Lab-Grown Brain Tumors Put Patient Immune Systems to the Test

Glioblastoma remains the most lethal primary malignant brain tumor in adults, and despite decades of surgical, radiotherapeutic, and chemotherapeutic refinement, median survival has barely moved. Immunotherapy, which has transformed the treatment of melanoma, lung cancer, and many other malignancies, has delivered strikingly little clinical benefit in this disease. A new review published in Cancer Immunology, Immunotherapy examines why the brain tumor immune landscape has proven so resistant to therapeutic manipulation, and argues that a relatively young technology—patient-derived glioblastoma organoids—may finally give researchers a realistic test bed for immune-based treatments before they ever reach a patient.

The review, authored by a team led by researchers at Chinese PLA General Hospital in Beijing together with colleagues at the Shenzhen Institutes of Advanced Technology and Tianjin University, lays out in systematic detail the barriers that have frustrated glioblastoma immunotherapy. The tumor is extraordinarily heterogeneous, both across patients and within a single lesion, with spatially distinct subclones carrying different antigenic profiles. Its immune infiltrate is dominated not by cytotoxic lymphocytes but by immunosuppressive myeloid cells, including tumor-associated macrophages and microglia, which actively dampen T-cell responses. Lymphocyte infiltration is restricted, antigens are unstable and evolve under treatment pressure, and the blood–brain barrier limits delivery of both antibodies and cells. Corticosteroid exposure, nearly universal in these patients, further suppresses immunity, and standard therapies themselves remodel the immune environment in ways that are difficult to model.

Conventional preclinical tools capture only fragments of this complexity. Established glioblastoma cell lines, propagated for years in serum, drift genetically and immunologically from the tumors they originally came from. Xenografts in immunocompromised mice cannot interrogate an intact immune system. Genetically engineered mouse models, while immunocompetent, do not reproduce the human tumor’s molecular and cellular idiosyncrasies. The result is a persistent translation gap: interventions that look promising in the laboratory repeatedly fail in the clinic because no model faithfully predicted how a specific patient’s tumor and immune system would interact with the drug.

Patient-derived glioblastoma organoids, or GBOs, have emerged as a partial answer. These three-dimensional cultures are generated directly from resected surgical specimens and, crucially, retain key histological, genetic, and phenotypic features of the parental tumor, including its cellular diversity and much of its microenvironmental architecture. Because they can be established rapidly and expanded in vitro, they allow functional therapeutic testing on a timescale that could theoretically inform clinical decisions. The review emphasizes that this combination of fidelity and tractability is what distinguishes organoids from earlier patient-derived systems such as neurospheres and spheroids.

The technical evolution of the platform is a central theme. Early organoid protocols produced relatively uniform tumor epithelial cultures, but newer approaches preserve immune cells within the organoid, generating what the authors describe as tumor-immune organoids that model the interaction between malignant cells and infiltrating myeloid and lymphoid populations. Co-culture strategies now allow researchers to introduce autologous or allogeneic immune cells, T cells engineered with chimeric antigen receptors, or immune checkpoint inhibitors, and to measure killing, cytokine release, and immune exhaustion in a controlled three-dimensional setting that approximates the tumor’s physical structure.

Applications span the full immuno-oncology repertoire. The review details how GBOs are being used to model responses to checkpoint blockade, to evaluate cellular therapies such as CAR-T and CAR-NK cells, to screen myeloid-targeting agents that reprogram immunosuppressive macrophages, and to test therapeutic vaccines. Multi-omic readouts—single-cell RNA sequencing, spatial transcriptomics, and epigenetic profiling applied to organoids before and after treatment—allow researchers to connect a treatment response to specific cell states and signaling pathways within the same patient’s tumor. Artificial intelligence-assisted image analysis is increasingly used to quantify spatial relationships between tumor cells and immune populations, extracting predictive features that would be invisible to manual assessment.

Particular emphasis is placed on oncolytic virotherapy, and specifically on engineered herpes simplex virus type 1 vectors. Oncolytic viruses are designed to infect and lyse tumor cells selectively while simultaneously provoking an antitumor immune response, and herpes-based vectors have the advantages of a large genetic payload capacity, well-characterized engineering tools, and clinical precedent. The review argues that GBOs are uniquely suited to interrogate the critical steps of viral therapy in human tissue: viral entry into tumor cells, the efficiency of intracellular replication, the spatial pattern of spread through the three-dimensional tumor mass, and the innate antiviral restriction mechanisms—such as type I interferon signaling—that can determine whether an oncolytic approach succeeds or fails in a given patient.

Beyond single-agent testing, organoids enable rational combination design. Immunogenic cell death triggered by viral lysis can release tumor antigens and danger signals, potentially converting a cold, lymphocyte-poor tumor into one that responds to checkpoint blockade. GBOs allow researchers to test such combinations—virus plus PD-1 inhibition, virus plus myeloid reprogramming, virus plus adoptive cell transfer—in the same specimen, identifying synergies and antagonisms before committing to a clinical regimen. The authors also outline staged, clinically integrated workflows in which organoid testing could be embedded within surgical and oncological care, with results returned in time to influence adjuvant treatment decisions.

The review is notably candid about limitations. Immune cells within organoids undergo attrition over time in culture, so experiments must be performed within narrow windows before the immune compartment degrades. Microenvironmental fidelity is incomplete: the blood–brain barrier, systemic immune circuits, and corticosteroid effects cannot be fully reproduced in a dish. Resource requirements are substantial, demanding rapid processing of fresh tissue, specialized culture infrastructure, and multidisciplinary expertise. Most importantly, the platform has not yet been validated prospectively—no trial has demonstrated that organoid-based testing reliably predicts clinical response. The authors call for small exploratory trials designed specifically to establish this predictive validity.

Even with those caveats, the trajectory is clear. As protocols standardize, as immune preservation improves, and as AI-driven analysis matures, patient-derived glioblastoma organoids are positioned to become a bridge between genomic profiling and functional precision medicine in neuro-oncology. For a disease in which nearly every immunotherapeutic advance has stalled at the bedside, the ability to rehearse a treatment against a living replica of a patient’s own tumor—complete with its immune defenders and saboteurs—represents one of the most consequential opportunities in the field. Whether that opportunity translates into longer survival will depend on the prospective studies the review’s authors say must now follow.

Subject of Research: Patient-derived glioblastoma organoids as models for tumor-immune interactions and oncolytic virotherapy in immuno-oncology

Article Title: Patient-derived glioblastoma organoids in immuno-oncology: tumor-immune modeling, oncolytic virotherapy, and translational opportunities

Article References: Liu, J., Liu, T., Wang, Y., Guo, B., Gao, Z., Su, S., Liu, Y., Xu, A., Li, W., Yang, Z., Li, M., Li, S., Wang, H., Lv, W., Xu, F., & Zhang, J. (2026). Patient-derived glioblastoma organoids in immuno-oncology: tumor-immune modeling, oncolytic virotherapy, and translational opportunities. Cancer Immunology, Immunotherapy. https://doi.org/10.1007/s00262-026-04530-9

Image Credits: AI Generated

DOI: 10.1007/s00262-026-04530-9

Keywords: glioblastoma, patient-derived organoids, immuno-oncology, oncolytic virotherapy, herpes simplex virus, tumor microenvironment, checkpoint blockade, CAR-T cells, myeloid cells, precision oncology, blood-brain barrier, immunotherapy