Retrospective comparisons of the dosimetric differences between coplanar and non‑coplanar volume‑modulated arc therapy in patients with pediatric brainstem glioma

Introduction

Brainstem gliomas are malignant primary tumors originating from glial cells within the brainstem and account for 10–20% of all pediatric central nervous system (CNS) tumors (1). Among these, nearly 80% are classified as diffuse midline gliomas [World Health Organization (WHO) grade IV] (2), which are marked by highly aggressive biological behavior and an exceptionally poor prognosis, with a 2-year survival rate of <10% (3–5). The remaining 20% are predominantly focal low-grade gliomas (WHO grades I–II), which tend to grow more slowly and are associated with more favorable outcomes, with reported 5-year survival rates ranging between 60 and 80% (6). The majority of brainstem gliomas (70–80%) occur in individuals <18 years of age, with an overall 5-year survival rate of ~30% in this pediatric population (7).

Radiotherapy remains the cornerstone of treatment for brainstem gliomas, particularly in high-grade tumors. In pediatric patients over the age of 3 years, clinical improvement is often observed within 1 to 2 weeks of radiotherapy (8). This modality enables temporary local tumor control in 50–70% of focal brainstem lesions and has been shown to extend median progression-free survival (PFS) time from 2.4 to 6 months, as well as median overall survival (OS) time from 3.4 to 9.8 months (9–13). Despite its benefits, radiotherapy is frequently regarded as a double-edged sword due to the potential for serious long-term side effects, including growth delays, neurocognitive impairment, hearing loss, visual deficits and a general decline in quality of life (14,15). As a result, radiation oncologists are actively investigating refined treatment strategies that maximize therapeutic benefit while reducing collateral damage, especially with the emergence of advanced techniques such as volumetric modulated arc therapy (VMAT). VMAT has been shown to surpass both three-dimensional conformal radiotherapy and intensity-modulated radiotherapy by enhancing target dose conformity, better sparing of normal tissues, reducing monitor units and shortening overall treatment time, improving both therapeutic outcomes and patient experience (10,16,17).

While VMAT has demonstrated promising potential in pediatric glioma treatment, most existing studies have focused on coplanar techniques (18). However, the dosimetric benefits of non-coplanar VMAT (NC-VMAT), particularly in the context of pediatric brain radiotherapy, remain underexplored, with limited evidence supporting the practical application of the technique. The key innovation of this study lies in being the first, to the best of our knowledge, to evaluate and validate the dosimetric advantages of NC-VMAT in the context of pediatric brainstem glioma. This study addresses a significant gap in the radiotherapy literature, where previous investigations have primarily focused on adult populations or non-brainstem pediatric tumors. By focusing on this anatomically and clinically sensitive region, the study provides novel insights into potential avenues for dose optimization and tissue sparing in pediatric CNS radiotherapy.

Materials and methods

Patients

The present study included 10 pediatric patients (aged <18 years) who were diagnosed with brainstem glioma and underwent postoperative VMAT at the Department of Radiotherapy, Fujian Children's Hospital (Fuzhou, China) between August 2022 and September 2024. The inclusion criteria were as follows: i) Histopathological confirmation of glioma; ii) availability of complete medical records; and iii) documented clinical administration of VMAT. All patients received standard coplanar VMAT as part of their routine treatment protocol. The study cohort consisted of 4 male and 6 female patients, with a median age of 5.5 years (range, 4–12 years).

For this study, NC-VMAT plans were retrospectively created using the same computed tomography (CT) datasets as the clinically implemented VMAT plans, allowing for direct dosimetric comparison. It is important to emphasize that NC-VMAT was not delivered to patients in clinical practice. Patient follow-up was conducted through a combination of regular outpatient visits, telephone interviews and, where applicable, death certificate verification. All follow-up data were reviewed and validated by attending clinical physicians to ensure accuracy and completeness.

Patient positioning and imaging

CT imaging was performed with patients positioned in a supine manner on an integrated immobilization system (Huayuxin HYX-UTS-CM; Jinan Huayu New Casting and Forging Materials Co., Ltd.) using a customized thermoplastic mask and headrest to ensure a stable and reproducible posture. Contrast-enhanced scans were acquired using a 16-slice, large-aperture CT simulator (Discovery RT590; GE Healthcare) following intravenous administration of iodinated contrast. Axial images were obtained at a slice thickness of 2.5 mm for precise anatomical delineation. The acquired images were then imported into the Eclipse treatment planning system (TPS) (version 15.5; Varian; Siemens Healthineers), where radiation oncologists delineated the target volumes and organs at risk (OARs) to develop the treatment plans. All radiotherapy treatments were delivered using a Varian TrueBeam linear accelerator (Varian; Siemens Healthineers).

Target volume delineation

To improve the accuracy of target and normal tissue delineation, both preoperative and postoperative contrast-enhanced T1-weighted and fluid attenuated inversion recovery magnetic resonance imaging (MRI) sequences were imported into the TPS for all patients. These MRI datasets were automatically aligned with the planning CT images using rigid registration within the TPS. Radiation oncologists manually reviewed and, if necessary, adjusted the image registration to ensure anatomical accuracy. The gross tumor volume (GTV) of the tumor bed was defined as the surgical resection cavity along with all regions of contrast enhancement observed on the preoperative T1-weighted MRI scans. The clinical target volume (CTV) was generated by expanding the GTV by 2 cm in all directions, followed by manual modification by the physician to conform to anatomical boundaries. A uniform 0.3-cm margin was then added to the CTV to create the planning target volume (PTV). Surrounding OARs were also carefully delineated with regard to the target area.

Treatment planning

All treatment plans were developed by the same team of radiation physicists using calibrated 6 MV X-rays within the TPS. For each patient, two distinct plans were retrospectively generated using the same CT dataset. In the control group (VMAT), a standard coplanar technique was employed, consisting of two full 360° arcs delivered with the treatment couch fixed at 0°. In the experimental group (NC-VMAT), the plan began with a single 360° arc at a couch angle of 0°, followed by four non-coplanar 180° half-arcs delivered at couch angles of 20°, 40°, 320° and 340°, respectively (schematic diagrams are shown in Fig. 1). Both planning approaches were optimized using the analytical anisotropic algorithm (19). To ensure a fair comparison focused on the dosimetric merits of each technique, identical optimization objectives and priority settings were applied across both plans. The prescription dose for all patients was uniformly set at a PTV of 50.4 Gy, delivered in 28 fractions of 180 cGy each [the dosing regimen was selected following National Comprehensive Cancer Network guidelines (20)], ensuring that 95% of the target volume received 100% of the prescribed dose. All plans were thoroughly reviewed and approved by senior radiation physicists and chief radiation oncologists to ensure consistency and clinical acceptability.

Figure 1. Schematic illustration of a non-coplanar irradiation field. (A) The radiation beam paths from a three-dimensional spatial perspective. (B) Distribution path of the beams relative to the target volume on a specific imaging plane.

Dosimetric evaluation

The dosimetric characteristics of the target dose distribution for the two treatment groups were evaluated using dose-volume histogram (DVH) analysis. Key parameters included the conformity index (CI) and homogeneity index (HI), calculated as follows: i) CI=TV95%/PTVtotal, where TV95% represents the volume receiving at least 95% of the prescribed dose, and PTVtotal denotes the total planning target volume. A CI value of 1 indicates ideal conformity. ii) HI=(D2-D98)/D50, where D2, D50 and D98 correspond to the doses received by 2, 50 and 98% of the target volume, respectively. An HI value approaching 0 reflects optimal dose uniformity.

In addition to these indices, comparisons were made between the two groups for the maximum dose delivered to critical structures, including the optic nerves, optic chiasm, lenses and cochleae, as well as the mean dose (Dmean) delivered to the temporal lobes (TLs) and surrounding normal brain tissue. The final dose level diagram for the plans in these two groups is shown in Fig. 2.

Figure 2. Comparison of dose coverage on the same level between two different plans: VMAT and NC-VMAT. NC-VMAT, non-coplanar volume-modulated arc therapy.

Plan verification and treatment duration

Plan verification for both treatment groups was conducted using the PTW OCTAVIUS 4D dosimetric validation system in conjunction with Verisoft 7.1 analysis software (PTW Freiburg GmbH). The γ passing rate was evaluated following AAPM Task Group 218 guidelines (21), using criteria of 3% dose difference and 2 mm distance-to-agreement, with a minimum dose threshold of 10% and an acceptance criterion of γ passing rate ≥95%. Treatment time was defined as the duration from beam-on to beam-off.

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics version 24.0 (IBM Corp.). Given the paired design of the study, where both treatment techniques were applied to the same patient cohort, continuous variables were first tested for normality using the Shapiro-Wilk test. Data following a normal distribution are presented as the mean ± standard deviation, and comparisons between techniques were conducted using paired-samples t-tests. For data that did not meet normality assumptions, values are presented as the median and interquartile range (IQR), with comparisons made using the Wilcoxon signed-rank test. A two-sided P-value of <0.05 was considered to indicate a statistically significant difference.

Results

The baseline characterestics of the 10 pediatric patients included in this study are summarized in Table I.

Table I. Baseline characteristics of the study participants.

Table I.

Baseline characteristics of the study participants.

SN	Age, years	Sex	PTVtotal (cc)	Histology	WHO grade
1	5	Female	121.10	Ependymoma	II
2	6	Female	132.70	Diffuse midline glioma	IV
3	4	Male	125.20	Anaplastic astrocytoma	III
4	7	Female	154.60	Diffuse midline glioma	IV
5	5	Female	176.20	Diffuse midline glioma	IV
6	4	Female	171.40	Ependymoma	II
7	7	Female	218.40	Diffuse midline glioma	IV
8	12	Male	179.00	Ependymoma	II
9	4	Male	165.40	Glioblastoma	IV
10	12	Male	155.70	Glioblastoma	IV

[i] PTV, planning target volume; WHO, World Health Organization; SN, serial number.

Dosimetric comparisons of the PTVs revealed that the NC-VMAT technique outperformed conventional VMAT in terms of Dmean, CI and HI, all of which showed statistically significant improvements (all P≤0.001). However, VMAT demonstrated a significantly shorter treatment time compared with NC-VMAT (P<0.001). The γ pass rates for the target volume were comparable between the two techniques, with no statistically significant differences observed (P>0.05). Detailed results are presented in Table II.

Table II. Dosimetric parameter comparisons for the planning target volume in the VMAT and NC-VMAT patient cohorts.

Table II.

Dosimetric parameter comparisons for the planning target volume in the VMAT and NC-VMAT patient cohorts.

Parameter	VMAT	NC-VMAT	t	P-value
Dmean, cGy	5271.90±20.91	5220.40±18.98	5.848	<0.001
HI	0.849±0.004	0.066±0.003	13.481	<0.001
CI	1.095±0.010	1.045±0.003	4.701	0.001
Treatment time, min	3.06±0.54	5.90±0.10	−32.111	<0.001
γ pass rate, %	99.10±0.30	98.80±0.20	0.896	0.394

[i] NC-VMAT, non-coplanar volume-modulated arc therapy; D_mean, mean dose; HI, homogeneity index; CI, conformity index.

Regarding radiation dose to OARs surrounding the target volume, NC-VMAT significantly reduced exposure to the lenses, optic nerves, cochlea, and optic chiasm compared to VMAT (all P<0.05). VMAT resulted in substantially lower radiation doses to the TLs and normal brain tissue relative to NC-VMAT (P<0.05). Corresponding DVH parameters and statistical analyses are provided in Table III and Fig. 3 for comprehensive comparison.

Figure 3. Radiation dose comparison of OARs between VMAT and NC-VMAT treatment plans. ***P<0.001, **P<0.01 and *P<0.05. NC-VMAT, non-coplanar volume-modulated arc therapy; L, left; R, right; OAR, organs at risk.

Table III. Comparative analysis of OAR dose exposure between two planning groups.

Table III.

Comparative analysis of OAR dose exposure between two planning groups.

	Mean radiation dose, cGy
OAR	VMAT	NC-VMAT	t	P-value
Lens-L	413.20±40.65	289.50±30.00	−6.147	<0.001
Lens-R	407.30±32.05	294.00±27.74	−4.487	0.002
Optic nerve-L	2,318.50±312.59	2,111.10±293.98	−3.274	0.010
Optic nerve-R	2,422.30±336.65	2,208.20±286.93	−3.661	0.005
Temporal lobe-L	1,977.70±155.11	2,491.20±88.48	4.480	0.002
Temporal lobe-R	1,961.50±144.14	2,477.00±86.29	4.688	0.001
Optic chiasm	5,063.70±40.71	4,511.10±62.58	−17.969	<0.001
Cochlea-L	4,203.70±221.98	3,520.20±197.45	−9.052	<0.001
Cochlea-R	4,218.00±154.47	3,671.90±164.12	−7.969	<0.001
Brain	1,398.50±10.28	1,540.30±17.79	8.242	<0.001

[i] OAR, organ at risk; L, left; R, right; NC-VMAT, non-coplanar volume-modulated arc therapy.

Discussion

The results of the present retrospective dosimetric analysis demonstrated that NC-VMAT radiotherapy provides superior target CI and HI compared with coplanar VMAT, aligning with findings from previous studies (22–24). These outcomes reinforce the potential of NC-VMAT to increase tumor dose coverage and reduce dose inhomogeneity, underscoring its dosimetric advantages over conventional VMAT techniques. NC-VMAT allows for flexible adjustment of beam angles through coordinated gantry and couch rotation, facilitating more precise dose modulation and enabling multidirectional beam delivery to the target volume. This approach enhances high-dose conformity to the tumor while effectively sparing surrounding healthy tissues, consistent with prior research (25,26). Despite the increased computational and mechanical demands associated with NC-VMAT planning and delivery, the comparable γ passing rates observed between the two techniques in the present study confirm that the TrueBeam linear accelerator can reliably and accurately implement NC-VMAT plans.

According to Radiation Therapy Oncology Group, American Society for Radiation Oncology and Pediatric Normal Tissue Effects in the Clinic guidelines, the risk of radiation-induced toxicity in pediatric patients significantly increases when the lens is exposed to doses >3 Gy, when the optic nerves or optic chiasm receive doses >45 Gy or when the cochlea is subjected to a Dmean >35 Gy (27,28). In the present study, dosimetric comparisons of OARs revealed that NC-VMAT achieved significantly better dose sparing of the lens, optic nerves, optic chiasm and cochlea than conventional VMAT (P<0.05), which was consistent with earlier findings (29,30). These results provide further support for the efficacy of NC-VMAT in reducing treatment-associated toxicities during radiotherapy for pediatric brainstem gliomas.

However, the present analysis also revealed that the Dmean delivered to the TLs and normal brain tissue was significantly higher in the NC-VMAT group compared with that in the VMAT group. This finding, which contrasts with a prior report (26), likely stems from the inherent design of NC-VMAT, which involves multiple beam paths that traverse these regions, increasing exposure to low-dose radiation. These observations underscore the need for careful clinical judgment when applying NC-VMAT, especially in pediatric patients with low-grade brainstem gliomas and longer projected survival times. As shown by Pokhrel et al (31) and Bertholet et al (32), fine-tuning key parameters of NC-VMAT, such as gantry angles, dose rate and beam delivery time, can help optimize dose distributions to minimize radiation exposure to critical structures. With the clinical adoption of NC-VMAT, radiation oncologists are afforded greater flexibility to individualize treatment parameters, improving the protection of surrounding OARs and reducing the risk of long-term complications.

Although NC-VMAT technology offers significant advantages, its clinical implementation presents several challenges. As demonstrated in the present study, treatment time for NC-VMAT was nearly double that of conventional VMAT, which may pose difficulties during radiotherapy sessions for young children. Prolonged treatment durations can reduce patient compliance, particularly in pediatric patients who may struggle to remain still, potentially compromising irradiation accuracy. Moreover, children requiring sedation may need increased doses of anesthetic agents. Previous studies by Guilcher et al (33) and Docking and Knijnik (34) have highlighted the risks associated with repeated sedation, noting that cumulative exposure to anesthetic agents in pediatric populations may elevate the risk of long-term adverse effects.

Further limitations to the widespread adoption of NC-VMAT include its higher equipment and operational costs, the complexity of its treatment planning process and the need for highly skilled technical personnel (35). These practical constraints underscore the importance of developing strategies to optimize the balance between treatment quality, efficiency and feasibility in pediatric settings.

Future studies are warranted to evaluate the clinical application of NC-VMAT further. These should include long-term follow-up to assess dose distribution, tumor control efficacy, patient survival outcomes, and treatment-related complications across various tumor types and patient subgroups. Whether the dosimetric benefits of NC-VMAT ultimately translate into meaningful clinical improvements remains an open question, and addressing this gap will be essential for informing evidence-based clinical guidelines.

The present study has several limitations. Firstly, as a single-center investigation with a relatively small sample size (n=10), the statistical power to detect subtle clinical differences may be limited. Although strict inclusion and exclusion criteria were applied to minimize potential confounding factors, larger multicenter studies with expanded cohorts are necessary to improve the generalizability and robustness of the findings. Furthermore, the current research lacks a comprehensive assessment of long-term neurocognitive outcomes in brain regions exposed to low-dose irradiation. Given the heightened radiosensitivity of the developing pediatric nervous system, the potential for delayed adverse effects from low-dose exposure warrants further investigation through extended follow-up.

Furthermore, none of the patients received NC-VMAT treatment in clinical practice. The NC-VMAT plans were generated retrospectively using the same CT images and planning data as the clinically delivered VMAT plans. These experimental plans were used solely for dosimetric comparison within the study framework.

In conclusion, NC-VMAT demonstrates significant advantages in the radiotherapeutic management of pediatric brainstem gliomas, particularly in terms of improved target dose conformity and enhanced protection of critical organs. Despite these benefits, its clinical implementation presents certain challenges, including longer treatment durations and increased radiation exposure to the TLs and surrounding brain tissue. However, the dosimetric strengths of NC-VMAT suggest that it holds potential for optimizing radiotherapy outcomes and improving the overall treatment experience for children with brainstem gliomas.

Acknowledgements

Not applicable.

Funding

This study was supported by the Natural Science Foundation of Fujian Province (grant nos. 2023J011305 and 2023J011299).

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

ZL, JJ, YS, BG, KS, YL and JY contributed to the study conception and design. ZL and YS were mainly responsible for study conception. BG and ZL were responsible for methodology. Formal analysis and investigation was performed by YL and KS. Original draft preparation was performed by ZL and YS. JJ and BG reviewed and edited the manuscript. Funding was acquired by ZL and BG. YS and JY supervised the study. ZL and BG confirm the authenticity of all the raw data. All authors commented on previous versions of the manuscript. All authors have read and approved the final manuscript.

Ethics approval and consent to participate

This study was conducted in strict adherence to the Ethical Review Measures for Biomedical Research Involving Human Subjects, the Declaration of Helsinki (as revised in 2013) and the International Ethical Guidelines for Health-related Research Involving Humans. All research activities involving human participants were approved by the Ethics Committee of Fujian Children's Hospital (Fuzhou, China; approval no. 2024ETKLRK10007). The ethics committee approved the waiver of parental consent due to the retrospective nature of the study and the anonymized nature of the data, and no identifiable MRI was included in the manuscript.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References