Abstract
Background/Aim: To assess the impact of the width of multileaf collimator (MLC) on dose distributions on HyperArc fractionated stereotactic irradiation for multiple (5-10) brain metastases. Patients and Methods: Twenty-one HyperArc (HA) plans were generated using the high definition (HD) MLC (2.5 mm) to deliver 30-35 Gy in 3-5 fractions (HA-HD). The HyperArc plans using Millennium (ML) MLC (5 mm) were retrospectively generated (HA-ML) using the same planning parameters with HA-HD. Dosimetric parameters between the planning target volume (PTV) and organs at risk (OARs) were compared. Results: The conformity index was significantly higher (p<0.0001) in the HA-HD plans (0.95±0.04) than that in the HA-ML plans (0.92±0.06). The HA-HD provided significantly lower (p<0.0001) gradient index (5.6±2.5) than HA-ML (6.2±3.5). For the brainstem and retina (right), a statistically significant difference (p<0.05) was observed between the HA-HD (12.8±10.9 and 2.8±1.7 Gy, for brainstem and retina, respectively) and HA-ML (13.6±11.1 and 3.0±1.8 Gy) plans. For the brain tissue, the HA-HD plans statistically significantly reduced dosimetric parameters (p<0.0001) in all evaluated dose range (V6Gy-V28Gy). Conclusion: The narrower MLC provided significantly higher conformity, steeper dose gradient, and better normal tissue sparing.
Owing to the improvement of systemic therapy and the advances in magnetic resonance imaging modality, the incidence of the brain metastases, which can be a direct cause of death, is increasing (1). Whole brain irradiation has been considered to be standard radiotherapy, while, approximately 90% patients showed impairment of one or more neurocognitive tests at baseline (2). The stereotactic irradiation (STI), which delivers high dose radiation in small fractions, is expected to reduce radiation-induced side effect without compromising treatment outcome (3-5). Yamamoto et al., showed that the overall survival for patients with 5-10 brain metastases was non-inferior to those with 2-4 brain metastases and they concluded that the STI might be a suitable alternative approach to the whole brain irradiation (6).
The volumetric imaging with the cone-beam computed tomography and the continuous movement of gantry and multileaf collimator (MLC) can achieve the precise and highly conformal STI using the C-arm linear accelerator (7, 8). Because brain metastases are surrounded by normal brain tissues, the rapid dose falloff from the target surface is required for the STI. For multiple brain metastases, the recent advanced treatment approach, named HyperArc, using the C-arm linear accelerator has advantages for of reducing the treatment time and delivering the conformity dose for targets compared with the other treatment units such as GammaKnife and CyberKnife (9-11).
To generate the complicated dose distribution, the width of leaves projected to the isocenter is quite important. Dhabaan et al. demonstrated that the narrower width of MLC provided higher conformity dose for targets and lower radiation dose for normal tissues in the dynamic conformal arc therapy for intracranial lesions (12). Similar results were reported by Abisheva et al. in volumetric modulated arc therapy (VMAT) for multiple brain metastases (13). Because HyperArc utilized more complex MLC patterns than conventional VMAT to generate the steep dose gradient for multiple targets (9), the narrower MLC has potential for providing the ideal dose distribution.
The aim of this study was to assess the impact of the width of MLC on the dosimetric parameters for targets and OARs in the HyperArc plans (fractionated STI) for patients with multiple (5-10) brain metastases.
Patients and Methods
Patients and simulation. This retrospective study included twenty-one patients with multiple (5-10) brain metastases who underwent fractionated STI, and was approved by the ethics committee of our institution. Table I lists the patients’ characteristics. For simulation, the patient was immobilized using the thermoplastic mask, and was scanned using a dual-energy computed tomography (DECT) system (Revolution HD; GE Medical Systems, Milwaukee, WI, USA). The following scanning parameters were used: tube voltage of 80/140 kVp, tube current of 550-600 mA, gantry rotation speed of 0.8 s/rot, helical pitch of 0.531:1. The virtual monochromatic images (VMIs) at 77 keV were reconstructed with a slice thickness of 1 mm and a field of view of 320 mm.
Patient characteristics.
Treatment planning. The VMIs were loaded into a treatment planning system (Eclipse, version 15.6; Varian Medical Systems, Palo Alto, CA, USA). Gross tumor volume (GTV) was delineated referring to a T1-weighted magnetic resonance image (slice thickness of 1 mm) with contrast enhanced medium (gadolinium). A planning target volume was formed by adding a 1 mm isotropic margin to the GTV. The clinical HyperArc plans were designed based on a TrueBeam STX or Edge linear accelerator equipped with a high-definition MLC with a leaf width of 2.5 mm (HA-HD) (14), and a photon energy of 6X (flattening filtered or flattening filter free) was used. The prescription dose of 30-35 Gy was delivered in 3-5 fractions to cover the 95% volume of the combined PTV. The position of isocenter, collimator angles and non-coplanar beam arrangement were automatically determined. In the optimization process, the inhomogeneity dose within the PTV was allowed and the doses to brain tissue were reduced as low as possible. Optimal resolution of 1.25 mm, dose calculation grid size of 1.25 mm and analytic anisotropic algorithm dose calculation algorithm were used.
Retrospectively, the clinical HyperArc plans (HA-HD) were modified to generate the HyperArc plans using a Millennium MLC with a leaf width of 5 mm (HA-ML) (15). Except for the width of MLC, the same treatment planning parameters (prescription dose, isocenter, collimator angles, optimization parameter, etc.) as the HA-HD plans were used for generating the HA-ML.
Data analysis. To assess the impact of MLC width on dose distributions, dosimetric parameters for the targets and organs at risk (OARs) between the HA-HD and HA-ML were compared. The conformity index (CI) was defined as follows: CI=TVpr/BVpr, where TVpr and BVpr indicate the volume of the target and body covered by the prescription dose, respectively (16). The gradient index (GI) was calculated as: GI=BV50%/BVpr, where BV50% represents the volume of 50% of the prescription isodose. The homogeneity index (HI) was calculated as: Dmax/Dpr, where respective Dmax and Dpr denote the maximum dose and prescribed dose (17). For OARs (brainstem, optic chiasm, optic nerves, retinas and lens), the Dmax was assessed. For brain tissue excluding the PTV, the volumes receiving a specific dose ranging from 6 Gy to 28 Gy (V6Gy-V28Gy), in 2 Gy increments, were evaluated. Subsequently, differences in dosimetric parameters for brain tissue were calculated for each patient.
The dosimetric parameters for the PTV and OARs between the HA-HD and HA-ML were compared using the paired Wilcoxon’ signed-rank test (SPSS, version 27; IBM, Armonk, NY, USA). A value of p<0.05 was considered to be statistically significant.
Results
The direct comparison of the physical characteristics of the individual treatment plans between the HA-HD and HA-ML are shown in Figure 1. The CI was statistically significantly higher (p<0.0001) in the HA-HD plans (0.95±0.04) than that in the HA-ML plans (0.92±0.06). Moreover, the HA-HD provided a statistically significantly lower (p<0.0001) GI (5.6±2.5) than HA-ML (6.2±3.5), which indicated that a steeper dose gradient was generated in the HA-HD plans. The comparable HI (1.7±0.3 vs. 1.7±0.2 for HA-HD and HA-ML plans, respectively, p=0.54) and MU (4,735±1,203 vs. 4,715 ±1,390 MU, p=0.77) were obtained between the two treatment planning approaches.
Direct comparison of the physical characteristics of the individual treatment plans between HyperArc (HA) plans using the High definition (HA-HD) multileaf collimator (MLC) and Millennium (HA-ML) MLC.
Figure 2 illustrates the comparison of the dose distributions between the HA-HD and HA-ML plans for patients #4, 14, and 21. HA-HD provided complicated shape of dose distribution, and the dose bridge in the 10-20 Gy isodose line was reduced (arrow). Table II summarizes the dosimetric parameters for the PTV and OARs. The dosimetric parameters for the PTV were comparable (p>0.1) between the HA-HD (54.8±11.3, 40.0±4.5 and 28.1±2.8 Gy for Dmax, Dmean and Dmin, respectively) and HA-ML (54.8±10.2, 40.2±4.3 and 27.9±2.6 Gy) plans. For the brainstem and retina (right), a statistically significant difference (p<0.05) was observed between the HA-HD (12.8±10.9 and 2.8±1.7 Gy, for brainstem and retina, respectively) and HA-ML (13.6±11.1 and 3.0±1.8 Gy) plans. The dosimetric parameters were comparable (p>0.05) for the optic chiasma, optic nerves, retina (left) and lens. For the brain tissue, the HA-HD plans statistically significantly reduced dosimetric parameters (p<0.0001) in all evaluated dose ranges (V6Gy-V28Gy). Figure 3 shows the differences in dosimetric parameters for brain tissue in each patient. For almost all cases, the narrower MLC resulted in better brain tissue sparing in all evaluated ranges. The differences in V10Gy and V20Gy were 10.8±11.2 and 1.8±2.6 ml, respectively.
Comparison of dose distributions between the HyperArc (HA) plans using the High definition (HD) multileaf collimator and Millennium (ML) multileaf collimator plans for patients #4, 14, and 21.
Comparison of dosimetric parameters for PTV and OARs.
Box plots of differences in dosimetric parameters between the HyperArc-high definition (HA-HD) and HyperArc-Millennium (HA-ML) plans for brain tissue. Boxes: median value, and upper and lower quartiles; Whiskers: maximum and minimum values within a 1.5× inter-quartile range; Dots: outliers located outside the whiskers.
Discussion
This study clearly demonstrated the impact of the width of MLC on dose distributions in fractionated STI using HyperArc for patients with multiple brain metastases. Because the thermoplastic mask is generally used for patient immobilization with the acceptable intra-fractional setup errors during the STI dose delivery using the C-arm linear accelerator, the dose fractionation can be performed easier than the invasive frame-based immobilization (18). The radiobiological advantage of the fractionated STI over the single fraction STI is that the dose fractionation may reduce the side-effects on normal tissue while maintaining tumor control (19). For large brain metastases (>2 cm), Minniti et al. reported that the 1-year cumulative local control rate was 77% in the single fraction STI (20% of patients experienced radionecrosis) while the local control rate was 91% in the fractionated STI (8% experienced radionecrosis) (20). Moreover, Lucia et al. demonstrated that the inhomogeneous dose distribution within the tumor in the fractionated STI resulted in better local control and lower risk of radionecrosis compared to homogeneous distribution (21).
The quality of dose distribution depends on the characteristics of the MLCs, and there are commercially available MLCs (leaf width of 2.5-10 mm) developed by the various vendors with specific design characteristics (22-26). Lafond et al. compared the dose distributions in VMAT plans generated using 10 mm MLC with those using 4 mm MLCs for patients with head and neck cancers (27). In that report, both MLCs achieved satisfactory dose distributions for complex target, while, the 4 mm MLC provided the better dose sparing especially for the brainstem and spinal cord. Moreover, Park et al. simulated the extremely narrow MLC (1.25 mm) in the VMAT plans for patients with prostate cancer, and treatment plans using the 1.25 mm MLC provided better dose homogeneity inside the target volume, better target conformity and less dose to normal tissue near the target volume than those using 2.5 mm MLC (28). These facts imply that a narrower MLC is suitable for generating the ideal dose distributions in the VMAT plans.
Historically, the STI for brain metastasis (single isocenter for single target) was performed using multiple static/arc conformal beams using the C-arm linear accelerator, and the narrower MLC could adjust the leaf aperture to the shape of the PTV in the beam’s-eye view resulting in better PTV conformity and surrounding tissue sparing (29, 30). Recent advances in the irradiation techniques such as VMAT allows simultaneous irradiation of multiple brain metastases, and is increasingly introduced in clinical practice owing to the acceptable target conformity and tissue sparing with its short treatment time (31). The HyperArc plan is one of the most complex irradiation techniques (non-coplanar irradiation, complex MLC movement and so on) in modern radiotherapy, and the HyperArc plans provides significantly higher conformity and rapid dose falloff with respect to the conventional VMAT plans (9). In this study, we firstly demonstrated that the width of MLC had significant impact on dose distributions in HyperArc plans, and narrower MLC could generate better treatment plan quality. In this regard, the narrower MLC is supposed to be better for tumor fitting in adjacent targets resulting in less dose bridge (Figure 2). Regarding radiation necrosis after the fractionated STI (three and five fractions), Inoue et al. reported that the dosimetric parameter of V14Gy for the brain tissue could be a useful indicator for the risk evaluation of radiation necrosis (32, 33). In the report by Minniti et al., V18Gy and V21Gy were the most predictive independent risk factors for radiation necrosis in the fractionated STI (three fractions) (34). Because the HA-HD significantly reduced doses to brain tissue in the range from V6Gy to V28Gy (Table II), the narrower MLC might be suitable for HyperArc planning for multiple brain metastases to minimize radiation-induced side effects.
Several limitations of this study warrant mention. First, although the brain metastases occur at various sites in the brain and vary enormously in size, in depth analysis of the effect of tumor size and distance between the metastases on dose distributions was not performed in this study due to the limited number of patients. Second, the same optimization parameters between the HA-HD (clinical plan) and HA-ML were used in this study, whereas better HA-ML plans might be generated using different optimization parameters. Finally, the impact of the width of MLC was assessed in patients with 5-10 brain metastases in this study, whereas in clinical practice, more than 10 brain metastases can be treated with the C-arm. Despite these limitations, our quantitative data can provide useful information for selecting the suitable treatment machine for patients with multiple brain metastases.
In conclusion, our results demonstrated that the HA-HD (2.5 mm MLC) provided significantly better tumor conformity (CI) and steeper dose gradient (GI) than those in the HA-ML plans (5 mm MLC). Moreover, doses for normal brain tissue were significantly reduced by using the narrower MLC.
Acknowledgements
This study was supported by JSPS KAKENHI Grant (Grant-in-Aid for Scientific Research (C) 21K07742).
Footnotes
Authors’ Contributions
All the Authors participated in the writing this article and take responsibility for its content. The Authors confirm that the content of the manuscript has not been published, or submitted for publication elsewhere.
Conflicts of Interest
The Authors have no conflicts of interest to declare in relation to this study.
- Received April 13, 2021.
- Revision received May 4, 2021.
- Accepted May 5, 2021.
- Copyright © 2021 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
