A clinically translatable 2.5D lung adenocarcinoma organoid platform derived from malignant pleural effusions for precision drug screening and tumor-immune-stromal modeling

Conventional three-dimensional (3D) organoid models are limited by high cost, technical complexity, and poor replication of tumor-immune interactions. To address these limitations, we developed two novel 2.5D systems from malignant pleural effusion (MPE) of patients with lung adenocarcinoma. The first uses MPE-derived adherent cells to create matrix-free niches, while the second combines MPE supernatant with DMEM-F12 and 10-30 % fetal bovine serum, with or without agarose, to form a thermoresponsive matrix that mimics dynamic tumor microenvironments. Autologous MPE supernatants were used to preserve native cytokines and extracellular matrix components. Genomic and immune microenvironment analyses demonstrated over 90 % concordance between 2.5D organoids and parental tumors, which aligned with the standards reported for existing 3D tumor-organoid cultures. The 2.5D organoids retained key functional characteristics, including mucus secretion, phenotypic plasticity, and angiogenesis. Organoid-based high-throughput screening enabled identification of candidate therapeutics, and drug sensitivity assays correlated strongly with patient outcomes, underscoring their utility for personalized medicine. The matrix-free design facilitated rapid organoid generation (3-10 days) and reduced costs by approximately 90 % relative to conventional 3D methods. These 2.5D models provide a clinically translatable platform that replicates tumor heterogeneity while overcoming cost and technical barriers, offering a promising tool for precision pharmacology and oncology.