Abstract
KRAS, once considered undruggable, has become actionable across specific alleles, with KRAS-G12C inhibitors now approved and next-generation approaches-including pan-KRAS/pan-RAS inhibitors, targeted degraders and RNA-based strategies-progressing rapidly. However, clinical benefit remains limited due to the frequent emergence of resistance. Escape mechanisms include on-target secondary mutations, pathway reactivation, epithelial-mesenchymal transition, lineage plasticity and metabolic rewiring within an immunosuppressive tumour microenvironment. Emerging evidence supports rational combination strategies, including parallel inhibition of epidermal growth factor receptor, protein tyrosine phosphatase non-receptor type 11 or SOS1 and vertical blockade of the mitogen-activated protein kinase-extracellular signal-regulated kinase or phosphatidylinositol 3-kinase-mechanistic target of rapamycin cascades; immunotherapies such as checkpoint blockade, T-cell receptor (TCR)-T cells, bispecific T-cell engagers or cytokine-armed oncolytic viruses; metabolic interventions targeting macropinocytosis or autophagy; as well as radiotherapy. Such combination therapies can transform primarily cytostatic effects into more durable antitumour responses, although with potential toxicity constraints. Precision approaches that integrate multiomics profiling with longitudinal circulating tumour DNA analysis enable biomarker-guided patient selection (eg, based on STK11 and KEAP1 comutations) and support therapeutic adaptations, including sequencing strategies and intermittent dosing. Thus, network-level KRAS interception combined with biomarker-driven, clonal evolution-informed trial design offers a path towards sustained control of KRAS-driven cancers.