Dataset Information

Development of an AI Model to Predict Tumor Immune Responsiveness in Colorectal Cancer Using Organoid–Immune Co-Culture platforms

ABSTRACT: This study aimed to functionally assess tumor immune responsiveness in colorectal cancer using a co-culture platform combining patient-derived organoids (PDOs) with autologous tumor-reactive T cells. The following samples were classified as immune-responsive based on significant T cell–mediated cytotoxicity and IFN-γ production in co-culture assays. RNA sequencing was performed to identify transcriptional signatures associated with immune activation. These immune-responsive PDOs served as the discovery cohort for biomarker identification, enabling the integration of functional immunogenicity data with transcriptomic profiling to develop explainable AI models for predicting immunotherapy response in colorectal cancer.

ORGANISM(S): Homo sapiens

PROVIDER: GSE297912 | GEO | 2026/05/25

REPOSITORIES: GEO

ACCESS DATA

Dataset's files

Source:

			Action	DRS
		Other

Items per page:

1 - 1 of 1

Similar Datasets

Project description:The interaction of neoplastic cells with their tumor microenvironment (TME) is required for cancer progression. However, in vitro cancer models, including recent in vitro 3-dimensional (3D) organoid cultures of primary human tumors, are typically comprised exclusively of neoplastic epithelium, with stromal and/or immune interactions requiring artificial reconstitution. As relevant to cancer immunotherapy, the unified co-culture of primary tumor epithelia with their endogenous, syngeneic tumor-infiltrating lymphocytes (TILs) as a cohesive unit has remained particularly elusive. Here, we used a single 3D air-liquid interface (ALI) methodology to successfully propagate 3D Patient-Derived Tumor Organoids (PDOs) as primary tumor epithelia together with native immune and myofibroblast stromal compartments without reconstitution. Derived from >100 diverse human surgically resected tumor samples or from murine tumors in syngeneic immunocompetent mice, PDOs preserved cancer histologic subtypes and mutational spectrum with endogenous T, B, NK cells and macrophages integrally embedded amidst the tumor epithelium. PDO-based TILs accurately recapitulated the T cell receptor (TCR) spectrum of the original tumors, as determined by a robust droplet-based immune profiling solution that links gene expression and immune repertoire in single cells. Anti-PD-1 or anti-PD-L1 treatment of organoids from murine tumors from syngeneic immunocompetent hosts induced activation and cytolytic activity of tumor antigen-specific TILs, indicating successful PDO modeling of immune checkpoint blockade (ICB) and anti-tumor immunity. Crucially, the anti-PD-1 antibody nivolumab activated TILs in PDOs from human lung, renal and melanoma clinical tumor resections. The organoid-based propagation of primary tumor epithelium en bloc with its endogenous immune stroma should facilitate mechanistic investigation of TME-specific local tumor immunity, with applications for functional testing of individualized patient immunotherapy responses.

Dataset Information

Development of an AI Model to Predict Tumor Immune Responsiveness in Colorectal Cancer Using Organoid–Immune Co-Culture platforms

Dataset's files

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets