Neural Remodeling Across Cancer Evolution: Temporal, Spatial, Cellular, and Functional Principles of Tumor–Nerve Ecosystems

Cancer neuroscience now asks not only whether tumors contain nerves, but how neural architecture, neuronal and glial states, and neural output change through cancer evolution. This critical narrative review integrates evidence from precursor lesions, primary tumors, invasion, metastasis, and treatment-associated states, while distinguishing neural remodeling from axonogenesis, neurogenesis, perineural invasion, and direct synaptic integration. We compare trophic, axon-guidance, metabolic, extracellular-vesicle, stromal, and neuroimmune mechanisms and emphasize that the direction of neural effects depends on stage, neuronal subtype, receptor context, and tissue immunity. The review operationalizes the proposed neural microdomain as a spatially bounded, measurable unit linking neural identity and topology to neighboring tumor, immune, stromal, and vascular states. It also separates neural abundance, neural state, and neural activity, and applies a consistent evidence taxonomy spanning anatomical, molecular, preclinical functional, human observational, and interventional evidence. Recent reports involving treatment-associated remodeling, systemic circuits, metastatic adaptation, and neural biomarkers are treated as emerging or preclinical where appropriate. We provide practical criteria for microdomain measurement, patient-level quantitative analysis, controls for denervation-related confounding, and distinctions among prognostic, predictive, and response biomarkers. The central translational conclusion is that neural interventions should be stage- and circuit-specific, biomarker-supported, and designed to preserve essential physiological neural functions.