast cancer.
Molecular Subtypes and Therapy Resistance
Breast cancer shows intratumoural heterogeneity across subtypes like luminal A/B, HER2- enriched, basal-like, and claudin-low, fueled by clonal evolution and non-genetic cell-state plasticity. Tumour cells dynamically interconvert subtypes, influenced by the tumour microenvironment (TME) including cancer-associated fibroblasts (CAFs), macrophages, immune cells, and hypoxia, promoting metastasis and resistance. Single-cell RNA-seq reveals hierarchical structures with multiple cancer stem cell (CSC) populations undergoing EMT and mesenchymal-epithelial transition (MET). For instance, ER+/HER2− primaries shift to HER2+ in metastases. ITH enables therapy escape, as seen in residual disease post- neoadjuvant treatment showing increased plasticity markers. Hybrid epithelial/mesenchymal states enhance invasiveness and stemness.
Cancer stem cells (CSCs) and Plasticity
Cellular plasticity in breast cancer drives tumour progression, metastasis, and therapy resistance through EMT and CSC properties. Cells toggle phenotypes without genetic changes, navigating the metastatic cascade-from invasion and circulation survival to colonization-via hybrid epithelial-mesenchymal states and stemness. Pathways like MAPK, PI3K/AKT, STAT3,Wnt, and Notch regulate this adaptability, worsened by signals from the microenvironment. Targeting plasticity can be promising against resistant metastatic disease. Breast cancer heterogeneity due to CSCs and linked to EMT is determined to be a key factor in cancer relapse and metastasis. This again compels for the need of targeted therapies. Research confirms breast cancer stem cells (BCSCs) are responsible for tumour initiation, progression, metastasis, resistance, and recurrence. Advances in identifying BCSCs and understanding their signaling pathways have enabled targeted therapies that disrupt quiescence, overcome treatment resistance, and inhibit stemness mechanisms.
Tumour microenvironment (TME)
The TME is now recognized as a critical contributor to breast cancer heterogeneity. Tu et al. (2025) analysed tumour microenvironments in 14,837 breast cancers and identified 7 distinct ‘TME types’. These TME types can independently predict clinical outcomes. They concluded B cell lineages as important prognostic factors. Their depletion was associated with metastasis, suggesting TME composition can guide personalized treatment strategies.
Diagnostic Challenges and Emerging Therapeutic Opportunities
Shihao Sun et al., 2025 conducted multi-scale analysis of triple-negative breast cancer, revealing significant cellular heterogeneity. They identified maternally expressed gene 3- positive (MEG3+) pre-cancer-associated fibroblasts subgroup and found BRCA1 wild-type patients had increased T-cell exhaustion and dendritic cell tolerance compared to BRCA1 mutants. They proposed Interferon-stimulated gene 15 (ISG15) as an immunoregulatory biomarker and developed a machine learning-based predictive system for immunotherapy response forecasting, though further validation is needed.
Recent integrative analyses further stress on the prognostic relevance of tumour microenvironment heterogeneity. A comprehensive bioinformatics and single-cell RNA sequencing study developed a tumour microenvironment related prognostic model that can classify breast cancer patients into different risk groups and also predict responsiveness to immunotherapy. The study identified endothelial cells as high-risk cellular populations associated with poor prognosis and demonstrated variations in drug sensitivity across cellular subtypes, highlighting the potential of TME-based molecular signatures for guiding personalized therapeutic strategies.
There are significant diagnostic challenges due to high inter-observer variability and low assay sensitivity when differentiating between HER2-low and HER2-null subtypes in heterogeneous tumours. This problem can be solved by using antibody-drug conjugates like trastuzumab deruxtecan as they can kill neighbouring cancer cells even when they vary in HER2 expression levels due to the bystander killing effect.
Conventional chemotherapy fails in many cases of triple-negative breast cancer (TNBC) mainly due to high heterogeneity, drug resistance and lack of specific targets. Molecular typing systems like Lehmann, Burstein, and Fudan typing have showed various subtypes with different molecular characteristics and treatment responses. This can help in the development of targeted therapies like PARP inhibitors, PI3K/AKT/mTOR pathway inhibitors, and antibody-drug conjugates for more precise and personalized treatment. Liquid biopsy and multi-omic biomarkers have transformed breast cancer diagnosis. These techniques are less invasive and real-time monitoring of ctDNA, CTCs, exosomes, and proteomic profiles is possible. These methods along with the integration of AI provide early detection, better therapy guidance and relapse prediction compared to traditional tissue biopsies. Challenges presented are standardization, costs, and inequality in acess.
Conclusion:
Clinical trial designs need to take into consideration tumour evolution and its innate complexity from the beginning. Multimodal molecular profiling, liquid biopsies, and dynamic treatment selection provide real-time monitoring of tumour changes and thus can help in adaptive therapy. Understanding biological heterogeneity in breast cancer is now essential for advancing personalized adaptive oncology to significantly improve long-term patient outcomes.
Upasana Pathak
DESTINY-Breast09: The Trial Changing the Future of HER2-Positive Breast Cancer
