Linking preclinical models with functional imaging and artificial intelligence to decode tumor angiogenesis

Tumor angiogenesis is a hallmark of cancer, influencing oxygen delivery, immune cell trafficking, and therapeutic response. Although anti-angiogenic therapies have achieved clinical relevance, their efficacy remains limited by tumor heterogeneity, adaptive resistance, and insufficient functional assessment of tumor vasculature. In this review, we evaluate preclinical models of tumor angiogenesis with a focus on their ability to generate quantitative and physiologically relevant readouts. We discuss established assays such as tube formation, fibrin bead assay, aortic ring, and chorioallantoic membrane (CAM) models, alongside emerging systems including vascularized organoids and organ-on-a-chip platforms, as well as zebrafish and murine tumor models. Particular emphasis is placed on advanced imaging modalities enabling longitudinal assessment of perfusion, permeability, and vascular remodeling, as well as on the growing role of artificial intelligence (AI) in automated and quantitative data interpretation. We further highlight the integration of single-cell and multi-omics approaches for resolving endothelial heterogeneity and tumor-stroma interactions. By linking biologically relevant models with functional imaging and computational analysis, this review outlines strategies to improve the predictive power and translational relevance of angiogenesis research.