Rédiger un PREreview

Breast cancer therapy has advanced substantially over the past two decades, yet disease recurrence and metastatic progression remain the leading causes of mortality. Resistance in hormone receptor-positive (HR+), HER2-positive, and Triple-negative (TNBC) disease is conventionally treated as three separate, each driven by its own escape route. This review suggests that resistance across all three subtypes instead converges on a shared, limited set of adaptive programs: epithelial-mesenchymal plasticity, cancer stem cell enrichment, metabolic rewiring, and microenvironment-mediated immune evasion. We trace the stage-wise clinical trajectory and dominant point of failure within each subtype, then examine how large-cohort genomics, single-cell transcriptomics, circulating tumour DNA, and machine learning have exposed this convergence, and the barriers still limiting its clinical use. To move beyond a purely mechanistic argument, we present an exploratory analysis stratifying overall survival within each subtype by which adaptive program dominates the resistant tumour. EMT enrichment in TNBC, immune evasion in HR-positive disease, and metabolic rewiring in HER2-positive disease. Although based on modest cohorts, this pattern suggests the four programs are genuinely different drawing on a shared cellular toolkit, rather than interchangeable descriptions of the same biology, and supports testing metabolic-program enrichment as a maintenance biomarker in HER2-positive disease. Closing the gap between mechanistic knowledge and clinical practice will require replacing subtype-exclusive resistance biomarkers with subtype anchored program-specific panels deployed alongside the standard clinical panel, shifting management from a model that reacts to resistance toward one that anticipates it.