Deconstructing cancer in 3D: models, mechanisms, and personalized solutions.

Three-dimensional (3D) cancer models, notably patient-derived organoids (PDOs), address the critical limitations of traditional preclinical systems, including two-dimensional (2D) monolayer cultures and patient-derived xenografts (PDXs), by better recapitulating physiological tumor architecture and patient-specific heterogeneity, thereby revolutionizing oncology research. We chart the complementary technological landscape, from high-fidelity PDOs and scalable spheroids to engineered systems that reconstruct the tumor microenvironment (TME) via co-culture, organ-on-a-chip, and 3D bioprinting. Beyond foundational biology, these tools are driving functional precision medicine, where PDO avatars predict clinical drug response, and accelerating drug discovery through physiologically relevant screening. A core focus is their unparalleled utility in modeling therapy resistance, enabling the induction and multi-omic dissection of resistant clones, deconstructing stroma-mediated protection, and testing rational combination therapies to overcome relapse. Despite challenges in standardization and complete TME integration, the convergence of 3D models with single-cell omics, CRISPR screening, and artificial intelligence heralds a new era of predictive oncology. Ultimately, rigorous validation and clinical translation of these models promise to bridge the gap between bench and bedside, enabling truly personalized and effective cancer therapies.