Stereotactic radiosurgery (SRS) is a noninvasive surgical technique that delivers a high dose of radiation in a single fraction. Enhanced precision in radiation therapy has led to the advancement of SRS techniques using a linear accelerator equipped with multileaf collimators (MLCs), instead of cones, allowing applications for primary and metastatic tumors in various organs. Numerous clinical studies have confirmed the effectiveness of this technique [1].

SRS for metastatic brain tumors was introduced before the development of volumetric modulated arc therapy (VMAT), and protocols to ensure treatment outcomes have been continuously refined and implemented in clinical practice [2].

Clinical studies have shown that compared with other treatment methods for metastatic brain tumors, such as resection and whole-brain radiation therapy, SRS can offer favorable prognostic factors under certain conditions, including improved survival rates, lower recurrence rates, better local control, and preserved cognitive function [2-8]. Based on these findings, the American Society of Clinical Oncology has developed and recommended guidelines for SRS applications in treating metastatic brain tumors [9].

Clinical studies have also examined the side effects of high doses, focusing on the correlation between the radiation dose delivered to normal brain tissue and severe complications, such as radiation necrosis [10-13]. To minimize such side effects of SRS, both the clinical and physical characteristics of the patients must be considered [2].

Recent studies have shown that, for single brain metastases, applying MLC-based dynamic conformal arc (DCA) therapy can reduce the dose delivered to normal brain tissue compared with the cone-based CyberKnife (Accuray, Inc.) [14]. Studies comparing DCA and VMAT in stereotactic radiation therapy (SRT) have shown no significant differences in the clinical characteristics between the two methods. However, in terms of physical characteristics, DCA demonstrated a higher planning target volume (PTV) coverage and gradient index, whereas VMAT showed a higher conformity index (CI) [15].

To maximize the effectiveness of SRS, the alignment between the tumor location defined in the planning and the patient’s position during treatment must be precise [16-18]. Unlike traditional SRS, which involves invasively securing the head with pins, in our institution, SRS is performed using a frameless fixation system. Therefore, to ensure positional accuracy, alignment was verified with cone-beam computed tomography (CBCT) before and after treatment.

Various indices have been developed and applied in clinical practice to quantitatively assess the quality of SRS treatment plans [19-21]. However, each index provides limited information depending on the parameters used, making it challenging to quantitatively assess the overall quality of a treatment plan with a single index. Therefore, the process of how each index can complement each other must be analyzed [19].

This study aimed to compare the dose characteristics of MLC-based VMAT, used at our institution for treating metastatic brain tumors, with DCA through a retrospective analysis of previously treated patients. Multiple metastases were excluded from the analysis because of the inherent difficulty in equitably comparing DCA and VMAT in such cases.

1. Patients

This study analyzed 12 patients with single metastatic brain tumors who were treated with VMAT at Kosin University Gospel Hospital. Although multitarget SRS cases were also treated, only single-target cases were selected for comparison with the DCA plan. In these patients, lesions were classified into two groups according to size, following different guidelines: small PTV (<2 cm in diameter) and large PTV (≥3 cm in diameter) [6,9]. In this study, no patients had medium PTVs (≥2 to <3 cm in diameter) (Table 1).

Table 1 . Patient characteristics.

Patient	PTV group	PTV in cc	PTV in cm (diameter)	Lesion site	Number of field
1	Small (<2 cm)	2.80	1.7	Left	4
2		1.12	1.3	Center	5
3		2.28	1.6	Center	5
4		2.34	1.6	Left	4
5		0.95	1.2	Right	4
6		1.47	1.4	Right	4
7		2.48	1.7	Left	4
8		1.46	1.4	Left	4
9		1.52	1.4	Left	4
10	Large (≥3 cm)	27.82	3.8	Center	5
11		17.51	3.2	right	5
12		23.15	3.5	Center	5
Mean	Small	1.82	1.5
	Large	22.83	3.5

To acquire CT images for treatment planning, scans were performed using Discovery RT (GE Healthcare Technologies, Inc.) with a CT slice thickness set to 1.25 mm. For SRS, the Solstice^TM SRS Immobilization System (CQ Medical Solutions) was utilized during image acquisition to minimize head movements from the start. A radiation oncologist delineated the gross tumor volume and PTV directly on the acquired treatment planning images.

2. Treatment planning

Identical parameters were applied for both the VMAT and DCA plans. The prescription dose was set based on tumor size, with 24 Gy for small and 15 Gy for large PTVs. Normalization was performed to ensure that the prescribed dose covered at least 90% of the PTV.

VMAT and DCA treatment plans were created in Eclipse V13.0 (Varian Medical Systems, Inc.), with a 0.125-cm dose calculation grid, using a single isocenter noncoplanar field (Fig. 1) and 6 flattening filter free (FFF) energy. The VMAT treatment plan was used for radiation delivery with the TrueBeam STx linear accelerator (Varian Medical Systems, Inc.), whereas the DCA plan was generated for comparison. The couch and gantry angles, along with the number of fields, and the field weights were optimized for each plan based on the tumor’s 3-dimensional position.

Figure 1. Example of the plan geometry for one of the 12 patients. For the centrally located PTV, both the VMAT and DCA plans utilized five evenly distributed arc fields optimized for uniform dose delivery. PTV, planning target volume; VMAT, volumetric modulated arc therapy; DCA, dynamic conformal arc.

For VMAT plans, inverse planning was employed to ensure that at least 90% of the prescribed dose covered the PTV while minimizing the dose to normal brain tissue. The plan evaluation was based on the criterion that the V_12Gy of the brain minus the PTV should remain within 10 cc.

The DCA plan was created using the same geometric parameters as the original VMAT treatment plan, excluding the MLC, namely, the isocenter, gantry and couch angles, 6FFF energy, number of fields, field weights, and definitions for the PTV and organs at risk (OARs). Unlike VMAT, where the MLC is modulated during treatment, DCA uses forward planning with a dynamic MLC that adjusts based on the PTV shape as the gantry rotates, ensuring that the middle position of each multileaf aligns with the 2-dimensional boundary of the PTV.

3. Evaluation and comparison

To quantitatively compare inverse-planned VMAT and forward-planned DCA techniques, the dose–volume histogram (DVH), brain minus PTV, total monitor unit (MU), and various indices commonly used for clinical evaluation in SRS treatment planning, including CI_LIM98, CI_ICRU, CI_RTOG, quality of coverage (QC)_RTOG, CI_SALT, healthy tissue conformity index (HTCI)_SALT, and CI_PADDIC, were analyzed (Table 2). These indices were statistically assessed to evaluate differences between VMAT and DCA. Considering the sample size of 12, the Shapiro–Wilk test was performed to examine normality. Because the data did not satisfy the normality assumption, the Wilcoxon signed-rank test was applied.

Table 2 . Various indices for evaluating plans.

Group	Index		Parameter
ICRU	Conformity index=CIICRU	V95%/PTV	V95%: volume of the 95% reference isodose
			PTV: planning target volume
LIM	Conformity index=CILIM98	TVPIV98%/PTV	TVPIV98%: target volume covered by the 98% reference isodose
			PTV: planning target volume
RTOG	Conformity index=CIRTOG	PIV/PTV	PIV: volume of the reference isodose
			PTV: planning target volume
	Quality of coverage=QCRTOG	D100%/PD	D100%: minimal isodose surrounding the target
			PD: prescribed isodose
SALT	Conformity index=CISALT	TVPIV/PTV	TVPIV: target volume covered by the reference isodose
			PTV: planning target volume
	Healthy tissue conformity index=HTCISALT	TVPIV/PIV	TVPIV: target volume covered by the reference isodose
			PIV: volume of the reference isodose
Other	Paddick’s conformity index=CIPADDIC	TVPIV2/PTV×PIV	PTV: planning target volume PIV: volume of the reference isodose

The results of the retrospective comparison of the VMAT and DCA treatment plans for patients treated with VMAT are shown in Table 3. For CI_SALT, which was proposed by the SALT group, both VMAT and DCA exhibited the same values of 0.9, as dose normalization was performed to ensure that 90% of the dose was delivered to the PTV during treatment planning.

Table 3 . Comparison of VMAT and DCA plans.

Patient	PTV group	VMAT/DCA
		CIICRU	CILIM98	CIRTOG	CISALT	HTCISALT	CIPADDIC	QCRTOG	V12Gy in cc (brain minus PTV)	MU
1	Small (<2 cm)	1.21/1.49	0.99/0.96	0.91/1.15	0.90/0.90	0.98/0.78	0.89/0.70	0.92/0.94	6.00/7.26	5,610/3,179
2		1.27/1.62	0.99/0.95	0.93/1.14	0.90/0.90	0.96/0.79	0.87/0.71	0.93/0.91	3.48/4.43	5,635/3,492
3		1.22/1.45	0.99/0.96	0.91/1.05	0.90/0.90	0.99/0.85	0.89/0.77	0.92/0.94	5.85/7.08	6,309/3,773
4		1.15/1.46	0.98/0.98	0.92/1.08	0.90/0.90	0.98/0.84	0.88/0.75	0.90/0.95	5.05/6.59	6,071/3,419
5		1.24/1.35	0.98/0.96	0.92/0.98	0.90/0.90	0.97/0.92	0.88/0.83	0.94/0.93	3.57/4.15	6,437/3,521
6		1.17/1.51	0.98/0.97	0.91/1.07	0.90/0.90	0.99/0.84	0.89/0.76	0.92/0.95	3.67/5.14	6,201/3,512
7		1.17/1.47	0.99/0.96	0.90/1.13	0.90/0.90	1.00/0.80	0.90/0.72	0.93/0.94	5.65/7.40	6,416/3,125
8		1.15/1.31	0.97/0.96	0.90/0.98	0.90/0.90	1.00/0.92	0.90/0.83	0.93/0.93	4.33/5.60	7,752/3,449
9		1.22/1.44	0.96/0.95	0.94/1.07	0.90/0.90	0.96/0.84	0.86/0.76	0.91/0.91	4.80/6.43	9,196/3,442
10	Large (≥3 cm)	1.07/1.32	0.95/0.95	0.92/1.12	0.90/0.90	0.98/0.80	0.88/0.72	0.87/0.91	9.23/15.18	3,972/2,090
11		1.13/1.39	0.99/0.95	0.93/1.16	0.90/0.90	0.96/0.78	0.87/0.70	0.94/0.89	7.66/14.44	4,451/1,996
12		1.10/1.26	0.99/0.97	0.90/1.02	0.90/0.90	1.00/0.88	0.90/0.79	0.91/0.93	9.14/11.76	4,209/2,202
Mean	Small	1.18/1.42	0.98/0.96	0.92/1.08	0.90/0.90	0.98/0.84	0.88/0.75	0.92/0.93	4.71/6.01	6,625/3,435
	Large								8.68/13.80	4,211/2,096
P-value		<0.001*	0.005*	<0.001*	NA	<0.001*	<0.001*	0.261	<0.001*	<0.001*

The Cl_LIM98, representing the ratio of the volume receiving 98% of the prescribed dose within the PTV, was 0.98 for VMAT and 0.96 for DCA, showing a significant difference.

The CI_ICRU and CI_RTOG indices, which quantify the ratio of the prescribed dose–volume to the PTV, were 1.18 and 1.42 for CI_ICRU (based on volume of the 95% reference isodose) and 0.92 and 1.08 for CI_RTOG (based onvolume of the 100% reference isodose) in VMAT and DCA, respectively. Accordingly, both CI_ICRU and CI_RTOG were significantly higher in DCA than in VMAT.

The mean values of HTCI_SALT, which provides indirect information on the dose delivered to normal brain tissue, were 0.98 and 0.84 for VMAT and DCA, respectively. This significant difference indicates that compared with VMAT, DCA delivered a higher dose to normal brain tissue.

For mean values of QC_RTOG, which represents the minimum dose delivered to the PTV volume, were 0.92 and 0.93 for VMAT and DCA, respectively, showing no significant difference.

For the brain minus PTV, representing normal brain tissue receiving doses >12 Gy, the mean V_12Gy values for VMAT and DCA were 4.71 and 6.01 cc for small PTV and 8.68 and 13.80 cc for large PTV, respectively. This indicates that the DCA plans for large PTVs showed a higher V_12Gy, exceeding 10 cc and failing to meet the criteria (<10 cc). Fig. 2 shows the DVH of the brain minus PTV for both VMAT and DCA with small PTVs.

Figure 2. Dose–volume histogram of the brain minus PTV in the absolute volume for patient 1. The black solid line represents VMAT, and the red dotted line represents DCA. V_12Gy values of the brain minus PTV are 6.00 cc for VMAT and 7.26 cc for DCA. PTV, planning target volume; VMAT, volumetric modulated arc therapy; DCA, dynamic conformal arc.

The MUs for VMAT plans were 6,625 for small PTVs and 4,211 for large PTVs, compared with 3,435 and 2,096 MU for DCA plans, respectively. Overall, the MU for VMAT plans was nearly twice as high as that of DCA plans.

The frameless fixation system used at our institution demonstrated errors within an acceptable range when comparing the pre- and posttreatment alignment. Minniti et al. [18] demonstrated that the positional accuracy of tumors between frame-based and frameless setups is within 1–2 mm, whereas He et al. [16] reported clinical findings showing no significant differences in treatment outcomes between the two methods. However, careful consideration is necessary when using the frameless approach, as deviations >3 mm have occasionally been reported [16-18].

Molinier et al. [22] compared VMAT and DCA to evaluate their dosimetric advantages for single lesions, multiple lesions, and lesions located near OARs. The study demonstrated that DCA provided better sparing of healthy brain tissue than VMAT for single metastases. However, VMAT was more advantageous for treating multiple metastases and targets located near OARs [22].

Torizuka et al. [23] compared VMAT created using both coplanar and noncoplanar beams and DCA plans created using noncoplanar beams for the treatment of single metastases. They found that VMAT with noncoplanar beams can save more normal brain tissue than DCA. However, the VMAT technique required a higher number of MUs, potentially increasing the workload for the medical staff [23].

Kuperman et al. reported that compared with DCA, VMAT demonstrated superior dosimetric outcomes in terms of the CI. However, no significant correlations were found between the CI of normal brain tissue and V_10Gy or V_12Gy. Therefore, DCA could be considered an alternative to VMAT in certain clinical situations [24].

At our institution, CI_LIM98 offers a simplified evaluation of SRS treatment plans by calculating the ratio of the volume enclosed by the 98% prescription isodose line to the PTV. This index, assessed with a reference value of 1, provides a clear measure of target coverage with a single metric. In this study, the CI_LIM98 of VMAT showed better dose coverage of the PTV compared with that of DCA.

CI_ICRU and CI_RTOG were calculated as the ratio of the volume enclosed by the prescription isodose line to the tumor volume. A value of 1–2 is considered suitable for treatment plan quality, whereas values <0.9 or >2.5 are deemed unsuitable. In this study, the values of CI_ICRU and CI_RTOG met the criteria for both VMAT and DCA. However, there are limitations. First, although the indices provide information on the dose coverage to the tumor, they do not allow for a precise correlation between the index values and clinical outcomes. Second, the exact reference isodose level for the contour, such as the 95% or 100% isodose line, is challenging to clinically determine for the volume of the reference isodose.

CI_SALT was calculated as the ratio of the tumor volume to the volume of the prescription isodose line within the tumor. A value of 1 is considered ideal for treatment plan quality, whereas values of ≤0.6 are deemed acceptable. However, this index does not provide information on the dose delivered to adjacent healthy tissues.

HTCI_SALT was calculated as the ratio of the prescription isodose volume within the tumor to the total prescription isodose volume. This index provides indirect information on the dose delivered to normal tissues. A value of 1 is considered ideal for treatment plan quality, whereas values of ≤0.6 are deemed unsuitable. In this study, HTCI_SALT showed that compared with VMAT, DCA delivered a higher dose to normal brain tissue, a result also reflected in the V_12Gy of the brain minus PTV. However, this index does not directly reflect the dose delivered to the tumor itself. For example, even if the index is calculated as 1, the dose delivered to the tumor may not be 100%.

CI_PADDIC is designed to provide a comprehensive analysis of the dose distribution to both the tumor and surrounding normal tissue. This index, which was proposed by the SALT group, is calculated by multiplying the dose coverage of the tumor by the dose distribution to normal tissue. A value of 1 indicates ideal treatment plan quality, whereas values of ≤0.6 are considered unsuitable. This index allows the indirect assessment of dose information for both the tumor and normal tissues. However, if the index is 0.6, it is unclear whether this reflects an underdose to the tumor with normal tissue sparing or an underdose to both the tumor and normal tissues. In this study, the values for this index were found to be acceptable for both VMAT and DCA.

QC_RTOG is calculated as the ratio of the minimum isodose level covering the entire tumor volume to the prescribed dose. In this study, this index was not able to distinguish between the two plans. The MU used in DCA can be reduced by approximately 40%–50% compared with that in VMAT, allowing for a shorter treatment time with DCA.

These indices alone should not be used to clinically assess the quality of treatment plans. In SRT (23.1 Gy/3 fractions), a comparison of clinical outcomes between the DCA and VMAT groups did not reveal differences in clinical outcomes (toxicity, local control, and overall survival) between the two methods; however, further research is needed to explore the clinical correlation between DCA and VMAT [15]. Similarly, for SRS, additional studies are likely necessary to investigate the clinical correlation between DCA and VMAT.

Our institution verified the patient setup using CBCT before and after treatment to confirm whether any movements occurred during the SRS session with a frameless fixation system. This process ensures the reliability of the frameless fixation system.

In most plan quality indices, significant differences were found between VMAT and DCA; however, which plan is superior in treatment outcomes based on specific CI values alone is challenging to determine. Therefore, a comprehensive review of the various indices is necessary.

For large PTVs, careful consideration is necessary when choosing a plan, as DCA can occasionally result in V_12Gy of a brain minus PTV exceeding 10 cc. Conversely, DCA provides the advantage of shorter treatment times because of its lower MU. This study emphasizes the importance of selecting an appropriate combination of indices for a robust quantitative assessment of treatment plans.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for profit sectors.

The authors have nothing to disclose.

All relevant data are within the paper and its Supporting Information files.

Conceptualization: Sangwook Lim. Data curation: Youngkuk Kim. Formal analysis: Sangwook Lim, Youngkuk Kim. Investigation: Sangwook Lim, Youngkuk Kim. Methodology: Sangwook Lim. Supervision: Sangwook Lim. Validation: Sangwook Lim, Youngkuk Kim, Kyung Ran Park, Ji Hoon Choi. Visualization: Sangwook Lim. Writing – original draft: Sangwook Lim, Youngkuk Kim. Writing – review & editing: Sangwook Lim, Youngkuk Kim, Kyung Ran Park, Ji Hoon Choi.