Treatment planning of VMAT and step‐and‐shoot IMRT delivery techniques for single fraction spine SBRT: An intercomparative dosimetric analysis and phantom‐based quality assurance measurements

Abstract Purpose To retrospectively compare clinically treated step‐and‐shoot intensity modulated radiotherapy (ssIMRT) and volumetric modulated arc therapy (VMAT) spine stereotactic body radiotherapy (SBRT) plans in dosimetric endpoints and pretreatment quality assurance (QA) measurements. Methods Five single fraction spine SBRT (18 Gy) cases — including one cervical, two thoracic, and two lumbar spines — clinically treated with ssIMRT were replanned with VMAT, and all plans were delivered to a phantom for comparing plan quality and delivery accuracy. Furthermore, we analyzed 98 clinically treated plans (18 Gy single fraction), including 34 ssIMRT and 29 VMAT for cervical/thoracic spine, and 19 ssIMRT and 16 VMAT for lumbar spine. The conformality index (CI) and homogeneity index (HI) were calculated, and QA measurement records were compared. For the spinal cord/cauda equina, the maximum dose to 0.03 cc (D0.03cc) and volume receiving 10 or 12 Gy (V10Gy/V12Gy) were recorded. Statistical significance was tested with the Mann–Whitney U test. Results Compared to ssIMRT, replanned VMAT plans had lower V10Gy/V12Gy and D0.03cc to the spinal cord/cauda equina in all five cases, and better CI in three out of five cases. The VMAT replans were slightly less homogeneous than those of ssIMRT plans. Both modalities passed IMRT QA with >95% passing rate with (3%, 3 mm) gamma criteria. With the 98 clinical cases, for cervical/thoracic ssIMRT and VMAT plans, the median V10Gy of spinal cord was 4.15% and 1.85% (P = 0.004); the median D0.03cc of spinal cord was 10.85 Gy and 10.10 Gy (P = 0.032); the median CI was 1.28 and 1.08 (P = 0.009); the median HI were 1.34 and 1.33 (P = 0.697), respectively. For lumbar spine, no significant dosimetric endpoint differences were observed. The two modalities were comparable in delivery accuracy. Conclusion From our clinically treated plans, we found that VMAT plans provided better dosimetric quality and comparable delivery accuracy when compared to ssIMRT for single fraction spine SBRT.


| INTRODUCTION
Recent studies 1-5 report that fast pain relief, excellent local control, and low toxicity are achievable with stereotactic body radiotherapy (SBRT) for the treatment of spinal metastatic diseases. The much larger biological effective dose of SBRT compared to that of conventional radiotherapy (RT) is more effective in overcoming radioresistance. 6 A complete course of SBRT often consists one to five fractions with 8 to 30 Gy per fraction. 7 This treatment is made possible with modern radiotherapy technology including inverse planning and optimization algorithms, patient specific quality assurance (QA), image guidance, high definition multi-leaf collimator (MLC), as well as advanced immobilization. When delivering SBRT on a linear accelerator (Linac) equipped with MLCs, two modalities are often usedstep-and-shoot intensity modulated radiotherapy (ssIMRT) and volumetric modulated arc therapy (VMAT).
Early studies 8-10 demonstrated the feasibility and benefits of using VMAT in conventionally fractionated radiation treatment. It has also been shown that VMAT reduced the treatment time for different sites including spine and lung SBRT, pediatric pelvic irradiation, and whole-abdominopelvic irradiation in an early case study. 11 Especially with the use of flattening filter-free (FFF) beams, increased dose rate further shortens SBRT treatment time. 12 Although VMAT has faster and easier delivery compared to ssIMRT, some concerns exist for the use of VMAT in spine SBRT cases where a steep dose fall-off is required between the boundary of the spinal cord and the tumor. Using the Eclipse treatment planning system in a study with ten patients treated with single fraction spine SBRT, 13 Wu et al. reported that single-arc VMAT provided less cord sparing compared to ssIMRT, while two-arc VMAT was only comparable to ssIMRT. In a study published by Huang et al., 14 it was found that VMAT plans had worse conformality than ssIMRT plans while the average D max of the spinal cord was not significantly different between VMAT plans (12.9 ± 1.3 Gy) and ssIMRT plans (12.5 ± 1.3 Gy). In their study, however, the VMAT plans were created using an early version of Pinnacle 9.0 (Philips) while ssIMRT plans were created using iPlan 4

2.B | Record review
Plan set-up, patient anatomy, target volume, target-OAR relationships, and clinical measurements were reviewed on plan and treatment records. A scoring method was used to evaluate the plan complexity: (a) Four scoring elements were defined: vertebral body, T A B L E 1 Plan acceptance criteria for target and organs at risks.

ROI
Acceptance criteria left transverse process and articular process, right transverse process and articular process, spinous process; (b) The involvement of each scoring element in the target volume adds one point to the plan complexity score.

2.C | Dosimetric analysis
Treatment plans were transferred from the TPS to MIM (MIM Software Inc., Cleveland, OH) for dosimetric analysis. Dose volume histograms (DVHs) for the target and spinal cord/cauda equina were extracted. For the spinal cord and cauda equina, the maximum dose to 0.03 cc (D 0.03cc ) and volume receiving 10 or 12 Gy (V 10Gy or V 12Gy ) were recorded. The plan conformality index (CI) and homogeneity index (HI) were calculated using the following equations, and here, V Rx is the volume that received the prescription dose, V target is the volume of the target, D max is the maximum dose, and D Rx is the prescription dose. Phantom QA records were reviewed for each plan, and gamma passing rates (gPR) were used for evaluating delivery accuracy. Plan quality was evaluated using the parameters including V 10Gy /V 12Gy and D 0.03cc for spinal cord or cauda equina, CI, and HI.
gPR was used to determine plan delivery accuracy. The gamma criteria was (3%, 3mm), and gamma < 1 was considered passing. gPR was the percentage of passing pixels. Statistical significance was tested using two-sided Mann-Whitney U test when applicable, and P < 0.05 was considered significant. 19

2.D | Replan-based comparison
Five ssIMRT plans (Table 2) with cervical (C), thoracic (T), and lumbar (L) spine lesions were selected from an Institutional Review Board (IRB) approved registry to be replanned with VMAT for head-to-head comparison. The five ssIMRT and five VMAT plans were measured using the electronic portal imaging device (EPID). EPID IMRT measurement results were analyzed using PerFraction TM (Sun Nuclear Corp., Melbourne, FL).

3.A | Replan-based comparison
Five clinically treated ssIMRT plans were replanned with VMAT. As shown in Table 2, the treatment locations included C, T, and L spines; the target volumes ranged from 29.78 to 280.03 cc; the complexity scores ranged from 2 to 4 for the five cases. Plan 1, 2, 3, and 5 had nine beams and plan 4 had 6 beams in the clinical ssIMRT plans. For the VMAT replans, plan 1 had one arc, and others had two arcs (  were no significant differences between ssIMRT and VMAT for any of the dosimetric endpoints, conformity, homogeneity, or delivery accuracy for L spine treatment plans (Table 5).

| CONCLUSION
Different from previously published works that were based on research plans using the same patient anatomy, this work was based on our clinically treated plans. We find that the plan quality of VMAT is better than that of ssIMRT for treating cervical and thoracic spine SBRT, achieving adequate target coverage, comparable delivery accuracy, better conformality, and lower dose to the spinal cord. With its improved delivery efficiency, VMAT should be considered a preferred treatment option for single fraction spine SBRT.

CONFLI CT OF INTEREST
None. T A B L E 5 Dosimetric endpoints and delivery accuracy comparison between step-and-shoot intensity modulated radiotheraphy (SSIMRT) and volumetric modulated arc theraphy (VMAT) for L spine treatment plans.