Skip to content
PhageCocktails
Oncology frontierViruses vs. cancer

Cancer, handled by a virus

Two completely different viruses are being turned against cancer: oncolytic viruses that infect and lyse tumor cells, and bacteriophages that never touch a human cell — they strip away the bacteria living inside tumors that drive growth, shield cancer from chemotherapy, and silence immunotherapy. Here is the honest, sourced state of play in 2026.

Cancer, handled by a virus

A 2026 briefing on how viruses — both the tumor-killing kind and the bacteria-killing kind — are being engineered into precision oncology agents.

Two very different viruses, one shared target

When people hear "a virus that fights cancer," they usually picture a single idea. In fact there are two distinct viral strategies, and conflating them is the most common error in this field.

  • Oncolytic viruses (OVs) infect human tumor cells directly. They replicate inside the cancer cell, burst (lyse) it, and spill tumor antigens plus danger signals that wake up the immune system. These are human-cell viruses — herpesvirus, adenovirus, reovirus, vaccinia — re-engineered for safety and immune punch.
  • Bacteriophages ("phages") cannot infect human cells at all. They infect bacteria. In oncology they are used a different way: to kill or reprogram the tumor-associated bacteria (the "oncobiome") that drive tumor growth, blunt chemotherapy, and dampen immunotherapy — and, separately, as molecular scaffolds (phage display, phage vaccines, phage-guided drug delivery).

One lyses the cancer cell. The other never touches the cancer cell — it targets the bacteria around and inside the tumor, or serves as an engineerable nanoparticle. Both are now real, clinical-stage tools. Here is the honest, sourced state of play in 2026.

Part 1 — Oncolytic virotherapy: viruses that lyse the tumor

T-VEC / talimogene laherparepvec (Imlygic) — the approved trailblazer

The field's proof of concept. T-VEC is a genetically modified herpes simplex virus type 1 (HSV-1), attenuated by deletion of neurovirulence genes (ICP34.5) and ICP47, and armed with two copies of GM-CSF to recruit dendritic cells. The FDA approved it on October 27, 2015 for unresectable, recurrent metastatic melanoma — the first oncolytic virus therapy ever approved in the United States. In the pivotal Phase 3 OPTiM trial, the durable response rate was 16.3% vs 2.1% for GM-CSF alone.

RP1 / vusolimogene oderparepvec (Replimune) — a cautionary tale on the bar for approval

RP1 is a next-generation HSV-1 oncolytic (RH1-based) expressing GM-CSF plus a fusogenic glycoprotein (GALV-GP-R−). Replimune sought approval in combination with nivolumab for anti-PD-1–failed advanced melanoma, supported by the IGNYTE Phase 1/2 trial (objective response rate ~32.9% in 140 patients). The regulatory path was rocky and is instructive about what "experimental" still means:

Cretostimogene grenadenorepvec (CG Oncology) — oncolytic adenovirus for bladder cancer

A serotype-5 oncolytic adenovirus engineered to replicate selectively in Rb-pathway-defective cells and to express GM-CSF, delivered intravesically. In the Phase 3 BOND-003 trial in high-risk BCG-unresponsive non-muscle-invasive bladder cancer (NMIBC) with carcinoma in situ, it produced a 74.5% complete response rate, with 12-month freedom from progression of 97.3% — and notably no grade ≥3 treatment-related adverse events.

Nadofaragene firadenovec (Adstiladrin, Ferring) — approved adenoviral gene therapy

A non-replicating adenovirus vector carrying the interferon alfa-2b gene, instilled into the bladder, turning the bladder wall into an "interferon microfactory." FDA-approved December 2022 — the first gene therapy approved for bladder cancer — for high-risk BCG-unresponsive NMIBC with CIS. In Phase 3, 51% achieved complete response at 3 months, 46% of whom remained recurrence-free at 12 months. (Technically gene therapy rather than a replicating oncolytic, but viral-vector and frequently grouped here.)

Pelareorep (Oncolytics Biotech) — a wild-type reovirus, no genetic engineering needed

Pelareorep is a proprietary formulation of reovirus serotype 3 Dearing, given intravenously. It naturally favors cells with activated Ras signaling and works largely as an immune converter, turning "cold" tumors "hot." The GOBLET platform trial (with PanCAN support) is testing it in GI cancers — including newly diagnosed metastatic pancreatic ductal adenocarcinoma with modified FOLFIRINOX ± atezolizumab — and it has shown signals in two randomized Phase 2 metastatic breast cancer studies. FDA Fast Track in both breast and pancreatic cancer; registration-enabling studies are the next step. Investigational.

Teserpaturev / DELYTACT (G47Δ; Daiichi Sankyo) — approved oncolytic for brain cancer (Japan)

A third-generation oncolytic HSV-1 (triple-mutated G47Δ), developed with Prof. Tomoki Todo (University of Tokyo). In June 2021 it received conditional, time-limited marketing approval in Japan for malignant glioma — the world's first oncolytic virus approved for primary brain cancer. In the Japanese Phase 2, 1-year survival was 92% (vs ~15% expected). Further trials are underway in prostate cancer, mesothelioma, and olfactory neuroblastoma. Approved in Japan; not FDA-approved.

Other oncolytic platforms in clinical development include oncolytic vaccinia viruses (e.g., pexastimogene devacirepvec/Pexa-Vec, which famously failed its Phase 3 PHOCUS liver-cancer readout — a reminder of the attrition rate) and numerous oncolytic adenovirus/HSV/vaccinia candidates in combination with checkpoint inhibitors.

Honest summary of Part 1: Three oncolytic/viral-vector products are approved somewhere — Imlygic (US, 2015), Adstiladrin (US, 2022, gene therapy), and DELYTACT (Japan, 2021). Everything else (RP1, cretostimogene, pelareorep) is still experimental, and RP1's two CRLs show the regulatory bar remains high.

Part 2 — Bacteriophages vs the oncobiome

Here the virus never touches the cancer cell. The target is the bacteria that live in and around tumors and rig the game in cancer's favor. The flagship villain is Fusobacterium nucleatum in colorectal cancer (CRC).

Why Fusobacterium matters

F. nucleatum binds tumor epithelium via the adhesin FadA (to E-cadherin) and the lectin Fap2 (to tumor Gal-GalNAc, which also lets Fap2 disarm NK and T cells). Mechanistically, it drives chemoresistance: by activating TLR4–MYD88 signaling and suppressing microRNAs miR-18a*/miR-4802, it switches on autophagy (via ULK1/ATG7), helping tumor cells survive oxaliplatin and 5-FU. Patients who relapse after chemotherapy have higher tumor Fusobacterium loads.

The Fna C2 clade — the 2024 plot twist (Nature)

It turns out not all F. nucleatum is the culprit. Bullman and colleagues (Nature, March 2024) showed that the subspecies animalis splits into two clades — Fna C1 and Fna C2 — and only Fna C2 is enriched in CRC, present in roughly 50% of tumors across cohorts. Fna C2 carries extra virulence factors (fap2, cmpA, fusolisin) and metabolically tilts the tissue toward pro-oncogenic, oxidative-stress conditions. This sharpens the therapeutic target enormously: you don't need to wipe out a species, you need to hit one pathogenic lineage.

Phages as guided missiles against tumor bacteria (Zheng et al., Nat. Biomed. Eng., 2019)

The landmark proof-of-concept. Zheng et al. (Nature Biomedical Engineering, 2019) isolated a phage from human saliva that inhibits F. nucleatum, covalently linked it (via azide click chemistry) to irinotecan-loaded dextran nanoparticles, and created a phage-guided biotic–abiotic hybrid. Orally or IV in mouse CRC models (and tested in piglets and patient samples), the system selectively suppressed F. nucleatum, augmented first-line chemotherapy, and — bonus — the dextran acted as a prebiotic to boost butyrate-producing Clostridium butyricum. Safety in piglets was clean.

This is the conceptual heart of the field: a phage that ignores human cells, depletes a cancer-promoting bacterium, and simultaneously delivers chemo where it's needed — precision targeting that systemic antibiotics (which would nuke the whole microbiome) cannot match.

Gut bacteria as modulators of chemo and immunotherapy

The oncobiome cuts both ways: some bacteria shield tumors (Fusobacterium → chemoresistance), others license immunotherapy to work (Part 4). The strategic insight is that editing the microbiome with phage precision — rather than blunt antibiotics or fecal transplant — could become a routine adjunct to standard oncology.

Honest summary of Part 2: This is preclinical-to-translational, not approved. The biology (Fusobacterium → chemoresistance; Fna C2 → CRC niche) is robust and published in Cell and Nature. Phage-guided anti-Fusobacterium therapy has been demonstrated in animals and ex vivo human samples, not yet in registered human trials.

Part 3 — Phage display in oncology

Phage display — fusing peptide or antibody libraries to phage coat proteins and "panning" for binders — is the most clinically validated phage technology of all, even though it produces drugs, not phage therapeutics.

The validated backbone

George Smith invented phage display in 1985; Gregory Winter adapted it to display antibodies. The two shared the 2018 Nobel Prize in Chemistry (with Frances Arnold). The technology produced adalimumab (Humira) — the first fully human antibody drug, via Cambridge Antibody Technology, and once the world's best-selling drug.

Tumor-homing peptide discovery

Pasqualini and Ruoslahti (1996) showed phage-displayed peptide libraries can home to specific organs and tumor vasculature in vivo. Since then, phage panning has yielded peptides that bind tumor cells and tumor blood vessels — which can be conjugated to imaging agents or cytotoxic drugs for affinity-based targeted delivery.

Phage-displayed imaging and targeting

Phage-selected peptides are being developed as cancer imaging probes (PET/optical) and targeting ligands, with a 2025 review surveying research and clinical applications in cancer imaging.

Phage as drug-delivery scaffold and cancer vaccine

The phage particle itself is an engineerable nanoparticle. M13 filamentous phage is being built into cancer vaccine platforms: displaying tumor antigens or mimotopes in high copy number gives intrinsic adjuvanticity (TLR9 signaling) and multivalent antigen presentation. Recent work includes HER2-displaying M13 phages inducing therapeutic anti-breast-cancer immunity, HER2 mimotope vaccines (Scientific Reports, July 2025), and co-delivery of the NY-ESO-1 tumor antigen with α-GalCer to expand tumor-specific CD8⁺ T cells.

Honest summary of Part 3: Phage display as a discovery engine is fully mature and FDA-validated (Humira and many antibodies came from it). Phage-display-derived tumor-homing peptides, imaging probes, and phage vaccines are mostly preclinical/early translational in oncology.

Part 4 — Phage + immunotherapy

Checkpoint inhibitors (anti-PD-1, anti-CTLA-4) only work in a minority of patients — and the gut microbiome is one reason why.

The microbiome gates checkpoint-inhibitor response

Multiple landmark studies established that specific gut bacteria determine whether immunotherapy works. Responders to anti-PD-1 in melanoma carry distinct, more diverse microbiomes (e.g., enriched Ruminococcaceae); Gopalakrishnan/Wargo et al. (Science, 2018) showed gut microbiome composition modulates anti-PD-1 response. Akkermansia muciniphila is repeatedly linked to response in lung and kidney cancers — and supplementing it restored immunotherapy response in antibiotic-treated mice.

Where phages come in

If "bad" bacteria (e.g., Fusobacterium, which uses Fap2 to inhibit NK/T cells) blunt immunotherapy, and "good" bacteria (Akkermansia, Ruminococcaceae) enable it, then precision microbiome editing becomes an immunotherapy adjunct. Phages are the natural tool: unlike antibiotics — which indiscriminately deplete the beneficial taxa immunotherapy depends on — phages can subtract a single response-blunting strain while sparing the rest. This is the logical convergence of Parts 2 and 4: depleting the oncobiome's immunosuppressive members to unblock checkpoint inhibitors. Phages also have intrinsic immunomodulatory effects (TLR9 engagement, antigen delivery) that overlap with the vaccine work in Part 3.

Honest summary of Part 4: The microbiome–immunotherapy link is strong and clinically observed; live-biotherapeutic and fecal-transplant trials to improve ICI response are underway. Using phages specifically to edit the microbiome for immunotherapy benefit is a logical, actively-researched but still preclinical frontier.

The opportunity

Conventional oncology blasts the patient. The viral approaches in this briefing do something categorically different: they are precision biological agents that exploit specificity — molecular recognition, not brute cytotoxicity.

  • Oncolytic viruses turn a tumor into its own vaccine factory — lysing cancer cells and converting "cold" tumors "hot" so the immune system finishes the job. The approvals (Imlygic, Adstiladrin, DELYTACT) prove the modality is real; the RP1 CRLs prove the bar is, rightly, high.
  • Bacteriophages attack cancer from an angle no other drug class can reach — the bacteria that drive tumor growth, shield tumors from chemo, and silence immunotherapy. A phage can delete one pathogenic Fusobacterium lineage (Fna C2) without collateral damage to the microbiome that keeps checkpoint inhibitors working.
  • And phage display has already given oncology validated drugs and an inexhaustible engine for tumor-homing peptides, imaging probes, and vaccines.

The frontier is the convergence: an oncolytic virus to lyse and inflame the tumor, a phage cocktail to strip away the bacteria that confer chemoresistance and immune evasion, and phage-display-derived ligands to steer drugs precisely to the tumor — each agent doing one specific thing exquisitely well, together rather than alone. In a discipline historically defined by toxicity, viruses — both the tumor-killing and the bacteria-killing kind — point toward an oncology of precision rather than poison.

Approval status as of June 2026. Approved: T-VEC/Imlygic (US, 2015), Adstiladrin (US, 2022), DELYTACT (Japan, 2021). Investigational: RP1 (two FDA CRLs, latest April 2026), cretostimogene (BLA in progress), pelareorep (Fast Track). All bacteriophage anti-oncobiome and phage-immunotherapy approaches are preclinical/translational; phage display is a mature, Nobel-recognized discovery platform.

The Fusobacterium / CRC deep dive →Phages & the microbiome
Scientific & educational content. As of 2026 no bacteriophage therapeutic is approved as a marketed drug in the United States or European Union. Phage therapy is available only through clinical trials, compassionate-use / expanded-access pathways, and national magistral frameworks. Nothing here is medical advice or an offer to sell a therapeutic.