Oncolytic Viruses Move Beyond Melting Tumors to Ignite Whole-Body Immunity

Oncolytic virotherapy has long been framed by a deceptively simple metaphor: convert immunologically cold tumors into hot ones by flooding them with inflammatory T cells. A new Perspective published in Nature Reviews Clinical Oncology argues that this cold-to-hot paradigm, while useful, captures only part of what engineered cancer-killing viruses can actually achieve. Writing from the vantage of both academic neurosurgery and industry development, William Jia, Ronghua Zhao, Howard L. Kaufman and Robert L. Martuza contend that oncolytic viruses, or OVs, should be understood not as local tumor-lysing agents with incidental systemic effects, but as systemic immune-reprogramming platforms that happen to be delivered locally. The distinction is more than semantic, because it reframes how these agents should be engineered, tested in the clinic and combined with checkpoint inhibitors.

The authors ground their argument in a fundamental limitation shared by immune checkpoint inhibitors and early-generation oncolytic viruses alike: both depend on pre-existing tumor-specific T cells to work. Checkpoint blockade releases the brakes on T cells that already recognize cancer, but it has little capacity to generate new tumor-reactive clones de novo. Similarly, first-generation OVs were designed primarily to replicate in and destroy tumor cells, releasing antigens in the hope that an antitumor response would follow. If a patient’s immune system has not already been sensitized to their cancer, the ceiling on efficacy is set by the existing T cell repertoire, an immunological constraint that clinical experience has repeatedly confirmed. Resistance rates for checkpoint inhibitors across tumor types remain substantial, and even approved OVs have produced durable responses mainly in a subset of patients.

The central conceptual advance proposed in the Perspective is that next-generation OVs overcome this ceiling by functioning as antigen-agnostic, in situ cancer vaccines. When an oncolytic virus infects a tumor, it triggers immunogenic cell death, a form of tumor destruction that releases not just one or two chosen antigens but potentially the full cancer proteome, all under adjuvant conditions created by pathogen-associated and damage-associated molecular patterns. Dendritic cells patrolling the inflamed tumor microenvironment engulf this debris and migrate to draining lymph nodes, where they present the entire antigenic spectrum of that patient’s cancer, including private neoantigens arising from mutations unique to the tumor. This process can broaden the T cell clonotype repertoire, effectively priming brand-new tumor-specific T cells rather than merely reactivating exhausted ones. In essence, each treated tumor becomes its own personalized vaccine factory, without needing to sequence a patient’s genome or manufacture an individualized product.

The authors organize the path forward around four pillars. The first is intratumoural vaccination as immunological ignition. They cite early clinical data demonstrating T cell clonotype broadening, regression of uninjected, so-called abscopal lesions, and survival benefit in patients whose disease had already failed checkpoint inhibitor therapy. These signals matter because abscopal responses, long considered rare curiosities of radiotherapy and immunotherapy, provide direct evidence that a locally delivered virus can reprogram immunity systemically. The second pillar is optimized, payload-driven immune priming. Modern OVs are increasingly engineered to carry transgenes encoding cytokines such as granulocyte-macrophage colony-stimulating factor, interleukin-12 and interleukin-15, or antibodies and nanobodies that block checkpoint pathways or hyperactivate antigen-presenting cells. The design goal is to convert the natural viral danger signals into a maximally productive priming event for dendritic cells and, through them, for naive T cells.

The third pillar addresses a quieter crisis in the field: how efficacy is measured. Standard RECIST criteria, which track the shrinkage of injected and measurable lesions, can systematically underestimate the delayed, nonlinear kinetics of immune-mediated tumor control. The authors call for revised evaluation frameworks incorporating immune-specific response criteria such as iRECIST, attention to durable response rates, and novel biological correlates such as circulating tumor DNA dynamics and T cell receptor repertoire diversification. Evidence already suggests that for some immunotherapies, including oncolytic agents, apparent stable disease can conceal a durable immune equilibrium that translates into extended overall survival even without dramatic radiographic regression. Regulatory acceptance of endpoints that capture these patterns will be essential if next-generation OVs are to reach patients.

The fourth and most ambitious pillar is the positioning of OVs as the foundational immuno-oncology platform, formalized in what the authors call the triple-A framework. Productive antitumor immunity requires three sequential gates: admission of T cells into the tumor, their availability in sufficient numbers with appropriate specificity, and their activation to effector function. Most therapeutic modalities satisfy only one or two of these conditions. Checkpoint inhibitors excel at activation but assume T cells are already present and tumor-specific. Adoptive cell therapies and bispecific T cell engagers supply availability and activation but struggle with physical admission into immunosuppressed, poorly vascularized tumor stroma. Payload-engineered next-generation OVs, the authors argue, are unique in satisfying all three prerequisites simultaneously: viral infection remodels the microenvironment to admit T cells, in situ vaccination generates and expands tumor-specific clones to ensure availability, and inflammatory danger signals plus engineered payloads drive activation.

The clinical landscape they survey is evolving rapidly. Talimogene laherparepvec, the first approved oncolytic virus in the United States, established proof of principle in melanoma, and combination trials with pembrolizumab and ipilimumab have tested whether viral priming can amplify checkpoint blockade. A strategy the authors describe as OV-prime, ICI-amplify is supported by trial data showing that vaccinating the immune system with a virus first and then removing inhibitory brakes with an antibody can yield benefit even in patients refractory to checkpoints alone. Newer agents illustrate the payload engineering trend: RP1, an oncolytic herpesvirus expressing GM-CSF and a fusogenic protein, has shown activity with nivolumab in anti-PD-1-failed melanoma and recently gained support from a US Food and Drug Administration advisory committee. In China, VG161, a multi-armoured oncolytic herpesvirus carrying multiple immunomodulatory transgenes, has demonstrated survival benefits in refractory hepatocellular carcinoma, while T3011, an oncolytic herpesvirus expressing both interleukin-12 and a PD-1 antibody, has entered first-in-human testing in advanced solid tumors.

Delivery logistics remain a genuine constraint that the authors confront directly. Most OVs are administered by intratumoral injection, which is straightforward for accessible cutaneous lesions but demanding for deep visceral metastases, although ultrasound-guided techniques are expanding the reachable set. Intravenous delivery, which would extend the approach to diffuse disease, is hampered by neutralizing antibodies, hepatic clearance and off-target sequestration, prompting engineering solutions ranging from cell carriage by mesenchymal stem cells and immune cells to polymer coating and tumor-specific promoter control of viral replication. Pre-existing antiviral immunity, once viewed purely as a barrier, may in some contexts enhance rather than diminish therapeutic efficacy by amplifying inflammatory recruitment to infected tumors. None of these obstacles is trivial, but the Perspective treats them as engineering problems rather than conceptual dead ends.

The broader significance of the argument lies in its reframing of therapeutic sequencing. If oncolytic viruses are truly systemic immune-reprogramming platforms, then the optimal role for an OV in a treatment regimen may be as the priming event, the ignition that creates the tumor-specific T cell pool, with checkpoint inhibitors, bispecifics or adoptive cells deployed afterwards to amplify and sustain the response. The authors acknowledge competing interests that come with their positions in oncolytic virotherapy companies, and the piece is explicitly a Perspective rather than a definitive clinical mandate. Still, the case they assemble, spanning mechanistic immunology, evolving trial data and a coherent framework for combination design, makes a credible argument that the field’s future lies not in making cold tumors hot, but in teaching the immune system, one infected tumor at a time, to recognize cancers it had never seen.

Historical context reinforces the authors’ argument that the field has been converging on this reframing for decades. The conceptual roots of oncolytic virotherapy stretch back more than a century to anecdotal reports of tumor regression after natural viral infections, but the modern era began in the early 1990s when Robert Martuza’s group described a genetically engineered herpes simplex virus mutant that could replicate in and destroy glioma cells. Landmark studies in the late 1990s and early 2000s established the tumor-selective logic of the field, exploiting activated ras signaling pathways, p53-deficient tumor cells, and defective interferon responses that characterize many cancers. Notably, the in situ vaccination concept itself was articulated in preclinical work as early as 1998, when replication-competent herpesviruses engineered to carry interleukin-12 were shown to induce local and systemic antitumor immunity, suggesting the current Perspective formalizes ideas whose time has finally arrived with enabling payload technology.

Safety data also support the platform’s maturation. Systematic reviews and meta-analyses of oncolytic virotherapy across malignancies have generally found favorable tolerability profiles, with most adverse events consisting of transient fever, injection-site reactions, and flu-like symptoms rather than the immune-related toxicities that complicate checkpoint blockade. This tolerability is clinically meaningful because it permits rational combination with other immunotherapies without prohibitive overlapping toxicity, a persistent challenge in immuno-oncology. At the same time, the empirical analysis of checkpoint inhibitor eligibility and response rates cited by the authors underscores the scale of unmet need: only a minority of patients who receive checkpoint inhibitors derive durable benefit, leaving a large population for whom antigen-agnostic priming strategies could be decisive. Whether the four-pillar framework translates into regulatory endorsements and standardized endpoints will likely determine how quickly next-generation oncolytic viruses move from promising biology to established cornerstone of cancer care.

Subject of Research: Next-generation oncolytic virotherapy as antigen-agnostic in situ cancer vaccination and a foundational immuno-oncology platform

Article Title: Beyond cold to hot: oncolytic virotherapy as the next cornerstone of immuno-oncology

Article References: Jia, W., Zhao, R., Kaufman, H. L., & Martuza, R. L. (2026). Beyond cold to hot: oncolytic virotherapy as the next cornerstone of immuno-oncology. Nature Reviews Clinical Oncology. https://doi.org/10.1038/s41571-026-01198-z

Image Credits: AI Generated

DOI: 10.1038/s41571-026-01198-z

Keywords: oncolytic virotherapy, immuno-oncology, immune checkpoint inhibitors, in situ cancer vaccination, tumor-specific T cells, immunogenic cell death, abscopal response, neoantigens, talimogene laherparepvec, dendritic cells, triple-A framework, cancer immunotherapy