Clinical and translational progress in oncolytic virotherapy for pediatric CNS tumors

Purpose

Pediatric central nervous system (CNS) tumors are the leading cause of cancer-related mortality in children. Development of more effective therapies for pediatric CNS tumors has been slow, underscoring an urgent need for novel and innovative approaches.

Methods

This review summarizes current pediatric clinical trials of oncolytic viruses for pediatric brain tumors including high-grade glioma (HGG), diffuse midline glioma (DMG), medulloblastoma (MDB), atypical teratoid rhabdoid tumors (ATRT), and other high-grade tumors, while highlighting limitations of early-phase data, exploratory biomarkers, imaging challenges, pseudoprogression, and future directions.

Results

Key platforms include HSV-based agents (HSV1716, G207, and M032); adenoviral vectors (DNX2401, Ad-TD-nsIL12, and ICOVIR-5); MV-NIS (measles virus); PVS-RIPO (poliovirus); and Reolysin (reovirus). We review trial status, innovations in viral engineering and delivery, combinatorial strategies and translational challenges to establish oncolytic virotherapy as part of the future standard care for pediatric brain tumors.

Conclusion

Oncolytic virotherapy or immunovirotherapy offers a promising strategy to selectively kill tumor cells and generate antitumor immune response while minimizing toxicity to healthy cells compared to conventional treatments. Although still emerging in pediatric neuro-oncology, preclinical studies and early-phase clinical trials show encouraging safety and efficacy signals.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11060-026-05480-z.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11060-026-05480-z.

Introduction

Pediatric central nervous system (CNS) tumors are the leading cause of cancer-related death in children [1], underscoring the need for more effective therapies. Oncolytic virotherapy offers a promising approach for treating CNS tumors. As adjuvant therapies, these approaches may allow for reduced doses of radiation and chemotherapy, potentially minimizing toxicity while increasing efficacy [2]. Oncolytic viruses (OVs) are biologic agents that exploit intrinsic viral properties and host interactions to selectively target and destroy tumor cells [3]. The field of oncolytic virotherapy continued to progress since the 1990s, culminating in the 2015 approval of oncolytic HSV (oHSV) talimogene laherparepvec for advanced melanoma by the U.S. Food and Drug Administration (FDA) [3]. Although employment of virotherapy is still in its early stages in pediatric neuro-oncology, substantial preclinical evidence and recent clinical trials results have been highly encouraging [2]. The conditional approval of G47Δ in Japan for adult glioblastoma provides a benchmark for potential broader clinical translation, including pediatrics [4]. Table 1 summarizes key oncolytic viruses currently under investigation. This review aims to provide an up-to-date overview of oncolytic virotherapy clinical trials for pediatric CNS tumors.

Table 1.

Key oncolytic viruses used in pediatric clinical trials for CNS tumors

Oncolytic virus	Viral backbone	Inserted genes or promoters	Therapeutic purpose of modification	Pediatric-specific considerations	Citations
HSV
HSV1716	Complete deletion of γ34.5 (ICP34.5)	None	Attenuation of neurovirulence	Limited pediatric enrollment; humoral response noted	[5]
G207	HSV-1(F) with γ34.5 deletions	UL39 inactivation via LacZ insertion	Restricted replication in dividing tumor cells (UL39); LacZ as marker	Enhanced pediatric tumor sensitivity; combined with radiation in trials	[6, 7]
M032*	Complete deletion of γ34.5 (ICP34.5)	Human IL-12	Recruits targeted immune response via IL-12 to kill tumor cells.	IL-12 for immune-cold tumors; upcoming DMG trial (NCT07076498)	[8]
Adeno
Delta-24-RGD (DNX-2401)	Deletion of E1A gene Rb-binding region	RGD-4 C motif insertion	Selectively replicates in Rb-deficient tumor cells; RGD motif enhances entry into CAR-negative tumors via integrin binding	RGD for integrin targeting CAR-low pediatric tumors	[9]
ICOVIR-5	Delta-24RGD backbone	Replacement of E1A promoter with E2F-1 promoter & insertion of consensus Kozak sequence upstream of E1A	Enhances tumor selectivity via E2F-1 control of E1A expression in pRB-defective cells; Kozak sequence boosts E1A translation	ESF-1 for Rb pathway defects common in pediatric tumors	[10]
Ad-TD-nsIL12	TD: E1A CR2, E1B-19 K, E3 gp19K (retains E3B)	Armed with Non-secretory IL-12	TD enhanced tumor selectivity and replication while local IL-12 boosts anti-tumor immunity with reduced systemic toxicity	Non-secretory IL-12 to minimize systemic toxicity	[11]
Measles
MV-NIS	Edmonston measles vaccine strain	Human NIS gene insertion	Same tumor-selective entry mechanism; NIS enables noninvasive imaging and radioiodine-mediated radiovirotherapy	Radioiodine potential; efficacy in MDB/ATRT models	[12]
Polio
PVSRIPO	Attenuated Sabin strain of PV type 1	Replacement of native IRES by HRV2 IRES	Loss of neurovirulence while retaining oncolytic potency via CD155-mediated tumor targeting	CD155 targeting; strong in recurrent HGG	[13]
Reo
Reolysin	Wild-type Reovirus type 3 Dearing strain	None	Unclear; associated with activation of the Ras signaling pathway, which is frequently dysregulated in many tumors including glioma	Ras pathway activation; limited pediatric responses	[14, 15]

Abbreviations: ATRT (atypical teratoid rhabdoid tumors), CAR (Coxsackie and Adenovirus Receptor), CD155 (Poliovirus Receptor), CD46 (Cluster of Differentiation 46), CEA (Carcinoembryonic Antigen), CR2 (Conserved Region 2 of E1A), CRAd (Conditionally Replicative Adenovirus), E1A (Early Region 1 A gene), E1B (Early Region 1B gene), E2F (E2F Transcription Factor), E. coli (Escherichia coli), GFP (Green Fluorescent Protein), GM-CSF (Granulocyte-Macrophage Colony Stimulating Factor), HSV (Herpes Simplex Virus), HRV2 (Human Rhinovirus Type 2), ICP (Infected Cell Protein), ICP34.5 (Neurovirulence protein encoded by γ34.5 gene), IL-12 (Interleukin-12), LacZ (β-galactosidase gene from E. coli), MDB (medulloblastoma), MHC (Major Histocompatibility Complex), NIS (Sodium Iodide Symporter), pRB (Retinoblastoma Protein), PV (Poliovirus), Rb (Retinoblastoma), RGD (Arginine-Glycine-Aspartic Acid motif), TD (Triple Deletion), UL (Unique Long region of HSV genome), US (Unique Short region of HSV genome). *A pediatric clinical trial is anticipated in 2026

Herpes simplex virus

HSV is a common human pathogen, with a global seroprevalence of 66% [16, 17]. HSV-1 is advantageous as an oncolytic virus due to its large genome that enables gene insertion, susceptibility to acyclovir, and low risk of insertional mutagenesis [18]. Seven distinct oHSVs have been investigated in clinical trials targeting CNS tumors: G207, HSV1716, G47Δ, M032, C134, rQNestin and OH2 [2, 4, 6, 19–24]. Most oHSVs are engineered with deletions in the γ₁34.5 neurovirulence gene (ICP34.5), a modification that limits their replication in normal cells. rQNestin (CAN-3110) restores one copy under a nestin promoter to enhance selective viral replication in tumor cells [2, 23]. Despite these modifications, cancer cells remain susceptible to oHSV due to oncogenic mutations and dysregulated signaling pathways that override normal antiviral defenses [25]. Additionally, oHSVs promote immunogenic tumor cell death by inducing the release of danger-associated molecular patterns (DAMPs), which activate innate and adaptive immune responses, even in settings of limited viral replication (Fig. 1) [18].

Fig. 1.

Schematic Representation of Oncolytic Virotherapy Mechanism in Pediatric CNS tumors. (1) Oncolytic viruses are delivered directly into the brain tumor via intratumoral injection. (2) These viruses selectively infect tumor cells by recognizing tumor-specific surface receptors, a property that may be enhanced in some viruses through viral genetic engineering. (3) Once inside tumor cells, the viruses replicate, causing cell lysis and release of progeny viral particles that can infect neighboring tumor cells, amplifying the therapeutic effect within the tumor microenvironment. (4–5) Lysis also releases tumor-associated antigens and danger signals such as DAMPs, which are captured by brain-resident antigen-presenting cells (APCs), including microglia, dendritic cells and infiltrating macrophages. (6) This process initiates innate and adaptive immune responses, recruiting and activating effector T cells that further mediate tumor cell killing and promote transformation of the tumor microenvironment from immunologically “cold” to “hot.” Created with BioRender.com

Pediatric high-grade brain tumor patient-derived xenograft cells were significantly more sensitive to oHSV than adult high grade glioma (HGG) cells, motivating the exploration of oHSV therapy in children [26]. G207 is a first-generation oHSV with deletions in both γ₁34.5 genes and inactivation of the UL39 ribonucleotide reductase gene by insertion of an E. coli lacZ gene, making it dependent on cellular ribonucleotide reductase, which is upregulated in dividing tumor cells but not in normal cells [27]. G207 was evaluated in adult HGG clinical trials that demonstrated safety and preliminary efficacy when administered intratumorally and when combined with a single radiation dose [7, 19, 28]. Radiation was added based on preclinical findings showing that a single 5 Gy dose can enhance G207 efficacy by increasing viral replication and spread [29]. The adult clinical data combined with preclinical evidence of increased susceptibility of pediatric CNS tumors to oHSV provided rationale for a phase 1 G207 clinical trial in children (NCT02457845), the first completed oncolytic virotherapy trial in pediatric CNS tumors [6, 26]; see Table 2; Fig. 2 for a summary of clinical trials in pediatric CNS tumors.

Table 2.

Summary of pediatric clinical trials of oncolytic viruses: delivery, safety, efficacy, and immune activation

Oncolytic virus	Delivery	Safety	Efficacy	Immune Activation	Citations
HSV
HSV1716	Single intratumoral injection	No DLT	Only 2 patients enrolled and had PD	Humoral antiviral immune response	[5]
G207	Stereotactic intratumoral catheter placement with continuous infusion over 6 h. ‘’’	No DLT, only grade 1 virus-attributable toxicities	Radiographic and clinical responses observed in 11/12 patients.	CD8⁺ tumor-infiltrating lymphocytes; HSV-1 serostatus and post-treatment seroconversion may correlate with outcome	[6, 7]
Adeno
Delta-24-RGD (DNX-2401)	Stereotactic intratumoral injection into brainstem tumors.	Virus-related AEs in 9/12 patients; predominantly grade 1–2, with one grade 3 event; no grade 4–5 toxicities.	8 SD, 3 PR, 1 PD (N = 12)	Increased systemic T-cell clonality; correlated with longer PFS	[9]
ICOVIR5-5/Celyvir	IV weekly x 6	No grade 2–5 toxicities reported; well tolerated	2 pediatric patients with NB had SD, 7 had PD including patient with MDB	PCR-detectable adenoviral replication in most pediatric patients (not adults); responders showed increased circulating B and CD4⁺/CD8⁺ T cells.	[10]
ICOVIR-5/AloCelyvir	IV weekly x8	No grade 2–5 toxicities reported; well tolerated	Limited benefit, 1 patient was alive 25 mo after diagnosis.	Adenoviral replication was detected in the plasma of 4/6 patients	[30]
Ad-TD-nsIL12	Intratumoral delivery via Ommaya reservoir	Virus-related grade 2 adverse events reported; no cytokine release syndrome or severe immune-related toxicities	Newly diagnosed DMG: 3 PR, 5 SD, 1 PD (n = 9); Progressive DMG: 3 SD, 3 PD (n = 6)	Responses observed despite pre-existing neutralizing antibodies.	[11]
Measles
MV-NIS	Intratumoral or intrathecal/subarachnoid dosing via lumbar puncture; single or repeat dosing	One DLT (grade 3 transaminitis); transient viral shedding in 5 patients; otherwise well tolerated	PF4 achieved in 7/10 patients in the disseminated recurrent MDB cohort	Peripheral blood profiling demonstrated early antiviral gene expression and immune cell activation following MV-NIS treatment.	[12]
Polio
PVSRIPO	Single intratumoral injection by CED	3 grade 3 events (2 headaches, 1 seizure); no treatment-related grade 4–5 events or deaths	3 SD and 5 PD (n = 8)	Peripheral immune activation with increased monocyte and CD4⁺ T-cell activation post-treatment	[13]
Reo
Reolysin	Intravenous infusion, with GM-CSF to mobilize CD11b⁺ cells	1 DLT (hyponatremia); grade 3 depressed consciousness and grade 4 confusion reported (likely disease-related); no treatment-related deaths	Pediatric CNS tumors (n = 6): No objective responses; all patients progressed within 60 days	Transient systemic immune changes; reovirus detectable in CD11b⁺ monocyte/macrophages despite neutralizing antibodies in adult studies.	[14, 15]

Abbreviations: AE (Adverse Event), CED (Convection-Enhanced Delivery), CRS (Cytokine Release Syndrome), DLT (Dose-Limiting Toxicity), DMG (Diffuse Midline Glioma), IV (Intravenous), NB (Neuroblastoma), PD (Progressive Disease), PF4 (Progression Free Survival), PR (Partial Response), SD (Stable Disease), TCR (T-Cell Receptor)

Fig. 2.

Completed, ongoing, and anticipated clinical trials of oncolytic viruses in pediatric CNS tumors. Each arrow represents a phase 1 or 2 trial and is annotated with the viral vector, tumor indication, and trial status. Green arrows that extend to the phase line indicate completed trials; those that do not reach the line represent ongoing trials. White-outlined arrows denote planned studies that have not yet opened to enrollment. Created with BioRender.com. Abbreviations: ATRT (atypical teratoid/rhabdoid tumor), CNS (central nervous system), d (days), DMG (diffuse midline glioma), GBM (glioblastoma), HSV (herpes simplex virus), IL-12 (interleukin-12), MDB (medulloblastoma), mo (months), mOS (median OS), NCT (National Clinical Trial number), OS (overall survival), PD (progressive disease), PFS (progression-free survival), pHGG (pediatric high-grade glioma), PR (partial response), pts (patients), SD (stable disease)

In this trial, 12 patients aged 7 to 18 years with progressive/recurrent pediatric HGG underwent stereotactic placement of 3–4 intratumoral catheters to facilitate direct G207 delivery by continuous infusion over 6 h. The study included four dose levels: 10⁷ plaque-forming units (PFU), 10⁸ PFU, and 10⁷ PFU or 10⁸ PFU combined with 5 Gy within 24 h of virus inoculation. The dose was not escalated beyond 10⁸ PFU because responses in the adult trials did not appear to be dose dependent [6, 19]. G207 was safe with only grade 1 toxicities attributable to virus. Most patients demonstrated radiographic and clinical responses with a median overall survival of 12.2 months, which compares favorably to the historical median survival of 5.6 months for recurrent pediatric high grade glioma (pHGG) [31, 32]. Importantly, matched pre- and post-treatment tumor tissues showed that G207 markedly increased cytotoxic CD8⁺ tumor-infiltrating lymphocytes [6, 33]. Results suggested that patients with baseline HSV-1 IgG antibodies had shorter survival compared to those who seroconverted after treatment; however the sample size was small, and results should be interpreted cautiously (Table 2) [6]. Contrarily, data from the rQNestin trial in recurrent HGG suggest that baseline HSV-1 seropositivity was associated with significantly longer survival in adults [23]. Given that most children are HSV-1 seronegative, further investigation is warranted to determine whether baseline HSV-1 serostatus or seroconversion after oHSV administration may serve as treatment response biomarkers [34].

Building on these encouraging findings, an ongoing phase 2 trial is evaluating G207 in children with recurrent pHGG (NCT04482933). This study involves intratumoral administration of G207 at 1 × 10⁸ PFU via stereotactically-placed catheters followed by 5 Gy of focal radiation within 24 h. In parallel, an ongoing phase 1 trial is evaluating G207 alone or combined with 5 Gy of radiation in children with recurrent/refractory cerebellar CNS tumors (NCT03911388) [35]. This trial was motivated by preclinical data suggesting that cerebellar inoculation is safe and embryonal tumors, such as medulloblastoma (MDB), may be even more sensitive to virus killing than pHGG [26, 36, 37].

Another upcoming oHSV pediatric trial will utilize M032, a next-generation oHSV engineered to express human interleukin-12 (IL-12), leading to physiologically relevant IL-12 production during viral replication [8]. IL-12 promotes a T helper 1 (Th1)-cell immune response and activation of natural killer cells. Preclinical studies demonstrated the safety of the IL-12 producing virus in animal models, including mice and nonhuman primates [38]. M032 was safe with preliminary evidence of efficacy in adults with recurrent or progressive malignant gliomas in a completed phase 1 trial (NCT02062827), leading to an ongoing phase 1 trial combining M032 with pembrolizumab in adult patients with recurrent or newly diagnosed malignant glioma (NCT05084430) [21]. Based on M032 safety in the completed adult clinical trial and promising preclinical data suggesting enhanced efficacy of the IL-12 producing virus [21, 26], a clinical trial evaluating M032 in children and adults (3 and older) with newly diagnosed supratentorial and pontine diffuse midline glioma (DMG) is expected to open later this year (NCT07076498). Patients will receive a single intratumoral infusion of M032 delivered 4–8 weeks after completion of standard-of-care radiation. This will be the first upfront study of oHSV in children.

Adenovirus

Adenoviridae is a family of nonenveloped viruses characterized by an icosahedral capsid enclosing a double-stranded DNA genome. The most widely used adenoviruses for therapeutic applications are derived from serotype 5 of the human adenovirus species C (Ad5). The Ad5 genome is highly amenable to genetic modification, allowing for the design of conditionally replicative vectors that selectively replicate in malignant cells [39]. Several adenovirus-based OVs have been investigated in clinical trials for CNS tumors, including, ONYX-015, ICOVIR-5, DNX-2401 (delta-24-RGD), DNX-2440, CRAd-S-pk7, and Ad-TD-nsIL12. ONYX-015 lacks the E1B 55-kDa gene, and its tumor selectivity is driven by cancer-associated alterations in viral RNA export rather than p53 inactivation [40]. The virus was found to be safe with preliminary efficacy in a phase 1 trial involving adults with recurrent HGG [41]. Subsequent generations of adenoviruses were engineered for enhanced tumor selectivity and potency. Delta-24 is an adenovirus with a 24-base pair E1A deletion that blocks retinoblastoma protein binding preventing E2F release and viral replication in normal cells but permitting selective replication in tumor cells that commonly exhibit retinoblastoma dysfunction and constant E2F activity [42].

To improve infection of tumors lacking the Coxsackie and Adenovirus Receptor (CAR), an RGD-4 C motif was added to the fiber knob, creating Delta-24-RGD (DNX-2401), which uses integrins for entry and expands tropism to CAR-negative tumors [39, 42]. DNX-2401 demonstrated potent antitumor activity in pHGG and DMG preclinical models and was synergistic with radiotherapy [43, 44]. Based on promising preclinical data, a first-in-human phase 1 clinical trial in 12 patients aged 3 to 18 years assessed safety and preliminary efficacy of intratumoral DNX-2401 in newly diagnosed DMG (NCT03178032) [9]. The first four patients received 1 × 10¹⁰ viral particles, while the remaining eight were given 5 × 10¹⁰. All but one patient subsequently underwent radiotherapy at a median dose of 54 Gy (range: 39.0–59.4 Gy). The treatment was generally well-tolerated. Three patients had partial responses and eight had stable disease. Median overall survival was 17.8 months (range: 5.9–33.5 months). At the time of reporting, two patients remained alive, one of whom was progression-free at 38 months of follow-up. These outcomes appear favorable compared to historical survival of ~ 12 months with radiation alone; however, because all patients received both radiation and DNX-2401, the specific contribution of the virus remains unclear [9, 45]. Of note, all enrolled tumors were in the brainstem, highlighting the unique application of oncolytic virotherapy in this challenging location. Post-DNX-2401 infusion, peripheral blood analyses showed decreased T-cell receptor diversity with increased clonality, correlating with longer progression-free survival (PFS), suggesting a tumor-focused T-cell response. However, these findings are hypothesis-generating and should be interpreted cautiously due to the small patient number. A phase 2 clinical trial of DNX-2401 for DMG is expected within the next year.

Ad-TD-nsIL12 is a genetically engineered oncolytic adenovirus with triple gene deletions (E1A-CR2, E1B-19 K, and E3gp19K) designed to enhance tumor selectivity, promote apoptosis, and stimulate anti-tumor immunity. It also encodes a membrane-bound form of IL-12 to localize immune activation and reduce systemic toxicity. Preclinical studies of Ad-TD-nsIL-12 demonstrated its ability to inhibit glioma progression and enhance intratumoral CD8 + T-cell infiltration [11]. Two parallel phase 1 trials evaluating Ad-TD-nsIL12 in pediatric DMG were recently completed [46]. Newly diagnosed patients (Group A; NCT05717712, n = 9) received multiple viral doses, starting with two injections before radiotherapy, followed by additional doses during and after radiotherapy; while progressive cases (Group B; NCT05717699, n = 6) received the same regimen without concurrent radiotherapy. To avoid repeated stereotactic brainstem injections, an Ommaya reservoir enabled bedside dosing. Group A achieved three partial responses, five with stable disease (median overall survival [mOS] of 10.3 months [mo] from first virus, 11.3 mo from diagnosis), and one with progression. Group B had three with stable disease and three with progression (mOS 6.4 mo and 12.7 mo, respectively). Ad-TD-nsIL12 showed a favorable safety profile without cytokine release syndrome or severe immune-related events. Treatment with Ad-TD-nsIL12 produced multiple virus-related grade 2 adverse events, all of which were manageable.

In the Ad-TD-nsIL12 trials, clinical responses were observed even in patients with pre-existing neutralizing antibodies, suggesting that baseline humoral immunity did not preclude benefit [46]. In contrast, the DNX-2401 trial suggested that lower post-treatment neutralizing antibody titers may be associated with longer median overall survival (21.2 vs. 12.5 months), implying that a less robust humoral response could enhance viral activity [9]. However, no definitive conclusions can be drawn regarding the impact of antiviral immunity on clinical outcomes due to small patient numbers.

Novel delivery strategies, including neural stem cells (NSCs) and mesenchymal stem cells (MSCs), may improve tumor tropism, help evade neutralizing antibodies, and facilitate dissemination within the brain parenchyma [2]. One such platform, Celyvir, uses autologous MSCs to deliver ICOVIR-5, an oncolytic adenovirus derived from Delta-24-RGD and further engineered for enhanced tumor selectivity and infectivity [10]. ICOVIR-5 incorporates several modifications: a deletion in the E1A region, insertion of an RGD motif, and inclusion of a tumor-specific E2F-1 promoter and Kozak sequence to enhance E1A transcription and translation.

Ruano et al. conducted a phase 1 study of Celyvir for pediatric and adult patients with recurrent or refractory solid tumors (NCT01844661) [10]. The cohort included nine pediatric patients—one with MDB and eight with extracranial solid tumors. Only grade 1 toxicities were observed. Adenoviral replication was detected in all but two pediatric patients and in none of the adults. Immune profiling revealed that certain circulating immune subsets, such as B lymphocytes and dendritic cells, were significantly higher in children compared to adults, suggesting potential age-related differences in immune responses. These findings are exploratory, and further studies are needed to confirm their reproducibility and clinical relevance.

Another phase 1 trial investigated an allogeneic MSC-based delivery of ICOVIR-5, AloCelyvir, in six pediatric patients (ages 5–15) with newly diagnosed DMG (NCT04758533) [30]. AloCelyvir was administered weekly as intravenous infusions simultaneously with radiation, with 8 doses administered at 0.5 × 10⁶ cells/kg/dose. Adverse events were all grade 1. Adenoviral replication was detected in the plasma of four patients. Although clinical benefit was limited in this trial, the detection of viral replication and evidence of prolonged survival in subsets of patients across multiple adenovirus-based platforms support continued optimization and investigation of adenovirus immunovirotherapy.

Measles virus

Measles virus (MV), a negative-sense RNA virus from the Paramyxoviridae family, has been engineered into oncolytic strains derived from the Edmonston B vaccine strain. These modified viruses preferentially infect tumor cells due to their selective tropism for CD46, a complement regulatory protein that is overexpressed on many cancer cells [2]. Preclinical data have demonstrated the efficacy of oncolytic MV in targeting both localized and disseminated models of MDB and atypical teratoid/rhabdoid tumors (ATRT) [47–49].

A phase 1 trial (NCT02962167) evaluating the safety of MV-NIS, a measles virus engineered to express the sodium iodide symporter (NIS) for non-invasive monitoring of viral distribution, was recently completed in pediatric and young adult patients (ages 1–39) with recurrent MDB or ATRT [12]. Patients with localized recurrence received a single dose directly into the tumor bed at the time of surgical resection (stratum A, n = 7). Patients with disseminated disease received subarachnoid dosing via lumbar puncture, either as a single dose (stratum B, n = 12) or as 2 doses administered on days 0 and 10 (Stratum C, n = 15). Among 34 enrolled patients (median age 8.5 years, range 2–31), 23 with MDB and 4 with ATRT were evaluable. The treatment was safe with only one patient experiencing a dose limiting toxicity (DLT); grade 3 transaminitis. Viral shedding occurred in five patients and resolved by treatment end. Analysis of peripheral blood samples showed upregulation of antiviral gene expression and activation of innate and adaptive immune cell populations, suggesting an antiviral immune response after MV-NIS treatment. These findings indicate a potential immunologic effect, but conclusions regarding efficacy or predictive biomarkers are limited by the small sample size. One cohort evaluated twice-weekly intrathecal dosing for patients with disseminated recurrent medulloblastoma; among 10 patients, 7 remained progression-free at 4 months. While these results compared favorably with historical data, the trial was not designed to assess survival outcomes, and interpretation is further limited by subsequent therapies administered [12].

Poliovirus

Poliovirus (PV), a non-enveloped RNA virus from the Picornaviridae family, naturally targets lower motor neurons. Its neurotropism is mediated by its internal ribosomal entry site (IRES) and the expression of CD155, which is frequently upregulated in various solid tumors, including gliomas [50, 51] and pediatric brain tumors including medulloblastoma and ependymoma [52]. The prototype virus PV1 (RIPO) was engineered by replacing the IRES of wild-type poliovirus type 1 with that of human rhinovirus type 2 (HRV2), effectively abolishing neurovirulence while preserving oncolytic activity. PVSRIPO builds on this design by using the attenuated Sabin strain of poliovirus type 1 as its backbone, incorporating the same HRV2 IRES to further enhance clinical safety [53]. Preclinical studies using glioma models showed that intratumoral PVSRIPO leads to efficient tumor regression [54].

A Phase 1 trial in 61 adults with recurrent glioblastoma, PVSRIPO The treatment was safe, with no grade 4 or 5 adverse events and only three grade 3 events (two headaches and one seizure). Overall survival was 21% at 24 months; patients with lower tumor mutational burden experienced better outcomes, highlighting potential relevance for pediatric CNS tumors which typically exhibit lower tumor mutational burden than adult tumors [2, 55]. Based on the promising adult trial, a pediatric phase 1 trial (NCT03043391) recruited eight children (median age 16.5 years; range 9–19) with recurrent malignant glioma [13]. Each patient received a single intratumoral dose of 5 × 10⁷ tissue culture infectious dose of 50% (TCID₅₀) by convection-enhanced delivery (CED). Median overall survival for multiply recurrent patients was 4.1 months, with one patient alive > 22 months. Post-treatment analyses revealed increased peripheral monocyte and CD4⁺ T-cell activation, suggesting an immunologic response. While these findings are consistent with the safety and potential immunologic activity of PVSRIPO in pediatric glioma, conclusions regarding clinical efficacy or predictive biomarkers remain limited by the small sample size and study design.

Reovirus

Reovirus is a non-enveloped, double-stranded RNA virus from the Reoviridae family [2]. The type 3 Dearing strain of mammalian orthoreovirus, commonly known as Reolysin or pelareolep, is the primary strain used in oncolytic applications due to its tumor selectivity. Although the exact oncolytic mechanism is not fully understood, it is associated with Ras pathway activation—which is frequently dysregulated in many human cancers, including gliomas [2, 39]. In a prior phase 1 trial of children and young adults with relapsed/refractory non-CNS solid tumors, Reolysin ± low-dose cyclophosphamide was well tolerated, demonstrated rapid viral clearance, and showed limited clinical activity with some disease stabilization [14]. In adult studies, reovirus was detected in CD11b⁺ monocyte/macrophages despite neutralizing antibodies, prompting the use of sargramostim (GM-CSF) to mobilize these cells [56, 57]. A phase 1 pediatric trial (NCT02444546) evaluating this strategy treated six patients (one MDB, two DMG, and three HGG) at two prespecified dose levels, 3 × 10⁸ and 5 × 10⁸ TCID₅₀ [15]. The first three patients received dose level 1; however, one developed grade 3 depressed consciousness and grade 4 confusion during the second cycle. Although this was likely due to disease progression, the final three patients were treated at dose level 1. Recruitment was halted after the sixth patient experienced hyponatremia, the only observed DLT, and the study was ultimately closed due to limited accrual. Overall, the regimen was well tolerated. Patients showed early increases in monocytes following GM-CSF and transient reductions in white blood cells, neutrophils, and platelets. Neutralizing anti-reovirus antibodies were detected within 12 days in the two patients who were assayed. Despite these immunological changes, no clinical responses were observed, and all patients experienced disease progression within 60 days, highlighting delivery and antiviral immunity as key barriers (Table 2). These findings support the feasibility and safety of intravenous Reolysin delivery, although no additional pediatric trials are currently in development. Again, conclusions regarding efficacy or predictive biomarkers are limited by the small cohort size and early-phase study design.

Comparative analysis

Across platforms, HSV-based agents have shown stronger evidence of tumor immune infiltration compared to other viral approaches, although delivery remains a universal barrier. Systemic platforms face poor blood brain barrier (BBB) penetration and rapid clearance, while intratumoral/CED methods yield transient replication but variable durability. Shared failure modes include antiviral host responses and tumor resistance, underscoring needs for combinations to enhance efficacy.

Limitations

The studies reviewed are early-phase trials with small sample sizes; therefore, observed clinical and immunologic outcomes should be considered hypothesis-generating and interpreted cautiously. Comparisons with historical controls are descriptive and may not provide definitive evidence of therapeutic benefit. Additionally, radiographic responses are complicated by pseudoprogression- immune-mediated inflammation, edema, and tumor cell death mimicking progression. This is particularly challenging post-CED, where changes post-infusion differ greatly from chemoradiation effects [13]. Variability in response criteria (IRANO, RAPNO) and some trials’ reliance on non-standardized MRI measurements further limits cross-trial comparisons and efficacy interpretation [6, 9, 13].

Future directions

Although early-phase trials demonstrate safety and transient immunologic changes across platforms, durable anti-tumor responses remain limited, and most patients ultimately progress. Negative or neutral outcomes, particularly in systemically administered platforms and carrier-based approaches (Celyvir, AloCelyvir, and Reolysin) [10, 14, 15, 30], highlight key constraints: poor BBB penetration, rapid immune clearance, transient viral replication, limited tumor tropism, and modest immunologic effects with minimal sustained anti-tumor activity [58–61]. Tumor-intrinsic factors and antiviral immune mechanisms further constrain efficacy [62]. These lessons emphasize the need for future studies to optimize the balance between anti-tumor and anti-viral immunity, improve delivery strategies, and explore combination approaches to enhance clinical benefit.

Oncolytic virotherapy is beginning to transition towards broader clinical use. The conditional approval of G47Δ in Japan for glioblastoma serves as an example of how engineered oncolytic viruses can achieve regulatory milestones and inform pediatric applications [4]. Evidence from early trials indicate oncolytic viruses, when combined with standard therapy or immune checkpoint blockade, can safely augment immune infiltration and reduce tumor immunosuppression [58].

Realistic integration into current treatment paradigms will likely involve combination with established modalities rather than standalone replacement. Intratumoral or CED could follow surgical resection or biopsy, with focal radiation to enhance viral replication and spread, as seen with oHSV [6]. The favorable safety profiles observed across multiple platforms suggest that intensive monitoring in highly specialized centers may not always be required, supporting potential broader implementation with appropriate surgical training and standardized protocols. However, feasibility and scalability face ongoing challenges, including manufacturing complexity, high costs, regulatory pathways for pediatric indications, and equitable access for diverse populations [63]. Future trials should prioritize inclusive designs to incorporate real-world evidence for practical adoption.

Ongoing and planned studies will provide critical insights: the Phase 2 trial of G207 with radiation in recurrent pHGG (NCT04482933), and the Phase 1 evaluation of M032 (NCT07076498) in newly diagnosed DMG post-radiation. Advancements in viral engineering, such as payload-armed viruses and improved selectivity [64], combined with multidisciplinary collaboration and larger, randomized trials, remain essential to translate these promising strategies into durable treatments for children with CNS tumors.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

A.E., A.E.F., M.F., J.H., and G.K.F. contributed to the conceptualization of the manuscript.A.E. and G.K.F. wrote the original manuscript draft and designed the figures and table.All authors (A.E., A.E.F., M.F., J.H., S.E.G., D.S.O., W.Z., L.Z., K-D.K., Z.Z., J.F., C.G-M., E.M.T., J.D.B., and G.K.F.) significantly contributed to reviewing and editing the language and intellectual content of the manuscript.G.K.F. supervised the work.All authors reviewed and approved the final manuscript.

Funding

Friedman laboratory’s efforts are supported by the Food and Drug Administration (FDA) of the U.S. Department of Health and Human Services (HHS) as part of a financial assistance award R01FD006368 totaling $750,000.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

JDB has an equity position in Treovir Inc., an oHSV clinical stage company and UpFront Diagnostics. JDB is also on the Centile Bioscience, QV Bioelectronics and NeuroX1 boards of scientific advisors and has equity positions in Barinthus Biotherapeutics. JMM has the following relationships which may pose or be perceived as posing a financial conflict of interest; he is a board and equity holding member, in Aettis, Inc. and may receive royalties. The company holds frozen oncolytic viral stocks. Mustang Bio Tech is licensing the Intellectual Property (IP) of C134 an oncolytic viral Therapy. Markert is blinded to the conditions for the C134 clinical trials. He is a shareholder for a privately held Small Business Innovation Research LLC, Treovir, Inc., concerning G207 oncolytic viral therapy now in clinical trial. Merck, Inc. provides industry grant support by providing Keytruda (pembrolizumab) for a clinical trial of M032 oncolytic virotherapy now in clinical trial. Additionally, JMM is a listee on IP- (1) related to a cancer immunotherapy system, and (2) to a novel immuno-virotherapeutic strategy targeting the glioma secretome. This IP has been filed by in8Bio (formerly Incysus, Ltd.) and has royalty earning potential. GKF holds an equity interest in Synaptive Medical, Inc. and a patent for methods and formulations related to the intrathecal delivery of oncolytic virus He received prior funding to his institution from Pfizer, Eisai, and E.R. Squibb & Sons for research unrelated to the content of this manuscript.