Skip to main content

Understand the population treatment leaves behind.

Glioblastoma contains malignant cells in different biological states. Published single-cell research shows how diversity and plasticity can shape those populations. Understanding a survivor therefore means examining its identity alongside the conditions and history that influenced it.

Our programme brings those elements into the same question. Which cells remain after treatment, how do they change over time and what can they do? The purpose is to connect a description of survivors with the behaviour that makes recurrence a development problem.

Two researchers in masks and safety glasses examine a sample using a benchtop microscope.

Examining a sample is one stage of an investigation. The model, preparation and measurement still determine what an observation can support. Illustrative laboratory photography.

LightFieldStudios · Scientists in Lab Coats Examining Sample Under MicroscopeEnvato Elements commercial licence

Find the contribution that changes regrowth.

A cell found after treatment may directly rebuild a tumour, support another population or accompany regrowth without causing it. These explanations lead to different intervention ideas and require comparisons capable of distinguishing them.

The causal question is whether changing a defined population changes the capacity for regrowth. A useful design must account for the relevant model, timing and conditions, then examine the result over an appropriate period. This is how a promising observation can become a clearer rationale for the next experiment.

Published science | A closer look

Read the biology.
Ask a sharper question.

Different cells can share a tumour and behave differently. Begin with one published comparison.

Models with different cell states.

Researchers compared two engineered mouse glioma models. Activating Notch in N1IC helped them study a relatively quiescent, astrocyte-like tumour-cell state.

Mouse glioma models | External published research

Complete Figure 1F: p53 and N1IC mouse glioma microscopy, with HA, Clu and DAPI labels, arrows and scale bars.
Banu et al. (2024), Fig. 1F | © The Author(s) | CC BY 4.0Panel crop | Source and image context

Glioblastoma programme | What the literature leaves open

Which bridge between the papers needs its own experiment?

Published cell-state and senescence studies can motivate a new question without testing the complete NEUVIOR proposal. Follow the connections between those starting points, a proposed intervention sequence and the outcomes that would matter. The visual makes the missing causal bridge explicit so a partner can assess what a first study should actually establish.

Glioma researchers, translational partners and research investors

Editorial decision model, not measured NEUVIOR results. Spatial positions indicate relationships, not measured quantities.

Separate what each experiment actually tested.

The starting papers address related biological questions in different experimental settings. Reading their models and endpoints separately prevents an attractive narrative from becoming a combined result that nobody measured. Their contribution is a reason to investigate a specified bridge, with the limits left visible.

Identify which endpoint belongs to which model before comparing the studies.

Fund a comparison that can change the direction.

A partner discussion should concentrate on whether the proposed sequence can be tested with reliable identity, sufficient follow-up and interpretable exposure and function. A first positive finding would justify another bounded question. It would not establish a clinical intervention, patient benefit or a finished therapeutic asset.

Define one feasible comparison and its stop or redesign criterion.

In focus

State-specific vulnerability

Published GPX4-related findings concern defined cell states and experimental conditions. Their role here is to motivate a precise dependency question. They do not establish that every treatment-surviving population shares that vulnerability or that the proposed sequence has already been tested.

  • Motivates a dependency
Evidence to keep in view

The relevant model, cell-state definition and actual endpoint in the published study.

A useful next step

Which part of the published vulnerability survives a change in treatment context?

The programme becomes more assessable when the experiment missing between the papers is visible.

The cells leave an aggregate through a particular environment

A recurrence programme needs questions that survive contact with the model. This published U87-MG aggregate time-lapse offers a visible example of dispersal under defined laboratory conditions, helping readers ask how matrix, treatment and cell identity shape an apparent invasive behaviour.

Time-lapse of dexamethasone-treated U87-MG aggregates from a published fibronectin-matrix study. This established cell-line model is not a patient tumour or evidence of NEUVIOR treatment efficacy.

Sabari J, Lax D, Connors D, Brotman I, Mindrebo E, Butler C, Entersz I, Jia D, Foty R · The cells leave an aggregate through a particular environmentCC BY 2.5 | Original research and methods

Source and image notes

H.264 viewing copy; complete frame retained with proportional downscaling where required; no colour enhancement, crop, speed change or added frames. Audio omitted; original source remains linked.

Follow the edge

The interaction between aggregate and surrounding surface is part of what is being measured.

Question the explanation

A change in dispersal can arise through multiple mechanisms. The original controls and independent comparisons matter.

Treat the tumour environment as part of the question.

Cells respond within a biological environment. Published studies show that ferroptosis-related effects can differ with cell state and the surrounding system. Some experimental work has linked tumour necrosis and ferroptosis with more aggressive behaviour.

This is why a development rationale needs to examine what the intervention changes beyond an immediate cell count. The proposed comparison must consider regrowth, remaining populations and relevant interactions. A selective effect is valuable when its biological context and later consequences are understood.

GLIO | Evidence into a research question

A question becomes a programme.

Follow the published starting points, the work NEUVIOR has documented and the comparison that could change what happens next.

Research design and evidence synthesisProgramme record: 8 October 2026

The reviewed programme record documents no original NEUVIOR experimental dataset. The proposed decision study has not run.

External findings

A starting point worth testing.

Glioblastoma contains different malignant-cell states, with experimental evidence that cells can change state (Neftel et al., 2019). That makes one universal treatment-surviving population an assumption to examine.

Specific experimental glioma states showed a GPX4-related vulnerability (Banu et al., 2024). Separate post-radiation senolytic work motivates asking whether some surviving cells support other cells (Tomimatsu et al., 2025). These studies have different models and endpoints; they do not establish the proposed NEUVIOR approach.

Conditional neuronal Gpx4 deletion caused delayed impairment in mice (Hambright et al., 2017). This is a genetic hazard signal, not a drug-specific safety estimate. It makes function and recovery important questions alongside malignant-cell effects.

The decision it informs

Which published finding transfers to the intended treatment context, and which needs a new comparison?

Bring a model, a method or a development question. Begin with what can be discussed openly; agree confidentiality before sharing unpublished material.

Discuss a research capability

Protect the functions a treatment would need to preserve.

The brain is the setting in which a future intervention would have to be useful. Adequate exposure, the infiltrating tumour margin and the preservation of normal neural function are part of the development challenge.

Our programme therefore connects target and cell-state questions with model qualification and functional assessment. A useful early result should explain both the intended effect and the relevant systems it may affect. Those considerations help determine which candidate or method is worth a more demanding next evaluation.

Connect two selected studies with one purpose.

The first selected study is “Characterising the Contribution of Treatment-Surviving Malignant Cell States to Glioblastoma Recurrence Following Chemoradiotherapy”. It focuses on the survivor populations and their contribution to regrowth.

The second is “Evaluating GPX4-Dependent Ferroptosis as a Vulnerability in Glioblastoma Cells Persisting After Senescent-Cell Depletion”. It examines a defined intervention sequence. Together, they link survivor identity, causality and a possible vulnerability within one programme, rather than presenting separate claims of discovery.

Let people’s priorities inform the research.

A useful future benefit would become part of a person’s life. Communication, independence, everyday function and treatment burden deserve attention alongside the biological question. We are developing patient and caregiver involvement to help those perspectives inform the programme.

Scientific development includes discussions with the University of Oxford’s Department of Oncology. The work currently centres on scientific design, model and candidate assessment and involvement planning. The aim is a focused next study that can teach us something worth acting on.

Make the next result informative.

The next evidence should clarify the population, mechanism or development decision being examined. A strong result needs an appropriate comparator, interpretable measurements and a clear account of uncertainty.

Programme updates will connect released work with those questions as it develops. NEUVIOR is not currently recruiting people to a glioblastoma clinical trial or offering an experimental treatment. For researchers and development partners, the opportunity is to help make the next scientific question answerable and its implications clearer.

A clearer question is real progress.

NEUVIOR's GLIO work asks what allows residual adult glioblastoma, IDH-wildtype, to regrow after treatment. The question connects malignant-cell identity, the contribution those cells make, the effect of a proposed intervention and the function of the surrounding neural system. Its value lies in making the next experiment more decisive: what must be measured together, what a promising signal leaves unanswered and which result should change the development decision.

Our documented progress is an evidence synthesis and research-design framework. The reviewed programme record dated 8 October 2026 brings together focused paper audits, relevant textbook findings, competing explanations and a proposed decision study. A critical-perspective manuscript is a working draft, not submitted or peer reviewed. No original NEUVIOR experimental dataset establishing regrowth benefit or a usable neural exposure window is documented in that record.

What the published science actually gives us.

Published work supplies specific starting points. Neftel and colleagues combined human tumour profiling with experimental approaches to examine diverse malignant-cell states and plasticity (Neftel et al., 2019). Banu and colleagues identified a GPX4-related vulnerability in particular glioma states across defined experimental systems (Banu et al., 2024). The microscopy shown here comes from their engineered murine models; it is not human tissue or a NEUVIOR experiment.

Tomimatsu and colleagues investigated post-radiation senescent-cell targeting in preclinical glioblastoma systems (Tomimatsu et al., 2025). That work motivates a distinct question about cells that support other cells. These studies do not test the entire NEUVIOR proposal, and their endpoints should not be combined into an invented effect estimate. Their useful contribution is a reason to test a precise causal bridge.

NEUVIOR Data Analysis

A source.
A question.
A clearer decision.

Useful intelligence keeps its evidence attached.

Our public analysis connects a defined question with named sources, reporting periods and methods. Each source below shows the work available today and the access still needed.

Published-data analysis, research hypotheses and evidence studies are distinct kinds of work. Their meaning should remain clear as the programme grows.

Open Targets | Quarterly platform releases

Open Targets Platform

Which published associations support a research hypothesis, and through what evidence?

Source identified

Official source

US National Library of Medicine | Source records updated by responsible parties; API refresh is source-controlled

ClinicalTrials.gov public study records

Which GLIO study designs and eligibility patterns merit scientific review?

Analysis available

Official source

Explore NEUVIOR Data Analysis

Surviving is not the same as restarting.

A cell can remain present after treatment without becoming the founder of later regrowth. It may support other malignant cells, enter a different state or carry a marker that is informative only under particular conditions. The proposed analysis separates these explanations. An image showing marker expression can help characterise a model, but cannot on its own establish which population causes recurrence or whether removing it creates durable control.

Our public-level framework therefore connects identity with contribution and outcome. A useful comparison would follow the relevant population over time, examine what changes after an intervention and test whether the consequence persists after withdrawal. The question should remain answerable even if the original explanation is wrong. A different mechanism or an informative negative finding can redirect the work without being relabelled as a breakthrough.

Four questions a study should answer together.

Identity: is the intended malignant population present in the model at the time of testing? Contribution: does it directly found regrowth, support another population or merely track a temporary condition? Durability: what happens after the intervention is withdrawn? Exposure and function: can the observed effect coexist with acceptable neural function and recovery at a meaningfully compared exposure?

These are proposed measurement principles, not a validated assay or a clinical surrogate. The comparator must include the relevant treatment background and justified alternatives. Independent biological sources need to remain distinct from repeated wells or images. Model feasibility, specialist methods and variance information should determine numerical thresholds before confirmatory results are examined.

Neural function belongs inside the question.

Hambright and colleagues reported delayed impairment and neurodegeneration after conditional neuronal Gpx4 deletion in mice (Hambright et al., 2017). That is a hazard signal from a genetic model, not a drug-specific safety estimate. It explains why a short observation of normal-cell viability would leave an important uncertainty unresolved.

The proposed programme links malignant-cell effects with measured exposure, neural function and recovery. A candidate that changes a tumour-related assay but creates persistent neural impairment at the required exposure would need to be stopped or redesigned. A laboratory comparison cannot establish preserved cognition in people, practical brain delivery or clinical benefit. Those require separate evidence and an appropriately qualified development path.

What would earn the next development commitment?

A first partner study should produce an interpretable account of model validity, mechanism, longitudinal response and functional comparison. A credible positive result would justify only the next bounded investigation. It would not establish an approved treatment, a licence value or readiness for a human study.

We would revise the design if the intended state disappears, controls fail or the estimates are too imprecise to answer the question. We would stop the tested route if a fair comparison excludes a worthwhile added effect or shows unacceptable persistent neural impairment. A simpler alternative that performs as well may be the stronger development decision. Access to materials, lawful rights, responsible analysis and independent challenge are part of the work needed to make those decisions credible.

Bring a capability. Begin a useful discussion.

Model specialists, quantitative biologists, pharmacology teams, neural-function researchers and development partners can each contribute a defined part of the next question. Start with the capability, the comparison it could support and the decision that the result would inform. A non-confidential enquiry is enough to begin that discussion.

Where a later conversation needs unpublished materials, we can agree the scope of a confidential briefing and whether an NDA is appropriate. Public disease context, published evidence, the broad research purpose and the proposal poster remain available without one. Detailed candidate information, enabling experimental designs, third-party confidential material and individual patient information do not belong in the public account.

Which contribution could reveal the next useful answer?

We welcome relevant scientific methods, model expertise and clearly scoped development contributions to the glioblastoma recurrence programme.

Continue the programme

  • Survivor-cell biology

    A proposed research collaboration to distinguish treatment survival from causal contribution to glioma regrowth, with model-specific tracking and measurable comparisons.

  • Target engagement and selectivity

    Explore a proposed GPX4-related mechanism study connecting engagement, exposure, lipid oxidation, selectivity and recovery before a development decision.

  • Neural-function preservation

    A proposed neuroscience collaboration examining activity, model relevance and recovery alongside malignant-cell effects, without inferring clinical neuroprotection.

  • From mechanism to development

    A proposed translational research collaboration connecting exposure, model choice, formulation questions and reproducibility with a staged feasibility decision.

NEWSPerspectives & features