Volumetric modulated arc therapy treatment planning of thoracic vertebral metastases using stereotactic body radiotherapy

Abstract Purpose/Objectives To retrospectively evaluate the plan quality, treatment efficiency, and accuracy of volumetric modulated arc therapy (VMAT) plans for thoracic spine metastases using stereotactic body radiotherapy (SBRT). Materials/Methods Seven patients with thoracic vertebral metastases treated with noncoplanar hybrid arcs (NCHA) (1 to 2 3D‐conformal partial arcs +7 to 9 IMRT beams) were re‐optimized with VMAT plans using three coplanar arcs. Tumors were located between T2 and T7 and PTVs ranged between 24.3 and 240.1 cc (median 48.1 cc). All prescriptions were 30 Gy in 5 fractions with 6 MV beams treated using the Novalis Tx linac equipped with high definition multileaf collimators (HDMLC). MR images were fused with planning CTs for target and OAR contouring. Plans were compared for target coverage using conformality index (CI), homogeneity index (HI), D90, D98, D2, and Dmedian. Normal tissue sparing was evaluated by comparing doses to the spinal cord (Dmax, D0.35, and D1.2 cc), esophagus (Dmax and D5 cc), heart (Dmax, D15 cc), and lung (V5 and V10). Data analysis was performed with a two‐sided t‐test for each set of parameters. Dose delivery efficiency and accuracy of each VMAT plan was assessed via quality assurance (QA) using a MapCHECK device. The Beam‐on time (BOT) was recorded, and a gamma index was used to compare dose agreement between the planned and measured doses. Results VMAT plans resulted in improved CI (1.02 vs. 1.36, P = 0.05), HI (0.14 vs. 0.27, P = 0.01), D98 (28.4 vs. 26.8 Gy, P = 0.03), D2 (32.9 vs. 36.0 Gy, P = 0.02), and Dmedian (31.4 vs. 33.7 Gy, P = 0.01). D90 was improved but not statistically significant (30.4 vs. 31.0 Gy, P = 0.38). VMAT plans showed statistically significant improvements in normal tissue sparing: Esophagus Dmax (22.5 vs. 27.0 Gy, P = 0.03), Esophagus 5 cc (17.6 vs. 21.5 Gy, P = 0.02), and Heart Dmax (13.1 vs. 15.8 Gy, P = 0.03). Improvements were also observed in spinal cord and lung sparing as well but were not statistically significant. The BOT showed significant reduction for VMAT, 4.7 ± 0.6 min vs. 7.1 ± 1 min for NCHA (not accounting for couch kicks). VMAT plans demonstrated an accurate dose delivery of 95.5 ± 1.0% for clinical gamma passing rate of 3%/3 mm criteria, which was similar to NCHA plans. Conclusions VMAT plans have shown improved dose distributions and normal tissue sparing compared to NCHA plans. Significant reductions in treatment time could potentially minimize patient discomfort and intrafraction movement errors. VMAT planning for SBRT is an attractive option for the treatment of metastases to thoracic vertebrae, and further investigation using alternative fractionation schedules is warranted.

Conclusions: VMAT plans have shown improved dose distributions and normal tissue sparing compared to NCHA plans. Significant reductions in treatment time could potentially minimize patient discomfort and intrafraction movement errors. VMAT planning for SBRT is an attractive option for the treatment of metastases to thoracic vertebrae, and further investigation using alternative fractionation schedules is warranted. Randomized data support the combination of surgery and radiation in cases of spinal cord compression; however, the optimal management for spinal metastases without cord compression is still unclear. 11 Radiation offers the advantage of sparing the patient from an invasive procedure, but conventional techniques are limited by the tolerance dose of the spinal cord. 12 Conventional single fraction treatments (BED~14 Gy 10 ) and multifraction treatments (BED~30-40 Gy 10 ) have demonstrated low clinical complete response rates and suboptimal tumor control. Retreatment is often required, especially for single fraction treatments, and may be precluded by prior radiation leaving surgery as the only option. [13][14][15][16] SBRT has emerged as an attractive method of dose escalation (BED~40-80 Gy 10 ) while respecting spinal cord tolerance through advanced planning techniques using image-guided (IG) intensitymodulated radiation therapy (IMRT). 17 While these methods produce highly conformal plans with a higher BED and shorter overall treatment time as compared to traditional two-dimensional palliative dose regimens, the individual treatments are lengthy and require a large number of total monitor units. 22 Methodology to minimize BOT is an area of interest for SBRT to the spine due to concerns over intrafraction motion and patient comfort. Volumetric Modulated Arc Therapy (VMAT) is an elegant technique of delivering IMRT that allows for shorter treatment times achieved by optimizing MLC positions and dose rate while the gantry rotates around the patient with the beam-on. [23][24][25] SBRT to the spine has been demonstrated to be feasible and safe in a phase I study; however, concerns over the possibility of vertebral compression fracture and radiation-induced myelitis remain and are an active area of investigation. [26][27][28] Techniques for planning and immobilization are of special interest in regards to limiting toxicity by keeping treatment times and intrafraction motion to a minimum. 29,30 In this report, we retrospectively evaluate VMAT plans for thoracic spinal metastases using SBRT in terms of plan quality, treatment efficiency, and accuracy. | 55 treatment plans were calculated using the pencil-beam algorithm with heterogeneity corrections turned on with 2.0 9 2.0 9 2.0 mm 3 dose grid sizes. A Monte Carlo algorithm is also available and is used for lung SBRT treatments in our clinic; however, based on clinical experience there are no significant differences with the Monte Carlo algorithm for spinal SBRT, and the pencil-beam algorithm is standardly used for these treatments. The treatment prescription was 30 Gy in 5 fractions with at least 90% of the PTV encompassing 100% of the prescription isodose. Immobilization was accomplished with a BodyFIX double-vacuum immobilization device (Elekta, Stockholm, Sweden) and abdominal compression. The ExacTrac system from BrainLAB was utilized for initial patient setup. Quality assurance checks were performed daily in order to ensure accurate target localization. Prior to each treatment, a pair of oblique kV x-ray images was acquired and automatic 2D/3D image registration was performed in the ExacTrac system. Cone beam CT scans were then performed with Varian onboard imaging (OBI). All quality assurance procedures were in compliance with the standard SBRT treatment delivery technique following AAPM guidelines. 27,28 Specifically, the Winston-Lutz test was performed daily before SBRT treatments confirming coincidence of the radiation isocenter and mechanical isocenter. 31

2.C | Evaluation of dose distribution
Dose volume histograms were generated for all hybrid and VMAT treatment plans in the Eclipse TPS for the PTV, spinal cord, esophagus, and heart. Dosimetric evaluation of these plans was performed by calculating conformality index (CI) and heterogeneity index (HI) using the DVH of the PTV.
The CI as defined per ICRU is: where V ip represents the treated volume enclosed by the prescrip- The HI as defined per ICRU is: where D 2% and D 98% correspond to the dose delivered to 2% and 98% of the PTV, respectively, and D median represents the median dose to the PTV. Smaller values of HI indicate better dose homogeneity within the target volume.
Each hybrid and VMAT plan was evaluated for PTV coverage and dose to the OARs (spinal cord, esophagus, and heart).

| RESULTS
Computed dose distributions for both hybrid (Hyb) and VMAT (Rapid) plans in coronal, sagittal, and axial views for one representative patient are shown in Fig. 2. The corresponding DVHs for both plans of the same patient are also shown in Fig. 3.   36 The authors reported that CyberKnife was the only modality that used noncoplanar beams and produced the lowest spinal cord doses and best conformality; however, these benefits came at the cost of much longer treatment times. In our study, we found that VMAT allowed for faster delivery of radiation compared to noncoplanar hybrid plans while actually improving conformality and normal tissue sparing. The average beam-on time was 4.7 min for VMAT compared 7.1 min for hybrid plans (not accounting for couch kicks). In actuality, effective treatment times for hybrid plans using noncoplanar arcs are much longer due to required repositioning of the table (couch kicks) by the therapist during treatment. For the patient population with spinal metastases undergoing palliative radiation, prolonged treatment times can cause significantly more pain and discomfort, which can also result in additional patient movement and aborted treatments. Our results demonstrate that VMAT allows for faster treatments through the use of coplanar arcs without sacrificing target coverage and OAR sparing.
Wu et al. have previously reported a series of 10 patients treated with SBRT to the spine using either static IMRT or VMAT plans (with one or two arcs). When only one arc was used, VMAT plans were significantly worse in terms of spinal cord dosing, but there was no difference with two arcs. The mean treatment times were improved with VMAT plans (6.38 min beam-on time for 2 arc plans).
In our study, we used three coplanar arcs for VMAT plans, and found similarly improved treatment efficiency without any significant difference in spinal cord sparing. 37

| CONCLUSION
In this paper, we have presented the results of our study investigating the feasibility of using VMAT for SBRT spine treatments instead of noncoplanar hybrid arcs. VMAT plans resulted in improved dose distributions and normal tissue sparing compared to NCHA plans.
Treatment times were significantly shorter with VMAT plans, which is advantageous in both clinical efficiency as well as minimizing patient discomfort and intrafraction movement error. VMAT planning using SBRT is an attractive option for the treatment of metastases to thoracic vertebrae, and further investigations using alternative fractionation schedules are warranted.

CONFLI CTS OF INTEREST
The authors do not have any conflicts of interest to declare.