Local Nanomedicine and Nano-Enabled Biomaterials After Glioblastoma Resection

Abstract

Despite maximal safe resection and chemoradiotherapy, glioblastoma almost invariably recurs near the resection margin. Incomplete tumor removal is only one contributor; infiltrative residual cells, resistant stem-like states, wound-healing responses, local immunosuppression, and limited drug access also promote regrowth. This review examines stand-alone nanocarriers and nano-enabled composites in which nanoscale components are incorporated into hydrogels, scaffolds, or implants at the postoperative cavity–margin interface. Biomaterials lacking a functional nanoscale component are included only as design or procedural comparators. Postoperative resection models are distinguished from intratumoral, unresected orthotopic, ex vivo, and in vitro studies, which provide indirect support. Most evidence of efficacy remains preclinical, whereas human studies mainly address feasibility, safety, pharmacodynamic activity, or workflow precedent. Translation depends on reproducible retention, margin coverage, biologically matched release, brain safety, scalable manufacturing, and neurosurgical compatibility. Local nanomedicine thus remains a conditional postoperative strategy whose clinical value has yet to be established.

Introduction

Glioblastoma (GBM) remains among the most lethal primary brain malignancies in adults; diffuse infiltration and marked molecular heterogeneity make durable control uncommon even in contemporary neuro-oncology.¹ For newly diagnosed disease, treatment continues to rely on maximal safe resection followed by radiotherapy and temozolomide (TMZ), with tumor-treating fields, molecularly guided trial enrollment, and supportive measures incorporated according to patient context. Surgical planning prioritizes maximal safety, as functional preservation, postoperative residual-tumor assessment, and extent-of-resection classification influence prognosis and subsequent treatment planning.² Attempts to intensify this backbone have not consistently improved outcomes; in the Phase 3 marizomib trial, adding marizomib to TMZ-based chemoradiotherapy did not improve overall or progression-free survival.³ Once recurrence develops, management is individualized across re-resection, reirradiation, systemic therapy, device-based interventions, and clinical trials, yet long-term control remains uncommon.⁴ Postoperative recurrence remains the central clinical bottleneck. Surgery creates a therapeutic opportunity, yet the surrounding tissue can still seed local failure.⁵

Postoperative relapse is difficult to prevent. The resection margin is a biologically distinct compartment, not simply the border left after debulking, and residual GBM cells may escape vulnerabilities inferred from the tumor bulk.⁶ Spatially resolved transcriptomic data reinforce this distinction, showing that infiltrative 5-aminolevulinic acid–metabolizing cells carry molecular programs associated with recurrence and poor survival and do not constitute a cleanly separable remnant of the primary tumor.⁷ Within this region, glioblastoma stem-like cell (GSC) populations form a durable reservoir through self-renewal, plasticity, and niche dependence; together with intratumoral heterogeneity, immune evasion, and therapeutic resistance, these states make recurrence irreducible to residual cell number alone.⁸ The surrounding tissue also evolves through wound repair, inflammatory signaling, vascular disturbance, and treatment-driven selection. A recent postoperative study showed that a biohybrid chiral hydrogel targeting pro-stemness cues could reduce regrowth in GBM resection models.⁹ The dynamic resection margin is therefore the relevant therapeutic target, where infiltrative residual cells, stem-like and treatment-resistant states, wound-repair and vascular responses, immune suppression, and delivery constraints converge. Incomplete removal and residual cell burden explain only part of postoperative failure.

Recurrence at the resection margin reflects both biological and delivery constraints. Residual disease persists in a narrow postoperative compartment where uneven disruption of the blood-brain barrier (BBB) and blood-brain tumor barrier (BBTB) can still restrict access to invasive cells.¹⁰ The restricted postoperative compartment favors direct placement of the drug source beside residual tissue. In a murine GBM resection model, lipid nanocapsules embedded in a hydrogel sustained delivery at the surgical site.¹¹ Postoperative immune delivery was tested with an exosome-cross-linked gel, which promoted local T-cell activation, whereas a Phase 1 chimeric antigen receptor (CAR) T-cell study in recurrent high-grade glioma established route feasibility and central nervous system bioactivity in a clinical setting not specific to tumor resection.^12,13 The studies address different questions. The preclinical platform study tests delivery after resection, whereas the clinical study establishes regional-delivery feasibility and bioactivity without demonstrating postoperative nanomedicine efficacy.

Previous reviews have separately addressed localized glioblastoma therapy and hydrogel-mediated delivery.^14,15 Nano-delivery systems and immune-directed nanomedicine have also been reviewed.¹⁶ This review examines postoperative locoregional nanomedicine where local drug delivery, residual-disease biology, and immune control intersect. Its scope includes nanoscale carriers and nano-enabled composite platforms intended for use after, or in direct relation to, tumor resection. Hydrogels, scaffolds, wafers, and implants without a functionally relevant nanoscale component are retained only when they clarify retention, margin coverage, release control, or surgical deployment. Findings from intratumoral, unresected orthotopic, ex vivo, and in vitro studies contribute mechanistic or design context; direct postoperative conclusions are drawn from resection models. Because clinical data specific to postoperative nanomedicine remain sparse, claims of therapeutic effect are confined to preclinical findings. The discussion proceeds from cavity retention and margin coverage to residual-disease control, immune remodeling, and translational evaluation, while making explicit when conclusions depend on indirect evidence or extrapolation.

The Postoperative Cavity–Margin Interface as a Local Therapeutic Window

Biological Features of the Postoperative Cavity–Margin Interface

For clarity, the postoperative cavity–margin interface comprises the surgically created cavity, its tissue-facing wall, and the adjacent resection margin; these anatomical components are related but not interchangeable. The resection cavity is a rapidly changing surgical compartment rather than a stable void left by tumor debulking. Postoperative hemorrhage is a recognized early complication after glioma surgery.¹⁷ Postoperative communicating hydrocephalus and later cerebrospinal fluid (CSF) diversion have also been documented after GBM resection, with ventricular opening among the reported risk factors.¹⁸ Serial imaging provides additional evidence, as changes in cavity position, volume, edema, and fluid-attenuated inversion recovery abnormalities between early postoperative magnetic resonance imaging (MRI) and radiotherapy-planning MRI can alter target delineation.¹⁹

The cavity wall is also abnormal, bordering a peritumoral brain zone already marked by radiological, cellular, molecular, and biomechanical changes that can influence residual disease and local intervention.²⁰ Surgery and standard therapy can continue to shape tumor evolution and select for treatment-adapted cellular and microenvironmental states.²¹ For local nanomedicine, the therapeutic target extends beyond the cavity lumen to the tissue-facing wall and adjacent resection margin. Together, these compartments form an evolving interface exposed to fluid exchange, brain deformation, and ongoing biological remodeling.

Spatial Distribution of Residual Tumor Cells

Residual tumor after GBM surgery is spatially organized. The peritumoral brain zone and resection margin serve as major reservoirs of recurrence-prone disease, and spatial multiomics identifies a shared infiltrative signature at the margin that supports treating it as a distinct residual compartment.^22,23 Recurrent GBM also shows perivascular niche organization, and invasive cells may extend beyond the cavity wall along white-matter tracts and perivascular routes.^24,25

Hypoxic and peri-necrotic regions add another layer of spatial complexity, with low-oxygen niches linked to aggressive phenotypes, stem-like maintenance, treatment resistance, and immunosuppressive signaling in GBM.²⁶ In a rat GBM resection model, stromal cell-derived factor 1 alpha (SDF-1α) presented from a macroscale silk fibroin–hyaluronic acid–heparin aerogel sponge attracted C-X-C motif chemokine receptor 4 (CXCR4)-positive glioblastoma cells and redirected tumor growth toward the resection cavity.²⁷ The study examines chemoattractive control at the cavity–margin interface with a biomaterial scaffold; nanoscale delivery was not evaluated. Cavity filling alone is insufficient; material properties, release behavior, and spatial reach determine whether postoperative platforms engage the infiltrative margin.

Local Recurrence as a Niche-Driven Process

After GBM resection, local recurrence reflects regrowth within an inflammatory, treatment-remodeled, and spatially constrained niche; residual cell counts alone do not capture this process. Longitudinal single-cell genomics further shows that recurrent disease evolves across malignant and nonmalignant compartments under the combined influence of treatment exposure and microenvironmental selection.²⁸ Tissue-repair and immune programs may reinforce this transition. Fibrotic extracellular-matrix remodeling has been linked to a recurrence-supportive environment, while glioblastoma-instructed astrocytes can suppress tumor-specific T-cell immunity.^29,30

Myeloid populations add another layer of resistance through tumor-cell-intrinsic DNA-damage responses, vascular barrier changes, and inflammatory signaling; in a postoperative GBM model, local inhibition of neutrophil extracellular trap formation reduced neutrophil-driven proliferation and migration of residual tumor cells.^31,32 Longitudinal ecosystem mapping, therapy-induced fibrotic remodeling, and postoperative neutrophil biology place the therapeutic target beyond the cavity lumen, at the postoperative cavity–margin interface where residual cells, repair responses, immune suppression, and treatment-induced resistance interact before clinically visible relapse.

Advantages and Risks of Local Nanomedicine

The rationale for local nanomedicine is primarily pharmacological. Regional placement may concentrate exposure at residual tissue while reducing reliance on systemic dosing.³³ A non-nanomedicine irinotecan implant provides preclinical evidence that immediate perioperative delivery can suppress recurrence-related growth, while sustained-release hydrogels and nanogels expand the range of payloads and release profiles that can be evaluated locally.^34,35 These findings define a design opportunity but do not show that local delivery alone can overcome the biological drivers of recurrence.

Local placement also concentrates risk within injured brain. Inflammation, edema, wound compatibility, local toxicity, and neurologic tolerability require direct assessment. In a small reoperation study, local irinotecan delivery—a clinical, non-nanomedicine comparator—was feasible and provided an initial safety signal, but it did not establish postoperative efficacy.³⁶ Injectable in situ-forming hydrogels are designed to conform to postoperative defects, but their local performance depends on gelation, degradation, and release behavior.³⁷ Regional molecular or metabolic changes may also precede anatomical response, and preclinical chemical exchange saturation transfer MRI data suggest that spatially sensitive readouts may be needed to interpret local effects.³⁸ Direct postoperative delivery studies, niche-interception models, and early human regional delivery point to a conditional development strategy centered on reproducible exposure, brain safety, and interpretable benefit, not efficacy claims.

Cavity Retention and Local Platform Material Design

Local Platform Classes for Cavity Retention

The physical form of a local system determines how it occupies postoperative space, making platform selection a prerequisite to judging payload potency.³⁹ Cavity-directed systems fall into three groups: stand-alone nanocarriers; nano-enabled composites in which a nanoscale component materially contributes to delivery; and macro-/microscale biomaterials used as comparators for cavity conformity, retention, and surgical handling. Injectable hydrogels and in situ-forming depots can conform to irregular postoperative spaces.⁴⁰ For composite platforms, the nanoscale component must materially affect cargo protection, targeting, release, or biological activity. Preformed wafers provide immediate intracavitary placement and geometric persistence, but their fixed geometry can limit conformity to irregular resection surfaces.⁴¹ Figure 1 summarizes the platform classes and their intended roles at the postoperative cavity–margin interface. Local implantation alone does not make a system nanomedicine.

Figure 1.

Local platform design for cavity retention, margin anchoring, and controlled release after GBM resection. Postoperative local platforms are shown at the cavity–margin interface. Hydrogels, scaffolds, and nanocarriers represent a conceptual continuum from deformable or in situ-forming matrices to preformed and nano-enabled composite depots. The left side depicts cavity filling and anti-washout retention; the right side depicts margin anchoring, release control, and spatial reach. OFF denotes baseline passive release in the absence of an activating trigger, whereas ON denotes additional stimulus-triggered release; OFF does not mean complete absence of release. The downward arrow indicates distribution toward the resection margin. Dashed circles are magnified release-state insets, and concentric lines group platform classes without indicating time, physical scale, or quantitative performance. Colored circles, squares, and surface-bound shapes denote generic cargos, targeting or adhesive groups, or responsive elements and are nonquantitative. The anatomical outline is schematic. The figure was created using Figdraw.

Abbreviation: GBM, glioblastoma.

Nanofiber scaffolds fall within the nanomedicine core when their nanoscale architecture or incorporated nanoparticles materially affect cargo delivery or tissue interaction.⁴² Preformed three-dimensional scaffolds without a functionally relevant nanoscale component are instead treated as biomaterial comparators; the continuous liquid interface production–printed poly(ethylene glycol) diacrylate/gelatin methacryloyl scaffold, for example, informs postoperative cavity occupancy and local cell-therapy delivery rather than nanomedicine efficacy.⁴³ Liposomes, polymeric nanoparticles, and related nanocarriers constitute the nanoscale delivery component and may be used alone or embedded within local matrices to improve cargo loading, protect labile agents, or support combination designs.⁴⁴ The classification separates nanomedicine from broader local biomaterial engineering while retaining the latter when it provides directly relevant information on cavity filling, anchoring, release, or surgical use. Representative platform designs supporting cavity retention are summarized in Table 1.

Table 1.

Representative Platform Designs Relevant to Local Retention at the Postoperative Cavity–Margin Interface in Glioblastoma

Platform Design	Model	Retention Basis	Evidence Level	Ref.
NFL-functionalized lipid nanocapsule hydrogel carrying GemC12	Murine orthotopic glioblastoma resection	Conformal cavity filling; margin-localized nanocapsules; sustained release	Direct postoperative	[11]
CLIP-printed PEGDA/GelMA scaffold carrying therapeutic cells	Mouse glioblastoma resection	Preformed defect occupancy; limited swelling; four-week structural persistence	Direct postoperative	[43]
SF/HA/Hep aerogel sponge carrying SDF-1α	Rat glioblastoma resection	Chemokine retention; chemoattractive cell trapping within the cavity	Direct postoperative	[27]
Injectable chitosan hydrogel carrying tumoricidal iNSCs	Mock resection-cavity mouse model	Cell encapsulation; local persistence beyond 196 days	Indirect cavity model	[45]
Electrospun chitosan scaffold containing SDF-1α-loaded PLGA nanoparticles	Healthy rat brain cavity	Nanoparticle reservoir; SDF-1α release for at least five weeks	Indirect healthy-brain model	[46]
CXCL12-releasing thiol–Michael hydrogel	Dual-layer glioblastoma cell-capture model	Sustained chemotactic capture near a modeled margin	Indirect in vitro	[47]

Notes: Evidence level distinguishes direct postoperative glioblastoma studies in vivo from mock-cavity, healthy-brain, or in vitro studies. Indirect evidence informs retention design but does not establish postoperative antitumor efficacy.

Abbreviations: CLIP, continuous liquid interface production; CXCL12, C-X-C motif chemokine ligand 12; GelMA, gelatin methacryloyl; GemC12, lipophilic gemcitabine prodrug; iNSC, induced neural stem cell; NFL, NFL-TBS.40–63 peptide; PEGDA, poly(ethylene glycol) diacrylate; PLGA, poly(lactic-co-glycolic acid); SDF-1α, stromal cell-derived factor 1 alpha; SF/HA/Hep, silk fibroin/hyaluronic acid/heparin.

Physical Retention, Cavity Filling, and Anti-Washout Design

Physical retention after GBM resection begins with residence. Before a payload can influence residual disease, the material must persist within a fluid-exposed, deforming resection cavity.^33,48 Here, physical retention refers to gelation-driven occupancy, cavity filling, and resistance to dilution or displacement. Spatiotemporally controllable glioma-delivery systems can add distinct mechanisms for localization and release control beyond residence alone.⁴⁹ In situ-forming systems meet this initial requirement by converting an injectable precursor into a local matrix, as shown by macroscale elastin-like polypeptide hydrogels evaluated in release and cell-based malignant glioma studies.⁵⁰ These experiments inform tunable residence and dosing, but not performance in a postoperative recurrence model. Hydrogel–nanoparticle systems are being developed to improve local retention and controlled release in brain tumors, yet evidence from actual postoperative cavities remains limited.⁵¹

In a mock resection-cavity mouse model, a macroscale thermoresponsive chitosan hydrogel carrying tumoricidal induced neural stem cells supported prolonged local cell persistence.⁴⁵ This cavity-like model informs retention rather than direct postoperative GBM control. A second macroscale platform, a macroporous chitosan/alginate hydrogel crosslinked with genipin, accumulated and retained GBM cells in an in vitro cell-trap system.⁵² Cavity filling is only the starting point; a retained depot may still fail to reach infiltrative cells beyond the cavity surface or the biological programs that sustain the infiltrative margin.⁵³ Local delivery is one component of multimodal GBM treatment, not a stand-alone solution.⁵⁴ Once residence is secured, attention shifts from cavity occupancy to margin anchoring and release control.

Bioadhesion and Margin Anchoring

The extracellular matrix surrounding GBM is structurally remodeled and contributes to the tumor microenvironment.⁵⁵ Glioma cells can actively reorganize this matrix in ways that favor recurrence and progression.⁵⁶ Bioadhesion and margin anchoring refer to sustained contact with tissue-facing surfaces, matrix-rich repair zones, and adjacent residual tissue; they are distinct from bulk gelation or prolonged cavity filling. Stable margin coverage in the postoperative brain depends on matching the mechanical and interfacial properties of injured tissue.⁵⁷ In a dual-layer in vitro model, C-X-C motif chemokine ligand 12 (CXCL12) released from a macroscale synthetic hydrogel drove chemotactic migration of residual GBM and GSC-like cells.⁴⁷ The experiment establishes interface-directed capture in vitro; postoperative efficacy was not assessed.

The rat resection study extended this concept in vivo: SDF-1α presented from a macroscale silk fibroin–hyaluronic acid–heparin aerogel sponge attracted CXCR4-positive glioblastoma cells and redirected tumor growth toward the implanted sponge.²⁷ Cancer cell-sticky hydrogels designed to engage GBM cell membranes also reduced invasive behavior in experimental GBM systems, although they lack direct postoperative validation.⁵⁸ These studies suggest that anchoring should be judged by durable, selective tissue engagement that does not compromise distribution, degradation, or release. Whether stronger adhesion improves postoperative tumor control remains unknown.

Controlled Release and Degradation Kinetics

After retention and anchoring are established, release is better characterized by measurable kinetic features—initial burst, duration of exposure, and relation to matrix degradation—than by the generic label of “sustained release.”^59,60 Rapid regional drug availability was achieved with a supramolecular postoperative implant.⁶¹ Embedding lipid nanocapsules in a hydrogel prolonged local delivery in a murine GBM resection model.¹¹ In healthy rat brain, incorporation of poly(lactic-co-glycolic acid) (PLGA) nanoparticles into a scaffold maintained SDF-1α release for at least five weeks.⁴⁶ Only the murine resection study directly addresses postoperative GBM delivery.

A depot’s functional lifetime depends on degradation, which is most useful when aligned with cargo release so that the material does not outlast its delivery role.⁶² Composite formulations may contain free and nanoencapsulated agents with different trajectories, making a single release descriptor inadequate.⁶³ The intended biological task defines an acceptable kinetic profile more precisely than a generic sustained-release label. Trigger-responsive hydrogels have been designed to modulate release in response to pH, reactive oxygen species (ROS), enzymes, hypoxia, light, ultrasound, magnetic fields, or radiotherapy.⁶⁴ The biological sequencing of these profiles is considered in Sequential Release and Spatially Programmed Delivery.

Stimuli-Responsive Retention and Release

Stimulus-responsive systems couple release to biochemical cues or externally applied procedures instead of relying on passive diffusion alone.⁶⁵ Reviews of stimulus-responsive hydrogels emphasize matching the trigger mechanism to the intended tumor microenvironment or external procedure.⁶⁶ Endogenous cues such as acidic pH, ROS, matrix metalloproteinase activity, glutathione gradients, and hypoxia can alter network integrity or cargo liberation, whereas light, ultrasound, magnetic fields, and radiotherapy allow operator-controlled activation.^67,68

Procedural responsiveness was evaluated using a histotripsy-compatible thiol-Michael hydrogel in cavity-like spaces and porcine brain; the study addressed in situ crosslinking, ultrasound visibility, and ablation compatibility, but not nanomedicine efficacy.⁶⁹ Radiation-responsive release was examined with a hydrogel containing radiopaque gold nanoparticles in an intratumoral xenograft model; the study assessed triggering and localization, not postoperative performance.⁷⁰ Neither study compared the responsive construct with a simpler passive system matched for residence, degradation, and payload kinetics.

Control of Residual Disease and Targeting of Infiltrative Resistant Cells

Chemotherapeutic Payloads for Postoperative Residual Cells

Polymeric local drug-delivery systems have been investigated as components of multimodal treatment for malignant glioma.⁷¹ In postoperative settings, conventional chemotherapeutic payloads provide a direct preclinical test of whether local exposure affects residual disease. These agents include TMZ, bis(2-chloroethyl) nitrosourea (BCNU), doxorubicin (DOX), paclitaxel, and camptothecin. After resection, their relevance lies in the exposure they create at residual-tumor interfaces.⁷² Local delivery may increase cytotoxic exposure near infiltrative cells while reducing systemic exposure, an approach particularly relevant to agents limited by low brain concentrations, short effective residence, dose-limiting toxicity, or poor penetration through residual BBB/BBTB-protected tissue.^10,73 In orthotopic human xenograft resection/recurrence models, incorporation of TMZ-loaded particles into a chitosan–beta-glycerophosphate thermogel maintained postoperative exposure and reduced recurrent GBM growth.⁷⁴ Figure 2 schematically links engineered payloads with sustained postoperative exposure, margin-directed delivery, and residual-disease control.

Figure 2.

Local nanomedicine strategies to target residual disease after GBM resection. Engineered payloads are incorporated into a postoperative local depot to prolong regional exposure and support distribution toward the cavity-adjacent residual niche. Numbers 1–4 mark conceptual steps rather than measured time points or scores: (1) engineered therapeutic payloads; (2) sustained postoperative exposure; (3) margin-directed delivery; and (4) intended control of GSC-like and therapy-resistant GBM cells. The term “elimination” denotes a therapeutic objective, not demonstrated complete eradication. Solid arrows indicate conceptual loading, local release or distribution, and engagement with residual tissue; dotted green leader lines connect numbered callouts and do not indicate transport. The vascular channel represents local microvasculature and does not imply systemic administration or a validated intravascular route. The right-hand tissue contour denotes the residual niche, including tumor, stromal, and immune components. Repeated carrier icons and colored particles are illustrative and nonquantitative. The figure was created using Figdraw.

Abbreivations: GBM, glioblastoma; GSC, glioblastoma stem-like cell; MDSC, myeloid-derived suppressor cell.

For DOX, whose systemic use is limited by toxicity and poor brain delivery, a lipid-nanocapsule hydrogel carrying a doxorubicin C12 prodrug improved local antitumor activity in postoperative GBM models; organotypic systems provided complementary support.⁷⁵ In a rat resection model, a ROS-sensitive nano-enabled composite combining PLGA nanoparticles with free BCNU and TMZ reduced recurrence-related tumor growth through early and delayed drug exposure.⁷⁶ Across the TMZ, DOX, and BCNU/TMZ models, local cytotoxic exposure was feasible and recurrence-related outcomes improved. None tested whether local delivery overcomes pathway-dependent resistance, cell-state plasticity, or lineage-specific survival, and none establishes clinical benefit.

Molecularly Targeted Agents and Pathway Inhibition

Molecularly targeted payloads provide another investigational route to residual-disease control by addressing pathway-dependent survival programs in residual GBM cells after surgery. The target families considered here include epidermal growth factor receptor and its variant III, the phosphoinositide 3-kinase/protein kinase B/mechanistic target of rapamycin axis, vascular endothelial growth factor and its receptor, poly(ADP-ribose) polymerase (PARP) and broader DNA-damage-response machinery, and histone deacetylase (HDAC)-regulated epigenetic states.⁷⁷ Local delivery may sustain exposure at the resection margin while limiting systemic dose. Systematic reviews nevertheless show that evidence for local brain-cancer delivery remains predominantly preclinical.⁷⁸ Target heterogeneity and adaptive pathway compensation remain independent biological constraints. Receptor-level experience illustrates the uneven benefit of systemic pathway inhibition, while intracellular kinase networks show how difficult it is to sustain suppression of adaptive survival states at the infiltrative margin.^79,80

DNA-damage-response targets may be particularly relevant when regional exposure is paired with radiation or alkylating stress.⁸¹ Epigenetic targets extend this strategy beyond repair pathways. A molecularly imprinted alginate–poly(N-isopropylacrylamide) hydrogel inhibited Janus kinase (JAK)–signal transducer and activator of transcription 3 (STAT3) signaling, increased apoptosis, and reduced migration in vitro, but it has not been validated after resection.⁸² The N²M² umbrella trial serves as a clinical, non-local-delivery comparator for rational pathway combinations, yet it does not establish local or postoperative efficacy.⁸³ Contrasting local JAK/STAT3 inhibition in vitro with biomarker-matched systemic combinations in N²M² underscores the need to keep pathway plausibility, delivery setting, and clinical evidence separate. GBM heterogeneity and plasticity motivate more programmable approaches but do not establish their postoperative relevance.

Nucleic-Acid Therapeutics and Gene Regulation

Nucleic-acid payloads enable programmable gene regulation in GBM, and localized administration may target the resection cavity or peritumoral tissue.⁸⁴ Small interfering RNA (siRNA), microRNA, antisense oligonucleotide (ASO), messenger RNA (mRNA), and clustered regularly interspaced short palindromic repeats (CRISPR) are distinct modalities, not a single “gene therapy” category.⁸⁵ RNA delivery to GBM remains constrained by extracellular stability, BBB/BBTB transport, cellular uptake, and endosomal escape, and targeted lipid-nanoparticle strategies are being developed to address these barriers.⁸⁶ Framework nucleic-acid nanoparticles have enhanced temozolomide sensitivity in unresected GBM models, while ASO strategies offer transcript-specific suppression or splice modulation; both currently provide target and carrier rationale rather than direct postoperative evidence.^87,88

Gene regulation may also target immune escape and recurrence-permissive signaling, including programmed death-ligand 1 (PD-L1), cluster of differentiation 47 (CD47), transforming growth factor β (TGF-β), and STAT3. Nucleic-acid immunotherapeutics and vaccines are being investigated in GBM to address poor immunogenicity and the immunosuppressive tumor microenvironment.⁸⁹ A postoperative hydrogel carrying complexes of functionalized polyamidoamine (PAMAM) and short hairpin RNA (shRNA) provides direct preclinical evidence that local CD47 knockdown can increase macrophage phagocytosis and support postoperative GBM control when combined with TMZ.⁹⁰ mRNA platforms extend this strategy to instructive expression.⁹¹ The available evidence, however, is not specific to postoperative local delivery. CRISPR platforms extend the strategy to genome editing; evidence still derives mainly from models without tumor resection and remains constrained by delivery, intracellular trafficking, off-target effects, durability, and variable in vivo activity.⁹² In vivo profiling of ASO and siRNA conjugates in GBM xenografts showed that tissue distribution and heterogeneous activity remain major constraints.⁹³ These findings do not yet establish that nucleic-acid platforms can control GSC-like reservoirs or the infiltrative margin after resection.

Targeting Glioblastoma Stem-Like Cells and Infiltrative Margins

GSC-like cells and the infiltrative margin shift residual-disease control from payload class to residual-cell hierarchy; the populations most likely to drive relapse are not necessarily the most susceptible to bulk cytotoxic exposure.⁹⁴ Quiescent human GSCs exemplify one such reservoir, with experimental evidence showing that these cells can survive chemotherapy, re-enter proliferative programs, and drive tumor initiation, expansion, and recurrence after treatment.⁹⁵ The infiltrative edge presents a related challenge. Slow-cycling or quiescent glioma subpopulations have been linked to invasive behavior, extending the problem beyond cells at the cavity surface.⁹⁶ Lymphatic endothelial-like cells can support GSC growth through cytokine-driven cholesterol metabolism, illustrating the niche dependence of these states.⁹⁷ Whether local nanomedicine can reach and alter such states depends on sustained exposure and margin coverage rather than a brief, spatially narrow drug pulse.

Direct postoperative evidence for a stemness-directed approach came from a biohybrid chiral hydrogel containing GSC-membrane-coated nanoparticles; multipronged interception of pro-stemness signaling reduced regrowth in GBM resection models.⁹ Ex vivo and in vitro work paired drugs against bulk GBM cells and GSC-like compartments in a dual-drug lipid-nanocapsule hydrogel. The experimental setting supports payload selection; postoperative efficacy was not assessed.⁹⁸ The macroscale CXCL12-releasing hydrogel described above also attracted residual GBM and GSC-like cells in a dual-layer in vitro model.⁴⁷ By contrast, localized curcumin–temozolomide nanogels were tested after resection in a chemoresistant GBM model and provided direct preclinical evidence of chemosensitization.⁹⁹ Recurrence-associated functional heterogeneity and GSC enrichment argue against treating any single stemness- or invasion-directed platform as sufficient.¹⁰⁰ The available evidence supports further preclinical testing of sustained, margin-directed, multipathway intervention.

Multimodal Strategies for Residual Disease Control

Postoperative residual disease is heterogeneous, and spatially adjacent cells differ in proliferation, stemness, metabolic vulnerability, immune visibility, and treatment tolerance. This diversity provides a rationale for multimodal designs.¹⁰¹ Chemotherapy may be paired with radiosensitization to increase regional DNA damage where residual cells persist while limiting additional systemic exposure.¹⁰² Gene silencing can also be combined with cytotoxic stress. B7 homolog 6-targeted siRNA plus temozolomide reduced stemness and migration features in GBM cell culture, providing mechanistic rather than postoperative evidence.¹⁰³ Postoperative biphasic TMZ/curcumin release was achieved with a nano-enabled in situ system that combined rapid cytotoxic exposure, sustained delivery, and curcumin-associated chemosensitizing or microenvironmental effects.¹⁰⁴ TMZ and erastin were delivered from a hydrogel–liposome implant, increasing ferroptotic pressure and reducing recurrence-related outcomes in preclinical postoperative models.¹⁰⁵

Redox and lipid-peroxidation pathways were targeted with a macroscale triptolide-loaded hydrogel in an orthotopic postoperative relapse model.¹⁰⁶ In a separate macroscale system, co-embedding anti-CD47 antibody and TMZ combined cytotoxic pressure with relief of phagocytic suppression after resection.¹⁰⁷ Physical modalities offer another combination route. Chemo-photothermal treatment was implemented with a mesoporous polydopamine-based hydrogel, a nano-enabled composite that coupled local chemotherapy with near-infrared-triggered heating in a postoperative GBM model.¹⁰⁸ Photodynamic and sonodynamic strategies add further options, but the evidence discussed here largely comes from unresected GBM settings and remains indirect for postoperative local therapy.¹⁰⁹ Theranostic studies provide imaging and guidance context rather than direct evidence of postoperative efficacy.¹¹⁰ Table 2 summarizes representative local therapeutic payloads targeting postoperative residual disease.

Table 2.

Representative Local Therapeutic Platforms Relevant to Residual Glioblastoma After Resection

Local Platform	Payload	Model	Target	Main Findings	Evidence Level	Ref.
Chitosan–βGP thermogel with TMZ-loaded particles	TMZ	Orthotopic human xenograft resection	Residual tumor cells	Reduced postoperative regrowth	Direct postoperative	[74]
Lipid nanocapsule hydrogel	DOXC12 prodrug	Postoperative GL261 model; organotypic slices	Infiltrative residual cells	Prolonged survival; reduced spheroid regrowth	Direct postoperative; ex vivo support	[75]
ROS-responsive hydrogel with PLGA nanoparticles	BCNU; TMZ	Rat model with approximately 90% resection	Residual tumor cells	Reduced recurrence; prolonged survival	Direct postoperative	[76]
Localized curcumin–TMZ nanogel	Curcumin; TMZ	Postoperative TMZ-resistant C6 model	Chemoresistance; stemness	Reduced recurrence; prolonged survival	Direct postoperative	[99]
PIDDS gel with PLGA nanoparticles	Free TMZ; free curcumin; nanoparticle-loaded drugs	Rat postoperative recurrence model	DNA repair; GSC-like cells; recurrence niche	Prolonged survival; lower systemic toxicity	Direct postoperative	[104]
Biohybrid chiral hydrogel with GSC-membrane-coated nanoparticles	Stemness-neutralizing GSNP system	Postoperative GL261, CT2A, and PDX models	Pro-stemness signaling	Suppressed regrowth and stemness signaling; prolonged survival	Direct postoperative	[9]
GelMA–liposome implant	TMZ; erastin	Modified postoperative recurrence model	Ferroptosis sensitization	Reduced recurrent burden; prolonged survival	Direct postoperative	[105]
TP@DNH hydrogel	Triptolide; Fe3⁺/TA network	Orthotopic U87MG-Luc relapse model	Ferroptosis pathways	Reduced recurrence; prolonged survival	Direct postoperative	[106]
HPCH thermogel	Anti-CD47 antibody; TMZ	Incomplete-resection GL261-Luc model	Residual cells; phagocytic escape	Reduced recurrence; prolonged survival	Direct postoperative	[107]
ECM-mimetic hydrogel with MPDA@MTIC-Co	MPDA@MTIC-Co	Orthotopic postoperative model	Residual cells; invasion signaling	Reduced recurrence; prolonged survival	Direct postoperative	[108]
Molecularly imprinted alginate–PNIPAm hydrogel	Ruxolitinib	Glioblastoma cell models	JAK/STAT3 signaling	Increased apoptosis; reduced migration	Indirect in vitro	[82]
GemC12 lipid nanocapsule hydrogel	GemC12; salinomycin or curcumin	Spheroids; patient-derived GSCs	Bulk tumor cells; GSCs	Reduced spheroid growth and GSC fraction	Indirect ex vivo/in vitro	[98]

Notes: Evidence level distinguishes direct postoperative in vivo studies from ex vivo or in vitro studies retained for mechanistic context. Outcomes were generated in different models and should not be compared quantitatively across rows. The table includes stand-alone nanocarriers, nano-enabled composites, and local biomaterial comparators; their composition is specified in the Local platform and Payload columns.

Abbreviations: βGP, beta-glycerophosphate; BCNU, bis(2-chloroethyl) nitrosourea; C6, rat glioma cell line; CD47, cluster of differentiation 47; Co, cobalt; CT2A, murine glioma cell line; DOXC12, doxorubicin C12 prodrug; ECM, extracellular matrix; Fe³⁺, ferric ion; GelMA, gelatin methacryloyl; GemC12, lipophilic gemcitabine prodrug; GL261, murine glioma cell line; GSNP, glioblastoma stem-like cell membrane-coated nanoparticle; GSC, glioblastoma stem-like cell; HPCH, hydroxypropyl chitin; JAK, Janus kinase; Luc, luciferase-labeled; MPDA, mesoporous polydopamine; MTIC, 3-methyl-(triazene-1-yl)-imidazole-4-carboxamide; PDX, patient-derived xenograft; PIDDS, postoperative in situ drug delivery system; PLGA, poly(lactic-co-glycolic acid); PNIPAm, poly(N-isopropylacrylamide); ROS, reactive oxygen species; STAT3, signal transducer and activator of transcription 3; TA, tannic acid; TMZ, temozolomide; TP@DNH, study-specific triptolide-preloaded hydrogel formulation; U87MG, human glioblastoma cell line.

Immune Remodeling and Reprogramming of the Postoperative Recurrence Niche

Immunosuppressive Features of the Postoperative GBM Niche

GBM is characterized by a strongly immunosuppressive microenvironment that can support tumor persistence and limit antitumor immunity.¹¹¹ In the postoperative setting, the same biology remains relevant at the cavity-adjacent residual niche. Tumor-associated macrophages (TAMs) and resident microglia form a dominant suppressive myeloid axis that shapes cytokine signaling, phagocytic restraint, antigen handling, and tissue-repair programs favoring residual tumor persistence.¹¹² Myeloid-derived suppressor cells (MDSCs) and regulatory T cells (Tregs) add interconnected suppressive circuits that reinforce immune escape.¹¹³ Exhausted or inert T-cell states further constrain immunity even when lymphocytes are present.¹¹⁴

Defective antigen presentation limits antitumor immunity by constraining antigen-specific T-cell responses.¹¹⁵ Hypoxia-associated metabolic suppression further limits nutrient availability and effector function.¹¹⁶ Postoperative inflammatory cells may also drive recurrence. In a GBM resection model, a copper selenide nanozyme hydrogel reduced neutrophil extracellular trap formation and dampened neutrophil-driven recurrence signals.³² Single-cell profiling has shown that the immune landscape evolves during GBM progression.¹¹⁷ Considered alongside astrocyte-mediated T-cell suppression and postoperative neutrophil biology, these data place the suppressive niche across resident, infiltrating, and tumor-instructed compartments rather than within a single immune-cell subset.

Reprogramming TAMs, Microglia, and MDSCs

Therapeutic targeting of myeloid cells is an active investigational strategy in GBM.¹¹⁸ After resection, its relevance reflects the abundance and plasticity of myeloid populations around residual disease. TAMs, resident microglia, and recruited MDSCs coordinate cytokine signaling, repair responses, antigen handling, and tumor–immune crosstalk.^112,119 Their heterogeneity and plasticity favor state reprogramming over indiscriminate depletion.¹²⁰ Macrophage-directed strategies in GBM include colony-stimulating factor 1 receptor- and STAT3-related approaches.¹²¹ TGF-β can reinforce wound-healing-like and immunosuppressive myeloid states in GBM.¹²²

Single-cell and functional analyses have identified distinct MDSC populations in human GBM, complicating uniform MDSC-directed intervention.¹²³ Myeloid checkpoints such as CD47/signal regulatory protein α (SIRPα) restrain phagocytosis and antitumor conversion.¹²⁴ Complementary postoperative approaches include a nanoporter-hydrogel that generated GSC-specific CAR macrophages within the resection cavity and a fibrin gel carrying M1-like macrophage nanovesicles loaded with mesoporous polydopamine–doxorubicin that shifted TAM phenotypes while reducing MDSC/Treg-associated suppression after subtotal resection.^125,126 Together with the T-cell-active exosome-cross-linked gel, these platforms indicate that myeloid remodeling is one component of local immune control, not a sufficient stand-alone strategy.

Restoration of T-Cell Antitumor Immunity

T-cell dysfunction is a major barrier to antitumor immunity in GBM.¹²⁷ At the postoperative cavity–margin interface, restoring T-cell-mediated activity cannot compensate for persistent myeloid suppression or inadequate antigen presentation. Effective recovery also requires entry of cluster of differentiation 8-positive (CD8+) T cells into the cavity-adjacent residual niche, as GBM-infiltrating CD8+ T cells may include clonally expanded effector populations that are spatially present yet still require functional support.¹²⁸ When antigen presentation remains weak, infiltration alone is insufficient; dendritic-cell priming and antigen-presenting-cell competence determine whether local T cells can recognize and expand against residual tumor antigens.¹²⁹ Outside local-delivery and resection settings, machine-learning-directed conversion of GBM cells toward dendritic cell-like antigen-presenting states provided mechanistic evidence for enhancing antigen presentation.¹³⁰ Postoperative delivery of an exosome-cross-linked gel designed as an artificial lymph-node-like structure promoted local T-cell activation, proliferation, and CD8+ T-cell infiltration and was associated with delayed recurrence.¹²

Even after entry and priming, T-cell exhaustion remains a separate barrier. Chronic stimulation produces a functionally constrained state, not merely a lack of lymphocytes.¹³¹ Checkpoint modulation may therefore need to relieve multiple exhaustion programs. Programmed cell death protein 1 (PD-1)/PD-L1 is central, but cytotoxic T-lymphocyte-associated protein 4, T-cell immunoglobulin and mucin-domain containing-3, and lymphocyte-activation gene 3 can mark additional or compensatory inhibitory programs that may limit durable single-axis blockade.¹³² A recurrent GBM nivolumab study was neither a local nanomedicine trial nor a postoperative study; it showed lesion exposure and T-cell pharmacodynamic activity while also revealing adaptive checkpoint upregulation.¹³³ A sequence-based interpretation of T-cell recovery therefore includes infiltration, antigen presentation, effector preservation, and context-specific checkpoint control.

Innate Immune Activation and Pattern-Recognition Receptor Strategies

Innate immune activation may link antigen release, antigen presentation, and T-cell engagement, although the strength and durability of this link remain model-dependent.¹³⁴ Signaling through cyclic GMP–AMP synthase and stimulator of interferon genes (STING) provides one route, as sensing tumor- or treatment-derived cytosolic DNA can induce type I interferon programs, dendritic-cell activation, antigen presentation, and cross-priming.¹³⁵ Toll-like receptor (TLR) agonists, including unmethylated cytosine–phosphate–guanine motifs acting through TLR9, offer a parallel route to innate priming by activating antigen-presenting-cell and dendritic-cell programs that convert local antigen release into productive T-cell engagement.¹³⁶ In postoperative mouse GBM models, a biodegradable acetalated dextran scaffold sustained local TLR7/8 agonist delivery and was associated with immune-mediated tumor clearance and protection against later challenge.¹³⁷ The acetalated dextran scaffold is a macro-/microscale biomaterial comparator.

Local stress biology offers a second route to innate activation. Mitochondrial stress and local innate immunity were induced with a nano-enabled hydrogel, which reduced recurrence-related outcomes in mouse postoperative GBM models.¹³⁸ Nanoscale self-assembly of a STING-agonist prodrug within a hydrogel altered innate–adaptive immune readouts and reduced recurrence-related outcomes in a postoperative model.¹³⁹ Vesicular and nanoparticle components were integrated into a bacterium–hydrogel superstructure that stimulated tumoricidal immune responses and reduced relapse-related outcomes in an orthotopic GBM resection model.¹⁴⁰ STING agonist 8803 was evaluated only in unresected or intratumoral preclinical GBM models, so its relevance here is limited to immune-mechanism selection.¹⁴¹

Local Delivery of Immunomodulatory Payloads

Local immunomodulatory payloads address one component of postoperative immune control. Biomimetic platforms have been explored to improve the delivery and tumor localization of GBM immunotherapies.¹⁴² Regional approaches such as convection-enhanced delivery are intended to increase local immunomodulatory exposure while limiting systemic distribution.¹⁴³ Biomaterial and nanotechnology platforms can carry diverse immunomodulatory cargos, including cytokines, antibodies, nucleic acids, and small molecules.¹⁴⁴ In an unresected preclinical GBM model, intratumoral interleukin-12 supported CAR T-cell activity.¹⁴⁵ The experiment was neither postoperative nor nanomedicine-based and is used only for mechanistic comparison. Postoperative glutamine-metabolism blockade was combined with copper–histidine nanoparticle-mediated chemodynamic therapy in a nano-enabled immunostimulatory hydrogel; the treatment altered local immune readouts and reduced recurrence-related outcomes in a postoperative GBM model.¹⁴⁶

PAMAM/shRNA complexes retained within a nano-enabled hydrogel maintained local CD47 gene silencing after surgery and provided direct preclinical postoperative evidence for an immune-regulatory payload.⁹⁰ The anti-CD47/TMZ system instead used a macroscale in situ hydrogel to pair phagocytosis-oriented checkpoint modulation with cytotoxic pressure after resection.¹⁰⁷ Studies of local antibody delivery in other solid tumors illustrate design options outside GBM.¹⁴⁷ Broader work on nanomaterial-driven immunomodulation provides general design context; efficacy after GBM resection remains untested.¹⁴⁸ Figure 3 summarizes the immune pathways and local payload strategies discussed in this section. Table 3 lists representative postoperative immune-remodeling strategies.

Figure 3.

Local immune remodeling in the postoperative GBM recurrence niche. Representative immune-remodeling pathways are shown in the cavity-adjacent niche after GBM resection. (A–D) are conceptual modules rather than a temporal sequence. (A) illustrates antigen transfer from GBM cells to DCs and subsequent priming of CD8+ T cells through pMHC–TCR recognition. (B) depicts NK-cell cytotoxicity through perforin release and death-receptor signaling involving Fas/FasL and TNFR-associated pathways. (C) shows reprogramming from M2-like toward M1-like TAM states, cytokine-mediated support of CD8+ T-cell activity, and inhibition of GBM cells. (D) summarizes locally delivered IL-12, shRNA, and anti-CD47-related payloads intended to reduce CD47/PD-L1/TGF-β-associated immune escape and T-cell exhaustion and to increase phagocytosis. The dashed outline around the central vesicle marks a conceptual delivery-vehicle inset; the listed cargos and immune-associated elements are representative and are not necessarily co-loaded in one platform. Solid arrows indicate directional transfer, activation, reprogramming, release, cytotoxic signaling, or downstream effects; blunt-ended lines indicate inhibition. Upward (↑) and downward (↓) arrows denote qualitative increases and decreases, respectively, not quantitative effect sizes or established clinical outcomes. CD8+ T cells are T lymphocytes expressing CD8. Colors and icon sizes distinguish schematic categories only and do not encode magnitude, evidence strength, or physical scale. The figure was created using Figdraw.

Abbreviations: GBM, glioblastoma; DC, dendritic cell; CD8, cluster of differentiation 8; pMHC, peptide–major histocompatibility complex; MHC, major histocompatibility complex; TCR, T-cell receptor; NK, natural killer; Fas, Fas receptor; FasL, Fas ligand; TNFR, tumor necrosis factor receptor; TAM, tumor-associated macrophage; IL-12, interleukin-12; shRNA, short hairpin RNA; CD47, cluster of differentiation 47; PD-L1, programmed death-ligand 1; TGF-β, transforming growth factor β; STING, stimulator of interferon genes; TLR, Toll-like receptor.

Table 3.

Representative Local Strategies for Immune Remodeling After Glioblastoma Resection

Local Platform	Intervention	Model	Mechanism	Immune Response	Tumor Outcome	Ref.
Exosome-cross-linked gel	DT-Exos; optional STING agonist	GL261-OVA and GL261-Luc resection models	CD8 priming; PD-1 restraint; antigen presentation	Increased proliferative CD8+ T cells; reduced PD-1; proinflammatory shift	Delayed recurrence; prolonged survival	[12]
Nanoporter hydrogel	In situ CD133-CAR macrophage program	GL261 and humanized PDX resection models	CAR macrophage generation; phagocytosis	Generated GSC-specific CAR macrophages; enhanced adaptive immunity	Suppressed regrowth; prolonged survival	[125]
Ace-DEX scaffold	Resiquimod	GL261 and CT2A resection models	TLR7/8 activation; antigen presentation	Activated myeloid antigen-presenting cells; limited exhausted T-cell states	Cleared residual tumor; protected against rechallenge	[137]
Peptide gel with Cu-His nanoparticles	CB-839; Cu-His nanoparticles	Luc-GL261 resection model	Metabolic blockade; immunogenic cell death	Increased mature dendritic cells, M1-like macrophages, and CD8+ T-cell infiltration	Reduced recurrence; prolonged survival	[146]
CuSe nanozyme hydrogel	CuSe nanozyme; doxorubicin	GL261 resection model; patient-linked analyses	NET formation; neutrophil signaling	Reduced NET formation and neutrophil-driven inflammatory signaling	Prolonged survival; preserved neurobehavior	[32]
INSTNA hydrogel	Mitochondrial-stress nanotherapeutic; optional TTF	GL261 and CT2A resection models	Innate activation; CD8 memory	Induced an immunostimulatory cavity state and CD8 memory	Suppressed regrowth; protected against rechallenge	[138]
STING prodrug hydrogel	CDA; GM-CSF; PTX hydrogelator	GL261-Luc resection model	STING activation; dendritic-cell maturation	Increased dendritic-cell recruitment and CD8 activity; reduced Tregs	Reduced recurrence-related outcomes; prolonged survival	[139]
PLGA–PEG–PLGA hydrogel with PAMAM/shRNA complexes	G5-BGG-shRNA871	Postoperative U87MG model	CD47 silencing; macrophage phagocytosis	Reduced CD47; increased macrophage phagocytosis	Prolonged survival with TMZ	[90]
Fibrin gel with M1 nanovesicles	MPDA–DOX; near-infrared irradiation	Subtotal-resection glioblastoma model	TAM repolarization; MDSC suppression	Increased M1-like TAMs; reduced MDSCs and Tregs	Inhibited postoperative regrowth	[126]
Bacterium–hydrogel superstructure	SDVs; SLINs	GL261 resection model	Pyroptosis; antigen presentation	Increased phagocyte recruitment and antigen presentation	Inhibited relapse; prolonged survival	[140]

Notes: All entries include postoperative in vivo models. Immune and tumor-control outcomes are summarized as reported and should not be ranked across models. Platform composition is specified in the Local platform and Intervention columns. Systems without a functional nanoscale component are presented as local biomaterial comparators rather than nanomedicine systems.

Abbreviations: Ace-DEX, acetalated dextran; CAR, chimeric antigen receptor; CB-839, telaglenastat; CDA, cyclic dinucleotide agonist; CD8, cluster of differentiation 8; CD47, cluster of differentiation 47; CD133, cluster of differentiation 133; CT2A, murine glioma cell line; Cu-His, copper-histidine; CuSe, copper selenide; DOX, doxorubicin; DT-Exos, dendritic cell–tumor cell fusion-derived exosomes; G5-BGG-shRNA871, study-specific short hairpin RNA delivery complex; GL261, murine glioma cell line; GM-CSF, granulocyte–macrophage colony-stimulating factor; GSC, glioblastoma stem-like cell; INSTNA, study-specific mitochondrial-stress nanotherapeutic formulation; Luc, luciferase-labeled; M1, classically activated macrophage-like phenotype; MDSC, myeloid-derived suppressor cell; MPDA, mesoporous polydopamine; NET, neutrophil extracellular trap; OVA, ovalbumin; PAMAM, polyamidoamine; PD-1, programmed cell death protein 1; PDX, patient-derived xenograft; PEG, poly(ethylene glycol); PLGA, poly(lactic-co-glycolic acid); PTX, paclitaxel; SDV, Salmonella delivery vehicle; shRNA, short hairpin RNA; SLIN, Salmonella lysis-inducing nanocapsule; STING, stimulator of interferon genes; TAM, tumor-associated macrophage; TLR, Toll-like receptor; TMZ, temozolomide; Treg, regulatory T cell; TTF, tumor-treating fields; U87MG, human glioblastoma cell line.

Integrated Design, Evaluation Models, and Translational Criteria

From Single-Function Depots to Multifunctional Local Platforms

The preceding sections distinguish three separable functions—retention and tissue reach, residual-disease control, and immune remodeling. Testing them separately makes the contribution of each function easier to interpret. One postoperative platform demonstrated penetrative delivery.¹⁴⁹ The mitochondrial-stress and STING-agonist hydrogels discussed in Innate Immune Activation and Pattern-Recognition Receptor Strategies examined innate and adaptive immune effects. No direct comparison has shown that combining the three functions improves outcome. Additional complexity is justified only when it produces a prespecified gain in distribution, biological effect, safety, or operative handling.

Sequential-release studies outside GBM show that distinct functions can be staged, but they remain materials precedents rather than evidence of postoperative benefit.¹⁵⁰ Their relevance depends on whether proposed functions can be ordered in space and time and then validated in postoperative models, endpoint frameworks, and clinical workflows.

Sequential Release and Spatially Programmed Delivery

Injectable hydrogels can be engineered for staged or personalized release of cancer immunotherapies.¹⁵¹ Release timing can be aligned with the intended order of biological events. Two-stage carboplatin hydrogel work in unresected glioma provides indirect support for temporal sequencing.¹⁵² Postoperative BCNU/TMZ systems show that free and nanoparticle-bound drugs can be staggered in resection models.⁷⁶ Staggered release was also reported with a postoperative TMZ/curcumin system.¹⁰⁴ Broader reviews of hydrogel-based local immunotherapy document controlled local delivery across multiple cancer settings.¹⁵³ Staged exposure is technically feasible; the sequence most likely to connect residual-cell killing with antigen release and immune engagement remains undefined.

Later immunomodulatory phases are supported mainly by non-GBM precedents, including hydrogel-driven tertiary lymphoid structure formation, and require direct postoperative validation.¹⁵⁴ Locoregional cancer delivery must balance depot retention with distribution beyond the placement site.¹⁵⁵ In postoperative GBM, the corresponding objective is to reach the cavity wall and infiltrative margin. A transferrin-targeted TMZ nanoparticle-in-gel system provides postoperative preclinical evidence that gel retention can be combined with targeted tissue delivery.¹⁵⁶ Hydrogel systems have been engineered for spatiotemporally controlled delivery of immunomodulators within tumor immune microenvironments.^151,157 When adapted to postoperative GBM, such systems may also need to distribute toward immune-cell-rich regions. Added programming is meaningful only when it improves a defined outcome over a simpler sustained-release control.

Postoperative Recurrence Models

Postoperative resection models provide the most direct preclinical test of temporal and spatial programming because they reproduce the surgical event that creates the cavity, injured margin, and residual-cell ecology targeted by local nanomedicine.¹⁵⁸ Findings from unresected orthotopic, conventional intratumoral, ex vivo, in vitro, and non-GBM systems contribute mechanistic or design information; none substitutes for clinical validation. Model setting is reported explicitly in the tables. Tables 1 and 2 separate direct postoperative from indirect evidence, Table 3 includes only postoperative in vivo studies, and Table 4 presents translational comparators separately from efficacy evidence. Models that omit resection may overstate therapeutic relevance. Murine work integrating surgery and corticosteroid exposure showed that perioperative variables can reshape local and systemic immune phenotypes along with tumor growth and outcome.¹⁵⁹ Orthotopic partial-resection models, resection-cavity implantation systems, and explicit postoperative recurrence models offer the most appropriate in vivo basis for testing cavity retention, margin coverage, material placement, and regrowth control.¹⁶⁰ For immune-remodeling claims, syngeneic immunocompetent models and carefully selected humanized models are particularly important for interpreting macrophage, T-cell, or cytokine readouts as efficacy-linked evidence.¹⁶¹

Table 4.

Translational Evidence Relevant to Postoperative Local Nanomedicine in Glioblastoma

Evidence Source	Domain	Study Setting	Core Measure	Interpretation	Ref.
EANS–EANO guideline	Surgical reporting	Adult diffuse glioma surgery	Maximum safe resection; residual-tumor class	Operative reporting framework; not evidence for local nanomedicine	[2]
Multicenter postoperative cohort	Risk stratification	Newly diagnosed IDH-wildtype glioblastoma	Resection class; MGMT; age; KPS	Baseline stratification; retrospective and intervention-independent	[162]
Multicenter recurrent cohort	Re-resection endpoints	Recurrent IDH-wildtype glioblastoma	Residual enhancing volume; resection class	Salvage-study stratification; retrospective and not local-therapy specific	[163]
Co-clinical organoid study	Response modeling	Patient-derived organoids; CAR T-cell trial	Cytolysis; antigen loss; cytokine release	Patient-linked screening; CAR T-cell-specific and not based on a resection model	[164]
Brain-barrier chip study	Delivery screening	Ultrasound-responsive BBB–glioblastoma chip	Barrier integrity; permeability; apoptosis	Delivery-variable screening; in vitro and not based on a resection model	[165]
Image-guided nanoprobe study	Margin evaluation	Orthotopic models; patient tissue	Pharmacokinetics; biodistribution; toxicity; specificity	Margin-imaging and safety precedent; no resection model	[166]
Confocal endomicroscopy study	Surgical usability	Intraoperative glioma margins	Diagnostic accuracy; reader agreement	Real-time margin interpretation; non-therapeutic with modest specificity	[167]
RANO 2.0 guideline	Imaging endpoints	Adult glioma response assessment	Post-radiotherapy baseline; progression confirmation	Standardized radiographic assessment; not local-therapy specific	[168]
First-in-human cavity-lining study	Cavity feasibility	Newly diagnosed glioblastoma resection	Procedural safety; serial MRI; survival	Direct workflow precedent; small single-arm non-nanomedicine study	[169]
Intraoperative photodynamic trial	Clinical endpoints	5-ALA-guided glioblastoma resection	PFS; OS; HRQOL; MRI; toxicity	Local-treatment follow-up template; small single-arm non-nanomedicine study	[170]

Notes: These sources inform surgical reporting, model selection, endpoint definition, workflow feasibility, and safety assessment. Except where explicitly stated, they do not establish the efficacy of postoperative local nanomedicine.

Abbreviations: 5-ALA, 5-aminolevulinic acid; BBB, blood-brain barrier; CAR, chimeric antigen receptor; EANS, European Association of Neurosurgical Societies; EANO, European Association of Neuro-Oncology; HRQOL, health-related quality of life; IDH, isocitrate dehydrogenase; KPS, Karnofsky Performance Status; MGMT, O⁶-methylguanine-DNA methyltransferase; MRI, magnetic resonance imaging; OS, overall survival; PFS, progression-free survival; RANO, Response Assessment in Neuro-Oncology.

Ex vivo brain-slice systems complement these models by preserving short-term tissue architecture and tumor–brain interactions. They can therefore screen local platform behavior, device–tissue contact, or margin-level responses before in vivo recurrence studies.¹⁷¹ Patient-derived GBM organoids provide a translationally relevant screening layer, but not clinical validation. In a co-clinical CAR T-cell trial workflow, organoid avatars were used to assess cytolysis, antigen loss, and cytokine release.¹⁶⁴ Organ-on-chip platforms address a separate gap. An ultrasound-responsive BBB–glioblastoma microfluidic chip quantified barrier integrity, nanocarrier permeability, and sonodynamic therapy-associated apoptosis under controlled microfluidic conditions, although postoperative in vivo validation remains necessary before such systems can support claims about local recurrence control.¹⁶⁵

Evaluation Endpoints

Endpoint design should reflect the function claimed for each postoperative local nanomedicine platform. Response Assessment in Neuro-Oncology (RANO) 2.0 provides a standardized framework for adult glioma response assessment.¹⁶⁸ Fluorescence-guided nanoprobe studies show that margin delineation, pharmacokinetics, biodistribution, specificity, and toxicity can be measured for imaging-oriented platforms.¹⁶⁶ Retention-focused evaluations are strengthened by direct measures of residence, release, degradation, and cavity or margin coverage. Imaging endpoints must also distinguish recurrence from treatment-related change.¹⁷² Clinical local-therapy studies such as INDYGO have reported progression-free survival, overall survival, health-related quality of life, serial MRI, and toxicity.¹⁷⁰ Preclinical antitumor evaluation can add residual-cell burden and time to local recurrence. Systemic and local immune suppression carry prognostic relevance in GBM.¹⁷³ Immune-remodeling platforms are therefore best evaluated with prespecified immune endpoints.

Peritumoral edema is a recognized complication in glioma and warrants dedicated assessment.¹⁷⁴ Local brain-delivery studies are most informative when they separately monitor neurotoxicity, excessive inflammation, behavioral effects, and imaging interference. Microfluidic BBB–GBM models can quantify barrier integrity and transport behavior.¹⁷⁵ These readouts complement recurrence, immune, and safety outcomes without replacing them. Automated glioma-margin systems illustrate that margin measurements can be standardized and assessed within surgical workflows.¹⁷⁶ Safety-related metrics likewise benefit from reproducible reporting. Operational guidance for RANO 2.0 supports standardized imaging interpretation across clinical settings.¹⁷⁷ Standardization is necessary but cannot by itself establish clinical benefit without comparative human testing.

Clinical Translation and Surgical Usability

Performance in a resection model does not establish human efficacy. Translation of nanomedicine also depends on reproducible manufacture, characterization, stability, scale-up, and regulatory readiness.¹⁷⁸ Deployment within neurosurgical workflows adds a separate procedure-specific requirement. Sterilization and storage stability are integral to development because they can alter hydrogel properties and in vivo performance before the material reaches a resection cavity.¹⁷⁹ Surgical fit is anatomy-dependent; cavity-directed systems must respect extent-of-resection standards and functional preservation.² They must also accommodate the irregular residual space created by maximal safe resection. Operative usability includes field control. An in situ chemo-immunotherapeutic hydrogel, a macroscale biomaterial evaluated in a postoperative glioma recurrence model, incorporated hemostatic function and indicated that local treatment may need to coexist with surgical bleeding management.¹⁸⁰

Real-time interpretability also shapes the clinical use of margin-facing tools. Intraoperative confocal endomicroscopy showed that diagnostic performance and user agreement can be measured within a practical glioma-margin workflow.¹⁶⁷ Postoperative compatibility extends beyond placement. A first-in-human vascularized flap trial lining the GBM resection cavity incorporated serial safety monitoring and MRI follow-up into its feasibility assessment.¹⁶⁹ A validated postoperative risk model for newly diagnosed GBM illustrates the value of setting-specific risk stratification.¹⁶² Recurrent GBM adds another challenge: re-resection outcomes depend on extent-of-resection classification, making endpoint interpretation inseparable from the surgical context in which a local platform is tested.¹⁶³ These non-nanomedicine clinical comparators define workflow, feasibility, and endpoint standards; they do not demonstrate nanomedicine efficacy. Current human evidence is insufficient to establish that postoperative local nanomedicine improves recurrence or survival, and comparative clinical evaluation remains necessary. Table 4 summarizes evaluation and translational criteria for postoperative local nanomedicine.

Conclusions and Future Perspectives

The evidence reviewed here shifts attention from the postoperative cavity as a passive drug reservoir to the postoperative cavity–margin interface as a biologically active therapeutic site. Recurrence emerges from interactions among infiltrative and stem-like tumor states, repair-associated remodeling, and local immune suppression; residual cell number alone is insufficient. Local nanomedicine is most informative when platform behavior matches a sequence of postoperative tasks: stable placement, margin coverage, appropriately timed release, control of residual disease, and, where relevant, immune remodeling. The resection-site framework used in this review separates direct postoperative evidence from indirect mechanistic or design support and applies the same criteria to stand-alone nanocarriers, nano-enabled composites, and local biomaterial comparators. The result is a conditional therapeutic rationale, not evidence of established clinical benefit.

The evidence base remains limited. Most efficacy claims derive from heterogeneous preclinical resection models. Unresected orthotopic, intratumoral, ex vivo, in vitro, and non-GBM studies can support mechanism or design, but not postoperative efficacy. Comparability is constrained by differences in surgical extent, residual burden, cavity geometry, immune competence, adjuvant treatment, release measurements, and recurrence endpoints. Material scale adds further heterogeneity. Nanoscale carriers, composite hydrogels, scaffolds, and implants are not equivalent simply because they are placed locally. Human evidence is largely limited to feasibility, safety, pharmacodynamic, or workflow precedents, without comparative proof that postoperative local nanomedicine delays recurrence or improves survival. Prolonged residence and multifunctionality may also introduce edema, inflammation, neurotoxicity, imaging interference, manufacturing variability, and operative complexity. Simple, well-controlled platforms therefore remain necessary comparators and should not be treated as obsolete designs.

Explicit, testable standards would make progress easier to assess. The most informative preclinical studies would use clinically credible orthotopic resection models that report the surgical procedure, residual disease, perioperative treatments, and immune context, with head-to-head comparisons against free drug, unloaded material, and the simplest functional platform. These studies would quantify cavity retention, tissue-facing contact, margin penetration, release and degradation, residual-cell control, immune effects, and recurrence-related outcomes while monitoring edema, neurobehavior, wound compatibility, systemic exposure, and imaging artifacts. Before clinical testing, formulation and device development need to establish reproducible manufacture, sterilization, storage stability, batch consistency, and practical deployment in irregular resection cavities without disrupting hemostasis or standard postoperative care. Early trials would be more informative with interpretable postoperative imaging, predefined local and neurologic safety monitoring, and setting-specific endpoints that distinguish technical feasibility from antitumor activity and, ultimately, patient benefit. A reproducible chain of evidence linking cavity fit and margin coverage to biological control, brain safety, surgical usability, and clinically meaningful outcomes will define the next generation of postoperative GBM nanomedicine; maximal material complexity alone will not.

Funding Statement

This work was supported by the Henan Provincial Health Commission Provincial-Ministerial Co-Construction Project (grant number SBGJ202102198).

Abbreviations

ASO, antisense oligonucleotide; BBB, blood-brain barrier; BBTB, blood-brain tumor barrier; BCNU, bis(2-chloroethyl) nitrosourea; CAR, chimeric antigen receptor; CD8, cluster of differentiation 8; CD47, cluster of differentiation 47; CRISPR, clustered regularly interspaced short palindromic repeats; CSF, cerebrospinal fluid; CXCL12, C-X-C motif chemokine ligand 12; CXCR4, C-X-C motif chemokine receptor 4; DOX, doxorubicin; GBM, glioblastoma; GSC, glioblastoma stem-like cell; HDAC, histone deacetylase; JAK, Janus kinase; MDSC, myeloid-derived suppressor cell; mRNA, messenger RNA; MRI, magnetic resonance imaging; PAMAM, polyamidoamine; PARP, poly(ADP-ribose) polymerase; PD-1, programmed cell death protein 1; PD-L1, programmed death-ligand 1; PLGA, poly(lactic-co-glycolic acid); RANO, Response Assessment in Neuro-Oncology; ROS, reactive oxygen species; SDF-1α, stromal cell-derived factor 1 alpha; shRNA, short hairpin RNA; siRNA, small interfering RNA; SIRPα, signal regulatory protein α; STAT3, signal transducer and activator of transcription 3; STING, stimulator of interferon genes; TAM, tumor-associated macrophage; TGF-β, transforming growth factor β; TLR, Toll-like receptor; TMZ, temozolomide; Treg, regulatory T cell.

Data Sharing Statement

Data sharing is not applicable to this article because no datasets were generated or analyzed during the current review.

Ethics Approval and Informed Consent

This review did not involve human participants, human data, human tissue or animals.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that they have no competing interests.