Despite major advances in radiation delivery, clinical outcomes remain constrained by tumor biology rather than technology. Conventional radiobiology has relied on reductionist two-dimensional (2D) systems that fail to capture the spatial, mechanical, and multicellular organization of tumors. Three-dimensional (3D) tumor models now resolve radiation response as a mechanobiological process coordinated across the extracellular matrix (ECM), adhesion signaling, cytoskeleton, and nucleus. This review focuses on a specific and we argue, underappreciated intersection: how 3D models expose integrin-mediated mechanotransduction as a determinant of the DNA damage response (DDR) and therapy resistance epitomized by cell adhesion-mediated radioresistance (CAM-RR) - an organizing principle we frame as the ECM-integrin-nucleus axis. We first delineate which model classes resolve which layer of this biology, distinguishing effects of three-dimensional organization from those of defined ECM-integrin signaling, and we treat the underlying mechanobiology quantitatively rather than descriptively. We then develop a mechanistic framework linking ECM architecture and stiffness to integrin-RTK crosstalk, cytoskeletal tension, LINC-mediated force transfer, and chromatin-dependent DNA repair, in which radiosensitivity emerges as a property of tissue context. From a translational standpoint, 3D models enable functional, radiation-specific assessment of context-dependent radiosensitivity and of mechanically targeted radiosensitization, while their integration with quantitative imaging and computational approaches further supports biomarker-guided and adaptive treatment strategies. We close with testable predictions that this framework generates, intended to guide the next phase of biology-driven radiation oncology.

Little is known as to whether there may be a patogenetic continuum among subsets of lung neuroendocrine neoplasms (NENs), including both neuroendocrine tumors (NETs) and neuroendocrine carcinomas (NECs). In this study, transcriptomic data from 205 pulmonary NETs and NECs were collected from four publicly available datasets and analyzed through supervised clustering of a curated 20-gene signature dealing with neuroendocrine differentiation, key oncogenic pathways, chromatin remodeling genes, and lineage-specific transcription factors. Two major clusters were identified, PNEN-A and PNEN-B. Compared to PNEN-A and independently of histological typing, PNEN-B demonstrated reduced neuroendocrine gene expression and increased activation of pulmonary epithelial cell lineages, such as glandular (club/AT2[+]), basal (ΔNp63[+]/p63[+]/CK5[+]), and tuft cell (POU2F3[+]) differentiation. PNEN-B tumors also exhibited features of inflamed but functionally constrained immune microenvironment, with increased expression of immunomodulatory neuropeptides (calcitonin, proopiomelanocortin, and gastrin-releasing peptide), alongside enrichment of T cell exhaustion markers. PNEN-B tumors significantly correlated with poor prognosis in a subset of cases. These findings reveal common transcriptional programs across the entire spectrum of pulmonary NENs, independent of histological typing. Attenuation of neuroendocrine differentiation, activation of alternative cell lineages, and modulation of microenvironment highlight subsets of tumors with potential pathogenetic and clinical implications.