Precision reirradiation in recurrent CNS tumors integrating biology techniques and multimodal care

Abstract

Reirradiation has re-emerged as a viable salvage modality for recurrent central nervous system (CNS) tumors in the modern era of precision radiotherapy. Advances in image guidance, biological understanding of normal-tissue repair, and integration of targeted systemic agents have expanded its therapeutic window. The 2025 ISRS meta-analysis (62 studies; n = 2640) established a median overall survival (OS) of ~ 10.2 months after re-irradiation (re-RT) of recurrent HGG, confirming that focal single-fraction SRS (16 Gy × 1) or hypofractionated SRS (25 Gy/5 fx) achieves optimal tumor control with acceptable neurologic toxicity (7% vs 4%). EQD₂ > 120–130 Gy markedly increases radionecrosis risk (p = 0.003). Integration of DNA-damage-repair (DDR) inhibitors, VEGF blockade, and tumor-treating fields (TTFields) is under active investigation. Re-RT in ependymoma, meningioma, and spinal tumors is also gaining renewed interest with proton and MR-guided adaptive techniques.

Highlights

• Focal reirradiation yields median OS ~ 10 months in recurrent HGG

• EQD2 > 120 Gy increases radionecrosis risk (ISRS 2025)

• DDR inhibitors, VEGF blockade, TTFields enhance re-RT efficacy

• MR-Linac & radiomics enable adaptive, personalized re-RT

• Multimodal, molecularly guided strategies optimize outcomes

Introduction

Radiotherapy remains a central component of multidisciplinary management for primary CNS tumors. However, recurrence within a previously irradiated field poses a major therapeutic challenge, and re-irradiation (re-RT) was historically avoided due to concerns regarding cumulative toxicity. Advances in conformal and image-guided techniques, together with a growing understanding of partial normal-tissue recovery, have progressively expanded the feasibility of re-RT in carefully selected patients.

Recent European data reflect increasing clinical adoption of this strategy but also highlight substantial practice heterogeneity. Re-RT is most frequently applied in high-grade gliomas, with stereotactic and hypofractionated approaches gaining preference over conventional techniques. Nonetheless, considerable variability persists in target delineation, margin selection, cumulative dose constraints, and organ-at-risk recovery assumptions. Concerns regarding radiation necrosis and the absence of harmonized guidelines remain key barriers to broader implementation [1].

These findings confirm that, while reirradiation has gained acceptance as a salvage strategy, significant variability persists regarding dose, target definition, and systemic combination approaches. Harmonization efforts through ESTRO, EANO, and ISRS are thus critical to define evidence-based dose constraints, integrate molecular prognostic factors, and optimize patient selection. The convergence of technological precision, biological stratification, and targeted radiosensitization is expected to transform reirradiation into a core component of precision neuro-oncology.

To address the clinical variability outlined above, this review proceeds by first examining the radiobiological rationale underlying re-irradiation and normal-tissue recovery. We then discuss patient selection frameworks integrating clinical, molecular, and imaging determinants of benefit and toxicity. Subsequent sections summarize contemporary treatment planning strategies, dose–fractionation paradigms, and multimodal combinations with targeted and immune-based therapies. Finally, we explore emerging technologies and applications across other CNS tumor entities, before outlining future research priorities and directions for standardization in precision re-irradiation.

Literature search strategy

To enhance transparency and comprehensiveness, a structured literature search was conducted across multiple electronic databases, including PubMed/MEDLINE, Embase, Scopus, Web of Science (WoS), and the Cochrane Library. The search covered studies published from January 1995 through March 2025, reflecting both foundational radiobiological work and contemporary advances in precision re-irradiation.

Search terms included combinations of the following keywords and Medical Subject Headings (MeSH): “re-irradiation,” “reirradiation,” “glioma,” “glioblastoma,” “radiation necrosis,” “pseudo-progression,” “perfusion MRI,” “rCBV,” “amino acid PET,” “FET,” “MET,” “DOPA,” “quality of life,” “neurocognition,” “proton therapy,” “MR-guided radiotherapy,” and “salvage radiotherapy.” Boolean operators (AND/OR) were applied to optimize sensitivity and specificity.

Eligible studies included prospective and retrospective clinical studies, randomized trials, meta-analyses, and relevant translational investigations published in English. Reference lists of key articles and recent consensus guidelines (ESTRO, EANO, ISRS) were manually screened to identify additional relevant publications. Conference abstracts were included when peer-reviewed full publications were not yet available but provided clinically meaningful data.

Given the narrative nature of this review, formal PRISMA methodology and quantitative meta-analysis were not performed; however, study selection prioritized methodological rigor, sample size, recency, and clinical relevance to re-RT practice.

Radiobiological rationale and normal-tissue repair

The feasibility of reirradiation of the brain is supported by both mechanistic and clinical evidence demonstrating partial recovery of normal tissue tolerance over time. Preclinical and translational studies demonstrate that irradiation induces dynamic responses in the neurovascular unit, with endothelial dysfunction and blood–brain barrier disruption followed by reparative remodelling extending over months, and astrocytic activation and structural remodelling persisting out to ~ 6–12 months post-irradiation in animal models, providing a biologically plausible temporal window for safe re-treatment approaches [2–4]. Ionizing radiation induces early endothelial apoptosis and microvascular rarefaction, resulting in blood–brain barrier disruption and tissue hypoxia; this hypoxic microenvironment promotes sustained activation of angiogenic pathways mediated by HIF-1α and VEGF, contributing to progressive radiation-induced brain injury. [5–7]. Concurrently, radiation-induced activation of microglia and astrocytes sustains chronic inflammatory signaling through tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β) pathways, which plays a critical role in the progressive evolution of radiation-induced necrosis. [8, 9] Figure-(1). Importantly, dose fractionation modulates these biological responses by limiting endothelial apoptosis and microvascular collapse while attenuating sustained inflammatory signaling, thereby providing a mechanistic rationale for the relatively lower risk of radiation necrosis observed with hypofractionated stereotactic regimens compared with single-fraction high-dose treatments. [10, 11].

Fig. 1.

Radiobiological mechanisms underlying Re-RT–associated brain injury

Clinically, the cumulative tolerance of normal brain tissue remains a central determinant of safety. Mayer et al. reported no radionecrosis when the cumulative equivalent dose (EQD₂, α/β = 2) remained below 100 Gy in patients receiving a second course of radiotherapy [12]. Size and dose effects were demonstrated in RTOG 90–05, where grade ≥ 3 toxicity occurred in 10% of tumors ≤ 2 cm treated with ≥ 24 Gy as a single fraction, 14% for 2.1–3.0 cm tumors treated with 18 Gy, and 20% for 3.1–4.0 cm tumors treated with 15 Gy [13]. Dose–volume metrics are equally critical: a V12 < 10 cm³ strongly predicts lower necrosis risk, while lesion location significantly influences toxicity, remaining < 5% in lobar locations but rising to ~ 20% in radiosensitive regions such as the brainstem, thalamus, and basal ganglia [14]. Hypofractionated data also indicate that larger tumor volumes and higher total doses markedly increase risk; radionecrosis rose from 23 to 36% when escalating from 20 Gy/4 fractions to 50 Gy/5 fractions, particularly for tumor volumes > 35 mL [15, 16]. Based on pooled toxicity estimates, Scoccianti et al. proposed dose–volume–fractionation strategies designed to maintain serious toxicities below 3.5%: 12–15 Gy SRS (EQD₂ < 65 Gy) for lesions ≤ 12.5 mL; 25 Gy/5 fractions (EQD₂ < 50 Gy) for 12.5–35 mL; and 36 Gy/18 fractions (EQD₂ < 36 Gy) for 35–50 mL [17].

These principles are consistent with contemporary recommendations from the 2025 International Stereotactic Radiosurgery Society (ISRS) guideline, which analyzed 62 studies (n = 2640) and supports focal reirradiation using either single-fraction SRS (16 Gy × 1) or hypofractionated SRS (24–25 Gy in 3–5 fractions) in appropriately selected recurrent high-grade glioma patients. The guideline further highlights that neurologic toxicity remains low (4–7%) when cumulative EQD₂ is contained below ~ 120–130 Gy, beyond which radionecrosis risk rises significantly [18]. Collectively, radiobiological mechanisms and clinical dose–volume data converge on a coherent conclusion: reirradiation of the CNS is technically feasible and biologically justified when adequate time has elapsed since initial treatment, lesion volume and anatomic location are carefully considered, and cumulative and fractionated doses are maintained within established safety thresholds.

While cumulative EQD₂ thresholds (e.g., 120–130 Gy) are frequently cited as reference ranges associated with increased radionecrosis risk Figure 2, these values should be interpreted with caution. Cumulative dose estimation relies on assumptions regarding the α/β ratio for late-responding brain tissue and on mathematical models of partial normal tissue recovery, both of which vary across institutions and treatment planning systems. Importantly, true normal tissue recovery kinetics after high-dose irradiation remain incompletely defined, and recovery may depend on factors such as time interval, irradiated volume, and vascular integrity. Therefore, reported EQD₂ thresholds should be considered pragmatic guidance rather than rigid universal limits, and individualized clinical judgment remains essential in re-irradiation planning.

Fig. 2.

Clinical decision flowchart for re-irradiation in recurrent CNS tumors

Schematic illustration of the key biological processes contributing to radiation-induced brain injury and cumulative toxicity following re-irradiation. Early endothelial damage results in blood–brain barrier disruption and microvascular rarefaction, leading to tissue hypoxia. This hypoxic microenvironment promotes angiogenic signaling via HIF-1α and VEGF pathways, contributing to abnormal vascular remodeling and edema Figure 3. Concurrently, persistent activation of microglia and astrocytes drives chronic neuroinflammation mediated by cytokines such as TNF-α and IL-1β. These processes culminate in oxidative stress, demyelination, and progressive tissue necrosis, clinically manifesting as radiation necrosis and neurological decline. Arrows indicate temporal progression from acute vascular injury to late-stage necrotic changes, highlighting the biological rationale for dose fractionation and cautious cumulative dose escalation in re-irradiation.

Fig. 3.

Graphical summary of disease-specific re-irradiation strategies for brain metastases, intracranial meningiomas, and pediatric/adult ependymomas, integrating clinical outcomes, toxicity risks, and practical dose–volume constraints. The figure highlights modality selection according to tumor histology, anatomical location, and prior treatment exposure. Key parameters include cumulative EQD₂ thresholds, normal brain V12 and V20 limits, and organ-at-risk tolerance for critical structures such as the brainstem and optic pathways. For brain metastases, focal approaches (SRS/FSRT) are emphasized to preserve neurocognitive function. In meningioma, treatment selection is guided by tumor grade, proliferative index, and proximity to critical neurovascular structures. In ependymoma, the role of maximal safe resection followed by conformal or proton-based re-irradiation is highlighted. The figure underscores the importance of individualized treatment planning integrating clinical, dosimetric, and biological factors to optimize therapeutic ratio

Patient selection: clinical and molecular decision tools

Re-RT is most frequently utilized in recurrent high-grade gliomas, and appropriate patient selection is paramount to treatment success. Several clinical, biological, and radiographic parameters have been investigated to define the subgroup most likely to benefit from a second course of radiotherapy.

In the pivotal series by Combs et al., including 233 patients (38% glioblastoma, 22% anaplastic astrocytoma, 40% low-gradeglioma), multivariate analysis identified age, histology, and interval between primary and salvage radiation as significant prognostic factors, which were subsequently incorporated into a validated prognostic scoring system [19]. However, inclusion of 40% low-grade gliomas with a median recurrence interval of 57 months limits the generalizability of the interval parameter across different histologies [19]. In the 2018 multicenter update by Combs et al., both the original and modified prognostic scores demonstrated robust stratification of overall survival (P < 0.001), with median survival ranging from ~ 6 months in unfavorable risk groups to over 13 months in favorable categories. Notably, performance status emerged as a dominant independent predictor of outcome, underscoring its clinical relevance in patient selection for re-RT [20]. Kessel et al. refined this stratification in a cohort of 209 patients, demonstrating that re-resection, planning target volume (≤ 47 mL vs > 47 mL), and Karnofsky performance score (KPS < 80 vs ≥ 80) were also independent predictors of survival [21]. Across pooled re-RT cohorts, larger tumor size, multifocality, eloquent-area involvement, and corticosteroid dependence emerge as reproducible markers of unfavorable prognosis. [20, 22, 23].

Dose–response modeling suggests that achieving a biological effective dose (BED) ≥ 40 Gy with stereotactic radiosurgery (SRS) or ≥ 45 Gy with fractionated stereotactic radiotherapy (FSRT) is associated with improved progression-free survival, extending median PFS from ~ 2 to 6 months [24]. Across modern re-irradiation strategies—including SRS, hypofractionated or conventionally fractionated stereotactic radiotherapy, and proton therapy (Table 1)—median overall survival (OS) in recurrent glioblastoma typically ranges from 7.6 to 13 months, with reported late toxicity rates between 0 and 36%.

Table 1.

Comparative characteristics of re-irradiation modalities in recurrent CNS tumors

Modality	Indications	Advantages	Limitations	Preferred clinical scenario
Stereotactic Radiosurgery (SRS)	Small-volume lesions (≤ 2–3 cm), well-demarcated tumors, non-eloquent locations	High dose conformity, steep dose gradient, short treatment duration (single fraction), excellent local control	Increased risk of radionecrosis in larger volumes or critical locations, limited suitability for eloquent regions	Solitary, small, well-circumscribed recurrences distant from critical structures
Fractionated Stereotactic Radiotherapy (FSRT)	Medium to large lesions (> 2–3 cm), lesions adjacent to critical structures, or previously irradiated regions requiring dose moderation	Improved normal tissue sparing through fractionation, reduced radionecrosis risk, safer for eloquent or deep-seated lesions	Longer treatment course, slightly lower conformity compared with single-fraction SRS	Lesions near brainstem, optic apparatus, or eloquent cortex; larger or irregularly shaped recurrences
Proton Therapy	Re-irradiation in the setting of high cumulative dose, large target volumes, pediatric patients, or anatomically complex cases	Reduced integral dose, superior normal tissue sparing beyond target (Bragg peak), potential reduction in cumulative toxicity	Limited availability, high cost, uncertainties in relative biological effectiveness (RBE), limited comparative prospective data	Large or previously heavily irradiated volumes, pediatric or long-survival patients, cases requiring maximal normal tissue preservation

This variability in survival outcomes is primarily driven by differences in patient selection and disease characteristics rather than treatment modality alone. Inferior outcomes are generally observed in patients with poor performance status, larger tumor volumes, shorter intervals from prior radiotherapy, and unfavorable molecular profiles, whereas improved survival is seen in highly selected patients with limited disease burden, smaller targets, longer treatment intervals (> 12 months), and favorable biology (e.g., MGMT methylation or IDH mutation). Consequently, reported survival outcomes should be interpreted within the context of baseline prognostic factors rather than as directly comparable results across studies.

Interpretation is further complicated by substantial heterogeneity in the available evidence. Selection bias is inherent, as re-irradiation candidates typically represent a favorable subgroup. In addition, variability in treatment protocols—including differences in target delineation (GTV vs GTV + FLAIR), dose and fractionation schemes, imaging integration, and use of systemic therapies—limits comparability. The evidence base is also predominantly composed of retrospective, single-institution series, with limited prospective validation and very few randomized data, thereby restricting the strength of causal inference.

When stratified by treatment modality, outcomes from hypofractionated and fractionated stereotactic radiotherapy (FSRT/HSRT) are largely derived from retrospective cohorts, demonstrating median OS of ~ 9–12 months, with outcomes influenced by tumor volume, dose selection, and treatment interval [25–31]. The addition of systemic therapies, particularly temozolomide or bevacizumab, may improve disease control, although these findings remain non-randomized [27, 32]. Larger multi-institutional datasets further support the prognostic significance of performance status, tumor burden, and interval from prior radiotherapy in shaping clinical outcomes [33–35].

In contrast, evidence for proton therapy re-irradiation is derived from both retrospective cohorts and prospective registry-based analyses, which suggest comparable or slightly improved survival outcomes (median OS ~ 10–12 months) with favorable toxicity profiles attributable to reduced integral dose, although definitive superiority over photon-based techniques has not been established [36, 37].

Importantly, systematic reviews and narrative syntheses highlight that cross-study comparisons remain unreliable due to the absence of standardized protocols and the predominance of heterogeneous retrospective data [38, 39]. Collectively, these limitations underscore that reported survival outcomes should be interpreted as descriptive benchmarks rather than evidence of modality equivalence, and emphasize the need for prospective, biomarker-driven trials to refine patient selection and optimize treatment strategies.

Contemporary evidence from the 2025 ISRS Practice Guideline (meta-analysis of 62 studies; n = 2640) emphasizes that optimal candidates typically have KPS ≥ 60–70, controlled edema, and limited-volume recurrence; single-fraction 16 Gy or 24–25 Gy in 3–5 fractions is recommended depending on proximity to critical structures, with cumulative dose constraints (EQD₂ > 120–130 Gy) strongly associated with radionecrosis risk [18].

Beyond clinical metrics, emerging data demonstrate that molecular profiles significantly influence outcomes of re-RT: IDH-mutant tumors exhibit more favorable responses and prolonged survival, MGMT promoter methylation confers better prognostic trajectory though not clearly predictive of radiosensitivity, while TERT promoter mutation correlates with aggressive recurrence and poorer outcomes [40–43]. Ongoing efforts incorporating genomic signatures, radiogenomic profiling, and advanced metabolic imaging (amino acid PET, perfusion-weighted MRI) aim to refine decision-making by distinguishing true recurrence from treatment effect and by identifying biologically plausible responders [44–47]. Taken together, effective patient selection for re-RT increasingly depends on integrating established clinical scoring systems with molecular biomarkers, neurofunctional status, and high-resolution imaging to guide personalized decision-making.

To date, commonly reported molecular alterations such as IDH mutation, MGMT promoter methylation, and TERT promoter mutation primarily provide prognostic information rather than validated predictive value for response to re-irradiation. While MGMT methylation may correlate with improved overall outcomes in recurrent glioma, particularly in the context of alkylating chemotherapy, its role as a predictor of radiosensitivity in the re-irradiation setting remains unproven [40]. Similarly, IDH mutation is consistently associated with more favorable disease biology but has not been validated as a biomarker guiding re-irradiation dose selection or modality choice.

Although molecular biomarkers such as IDH mutation, MGMT promoter methylation, and TERT promoter mutation are increasingly incorporated into the management of gliomas, their role in the re-irradiation setting remains primarily prognostic rather than predictive. IDH mutation and MGMT methylation are consistently associated with improved overall outcomes, whereas TERT promoter mutations correlate with more aggressive disease biology; however, none of these markers have been prospectively validated to guide re-RT selection, dose, or modality. Consequently, treatment decisions in re-RT continue to rely predominantly on clinical, volumetric, and dosimetric factors, underscoring the need for biomarker-driven prospective studies [40–43].

Table 2 summarizes patient selection determinants, including clinical performance status, tumor characteristics, molecular markers, and treatment interval parameters that influence the likelihood of clinical benefit from re-RT.

Table 2.

Clinical and molecular determinants of patient selection for Re-RT in recurrent high-grade gliomas

Category	Variable	Favorable Prognostic Group	Unfavorable Prognostic Group	Evidence/Comments
Clinical performance	KPS	≥ 80	< 80	Consistent predictor of OS
Age	Chronological age	< 50 years	≥ 50 years	Associated with survival in multivariate models
Neurologic status	Symptomatic burden	Mild/no symptoms	Progressive focal deficits	Higher toxicity risk and poorer outcomes
Steroid requirement	Corticosteroid use	None/minimal	Chronic or high-dose steroids	Strong negative prognostic factor
Tumor volume	Gross tumor volume or PTV	≤ 47 mL	> 47 mL	Larger volumes → increased necrosis risk
Lesion number	Solitary vs multifocal	Single lesion	Multifocal disease	Multifocality worsens survival
Tumor location	Proximity to eloquent/critical structures	Non-eloquent	Critical regions (brainstem/thalamus)	Limits safe dose; risk of necrosis ↑
Time interval from prior RT	Interval between RT courses	≥ 12 months	< 12 months	Integrated in Combs and Kessel scoring
Histology	WHO grade/type	AA, LGG recurrence	GBM recurrence	Histology remains strong determinant
Previous surgery	Cytoreduction before re-RT	Repeat resection performed	No viable surgical option	Survival improved with re-resection
Dose–effect	Biological effective dose (BED)	BED ≥ 40 Gy (SRS) or ≥ 45 Gy (SRT)	BED below thresholds	Improves PFS from 2 to 6 months
Re-RT technique	Modality	SRS/SRT/Proton	Conventional 3D	Higher control rates with more precise techniques
Molecular markers	IDH mutation	Mutant	Wild-type	Predicts prolonged survival after re-RT
	MGMT methylation	Methylated	Unmethylated	Prognostic but not radiosensitivity-predictive
	TERT promoter mutation	Wild-type	Mutant	Associated with aggressive recurrence
Imaging biomarkers	MRI/PET parameters	Amino acid PET +/rCBV	MRI-only progression	Improves lesion delineation
Guideline-based risk	ISRS 2025	EQD2 ≤ 120 Gy	EQD2 > 130 Gy	Necrosis risk increases beyond threshold

KPS Karnofsky performance score, AA Anaplastic astrocytoma, LGG Low-grade glioma, GBM Glioblastoma, re-RT Re-irradiation, SRS Stereotactic radiosurgery, SRT Stereotactic radiotherapy

While prior systematic reviews established the safety and modest survival benefit of re-irradiation, they largely reflect a pre-molecular era characterized by anatomical target definition and uniform fractionation strategies. The current generation of precision re-irradiation differs fundamentally by incorporating molecular stratification, advanced functional imaging, and biologically rational systemic combinations. Thus, direct survival comparisons across eras must be interpreted cautiously, as improved patient selection rather than dose intensification alone may account for observed outcome gains.

Clinical decision flowchart outlining a structured, multidisciplinary approach to patient selection and treatment planning for re-irradiation. Initial stratification is based on performance status, with patients having Karnofsky Performance Status (KPS) ≥ 70–80 considered candidates for active treatment, while those with lower KPS are typically managed with supportive care. Subsequent decision nodes integrate tumor volume, anatomical location, interval from prior radiotherapy, and cumulative dose constraints. Small, well-demarcated lesions (< 2 cm) in non-eloquent regions are preferentially treated with single-fraction stereotactic radiosurgery (SRS), whereas intermediate-sized lesions (2–4 cm) or tumors adjacent to critical structures favor fractionated stereotactic radiotherapy (FSRT). Larger tumors (> 4 cm), previously heavily irradiated regions, or cases requiring maximal normal tissue sparing may be considered for FSRT or proton therapy. Additional constraints include maintaining cumulative EQD₂ ≤ 120 Gy and limiting normal brain V12 to < 10 cm³ to mitigate radionecrosis risk. Final treatment decisions are made within a multidisciplinary framework, incorporating surgical options and systemic therapies, with the goal of optimizing tumor control while preserving neurological function and quality of life.

Treatment planning and dose–fractionation strategies

Treatment planning for re-RT of central nervous system tumors remains heterogeneous, and no universal consensus exists regarding optimal target delineation. In most contemporary stereotactic re-RT series, target volumes are limited to the contrast-enhancing tumor on T1-weighted MRI, with no additional clinical target volume margin, effectively resulting in a GTV equal to the CTV [19, 22, 48]. Depending on tumor location and proximity to critical structures, planning target volume (PTV) margins of 1–5 mm are applied to accommodate setup variability and motion uncertainty.

While conventional GTV delineation relies primarily on post-contrast T1-weighted MRI, biologically informative imaging modalities are increasingly incorporated into target definition. Perfusion-weighted MRI (rCBV), diffusion-weighted imaging (ADC), and MR spectroscopy improve discrimination of metabolically active tumor from therapy-related changes. Molecular imaging with amino acid PET tracers, including ¹⁸F-DOPA and ¹¹C-methionine, enables visualization of infiltrative tumor extensions not apparent on conventional MRI, whereas SPECT imaging may assist in differentiating tumor recurrence from radiation necrosis [49–53]. Although a simplified GTV-only approach is common, selective CTV expansion to include FLAIR abnormalities has been explored in diffusely infiltrative recurrences. In a retrospective series of 53 patients with recurrent glioblastoma treated with fractionated radiosurgery, Kim et al. compared 32 Gy/4 fractions to contrast enhancement alone versus simultaneous differential dosing of 32 Gy/4 fractions to enhancement and 24 Gy/4 fractions to the FLAIR abnormality; inclusion of the FLAIR volume was associated with a lower local/regional failure rate (52% vs 86.7%; p = 0.003), albeit with higher distant failure, suggesting a potential role for extended-volume planning in carefully selected cases. [54].

Dose–fractionation strategies vary widely by institutional preference, tumor volume, and anatomic constraints. The literature demonstrates that hypofractionated stereotactic approaches offer superior conformity and normal-tissue sparing relative to conventional fractionation. Table 3 summarizes representative contemporary outcomes for recurrent GBM across commonly utilized techniques.

Table 3.

Representative outcomes of Re-RT in recurrent high-grade gliomas

Study	Year	n	Technique	Dose/fractionation	Median OS (months)
Fogh et al. [48]	2010	147	FSRT	35 Gy/10 fx	11.0
Yazici et al. [55]	2014	37	FSRT	30 Gy/5 fx	10.6
Dincoglan et al. [16]	2015	28	FSRT	25 Gy/5 fx	10.3
Imber et al. [56]	2017	174	SRS	16 Gy × 1	10.6
Saeed et al. [36]	2020	45	Proton RT	46.2 Gy/23 fx	14.2
NRG/RTOG 12–05 [57]	2022	170	HF-SRT	35 Gy/10 fx	10.1
ISRS Practice Guideline [18]	2025	2640	SRS/HF-SRT (meta-analysis)	16 Gy × 1 or 25 Gy/5 fx	10.2

OS Overall survival, FSRT Fractionated stereotactic radiotherapy, SRS Stereotactic radiosurgery, HF-SRT Hypofractionated stereotactic radiotherapy

It is important to emphasize that survival outcomes reported across different re-RT modalities should not be interpreted as directly comparable. The available data derive predominantly from heterogeneous retrospective series with substantial variability in tumor volume, performance status, molecular characteristics, interval from prior radiotherapy, and treatment intent. Modality selection (e.g., SRS, hypofractionated SRT, conventionally fractionated RT, or proton therapy) is typically driven by these patient- and disease-specific factors, which inherently confound cross-modality comparisons. Therefore, pooled survival ranges should be viewed as descriptive benchmarks rather than evidence of therapeutic equivalence (Table 4).

Table 4.

Consolidated dose–volume constraints and toxicity risk parameters in CNS Re-irradiation

Parameter	Threshold/Range	Clinical relevance	Implication for practice
Cumulative EQD₂ (α/β = 2)	≤ 100–110 Gy (preferred) 120–130 Gy (upper limit)	Strong predictor of radionecrosis risk	Maintain cumulative EQD₂ ≤ 110–120 Gy when feasible; caution above 120 Gy
Single-fraction dose (SRS)	12–16 Gy (depending on size/location)	Higher doses increase necrosis risk, especially in larger lesions	Use lower doses for lesions > 2–3 cm or near critical structures
Hypofractionated dose (FSRT)	24–30 Gy in 3–5 fractions	Improved safety via fractionation	Preferred for larger lesions or eloquent locations
Brain V12 (SRS)	< 10 cm3 (optimal) > 10–15 cm3 (increased risk)	Strongest predictor of symptomatic radionecrosis	Limit high-dose volume; avoid SRS if V12 exceeds threshold
Brain V20 (Hypofractionation)	< 20 cm3	Associated with acceptable toxicity in FSRT	Use as planning constraint for hypofractionated regimens
Lesion volume	≤ 12.5 mL (ideal for SRS) > 35 mL (high risk)	Larger volumes correlate with toxicity and radionecrosis	Prefer FSRT or proton therapy for larger lesions
Time interval between RT courses	≥ 12 months (preferred)	Reflects partial normal tissue recovery	Delay re-RT when feasible to reduce toxicity
Brainstem (cumulative EQD₂)	< 100 Gy	High sensitivity to radiation injury	Strict constraints required; consider FSRT/proton
Optic pathways (cumulative EQD₂)	< 75 Gy	Risk of optic neuropathy	Avoid high-dose SRS near optic apparatus
Whole brain (cumulative EQD₂)	≤ 100–120 Gy	Global toxicity and neurocognitive risk	Important in prior WBRT settings

A major practical challenge in the re-RT setting is the distinction between true tumor progression and treatment-related effects, including radiation necrosis and pseudo-progression. Conventional contrast-enhanced MRI alone is frequently insufficient, particularly in previously irradiated tissue [58, 59]. Therefore, integration of advanced imaging modalities into treatment planning is strongly recommended.

Dynamic susceptibility contrast (DSC) perfusion MRI provides quantitative assessment of relative cerebral blood volume (rCBV), which has emerged as a validated biomarker for distinguishing viable tumor from treatment effect [60–62]. Recurrent tumor typically demonstrates elevated rCBV due to neovascularization, whereas radiation necrosis is characterized by reduced perfusion. Several studies suggest that an rCBV ratio > 1.75–2.0 (lesion-to-contralateral white matter) favors tumor recurrence, whereas values < 1.0–1.2 are more consistent with treatment-related change [60, 61]. Thresholds may vary depending on acquisition protocol and post-processing technique; therefore, institutional standardization is essential. Importantly, focal regions of high rCBV within an enhancing lesion may guide dose escalation or target refinement in re-RT planning.

Amino acid PET imaging further enhances diagnostic accuracy in this context. Tracers such as ^18F-FET, ^11C-MET, and ^18F-DOPA have demonstrated superior sensitivity and specificity compared with conventional MRI for differentiating recurrence from radiation necrosis [63–65]. Tumor recurrence typically exhibits increased tracer uptake with tumor-to-background ratios (TBRmax) ≥ 1.6–2.0, whereas treatment-related changes show low or stable uptake [63, 64]. Dynamic FET-PET analysis, including time–activity curve patterns, may further improve discrimination, with early peak and washout kinetics favoring recurrent tumor [64]. Integration of amino acid PET into target delineation can reduce geographic miss and minimize unnecessary irradiation of necrotic tissue, thereby potentially improving the therapeutic ratio [65].

In clinical practice, a multimodal approach combining structural MRI, perfusion imaging (rCBV mapping), diffusion-weighted imaging, and amino acid PET provides the highest diagnostic confidence [59, 65]. When imaging findings remain equivocal, multidisciplinary discussion and, in selected cases, histopathologic confirmation should be considered prior to high-dose re-RT.

Targeted therapies and radiosensitizers in RE-RT

Re-RT is increasingly combined with systemic therapies to improve disease control in recurrent high-grade gliomas. Among cytotoxic agents, temozolomide (TMZ) is the most frequently used radiosensitizer in this setting; however, the available evidence largely derives from single-arm clinical studies and institutional cohorts rather than randomized comparisons. High-precision FSRT/HSRT with concurrent TMZ has been reported to yield median overall survival after re-RT in the range of ~ 8–10 months, with median PFS around ~ 5 months in recurrent glioblastoma series, and generally acceptable acute toxicity profiles [27, 66–68].

Collectively, these outcomes appear broadly comparable to those reported with re-RT alone in similarly selected patients, and any incremental benefit of TMZ must be weighed against hematologic toxicity risk, which—although often manageable in published cohorts—remains clinically relevant in heavily pretreated populations [69]. Accordingly, TMZ is best viewed as a selective systemic partner rather than a universal adjunct, with potential enrichment in patients with favorable prognostic/molecular features (e.g., MGMT promoter methylation reported as prognostic in some re-RT cohorts) [27].

Bevacizumab (BEV), a monoclonal VEGF-A inhibitor, remains the most widely investigated targeted agent combined with re-RT. Early retrospective studies suggested improved survival and enhanced tolerance by mitigating edema and symptomatic radionecrosis. In the series by Niyazi et al., adding BEV to re-RT increased median overall survival from 6 to 12 months and median PFS from 5 to 8 months [70]. Cabrera et al. reported that a biologically informed approach using lesion-size–adapted regimens (24 Gy single-fraction SRS + BEV for < 2 cm lesions; 18 Gy SRS + BEV for 2–2.9 cm lesions; and 25 Gy in 5 fractions + BEV for 3–5 cm lesions) resulted in 14.4 months median overall survival and PFS of 4 months [71]. The Phase II RTOG 1205 randomized 170 recurrent GBM patients ≥ 6 months from primary chemoradiation to re-RT (35 Gy in 10 fractions) plus BEV versus BEV alone. Updated results showed no significant improvement in overall survival (10.1 vs 9.7 months, HR = 0.98; p = 0.46), but a clinically meaningful prolongation in PFS (7.1 vs 3.8 months; HR = 0.73; p = 0.05) and significantly higher 6-month PFS rates (54.3% vs 29.1%; p = 0.001), with very low rates of grade ≥ 3 CNS toxicity and no delayed high-grade adverse events [57]. Thus, although not life-prolonging, re-RT + BEV offers substantial benefit in radiographic stability, edema control, and neurological symptom management. Potential toxicities warrant monitoring, including wound-healing complications, hemorrhage, hypertension, thromboembolic events, and cytopenias.

Multimodal treatment strategies have gained interest due to synergistic biological effects. In a secondary analysis of the RTOG 0525 dataset (637 patients), salvage outcomes after recurrence were compared among best supportive care, chemotherapy alone, re-RT alone, and re-RT plus systemic therapy. The longest median overall survival was observed in the combined re-RT + systemic therapy group (12.2 months), compared with chemotherapy alone (10.6 months), re-RT alone (8.2 months), and supportive care (4.8 months). Although only 14% received re-RT and randomization was not designed to address this question, these findings support the principle that re-RT provides maximal benefit when embedded within combined modality treatment rather than used in isolation [72].

Reported survival outcomes in this section (median OS ranging from ~ 8 to 12.2 months) reflect heterogeneous patient populations and treatment approaches. Lower survival estimates (8–10 months) are generally derived from unselected or heavily pretreated cohorts, whereas higher values (10.6–12.2 months) are typically observed in studies including selected patients treated with stereotactic or hypofractionated techniques, often combined with systemic therapies such as temozolomide or bevacizumab. Differences in histology, tumor burden, and prior treatment exposure further contribute to variability across studies, limiting direct comparisons.

Surgical resection followed by re-RT represents an important combined salvage strategy in recurrent high-grade glioma. In a cohort of 84 patients, Chun et al. demonstrated that re-operation followed by re-RT was associated with a significantly longer progression-free survival compared with surgery alone (9.0 vs 3.5 months; p = 0.025), with a strong trend toward improved overall survival (28.1 vs 12.7 months; p = 0.066). Advanced age (> 50 years), WHO grade IV histology, and unmethylated MGMT promoter status were identified as adverse prognostic factors, and the survival benefit of combined treatment was most pronounced in lower-risk patients harboring fewer adverse features. Importantly, no cases of radionecrosis were observed among patients with cumulative EQD₂ (α/β = 2) below ~ 99 Gy [73]. The ongoing phase II NOA-17 (GlioCave) trial, comparing postoperative re-RT doses of 46 Gy in 23 fractions versus 36 Gy in 12 fractions to the resection cavity, is expected to further refine optimal dose–fractionation strategies in this setting [74].

Beyond TMZ and BEV, expanding efforts have explored multiple classes of molecularly targeted agents. EGFR tyrosine kinase inhibitors (erlotinib, afatinib, osimertinib) [75–77] and VEGFR/multikinase inhibitors (regorafenib) [78] have been investigated in recurrent glioblastoma; however, randomized Phase III trials of EGFR-targeted therapies in unselected GBM populations have failed to demonstrate a survival benefit. Accordingly, current evidence does not support routine use of EGFR inhibition in unselected recurrent GBM outside biomarker-driven or investigational contexts. PARP inhibitors (olaparib, veliparib) exploit DNA repair vulnerabilities and may enhance radiosensitivity, with early-phase studies demonstrating acceptable toxicity profiles and preliminary signals of metabolic or radiographic response in biomarker-selected tumors [79–81], although definitive survival benefit remains unproven.

Similarly, immune checkpoint inhibitors (nivolumab, pembrolizumab) have not improved overall survival in unselected recurrent GBM, with randomized data showing no advantage of PD-1 blockade compared with standard salvage approaches [82–84]. As a result, interest has shifted toward rational combinatorial strategies, including integration with hypofractionated re-RT, based on the capacity of radiation to induce immunogenic cell death and modulate the tumor microenvironment. Across these approaches, an emerging priority is biologic stratification: integration of molecular drivers (IDH mutation, MGMT methylation, EGFR amplification/EGFRvIII, TERT promoter mutation) with advanced imaging biomarkers (amino acid PET, MR perfusion) to identify patients most likely to benefit, while maintaining strict cumulative dose constraints to mitigate radionecrosis risk [85–88].

In summary, targeted and systemic agents represent increasingly utilized adjuncts to re-RT for recurrent high-grade gliomas. Current evidence supports bevacizumab-based combinations primarily for progression-free survival benefit and symptom control, while rational combinations involving DNA repair inhibitors, anti-angiogenic agents, and immunomodulatory strategies may enhance radiosensitivity within molecularly selected subgroups. Integration of genomic classification, metabolic imaging, and conformal re-RT is expected to shape future treatment paradigms, and prospective biomarker-stratified trials will be essential to identify the patient populations most likely to derive meaningful survival benefit.

RE-RT in other CNS tumors

Although recurrent gliomas comprise the majority of re-RT scenarios, secondary CNS tumors including brain metastases, intracranial meningiomas, and pediatric ependymomas also benefit from carefully selected re-RT strategies. The increasing survival of cancer patients, driven by systemic therapy advances, has led to a growing need for late-course salvage radiotherapy in previously irradiated regions.

Brain metastases

Brain metastases constitute the most common non-glial indication for intracranial re-RT. Historically, retreatment was predominantly performed with whole-brain radiotherapy (WBRT), yielding symptomatic improvement in approximately one-third to two-thirds of patients but conferring limited survival, typically on the order of 3–5 months, and a substantial risk of cumulative neurocognitive toxicity [89, 90]. In contemporary practice, focal techniques such as stereotactic radiosurgery (SRS) and fractionated stereotactic radiotherapy (FSRT) have largely replaced WBRT in the salvage setting, owing to improved local control, avoidance of diffuse brain irradiation, and more favorable neurocognitive profiles. Across pooled series, salvage SRS or FSRT has been associated with 1-year local control rates of ~ 60–88% and median overall survival ranging from 8 to over 20 months, with symptomatic radionecrosis reported in 4–23% of cases and consistently lower toxicity observed with fractionated approaches. Widely accepted cumulative tolerance thresholds include EQD₂ values of approximately ≤ 100–120 Gy for the whole brain, < 100 Gy for the brainstem, and < 75 Gy for the optic apparatus [91, 92].

Radionecrosis risk is strongly associated with cumulative dose–volume parameters, particularly V12 of normal brain, with symptomatic radionecrosis rates increasing substantially when cumulative V12 exceeds ~ 10–15 cc [10, 93, 94]. In hypofractionated approaches, maintaining brain V20 below ~ 20 cc has been associated with acceptable toxicity profiles [94]. Importantly, extracranial disease control and systemic molecular drivers (e.g., EGFR, ALK, BRAF mutations) significantly influence overall prognosis and should inform intracranial retreatment decisions [94]. Integration of advanced imaging modalities is critical to distinguish radionecrosis from true progression, particularly in patients receiving concurrent immunotherapy, which may increase treatment-related imaging changes [94].

The posterior fossa represents a particularly high-risk compartment for salvage irradiation, where limited intracranial reserve, mass effect, and hydrocephalus frequently complicate management. In a recent compartment-specific analysis focusing on cerebellar metastases, stereotactic re-RT achieved radiographic response in more than 90% of treated lesions and symptomatic improvement in ~ 75% of patients, with a low incidence of symptomatic radionecrosis (~ 1–2%); six-month overall survival approached 50% [95]. In parallel, outcomes from large retrospective series evaluating salvage stereotactic radiosurgery after prior whole-brain radiotherapy indicate that younger age and longer intervals between treatment courses are associated with improved outcomes. Early re-salvage SRS following WBRT has been associated with 1-year local control rates of ~ 68–75% and median overall survival of around 10 months, with asymptomatic radionecrosis reported in roughly 5–12% of patients. Collectively, these data support a paradigm favoring focal, region-directed salvage irradiation over repeat whole-brain exposure when anatomically and clinically feasible [96, 97].

In the setting of recurrent brain metastases, repeat stereotactic radiosurgery (SRS) has emerged as an effective salvage strategy in carefully selected patients. In a large multi-institutional analysis, repeat SRS achieved favorable local control rates of 79% at 1 year and 72% at 2 years, with symptomatic radiation necrosis observed in only 7% of cases, particularly when dosimetric constraints such as V12 ≤ 9 cm³ were respected [98]. Similarly, single-institution data demonstrate that tumor control following repeat SRS is strongly influenced by lesion volume and delivered dose, with improved outcomes observed in smaller tumors and with prescription doses ≥ 16 Gy [99]. Collectively, these findings support repeat SRS as a feasible and clinically meaningful option for locally recurrent brain metastases, although outcomes remain inferior to primary SRS and careful patient selection is essential.

Intracranial meningiomas

Re-RT has emerged as a relevant salvage strategy for recurrent meningiomas, particularly for tumors arising from the skull base or in close proximity to critical neurovascular structures where repeat surgery is often limited. In a large Italian multicenter cohort including 181 patients treated with stereotactic radiosurgery, fractionated stereotactic radiotherapy, proton therapy, or conventional radiotherapy, re-RT achieved a 3-year second progression-free survival of 51.6% and a 3-year overall survival of 72.5%. Outcomes were strongly influenced by patient selection and tumor biology, with improved local control observed in patients treated with stereotactic techniques, longer intervals between radiation courses (notably > 24–36 months), and higher delivered biologically effective dose, while higher WHO grade and increased proliferative activity (Ki-67 > 5%) were associated with an elevated risk of progression [100, 101].

Treatment-related toxicity was generally acceptable but correlated with target volume and cumulative dose, particularly in anatomically constrained regions such as the skull base, where critical structures including the optic apparatus and brainstem limit dose escalation [102, 103]. In this context, conservative cumulative EQD₂ constraints (< 100–110 Gy for critical organs) are commonly recommended. Hypofractionated regimens (e.g., 25–30 Gy in 5 fractions) are frequently favored for small-volume recurrences due to a more favorable therapeutic ratio and manageable radionecrosis risk [103, 104]. Although particle and proton therapy are increasingly explored for their dosimetric advantages in complex anatomical settings, robust comparative evidence demonstrating superiority over photon-based stereotactic techniques remains limited [105].

Notably, re-RT efficacy appears more limited in higher-grade (WHO grade 2–3) meningiomas, underscoring the importance of individualized, risk-adapted treatment strategies [106]. Emerging molecular data further suggest that alterations such as TERT promoter mutation or CDKN2A/B loss may be associated with more aggressive and potentially radioresistant disease, indicating that selected high-risk patients may be better suited for clinical trial enrollment rather than repeated dose escalation alone [107]. Overall, current evidence supports a multidisciplinary approach integrating tumor biology, anatomical constraints, treatment interval, and cumulative dose considerations to optimize outcomes while minimizing toxicity.

Pediatric ependymoma

Re-RT has become an increasingly important salvage option for recurrent pediatric ependymoma, in which relapse frequently occurs within previously treated volumes and local control remains a dominant determinant of outcome. Contemporary multi-institutional data support a strategy of maximal safe resection followed by re-RT in appropriately selected patients. In a North American multi-center cohort of 35 children treated with re-RT, post-re-RT outcomes were favorable (median OS 65 months; median PFS 33 months; 5-year post-re-RT OS 52.5%), and gross total resection at recurrence was strongly associated with improved survival endpoints, particularly for infratentorial primaries [108]. These findings are concordant with earlier high-impact institutional experience demonstrating that durable disease control can be achieved in a subset of patients after surgery plus re-RT, including with craniospinal irradiation (CSI) when patterns of failure suggest neuraxis risk [109]. Larger outcome analyses further reinforce the potential value of CSI as part of re-RT in selected relapse patterns—particularly for patients with distant-only failure after initial focal RT—while underscoring that tumor biology and recurrence features remain key modifiers of benefit [110]. Importantly, modern re-RT series consistently emphasize the need to balance disease control against late toxicity; available data suggest that full-dose repeat irradiation can be delivered with acceptable tolerability in many patients, with careful attention to organs-at-risk, fractionation, and cumulative dose constraints [111]. Ongoing prospective efforts, including trials explicitly testing surgery plus fractionated re-RT and risk-adapted use of CSI, are expected to refine patient selection and standardize volume/dose strategies in this high-need population [112].

In pediatric ependymoma, re-irradiation has demonstrated meaningful progression-free survival benefit, particularly when preceded by gross total resection and delivered using highly conformal or proton-based techniques [113, 114]. A contemporary pediatric series evaluating proton-based re-irradiation for recurrent ependymoma with long-term follow-up reported a 6-year overall survival of 55.5% and progression-free survival of 25.5%, supporting meaningful disease control in a heavily pretreated population. Importantly, treatment-related toxicity was acceptable, with only 13.3% experiencing grade 2 and 3.3% grade 3 radiation-associated changes, reinforcing proton therapy as a feasible and relatively safe salvage strategy in carefully selected patients [114]. In contrast, adult ependymoma data remain limited and largely retrospective, with outcomes influenced by molecular subtype (e.g., RELA fusion–positive supratentorial tumors) and extent of resection [115]. Across both age groups, comprehensive staging—including evaluation for cerebrospinal fluid dissemination—is essential prior to focal re-irradiation. Given the prolonged survival potential in selected patients, structured long-term neurocognitive monitoring is recommended as part of survivorship care [113–115].

Emerging technologies

The following section summarizes exploratory and early-phase therapeutic strategies that remain investigational in the re-irradiation setting. While biologically compelling, current evidence for CAR-T therapy, oncolytic virotherapy, and combinations is largely derived from early-phase clinical trials or preclinical studies, and these approaches should not yet be considered established components of routine re-irradiation practice.

MRI-guided adaptive radiotherapy (MR-Linac)

MRI-based linear accelerator (MR-Linac) platforms enable online adaptive radiotherapy with superior soft-tissue contrast and real-time visualization, representing a substantial technological advance for re-RT of recurrent gliomas. These capabilities are particularly relevant in the salvage setting, where treatment is challenged by fluctuating edema, evolving target boundaries, and close proximity to eloquent brain structures. Early feasibility and dosimetric studies have demonstrated that MR-guided adaptive radiotherapy allows for daily plan re-optimization, reduction of high-dose exposure to uninvolved brain, and the use of smaller planning target volume margins through improved target delineation and intrafraction motion monitoring [116]. Moreover, MR-Linac systems offer the potential to incorporate biologically informed adaptation, including response- or hypoxia-guided dose modulation, which may enhance the therapeutic ratio in previously irradiated tissue. Initial clinical experiences in recurrent glioblastoma have reported the safe delivery of hypofractionated MR-guided schedules—commonly in the range of 25–30 Gy in five fractions—with local control comparable to conventional stereotactic radiosurgery, alongside improved dosimetric sparing of organs at risk and encouragingly low rates of symptomatic radionecrosis [117]. While these early results support the feasibility and potential advantages of MR-Linac–guided re-RT, longer-term prospective data are required to define its impact on durable tumor control, neurotoxicity, and overall survival.

Proton therapy

Proton therapy has gained increasing attention as a re-RT modality in recurrent gliomas because of its favorable dose distribution and potential to reduce cumulative normal brain exposure. Early clinical series of fractionated proton re-RT for recurrent gliomas have demonstrated feasibility, with survival outcomes comparable to or slightly better than those reported with advanced photon techniques, despite high cumulative doses. In a cohort from Indiana University, Galle et al. treated 20 patients with recurrent gliomas (WHO grades II–IV) using fractionated proton therapy; median re-RT doses were 54–59.4 Gy (RBE), with cumulative doses often exceeding 100 Gy EQD₂. Median overall survival from completion of proton re-RT was 10.2 months for grade III gliomas and 8.2 months for glioblastoma, and the rate of clinically relevant radionecrosis was ~ 10%, which was considered acceptable given the high cumulative dose burden [118].

More contemporary data from a larger series of 44 patients with unresectable recurrent gliomas treated with intensity-modulated proton therapy further support the safety and efficacy of proton re-RT. Gulidov et al. reported a median proton EQD₂ (α/β = 10) of 55 Gy to the recurrent tumor delivered after a median interval of 28 months from prior RT, with concomitant or adjuvant chemotherapy in nearly 87% of patients. Median overall survival and progression-free survival from proton re-RT were 12 and 9 months, respectively, with 1- and 2-year overall survival rates of 49.6% and 35.1%. The incidence of radiation-induced necrosis was low (6.8%), despite relatively large target volumes and prior high-dose irradiation, and was largely asymptomatic or manageable with medical therapy [119]. These data suggest that proton re-RT can be delivered with curative intent, even in unresectable cases, when careful target definition, cumulative dose summation, and organ-at-risk constraints are observed.

Beyond conventional proton techniques, proton minibeam radiation therapy (pMBRT) has emerged as a novel strategy to further widen the therapeutic window in CNS re-RT. Preclinical work in rat brain models has shown that, compared with conventional re-RT, pMBRT markedly reduces neurotoxicity while maintaining or improving tumor control. In a recent study, rats receiving a second course of conventional RT exhibited severe skin toxicity, extensive hippocampal and white-matter necrosis, and significant impairments in locomotor activity, recognition memory, and spatial learning. By contrast, animals re-irradiated with pMBRT after an initial photon course had substantially fewer anatomical lesions (limited mainly to focal fimbria necrosis) and largely preserved cognitive and behavioral functions on open-field, object recognition, and Morris water maze testing [120, 121]. These findings support pMBRT as a potentially transformative technique for brain re-RT, combining the physical advantages of protons with spatially fractionated dose delivery to spare normal tissue.

Together, these clinical and preclinical data indicate that proton therapy—particularly when coupled with advanced imaging and biologic selection—represents a promising avenue to safely escalate or re-deliver dose in previously irradiated CNS tumors. As prospective trials mature and technologies such as pMBRT transition toward clinical implementation, proton-based re-RT is likely to play an increasingly central role in multimodal salvage strategies for recurrent gliomas and other CNS malignancies.

Despite their promising dosimetric and adaptive advantages, the real-world implementation of advanced technologies such as MR-Linac and proton therapy remains limited by substantial infrastructure requirements, high capital and operational costs, and restricted global accessibility. MR-Linac platforms require specialized facilities, extended treatment times, and workflow adaptation, while proton therapy availability is largely confined to high-resource centers, contributing to disparities in access. Furthermore, the absence of definitive randomized evidence demonstrating superiority over advanced photon techniques raises ongoing questions regarding cost-effectiveness in the re-irradiation setting [116–119].

Laser interstitial thermal therapy (LITT) as surgical–radiation hybrid

LITT (Laser Interstitial Thermal Therapy) is increasingly applied both as a cytoreductive modality and as a means of improving radiotherapeutic efficacy after recurrence. By inducing focal coagulative necrosis, LITT reduces tumor burden and decreases edema, enabling safer re-RT margins. In recurrent glioblastoma, LITT combined with salvage radiotherapy has been associated with median survival of 9–11 months and radiographic response stabilization [122]. Treatment is well suited for deep, non-resectable lesions, though thermal damage risk must be balanced against proximity to white-matter tracts.

CAR-T Cell immunotherapy

Chimeric antigen receptor (CAR) T-cell therapy is rapidly emerging as one of the most promising immunologic strategies for treatment-resistant gliomas; however, its clinical efficacy remains constrained by key biological barriers, including an immunosuppressive tumor microenvironment, antigen heterogeneity, limited CAR-T persistence, and restricted intratumoral trafficking. A recent phase I trial evaluating intraventricular administration of bivalent CAR-T cells targeting EGFR and IL13Rα2 demonstrated clinical feasibility and measurable biological activity in recurrent glioblastoma, with radiographic tumor reduction observed in ~ 62% of evaluable patients and durable disease stabilization exceeding 16 months in a subset, albeit with notable neurotoxicity [123]. These findings represent one of the most encouraging anti-tumor signals reported to date for CAR-T therapy in glioblastoma and provide a compelling rationale for rational combinatorial strategies.

In this context, integration with focal, image-guided re-irradiation (reRT) has gained increasing attention. Radiotherapy may enhance CAR-T efficacy through multiple complementary mechanisms. Ionizing radiation induces immunogenic cell death characterized by calreticulin exposure, HMGB1 release, and ATP secretion, thereby promoting dendritic cell activation and antigen presentation. Radiation can also increase tumor-associated antigen expression and upregulate major histocompatibility complex (MHC-I) molecules, improving recognition by engineered T cells. Additionally, radiotherapy modulates the tumor microenvironment by reducing regulatory T-cell activity, altering myeloid-derived suppressor cell populations, and promoting pro-inflammatory cytokine signaling. Sublethal irradiation may further increase vascular permeability and facilitate CAR-T trafficking across the blood–brain barrier [124]. As highlighted in recent translational analyses, including the comprehensive review by Hovhannisyan and colleagues [125], spatially confined irradiation and optimized sequencing strategies may enhance intracranial CAR-T persistence while limiting systemic toxicity. Collectively, these data provide a strong biological and translational rationale for prospective trials combining CAR-T therapy with focal reRT in selected patients with well-defined enhancing disease, while underscoring the need for rigorous toxicity monitoring and biomarker-driven patient selection.

Oncolytic viral therapies

Oncolytic virotherapy has advanced substantially through rational engineering of herpes simplex virus type-1 (HSV-1) backbones to enable selective replication within malignant glioma cells while concurrently modulating the antitumor immune microenvironment. Preclinical and translational studies using genetically attenuated oHSV constructs in orthotopic glioma models have consistently demonstrated significant survival prolongation, with a subset of treated animals achieving long-term survival and rejection of subsequent tumor rechallenge, consistent with the induction of durable adaptive immune memory [126, 127]. These antitumor effects extend beyond direct oncolysis and are accompanied by pronounced remodeling of the tumor microenvironment, including enhanced infiltration and activation of CD4⁺ and CD8⁺ effector T cells and improved antigen presentation [128]. Clinical translation of these principles has been exemplified by next-generation oHSV platforms such as G47Δ, which demonstrated safety and encouraging antitumor activity in patients with recurrent or residual glioblastoma in a phase II trial [129]. Collectively, these data position modern oHSV vectors as promising partners for precision re-RT strategies, where synergistic effects may be achieved through radiation-induced immunogenic cell death, augmented tumor antigen release, and immune priming, supporting further exploration of combined stereotactic radiotherapy and immunotherapeutic approaches.

Tumor-treating fields

Tumor-treating fields (TTFields) represent a noninvasive, regionally targeted modality that applies alternating electric fields through scalp-mounted arrays to disrupt proliferating glioma cells. Mechanistically, TTFields exert antitumoral effects through mitotic spindle disruption, interference with cytokinesis, destabilization of cytoskeletal architecture, impaired DNA repair, and induction of immunogenic cell death, generating strong biological rationale for synergy with focal irradiation and systemic agents.

The EF-14 randomized clinical trial established TTFields with maintenance temozolomide (TMZ) as a backbone treatment for newly diagnosed GBM, extending median overall survival from 16.0 to 20.9 months (HR 0.63; p < 0.001). Durable benefit was seen across age and MGMT subgroups, and 5-year survival improved significantly (13% vs 5%). Real-world and meta-analytic datasets corroborate these findings, reporting median OS consistently in the 20–31 month range [130].

Biologically, TTFields enhance the radiosensitivity of glioma cells via replication stress and suppression of DNA-repair machinery, while concurrently increasing blood–brain barrier permeability and promoting radiation-induced immunogenic cell death [131]. These attributes provide a compelling rationale for integration with re-RT, especially in hypofractionated schedules that amplify antigen release and inflammatory signals. TTFields may also potentiate synergistic immune-based strategies, including checkpoint inhibitors and viral immunotherapy frameworks.

Prospective combinations relevant to re-RT are under evaluation. Early-phase trials initiating TTFields during chemoradiation have demonstrated feasibility and median PFS of ~ 9 months, while the TRIDENT EF-32 phase III trial will clarify optimal sequencing with radiation at recurrence [131].

Toxicity is minimal, dominated by dermatologic irritation at array sites, and systemic toxicity has not been observed. This favorable safety profile enables use alongside re-RT and targeted therapies without compromising dose-escalation strategies or requiring steroid intensification.

Emerging technologies: evidence level and clinical limitations

While MR-Linac, proton therapy, laser interstitial thermal therapy (LITT), and CAR-T–based immunotherapy represent promising advances in the management of recurrent CNS tumors, the current level of evidence remains largely limited to retrospective series, early-phase trials, and single-institution feasibility studies. No randomized controlled trials have yet demonstrated a definitive survival advantage of these technologies specifically in the re-irradiation setting.

MR-guided adaptive radiotherapy (MR-Linac) offers superior soft-tissue visualization and online plan adaptation; however, its clinical evidence in re-irradiation primarily derives from small prospective cohorts and dosimetric studies. Long-term neurotoxicity data and randomized comparative outcomes are lacking. In addition, implementation requires significant infrastructure investment, workflow adaptation, and institutional expertise, which may limit widespread accessibility.

Similarly, proton therapy provides theoretical dosimetric advantages through reduced integral dose and improved normal-tissue sparing. Nevertheless, comparative survival benefits over advanced photon-based stereotactic techniques have not been conclusively demonstrated in recurrent glioma. Existing evidence consists mainly of retrospective cohorts with heterogeneous patient populations and limited follow-up. Cost, access inequity, and uncertainty regarding relative biological effectiveness (RBE) modeling in re-irradiation contexts remain ongoing challenges.

LITT has emerged as a minimally invasive option for focal recurrences, particularly in deep-seated or surgically high-risk lesions. However, current data are based on small non-randomized series, and its integration with re-irradiation lacks standardized protocols.

CAR-T cell therapy represents a biologically compelling strategy in recurrent gliomas, yet clinical efficacy has been modest in early-phase trials. Challenges include antigen heterogeneity, limited CAR-T persistence, immunosuppressive tumor microenvironment, and risk of neurotoxicity. Optimal sequencing and integration with radiotherapy remain investigational.

Overall, although these technologies expand the therapeutic armamentarium, their incorporation into routine re-irradiation practice should be guided by careful patient selection, institutional expertise, and prospective data generation.

Toxicity mitigation and monitoring

Toxicity remains the principal limiting factor in CNS re-RT and requires careful risk stratification integrating lesion size, anatomical location, dose–volume metrics, and systemic therapy. The most clinically relevant late adverse event is radiation necrosis (RN), whose incidence ranges from 4–36%, depending on volume, fractionation, and cumulative dose constraints. The 2025 ISRS guideline notes that cumulative EQD₂ exceeding ~ 120–130 Gy significantly increases RN risk, whereas values ≤ 110 Gy are associated with acceptable safety profiles [18].

Lesion volume is highly prognostic: volumes ≥ 35 mL and prescriptions exceeding 40–45 Gy (in 5 fractions) markedly elevate necrosis risk [12, 13]. Anatomical sensitivity must be considered; RN is < 5% in superficial lobar regions but approaches 15–20% in thalamic/basal ganglia and 30% near the brainstem [14, 27]. Hypofractionated regimens (e.g., 25–30 Gy in 5 fractions) typically offer lower toxicity compared with single-fraction SRS for larger lesions due to improved repair kinetics and reduced high-dose gradients.

Reported rates of radiation necrosis (RN) vary widely (~ 4%–36%) depending on treatment technique, target volume, anatomical location, and cumulative dose exposure. Lower rates (< 5–10%) are typically observed in small-volume lesions treated with stereotactic approaches (V12 < 10 cm³), whereas higher rates (15–30% or more) are associated with larger lesions, eloquent or deep-seated locations, and higher cumulative dose exposure (e.g., elevated V12 or maximum dose ≥ 40 Gy). Importantly, RN risk is more strongly correlated with dosimetric parameters such as V12 and cumulative EQD₂ than with treatment modality alone, underscoring the importance of careful treatment planning and patient selection.

To improve interpretability, cumulative EQD₂ thresholds have been categorized into clinically relevant risk tiers Table (5), facilitating more consistent application in treatment planning.

Table 5.

EQD₂—Convert into risk tier system

Risk tier	Cumulative EQD₂ (α/β = 2)	Estimated risk of toxicity	Clinical interpretation
Low risk	≤ 100 Gy	Low risk of radionecrosis	Generally safe in well-selected patients
Intermediate risk	100–120 Gy	Moderate risk	Acceptable with caution and strict constraints
High risk	120–130 Gy	Increased risk	Requires careful patient selection and planning
Very high risk	> 130 Gy	High risk of severe toxicity	Avoid if possible; consider alternative strategies

Bevacizumab (BEV) plays a major role in toxicity mitigation, decreasing edema, improving steroid-dependence, and reducing symptomatic RN. In RTOG 1205, high-grade toxicity was rare, and delayed high-grade events were not observed, demonstrating BEV’s safety when combined with re-RT [57]. Adverse events with BEV itself include hypertension, thrombosis, impaired wound healing, and hemorrhage, necessitating judicious patient selection. Systemic agents such as TMZ contribute to hematologic toxicity, particularly lymphopenia and thrombocytopenia in up to 30–35% of patients [69].

Routine monitoring should emphasize longitudinal MRI with radiomics/rCBV maps, amino-acid PET when available, and multidisciplinary collaboration for early intervention. Steroid dose tapering, BEV rescue, hyperbaric oxygen therapy, and surgical debulking remain the main approaches for symptomatic RN.

Beyond traditional oncologic endpoints such as overall survival and progression-free survival, preservation of neurological function and health-related quality of life (QoL) represents a central therapeutic objective in the palliative setting of brain re-irradiation. Available evidence suggests that modern focal re-RT strategies—particularly conventionally fractionated conformal techniques and stereotactic approaches—are generally associated with stabilization of global QoL and maintenance of functional domains when cumulative dose constraints are respected. In a pooled analysis of 13 studies including ~ 300 patients with recurrent glioblastoma, Nieder et al. reported clinical improvement in 24–45% of patients, reduction in steroid requirement in 20–60%, and acceptable late toxicity rates, underscoring the palliative benefit of re-RT beyond survival metrics [132]. Prospective hypofractionated stereotactic series have similarly demonstrated stable QoL in the majority of patients with limited-volume recurrence [133], while randomized phase II data combining re-RT with APG101 showed no clinically meaningful deterioration in QoL domains compared with re-RT alone, suggesting preservation of functional status over time [134]. In a prospective study incorporating concurrent temozolomide and bevacizumab, a transient decline in QoL immediately following treatment was observed, with subsequent recovery toward baseline during adjuvant therapy [135]. Importantly, proton-based re-RT has also been associated with substantial stability across EORTC QLQ-C30 and BN20 domains, without significant long-term impairment in global health status or symptom burden [136].

Importantly, the most robust prospective longitudinal data to date come from Maitre et al., who evaluated 60 patients treated with high-dose salvage re-RT using serial EORTC QLQ-C30, BN20, and modified Barthel Index assessments [137]. Over a 12-month follow-up, significant improvements were observed in physical (p < 0.001), emotional (p = 0.002), cognitive (p = 0.009), and social functioning (p = 0.047), while global health status remained stable. Fatigue scores improved significantly, and activities of daily living demonstrated a meaningful increase over time (p = 0.001), reflecting enhanced functional independence. Although approximately one-third of patients developed post-treatment radiologic changes suggestive of pseudo-progression—particularly at higher cumulative EQD₂ doses—QoL deterioration was more closely associated with true tumor progression than with treatment-related toxicity.

In a recent systematic review and meta-analysis, combination therapy demonstrated survival benefits without a clear increase in grade ≥ 3 toxicities, and bevacizumab-containing regimens were associated with reduced radiation necrosis, a complication strongly linked to neurological deterioration and QoL decline [138]. Although formal meta-analysis of QoL outcomes was limited due to heterogeneous reporting, available data suggest that steroid-sparing effects and improved intracranial control may translate into meaningful symptomatic benefit. Future prospective trials should incorporate validated neurocognitive and patient-reported outcome measures to better define the functional impact of precision re-RT strategies.

Beyond survival outcomes, accumulating evidence indicates that modern focal re-irradiation strategies are generally associated with preservation or stabilization of neurocognitive function and global quality of life, particularly when appropriate patient selection and dose constraints are applied [132–137]. Moreover, prospective and randomized data suggest that QoL deterioration is more strongly related to tumor progression rather than treatment-related toxicity, supporting the clinical value of re-RT in maintaining functional outcomes despite modest survival benefits [134, 137, 138].

Collectively, contemporary data suggest that toxicity after re-RT is generally manageable when dose-constraint principles are respected, systemic agents are integrated thoughtfully, and biological imaging supports differential diagnosis between recurrence and RN (Table 6).

Table 6.

Dose–volume parameters and treatment-related toxicity risk factors in CNS Re-RT

Factor	Evidence	Clinical Implication
Cumulative dose (EQD2)	RN increases sharply > 120–130 Gy	Keep cumulative EQD2 ≤ 110–120 Gy
Lesion volume	Risk ↑ when ≥ 35 mL	Favor HF-SRT rather than SRS
Single-fraction SRS	Higher RN risk in large/central lesions	Avoid if V12 > 10–12 cm3
Hypofractionation	Lower necrosis for 25–30 Gy/5 fx	Preferred for > 2–3 cm lesions
Critical structures	RN 15–30% near thalamus/brainstem	Tight constraints, consider PT
Time interval	≥ 12 mo reduces toxicity	Favor re-RT after ≥ 12 months
Systemic therapy (TMZ)	Myelosuppression up to 30%	CBC monitoring required
BEV co-therapy	Reduces edema/RN; minimal delayed toxicity	Useful both prophylactically and as salvage
Imaging biomarkers	PET/rCBV distinguish RN vs progression	Early detection; avoids overtreatment

RN Radiation necrosis, EQD₂ Equivalent dose according to alfa/beta ratio 2, HF-SRT Hypofractionated stereotactic radiotherapy, SRS Stereotactic radiosurgery, PT Proton therapy, CBC Complete blood count, TMZ Temozolamide, BEV Bevacizumab

Ongoing clinical trials

Re-irradiation of central nervous system (CNS) tumors represents the most actively investigated domain within the broader re-RT research landscape. In a recent systematic review of ClinicalTrials.gov, 40% of all ongoing prospective re-irradiation studies were focused on CNS malignancies, underscoring the high clinical priority of this setting. The majority of these trials address recurrent glioma—particularly glioblastoma—reflecting the persistent therapeutic challenges in this population [139]. Notably, randomized designs are increasingly being implemented. The phase III LEGATO trial (NCT05904119; (completion date: February 2028, sample size:411) evaluates lomustine with or without re-RT in recurrent glioblastoma and includes overall survival and quality-of-life endpoints, aiming to generate practice-informing data [140]. Additional randomized trials investigate re-RT strategies in meningioma (NCT03604978; (completion date: March 2027, sample size: 38) [141] and brain metastases (NCT05124912; completion date: October 2028, sample size: 261) [142], further expanding the scope of prospective evaluation.

From a technical perspective, high-precision approaches predominate, including stereotactic radiotherapy and particle therapy (proton and carbon ion modalities). Dose-escalation studies are ongoing, and a substantial proportion of trials combine re-RT with systemic agents, particularly immune checkpoint inhibitors. Although toxicity and local control remain the most common primary endpoints, progression-free survival and patient-reported quality-of-life measures are increasingly incorporated. Collectively, these efforts reflect methodological maturation in the field; however, the limited number of phase III randomized trials highlights the continued need for high-level evidence to define optimal patient selection, dose strategies, and multimodal integration in CNS re-RT.

Conclusion

Re-RT has evolved from a historically cautious salvage option to a biologically and technologically refined component of precision neuro-oncology. Contemporary data support its feasibility in carefully selected patients when cumulative dose constraints, lesion volume, anatomical considerations, and performance status are rigorously integrated into decision-making. Advances in stereotactic techniques, adaptive image guidance, molecular profiling, and multimodal integration have collectively expanded the therapeutic window while maintaining acceptable toxicity profiles.

However, substantial heterogeneity persists in patient selection criteria, dose–fractionation strategies, and systemic combinations. Future research should prioritize biomarker-stratified prospective trials that integrate molecular signatures (e.g., IDH status, MGMT methylation, radiogenomic profiles) with advanced imaging biomarkers to better define responders. Randomized studies comparing focal re-RT with or without rational systemic partners—including anti-angiogenic agents, DNA repair inhibitors, and immunotherapeutic strategies such as CAR-T—are particularly needed. Additionally, prospective evaluation of adaptive MR-guided re-RT protocols and harmonized cumulative dose-constraint models will be critical to standardize practice.

In routine clinical care, incorporation of structured patient selection algorithms, rigorous cumulative EQD₂ assessment, advanced imaging to distinguish recurrence from treatment effect, and multidisciplinary discussion should form the foundation of responsible re-RT practice. Ultimately, the integration of biological stratification, technological precision, and carefully designed clinical trials will determine whether re-RT becomes a consistently evidence-based standard rather than an individualized salvage strategy.

Author contributions

The author was responsible for the conception, literature review, analysis, and writing of the manuscript.

Funding

None.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.