Impact of field number and monitor units per segment on magnetic resonance-guided hypofractionated stereotactic radiotherapy for brain metastases: plan quality, deliverability, and robustness trade-offs

Introduction

Brain metastases (BMs) represent the most common intracranial malignancies in adults (1-5). For BMs larger than 2 cm, hypofractionated stereotactic radiotherapy (HSRT) is typically favored, as it provides effective local control while lowering the risk of radiation necrosis (RN) compared with single-fraction stereotactic radiosurgery (SRS) (6-9). Recently, magnetic resonance-guided linear accelerators (MR-Linacs) have been introduced into clinical practice (10-13), providing enhanced soft-tissue visualization and facilitating adaptive radiation therapy (ART). This advancement may further improve HSRT outcomes for BMs (14).

However, despite these advantages, technical constraints limit the widespread clinical implementation of cranial HSRT using current MR-Linac platforms. For instance, the Elekta Unity system only permits coplanar, step-and-shoot intensity-modulated radiotherapy (IMRT) and does not support arc-based delivery (15), such as volumetric modulated arc therapy (VMAT), potentially resulting in prolonged beam-on time (BOT). Furthermore, geometric and mechanical limitations often necessitate off-isocenter beam delivery, potentially reducing plan quality. Beyond concerns regarding deliverability and plan quality, the robustness of dose calculation presents an additional challenge in MR-Linac-based adapt-to-shape (ATS) workflows. Since ATS dose calculation primarily depends on bulk density assignment (16,17), accurate dose computation relies heavily on the precise delineation of anatomical structures. This accuracy is particularly critical when high-density regions, such as the skull, are involved. Although magnetic resonance imaging (MRI) demonstrates superior soft-tissue contrast compared with computed tomography (CT) and facilitates lesion identification, the accurate delineation of cranial bone boundaries on MRI can occasionally be difficult, leading to potential density misassignment (Figure S1). Given that bone exhibits a substantially higher electron density than brain tissue (Figure S2), even slight skull mislocalization can cause notable dose deviations in HSRT.

Given the steep dose gradients and high fractional doses characteristic of HSRT for BMs, planning strategies within ATS workflows must balance dosimetric quality, treatment efficiency, and robustness of dose calculation. In IMRT, the number of fields and the minimum monitor unit (MU)/segment are critical adjustable parameters influencing both delivery performance and plan quality (18-20). Increasing the number of fields may enhance plan quality but often results in increased segmentation and prolonged BOT (18-20). Conversely, increasing the minimum MU/segment reduces segmentation, improving treatment efficiency but potentially limiting modulation capability (21). Therefore, combining a higher number of fields with carefully selected MU/segment constraints may offer an optimal trade-off between efficiency and quality for MR-guided workflows. However, this balance remains unexplored in MR-guided cranial HSRT.

Accordingly, this study compared four clinically relevant IMRT configurations for MR-guided online adaptive HSRT of solitary medium-sized BMs: 9-field IMRT with 15 MU/segment (9FL-IMRT), 9-field IMRT with 5 MU/segment (9FS-IMRT), 15-field IMRT with 15 MU/segment (15FL-IMRT), and 15-field IMRT with 5 MU/segment (15FS-IMRT). The trade-offs between plan quality, delivery performance, and robustness of dose calculation were systematically assessed to provide practical guidance for efficient and reliable planning strategies in MR-guided cranial HSRT. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1-2798/rc).

Methods

Patients, imaging, contouring, and prescription dose

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the ethics committee of West China Hospital (approval No. 2024124) and individual consent for this retrospective analysis was waived. Twenty patients with solitary medium-sized (2–3 cm) BMs were retrospectively included at West China Hospital, Sichuan University. The median age was 56 years (range, 35–81 years). Among the patients, 13 (65%) were male, and 7 (35%) were female. All patients had intact (non-resected) BMs with a median gross tumor volume (GTV) of 3.14 cm³ (range, 1.65–7.99 cm³) and a median planning target volume (PTV) of 5.94 cm³ (range, 4.07–13.86 cm³). Patient characteristics are summarized in Table 1. Simulation imaging included high-resolution CT and MRI, acquired in the treatment position (supine, immobilized with a thermoplastic mask). The GTV was delineated on T1-weighted contrast-enhanced MRI and subsequently transferred to CT. Nearby organs at risk (OARs), including lenses, brainstem, and optic apparatus, were contoured. Additionally, high-density cranial bone structures were outlined. The PTV was generated by applying a uniform 2-mm margin around the GTV. All patients received a prescription dose (D_p) of 30 Gy delivered in three fractions.

Planning design

MRI images were not utilized for plan optimization or dose calculation. Instead, CT images with density overrides were employed to simulate MRI-based planning conditions within the ATS workflow. This approach avoids geometric discrepancies between CT and MRI images and eliminates MRI-related geometric distortions and volumetric uncertainties, while enabling a direct comparison of dose distributions.

All simulated plans were generated using Monaco version 5.4 on a 1.5-T Elekta Unity MR-Linac (Elekta AB, Stockholm, Sweden), which features a gantry rotation speed of 6.0 rpm, relevant for evaluating treatment efficiency. For each patient, four IMRT plans were created by systematically varying number of fields and minimum MU/segment. Based on clinical planning experience, the maximum number of beam directions was set at 15, as further increasing beam numbers could significantly increase computational load and compromise feasibility on the Unity MR-Linac. Specifically, the plans used either 9 or 15 coplanar fields evenly spaced within a predefined angular range, excluding beam angles intersecting MRI cryostat or radiofrequency coil structures.

For each beam setting, two minimum MU/segment values, 5 and 15 MU, were evaluated. These values were selected based on manufacturer recommendations, and prior evidence on IMRT plan complexity and deliverability (21,22). The 5 MU threshold was chosen as the lower practical limit to maintain modulation flexibility while avoiding excessive small-segment generation, which may increase plan complexity and dosimetric uncertainty. Conversely, the 15 MU threshold was chosen as the upper practical limit to reduce segmentation and improve delivery efficiency without overly restricting modulation capability in high-gradient HSRT treatments. Combining two beam-number choices (9 or 15 fields) with two MU/segment settings produced four distinct IMRT configurations. The naming convention indicates the number of fields followed by “L” (large minimum MU) or “S” (small minimum MU) per segment (i.e., 9FL-IMRT, 9FS-IMRT, 15FL-IMRT, and 15FS-IMRT).

For all configurations, optimization and dose calculations used accurately positioned anatomical structures with average electron density overrides, simulating bulk density-based optimization in ATS workflows. The maximum segment number was restricted to 125, enabling automatic beam sequencing on the Unity MR-Linac. Minimum segment area was set uniformly at 2 cm². Identical objectives and optimization priorities were applied across all plans. Table S1 summarizes planning parameters and differences among the four configurations. Dose normalization ensured that 98% of the PTV received the prescription dose (PTV D_98% =30 Gy). The dose calculation grid size was 1 mm. The maximum point dose within the PTV was limited to ≤150% of the prescribed dose. To avoid inter-planner variability, all treatment plans were designed by the same experienced medical physicist.

Dosimetric evaluation

Plan quality was evaluated using dose-volume histogram (DVH) metrics for the target and normal brain tissue (NBT). For each plan, the maximum dose (D_max) and near-minimum dose (D_98%) to the GTV were recorded to assess target coverage. The conformity index (CI) (23) was calculated as the ratio of the volume enclosed by the 100% prescription isodose to the PTV. The gradient index (GI) (24) was defined as the ratio of the volume receiving the 50% isodose to the volume receiving the 100% isodose. The homogeneity index (HI) (25) was calculated as the ratio of doses covering 5% (D_5%) and 95% (D_95%) of the PTV, obtained from the DVH. For NBT (defined as brain volume excluding the PTV), evaluated dosimetric parameters included mean dose (D_mean) and volumes receiving doses ≥18 Gy (V_18Gy), 21 Gy (V_21Gy), and 23 Gy (V_23Gy).

Delivery efficiency and quality assurance (QA)

To compare treatment efficiency, the total number of segments, BOT, and total MU were recorded for each plan. BOT included gantry rotation time and segment transition intervals. Each plan underwent pretreatment QA using an MR-compatible ArcCHECK (Sun Nuclear, Melbourne, FL, USA) on the MR-Linac system. Gamma analysis was applied for patient-specific dose verification, using criteria of 2%/1 mm, 2%/2 mm, and 3%/2 mm with a global dose threshold of 10%. A GPR (3%/2 mm) of ≥95% was considered clinically acceptable for HSRT at West China Hospital.

Dose calculation robustness

To evaluate robustness of dose calculation against spatial density misalignment, each plan was reoptimized after deliberately introducing positional uncertainties in skull placement. Ulin et al. (26) evaluated CT-MRI registration accuracy in cranial imaging by distributing MR and CT datasets to 45 institutions, each performing registration using their routine clinical methods. They reported a mean inherent registration uncertainty of 2.2 mm [one standard deviation (SD)]. In addition, the Elekta Unity online treatment planning system supports only three-degree-of-freedom (3DoF) registration, which corrects translational misalignment but does not account for rotational errors. Based on these considerations, the ±2 mm translational offsets simulated in the present study were selected to represent clinically realistic rigid registration uncertainties in cranial radiotherapy. Skull misalignment was simulated through randomly generated translational offsets ranging from −2 to 2 mm along the left-right, anterior-posterior, and superior-inferior axes. For each patient, identical offsets were consistently applied across the four planning configurations (9FL-IMRT, 9FS-IMRT, 15FL-IMRT, and 15FS-IMRT) to ensure a fair comparison between techniques, whereas offsets varied among patients.

Plan optimization continued until achieving target coverage equivalent to the original plan or until further optimization showed no meaningful improvement. The reoptimized plans were recalculated on the original CT dataset using voxel-based electron density as the reference standard (gold standard). Dose calculation robustness was quantified by global gamma analysis, comparing bulk density-based dose distributions under spatial misalignment to the CT voxel-based reference dose. GPRs were assessed using criteria of 3%/2 mm, 2%/2 mm, 2%/1 mm, and 1%/1 mm, employing a 10% low-dose cutoff.

Statistical analysis

All statistical analyses were performed using SPSS software (version 25.0; IBM, New York, USA). Quantitative data are expressed as mean ± standard deviation. Data normality was tested using the Shapiro-Wilk test. Parameters meeting normality (P>0.05) were analyzed using repeated-measures analysis of variance (RM-ANOVA) to assess differences among techniques. For parameters violating normality, the non-parametric Friedman test was used, followed by pairwise comparisons conducted with the Wilcoxon signed-rank test employing Bonferroni correction. A two-sided P value <0.05 was considered statistically significant.

Results

All 80 IMRT plans developed in this study produced conformal dose distributions, meeting established clinical criteria for acceptability. Target coverage and overall plan quality were consistently high among the four evaluated configurations (9FL-IMRT, 9FS-IMRT, 15FL-IMRT, and 15FS-IMRT). As indicated in Table 2, there were no statistically significant differences observed in GTV D_98%, GTV D_max, or the PTV-related indices (CI, GI, HI) across the four techniques (all P>0.05). Typical dose distributions in axial, coronal, and sagittal views, along with DVHs from one representative patient, appear in Figure 1, illustrating nearly identical dose patterns for all IMRT configurations.

Figure 1 Representative dose distributions and DVHs for a patient with a solitary medium-sized brain metastasis under four IMRT planning configurations. (A) Axial, coronal, and sagittal views of dose distributions for 15FL-IMRT, 15FS-IMRT, 9FL-IMRT, and 9FS-IMRT plans. The 100% and 50% isodose lines, displayed as color-wash distributions, demonstrate highly comparable target coverage and dose fall-off characteristics across all four configurations. (B) Corresponding DVHs for the PTV and NBT illustrating nearly overlapping PTV coverage and similar dose fall-off among the four IMRT techniques. 9FL-IMRT, 9-field IMRT with a larger minimum MU per segment (15 MU/segment); 9FS-IMRT, 9-field IMRT with a smaller minimum MU per segment (5 MU/segment); 15FL-IMRT, 15-field IMRT with a larger minimum MU per segment (15 MU/segment); 15FS-IMRT, 15-field IMRT with a smaller minimum MU per segment (5 MU/segment); DVHs, dose-volume histograms; IMRT, intensity-modulated radiotherapy; MU, monitor units; NBT, normal brain tissue; PTV, planning target volume.

Although certain NBT dose parameters (V_18Gy–V_23Gy) demonstrated statistical significance in pairwise comparisons, the absolute differences remained minimal (≤1 cm³; Table 2), thus considered unlikely to have clinical relevance. However, differences were evident regarding treatment efficiency among the strategies. The 9FL-IMRT plans had the lowest MU (2,037±174 MU) and shortest BOT (6.69±0.92 min). Notably, the 15FL-IMRT maintained a similarly favorable efficiency profile, with only a minor increase in BOT (7.59±1.18 min) and a low segment count (33±7), ranking second to 9FL-IMRT without a statistically significant difference (Table 2). Conversely, 15FS-IMRT displayed substantially higher segment counts (up to 58±14) and extended BOTs (up to 8.94±1.68 min).

Despite these efficiency variations, delivery accuracy remained consistently high across all techniques. Patient-specific QA indicated mean GPRs above 95% under standard criteria (3%/2 mm), without significant inter-technique differences. These findings align closely with representative measurement–calculation comparisons presented in Figure 2. Robustness analyses for dose calculation further distinguished the planning strategies. Under spatial density misalignment, global GPRs comparing bulk density-based and CT voxel-based (gold standard) dose distributions remained high under relaxed criteria (3%/2 mm and 2%/2 mm), with minimal differences among techniques (Table 3). However, significant differences appeared under more stringent criteria (2%/1 mm and 1%/1 mm). Specifically, the 15FL-IMRT configuration demonstrated the highest GPRs (96.02%±2.93% at 2%/1 mm), significantly superior to 9FL-IMRT and 9FS-IMRT following Bonferroni correction (Table 3). Line-profile comparisons also indicated closer agreement between bulk density-based and CT voxel-based calculations for 15FL-IMRT (Figure 3), highlighting greater resilience to density-related spatial uncertainties.

Figure 2 Patient-specific delivery QA results for the four IMRT configurations, measured using ArcCHECK. Global gamma analysis was conducted with a 3%/2 mm criterion and a 10% low-dose threshold. All four planning configurations exhibited excellent agreement between measured and calculated dose distributions, with GPRs exceeding 98%, indicating consistently high delivery accuracy irrespective of field number or minimum MU per segment. 9FL-IMRT, 9-field IMRT with a larger minimum MU per segment (15 MU/segment); 9FS-IMRT, 9-field IMRT with a smaller minimum MU per segment (5 MU/segment); 15FL-IMRT, 15-field IMRT with a larger minimum MU per segment (15 MU/segment); 15FS-IMRT, 15-field IMRT with a smaller minimum MU per segment (5 MU/segment); GPRs, gamma passing rates; IMRT, intensity-modulated radiotherapy; MU, monitor units; QA, quality assurance.

Figure 3 Comparison of dose calculation robustness under simulated skull displacement between bulk density-based and CT voxel-based (gold standard) dose calculations for the four IMRT configurations. (A-D) Two-dimensional dose distributions for bulk density-based plans (left) and corresponding CT voxel-based recalculations (right) for 15FL-IMRT, 15FS-IMRT, 9FL-IMRT, and 9FS-IMRT, respectively. (E-H) Representative dose line profiles comparing bulk density-based and CT voxel-based dose distributions, evaluated using a 2%/1 mm gamma criterion with a 10% low-dose threshold. A closer agreement between bulk density-based and CT voxel-based dose profiles was observed for 15FL-IMRT, indicating enhanced robustness against density-related spatial uncertainties compared to other configurations. 9FL-IMRT, 9-field IMRT with a larger minimum MU per segment (15 MU/segment); 9FS-IMRT, 9-field IMRT with a smaller minimum MU per segment (5 MU/segment); 15FL-IMRT, 15-field IMRT with a larger minimum MU per segment (15 MU/segment); 15FS-IMRT, 15-field IMRT with a smaller minimum MU per segment (5 MU/segment); CT, computed tomography; GPR, gamma passing rate; IMRT, intensity-modulated radiotherapy; MU, monitor units.

Discussion

MR-guided radiotherapy represents a significant advancement in precision radiotherapy (27). Several studies have demonstrated the feasibility and clinical potential of MR-Linac-based SRS and HSRT for BMs (14,28-31). However, few investigations have specifically explored planning optimization tailored to MR-guided workflows for HSRT in BMs. This study systematically evaluated four clinically relevant IMRT configurations (15FS-IMRT, 15FL-IMRT, 9FL-IMRT, and 9FS-IMRT) using the 1.5 T Elekta Unity MR-Linac for solitary medium-sized BMs. Among these, 15FL-IMRT provided the most favorable balance by improving dose calculation robustness while maintaining comparable plan quality, treatment efficiency, and delivery accuracy. By incorporating robustness into traditional plan assessment, the current findings offer practical, platform-specific guidance for optimizing cranial HSRT within MR-guided adaptive workflows.

For smaller lesions, the impact of field number and MU/segment may be less pronounced, as plan quality is generally less sensitive to beam arrangement due to simpler geometry and smaller target volumes. In contrast, for larger targets, increasing the number of fields may offer greater dosimetric benefits in terms of CI and GI, although the influence of density-related uncertainties may become more complex. For multiple BMs, particularly in single-isocenter multi-target treatments, geometric uncertainties—such as rotational errors and off-isocenter effects—may be amplified (32,33). In this context, planning strategies that enhance robustness, including the use of higher field numbers, may be more critical, although the trade-off with delivery efficiency must be carefully balanced. Overall, these considerations are hypothesis-driven, and dedicated studies are required to validate the applicability of the present findings across different clinical scenarios.

Although cone beam computed tomography (CBCT)-guided VMAT remains a widely used and highly efficient approach for cranial HSRT (34,35), MR-guided HSRT may offer distinct advantages in selected clinical scenarios. Because the skull is relatively rigid, skull-based CBCT registration can generally provide reliable setup correction (36). However, CBCT has limited soft-tissue contrast and often cannot directly visualize BMs or peritumoral changes. In contrast, MRI enables superior visualization of intracranial soft tissue, tumor boundaries, and edema (37-40). Importantly, previous studies have demonstrated that BMs may exhibit spatial shifts over time, which are associated with changes in peritumoral edema (38). Therefore, MR-guided adaptive workflows may be particularly valuable in situations where edema evolution, tumor-volume variation, or subtle anatomical changes affect target position or dose coverage. Furthermore, unlike conventional CBCT-guided VMAT workflows, MR-guided ATS strategies enable online contour modification and replanning based on daily anatomy, providing a mechanism to account for such dynamic changes. This adaptive capability represents a key advantage of MR-guided HSRT.

However, these advantages must be considered in the context of the inherent technical constraints of current MR-Linac systems. Unlike conventional Linacs equipped with VMAT, the Elekta Unity MR-Linac only supports step-and-shoot IMRT delivery at a relatively low dose rate (425 MU/min). These technical limitations can significantly extend BOT, thus increasing the potential for intrafractional uncertainties during HSRT of BMs. Therefore, strategies that enhance treatment efficiency without compromising plan quality are essential in MR-guided adaptive workflows. Increasing beam directions in IMRT planning is generally recognized as a means to enhance dosimetric quality. Previous studies demonstrated that increased field numbers improve target conformity and OAR sparing across various disease sites (18-20), including nasopharyngeal, pelvic, and head-and-neck cancers, though often with increased MU and prolonged delivery times. Consistent with these studies, our results revealed slightly better NBT sparing with 15FS-IMRT compared to 9FS-IMRT (Table 2). However, the results also indicated that increasing beam directions alone does not necessarily achieve an optimal balance between treatment efficiency and plan quality. Additionally, previous studies reported that increasing the minimum MU/segment could reduce total segment numbers (21), thereby shortening BOT. Motivated by these observations, we combined increased beam directions with higher minimum MU/segment (15FL-IMRT). This approach maintained plan quality and delivery accuracy, while significantly improving treatment efficiency compared to 15FS-IMRT (Table 2, Figure 2).

Despite advances in HSRT, RN remains the predominant late toxicity in patients with BMs. Previous studies have shown a strong correlation between RN risk and the volume of NBT receiving high-dose radiation (41-43), particularly at V_18Gy, V_21Gy, and V_23Gy. In our study, absolute differences in V_18Gy–V_23Gy among the four planning strategies were minimal (Table 3; ≤1 cm³) and therefore unlikely to be clinically significant. These findings suggest the 15FL-IMRT approach, combining higher beam numbers with increased minimum MU/segment, does not compromise NBT sparing. The limited dosimetric differences observed might be attributed to the near-spherical geometry of the target, which reduces sensitivity to variations in beam arrangement.

Regarding delivery accuracy, all four strategies exhibited excellent patient-specific QA outcomes, with mean global GPRs exceeding 97% at 3%/2 mm and 96% at 2%/2 mm, consistently meeting institutional acceptance standards (Table 2). The uniformly reduced GPRs observed under stricter criteria (2%/1 mm) likely reflect the relatively coarse detector spacing of ArcCHECK and inherent measurement variability. Supporting this interpretation, Lee et al. (44) reported higher GPRs using SRS MapCHECK compared to ArcCHECK across multiple gamma criteria for single-lesion SRS. Similarly, James et al. (45) found that ArcCHECK GPRs decreased rapidly with tighter gamma criteria, suggesting increased sensitivity to stringent thresholds during stereotactic verification. Nevertheless, several limitations of ArcCHECK-based QA should be considered. Because measurements occur in a homogeneous phantom, ArcCHECK primarily evaluates plan deliverability and agreement between calculated and measured doses under idealized phantom conditions rather than patient-specific anatomical dose distributions.

In the context of MRI-based treatment planning, additional uncertainties must be considered, including geometric inaccuracies and segmentation errors arising from voxel ambiguity, partial volume effects, and image quality (46). Consequently, conventional verification methods may not fully capture errors associated with electron density assignment, anatomical heterogeneity, and structure propagation in adaptive workflows. To address this issue, the present study specifically evaluated the robustness of dose calculation against density-related spatial misalignment. This analysis demonstrated that both 15FL-IMRT and 15FS-IMRT configurations achieved higher GPRs compared to 9FL-IMRT and 9FS-IMRT under the stringent 1%/1 mm criterion. This improvement is likely due to the increased beam directions in the 15-field configurations, which mitigate dosimetric disturbances caused by local tissue heterogeneities within the brain by averaging density-related perturbations across multiple beam paths. Notably, simulated skull displacement was integrated into the plan optimization using bulk density override. These optimized plans were then recalculated without density override to produce CT voxel-based reference doses. This approach closely reflects clinical MR-guided adaptive radiotherapy, in which bulk density assignment is inherently involved in both optimization and online dose calculation, while the true electron density distribution remains unknown. Optimization error, defined as inaccuracies introduced during inverse planning from incorrect electron density, is often overlooked in robustness assessments. Previous work (47) has indicated that such errors could exceed dose calculation errors alone, particularly in structures with lower optimization priority.

In MR-guided adaptive workflows, image registration may vary between therapists due to differences in experience and interpretation of anatomical landmarks (48). Such variability can introduce subtle discrepancies in structure propagation, which may ultimately affect dose distribution, particularly when bulk density-based dose calculation is employed. Under simulated translational perturbations within ±2 mm, dose calculation robustness remained consistently high across all planning configurations under conventional gamma criteria (3%/2 mm and 2%/2 mm). However, under more stringent criteria (2%/1 mm and 1%/1 mm), differences among techniques became apparent, with 15FL-IMRT consistently achieving the highest GPRs. Taken together, these findings suggest that the improved robustness observed under simulated translational perturbations may render 15FL-IMRT less sensitive to variations in structure propagation. Consequently, enhanced robustness may contribute to more consistent target coverage across different operators and adaptive sessions, thereby improving the overall reliability of MR-guided HSRT. Notably, applying identical spatial offsets across all planning configurations within each patient effectively isolated technique-related effects from random variations, thereby improving the fairness and interpretability of robustness comparisons.

Several limitations of this retrospective study should be acknowledged. First, the sample size was relatively small (n=20), and clinical outcome data, such as local control and treatment-related toxicity, were not assessed. However, it should be noted that this study employed a within-patient comparative design, in which multiple planning configurations were generated and evaluated for each individual patient. This approach reduces inter-patient variability and enables robust relative comparisons between different planning strategies. Nevertheless, the limited cohort size may still restrict the generalizability of the findings, and further validation in larger patient populations is warranted. Second, robustness analysis focused only on rigid skull translations, while clinical workflows might also involve rotational components and subtle local deformations; future studies should include full six-degree-of-freedom perturbations. Third, the analysis concentrated mainly on density uncertainties related to high-density cranial bone. Uncertainties associated with low-density structures were not explicitly examined and require further study. Fourth, despite the higher delivery efficiency typically associated with VMAT, a direct comparison between MR-guided IMRT and CBCT-guided VMAT was beyond the scope of the present study and warrants investigation in future dedicated studies. Finally, the current study was limited to solitary, medium-sized BMs (2–3 cm), and the applicability of this strategy to smaller lesions, larger targets, or multiple BMs remains unclear. Despite these limitations, the findings provide valuable insights for optimizing efficient and robust MR-guided HSRT planning on the Unity MR-Linac platform.

Conclusions

In summary, 15FL-IMRT achieved a favorable balance among plan quality, treatment efficiency, and dose calculation robustness for MR-guided HSRT of solitary medium-sized BMs. This strategy may facilitate more efficient and reliable clinical implementation of MR-guided adaptive HSRT. However, as this study focused exclusively on solitary medium-sized lesions (2–3 cm), the applicability of these findings to other clinical scenarios requires further validation. Further prospective studies with larger and more diverse patient populations are warranted to validate their generalizability and clinical relevance.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1-2798/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1-2798/dss

Funding: This work was supported by 1·3·5 Project for Disciplines of Excellence-Clinical Research Fund, West China Hospital, Sichuan University (grant No. 2025HXFH036), Sichuan University “From 0 to 1” Innovative Research Program (grant No. 2023SCUH0035), Sichuan Science and Technology Program (grant No. 2024YFFK0147), and Chengdu Science and Technology Program (grant No. 2024-YF05-01040-SN).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1-2798/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the ethics committee of West China Hospital (approval No. 2024124) and individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Villarreal-Garza C, de la Mata D, Zavala DG, Macedo-Perez EO, Arrieta O. Aggressive treatment of primary tumor in patients with non-small-cell lung cancer and exclusively brain metastases. Clin Lung Cancer 2013;14:6-13. [Crossref] [PubMed]
2. Sperduto PW, Yang TJ, Beal K, Pan H, Brown PD, Bangdiwala A, et al. Estimating Survival in Patients With Lung Cancer and Brain Metastases: An Update of the Graded Prognostic Assessment for Lung Cancer Using Molecular Markers (Lung-molGPA). JAMA Oncol 2017;3:827-31. [Crossref] [PubMed]
3. Lin X, DeAngelis LM. Treatment of Brain Metastases. J Clin Oncol 2015;33:3475-84. [Crossref] [PubMed]
4. Barnholtz-Sloan JS, Sloan AE, Davis FG, Vigneau FD, Lai P, Sawaya RE. Incidence proportions of brain metastases in patients diagnosed (1973 to 2001) in the Metropolitan Detroit Cancer Surveillance System. J Clin Oncol 2004;22:2865-72. [Crossref] [PubMed]
5. Brown PD, Ballman KV, Cerhan JH, Anderson SK, Carrero XW, Whitton AC, Greenspoon J, Parney IF, Laack NNI, Ashman JB, Bahary JP, Hadjipanayis CG, Urbanic JJ, Barker FG 2nd, Farace E, Khuntia D, Giannini C, Buckner JC, Galanis E, Roberge D. Postoperative stereotactic radiosurgery compared with whole brain radiotherapy for resected metastatic brain disease (NCCTG N107C/CEC·3): a multicentre, randomised, controlled, phase 3 trial. Lancet Oncol 2017;18:1049-60.
6. Fokas E, Henzel M, Surber G, Kleinert G, Hamm K, Engenhart-Cabillic R. Stereotactic radiosurgery and fractionated stereotactic radiotherapy: comparison of efficacy and toxicity in 260 patients with brain metastases. J Neurooncol 2012;109:91-8. [Crossref] [PubMed]
7. Lee EJ, Choi KS, Park ES, Cho YH. Single- and hypofractionated stereotactic radiosurgery for large (> 2 cm) brain metastases: a systematic review. J Neurooncol 2021;154:25-34. [Crossref] [PubMed]
8. Harikar MM, Venkataram T, Palmisciano P, Scalia G, Baldoncini M, Cardali SM, Umana GE, Ferini G. Comparison of Staged Stereotactic Radiosurgery and Fractionated Stereotactic Radiotherapy in Patients with Brain Metastases > 2 cm without Prior Whole Brain Radiotherapy: A Systematic Review and Meta-Analysis. World Neurosurg 2023;178:213-232.e6. [Crossref] [PubMed]
9. Minniti G, Scaringi C, Paolini S, Lanzetta G, Romano A, Cicone F, Osti M, Enrici RM, Esposito V. Single-Fraction Versus Multifraction (3 × 9 Gy) Stereotactic Radiosurgery for Large (>2 cm) Brain Metastases: A Comparative Analysis of Local Control and Risk of Radiation-Induced Brain Necrosis. Int J Radiat Oncol Biol Phys 2016;95:1142-8. [Crossref] [PubMed]
10. Alongi F, Rigo M, Figlia V, Cuccia F, Giaj-Levra N, Nicosia L, Ricchetti F, Sicignano G, De Simone A, Naccarato S, Ruggieri R, Mazzola R. 1.5 T MR-guided and daily adapted SBRT for prostate cancer: feasibility, preliminary clinical tolerability, quality of life and patient-reported outcomes during treatment. Radiat Oncol 2020;15:69. [Crossref] [PubMed]
11. Boeke S, Mönnich D, van Timmeren JE, Balermpas P. MR-Guided Radiotherapy for Head and Neck Cancer: Current Developments, Perspectives, and Challenges. Front Oncol 2021;11:616156. [Crossref] [PubMed]
12. Crockett CB, Samson P, Chuter R, Dubec M, Faivre-Finn C, Green OL, Hackett SL, McDonald F, Robinson C, Shiarli AM, Straza MW, Verhoeff JJC, Werner-Wasik M, Vlacich G, Cobben D. Initial Clinical Experience of MR-Guided Radiotherapy for Non-Small Cell Lung Cancer. Front Oncol 2021;11:617681. [Crossref] [PubMed]
13. Eijkelenkamp H, Boekhoff MR, Verweij ME, Peters FP, Meijer GJ, Intven MPW. Planning target volume margin assessment for online adaptive MR-guided dose-escalation in rectal cancer on a 1.5 T MR-Linac. Radiother Oncol 2021;162:150-5. [Crossref] [PubMed]
14. Zheng S, Ding S, Liu B, Xiong Y, Zhou R, Zhang P, Liu F, Liu Y, Chen M, Situ Y, Wang M, Huang X, Yin S, Fang W, Mou Y, Qiu B, Wang D, Liu H. Efficacy and safety of online adaptive magnetic resonance-guided fractionated stereotactic radiotherapy for brain metastases in non-small cell lung cancer (GASTO-1075): a single-arm, phase 2 trial. EClinicalMedicine 2025;82:103189. [Crossref] [PubMed]
15. Torizuka D, Uto M, Takehana K, Mizowaki T. Dosimetric comparison among dynamic conformal arc therapy, coplanar and non-coplanar volumetric modulated arc therapy for single brain metastasis. J Radiat Res 2021;rrab092.
16. Kristensen BH, Laursen FJ, Løgager V, Geertsen PF, Krarup-Hansen A. Dosimetric and geometric evaluation of an open low-field magnetic resonance simulator for radiotherapy treatment planning of brain tumours. Radiother Oncol 2008;87:100-9. [Crossref] [PubMed]
17. Jonsson JH, Karlsson MG, Karlsson M, Nyholm T. Treatment planning using MRI data: an analysis of the dose calculation accuracy for different treatment regions. Radiat Oncol 2010;5:62. [Crossref] [PubMed]
18. Ibrahim MS, Attalla EM, El Naggar M, Elshemey WM. Is 9-field IMRT superior to 7-field IMRT in the treatment of nasopharyngeal carcinoma? Indian J Cancer 2020;57:388-92. [Crossref] [PubMed]
19. Alfitra LS. Suwandi, Arif I. IMRT Planning Using 5, 7 and 9 Radiation Field Directions for Nasopharyngeal Cancer Cases. Journal of Physics: Conference Series 2024;2734:012001.
20. Phat LT, Thao MT, Kien TT, Tai DT, Sandwall P, Sulieman A, Tamam N, Yani S, Chow JCL. Dosimetric analysis of beam number variations in IMRT for head-and-neck, breast, and pelvic cancers using Halcyon. Radiation Physics and Chemistry 2025;236:112755.
21. Chen X, Cheng Y, Wu W, He Y, Zhang Y, Han L, He G, Li Z. Effects of minimum monitor unit per dynamic control point on intensity-modulated radiotherapy planning for nasopharyngeal carcinoma: A retrospective study. Med Dosim 2025;50:96-101. [Crossref] [PubMed]
22. McNiven AL, Sharpe MB, Purdie TG. A new metric for assessing IMRT modulation complexity and plan deliverability. Med Phys 2010;37:505-15. [Crossref] [PubMed]
23. Shaw E, Kline R, Gillin M, Souhami L, Hirschfeld A, Dinapoli R, Martin LRadiation Therapy Oncology Group. radiosurgery quality assurance guidelines. Int J Radiat Oncol Biol Phys 1993;27:1231-9. [Crossref] [PubMed]
24. Paddick I, Lippitz B. A simple dose gradient measurement tool to complement the conformity index. J Neurosurg 2006;105:194-201. [Crossref] [PubMed]
25. Wang X, Zhang X, Dong L, Liu H, Gillin M, Ahamad A, Ang K, Mohan R. Effectiveness of noncoplanar IMRT planning using a parallelized multiresolution beam angle optimization method for paranasal sinus carcinoma. Int J Radiat Oncol Biol Phys 2005;63:594-601. [Crossref] [PubMed]
26. Ulin K, Urie MM, Cherlow JM. Results of a multi-institutional benchmark test for cranial CT/MR image registration. Int J Radiat Oncol Biol Phys 2010;77:1584-9. [Crossref] [PubMed]
27. Otazo R, Lambin P, Pignol JP, Ladd ME, Schlemmer HP, Baumann M, Hricak H. MRI-guided Radiation Therapy: An Emerging Paradigm in Adaptive Radiation Oncology. Radiology 2021;298:248-60. [Crossref] [PubMed]
28. Zou S, Li A, Yu H, Xiao Q, Zhu S, He J, Wei W, Zeng X, Lai J. Evaluation of single-isocenter multiple-target IMRT using 1.5T Elekta Unity MR-guided linear accelerator for stereotactic radiosurgery of multiple brain metastases. Journal of Radiation Research and Applied Sciences 2025;18:101245.
29. Guerini AE, Nici S, Magrini SM, Riga S, Toraci C, Pegurri L, Facheris G, Cozzaglio C, Farina D, Liserre R, Gasparotti R, Ravanelli M, Rondi P, Spiazzi L, Buglione M. Adoption of Hybrid MRI-Linac Systems for the Treatment of Brain Tumors: A Systematic Review of the Current Literature Regarding Clinical and Technical Features. Technol Cancer Res Treat 2023;22:15330338231199286. [Crossref] [PubMed]
30. Tseng CL, Eppinga W, Seravalli E, Hackett S, Brand E, Ruschin M, Lee YK, Atenafu EG, Sahgal A. Dosimetric feasibility of the hybrid Magnetic Resonance Imaging (MRI)-linac System (MRL) for brain metastases: The impact of the magnetic field. Radiother Oncol 2017;125:273-9. [Crossref] [PubMed]
31. Seravalli E, Sierts M, Brand E, Maspero M, David S, Philippens MEP, Voormolen EHJ, Verhoeff JJC. Dosimetric feasibility of direct post-operative MR-Linac-based stereotactic radiosurgery for resection cavities of brain metastases. Radiother Oncol 2023;179:109456. [Crossref] [PubMed]
32. Lai J, Liu S, Liu J, Fu M, Jiang M, Li A, Li B, Li X, Cheng X, Zhou L. Going from 3D/3D to 2D/3D registration for noncoplanar setup verification in intracranial single-isocenter multiple-target hypofractionated stereotactic radiotherapy: comparison between kilo-voltage/mega-voltage image pairs and noncoplanar cone-beam computed tomography. Quant Imaging Med Surg 2023;13:8094-106. [Crossref] [PubMed]
33. Roper J, Chanyavanich V, Betzel G, Switchenko J, Dhabaan A. Single-Isocenter Multiple-Target Stereotactic Radiosurgery: Risk of Compromised Coverage. Int J Radiat Oncol Biol Phys 2015;93:540-6. [Crossref] [PubMed]
34. Lai JL, Liu SP, Jiang XX, Liu J, Li A, Li B, Li XK, Ye XJ, Lei KJ, Zhou L. Can Optical Surface Imaging Replace Non-coplanar Cone-beam Computed Tomography for Non-coplanar Set-up Verification in Single-isocentre Non-coplanar Stereotactic Radiosurgery and Hypofractionated Stereotactic Radiotherapy for Single and Multiple Brain Metastases? Clin Oncol (R Coll Radiol) 2023;35:e657-65. [Crossref] [PubMed]
35. Chen H, Li Z, Liao G, Wang L, Deng X, Zhong H. HyperArc performance in radiotherapy and its correlations with number and volume of multiple brain metastases. Radiation Medicine and Protection 2024;5:139-44.
36. Guckenberger M, Baier K, Guenther I, Richter A, Wilbert J, Sauer O, Vordermark D, Flentje M. Reliability of the bony anatomy in image-guided stereotactic radiotherapy of brain metastases. Int J Radiat Oncol Biol Phys 2007;69:294-301. [Crossref] [PubMed]
37. Baris MM, Celik AO, Gezer NS, Ada E. Role of mass effect, tumor volume and peritumoral edema volume in the differential diagnosis of primary brain tumor and metastasis. Clin Neurol Neurosurg 2016;148:67-71. [Crossref] [PubMed]
38. Hessen ED, van Buuren LD, Nijkamp JA, de Vries KC, Kong Mok W, Dewit L, van Mourik AM, Berlin A, van der Heide UA, Borst GR. Significant tumor shift in patients treated with stereotactic radiosurgery for brain metastasis. Clin Transl Radiat Oncol 2017;2:23-8. [Crossref] [PubMed]
39. Kim MS, Park SH, Park ES, Park JB, Kwon SC, Lyo IU, Sim HB. Quantitative analysis in peritumoral volumes of brain metastases treated with stereotactic radiotherapy. J Neuroradiol 2018;45:310-5. [Crossref] [PubMed]
40. Maziero D, Straza MW, Ford JC, Bovi JA, Diwanji T, Stoyanova R, Paulson ES, Mellon EA. MR-Guided Radiotherapy for Brain and Spine Tumors. Front Oncol 2021;11:626100.
41. Gondi V, Bauman G, Bradfield L, Burri SH, Cabrera AR, Cunningham DA, Eaton BR, Hattangadi-Gluth JA, Kim MM, Kotecha R, Kraemer L, Li J, Nagpal S, Rusthoven CG, Suh JH, Tomé WA, Wang TJC, Zimmer AS, Ziu M, Brown PD. Radiation Therapy for Brain Metastases: An ASTRO Clinical Practice Guideline. Pract Radiat Oncol 2022;12:265-82. [Crossref] [PubMed]
42. Patwe PT, Deshpande SS, Mahajan GR. Stereotactic and fractionated stereotactic radiosurgery for single and multiple brain metastases: Results of multicenter planning studies. Phys Med 2025;132:104950. [Crossref] [PubMed]
43. Upadhyay R, Ayan AS, Jain S, Klamer BG, Perlow HK, Zoller W, Blakaj DM, Beyer S, Grecula J, Arnett A, Thomas E, Chakravarti A, Raval RR, Palmer JD. Dose-Volume Tolerance of the Brain and Predictors of Radiation Necrosis After 3-Fraction Radiosurgery for Brain Metastases: A Large Single-Institutional Analysis. Int J Radiat Oncol Biol Phys 2024;118:275-84. [Crossref] [PubMed]
44. Lee YC, Kim Y. A patient-specific QA comparison between 2D and 3D diode arrays for single-lesion SRS and SBRT treatments. J Radiosurg SBRT 2021;7:295-307.
45. James S, Al-Basheer A, Elder E, Huh C, Ackerman C, Barrett J, Hamilton R, Mostafaei F. Evaluation of commercial devices for patient specific QA of stereotactic radiotherapy plans. J Appl Clin Med Phys 2023;24:e14009. [Crossref] [PubMed]
46. Hsu SH, Zawisza I, O'Grady K, Peng Q, Tomé WA. Towards abdominal MRI-based treatment planning using population-based Hounsfield units for bulk density assignment. Phys Med Biol 2018;63:155003. [Crossref] [PubMed]
47. Whelan B, Kumar S, Dowling J, Begg J, Lambert J, Lim K, Vinod SK, Greer PB, Holloway L. Utilising pseudo-CT data for dose calculation and plan optimization in adaptive radiotherapy. Australas Phys Eng Sci Med 2015;38:561-8. [Crossref] [PubMed]
48. Faccenda V, Panizza D, Pinzi V, Carminati S, Colciago RR, Mascellani SA, Bordigoni B, Manno I, Mombelli M, Pierini VE, Mateo AR, De Martin E, De Ponti E, Arcangeli S. Expert variability as a benchmark for validating automatic CT-MRI registration in brain stereotactic radiosurgery. Phys Med 2026;145:105765. [Crossref] [PubMed]