Evaluation of mixed energy partial arcs for volumetric modulated arc therapy for prostate cancer

Abstract Purpose The purpose of this work was to investigate the dosimetric impact of mixed energy (6‐MV, 15‐MV) partial arcs (MEPAs) technique on prostate cancer VMAT plans. Methods This work involved prostate only patients, planned with 79.2 Gy in 44 fractions to the planning target volume (PTV). Femoral heads, bladder, and rectum were considered organs at risk. This study was performed in two parts. For each of the 25 patients in Part 1, two single‐energy single‐arc plans, a 6 MV‐SA plan and a 15 MV‐SA plan, and a third MEPA plan involving composite of 6‐MV anterior–posterior partial arcs and a 15‐MV lateral partial arc weighted 1:2 were created. The dosimetric difference between MEPA(6/15 MV 1:2 weighted) and 6 MV‐SA plans, and MEPA(6/15 MV 1:2 weighted) and 15 MV‐SA plans were measured. In the Part 2 of this study, a second MEPAs plan (6 MV anterior–posterior arcs and 15 MV lateral arcs weighted 1:1), (MEPA 6/15 MV 1:1 weighted), was generated for 15 patients and compared only with two single‐energy partial arcs plans, a 6 and a 15 MV‐PA, to investigate the influence of the energy only. Dosimetric parameters of each structure, total monitor‐units (MUs), homogeneity index (HI), and conformity number (CN) were analyzed. Results In Part 1, no statistically significant differences were observed for mean dose to PTV and CN for MEPAs (6/15 MV 1:2 weighted) vs 6 and 15 MV‐SA. MEPAs (6/15 MV 1:2 weighted) increased HI compared to 6 and 15 MV‐SA (P < 0.0005; P < 0.0005). MEPAs (6/15 MV 1:2 weighted) produced significantly lower mean doses to rectum, bladder, and MUs/fraction, but higher mean doses to femoral heads, compared to 6 MV‐SA (P < 0.0005) and 15 MV‐SA (P < 0.0005). The results of Part 2 of this study showed that, in comparison to 6 and 15 MV‐PA, MEPAs (6/15 MV 1:1 weighted) plans significantly improved CNs (P < 0.0005; P < 0.0005) and produced significantly lower mean doses to the rectum and bladder (P < 0.0005; P < 0.0005). While mean doses to the PTV and femoral heads of MEPAs (6/15 MV 1:1 weighted) plans were statistically comparable to 6 MV‐PA (P > 0.05), MEPAs (6/15 MV 1:1 weighted) increased mean doses to left (P = 0.04) and right (P = 0.04) femoral heads compared to 15 MV‐PA. MEPAs (6/15 MV 1:1 weighted) resulted in significantly lower total MUs compared to 6 MV‐PA (P < 0.0005) and 15 MV‐PA (P = 0.04). Conclusion The study for prostate radiotherapy demonstrated that a choice of MEPAs for VMAT has the potential to minimize doses to OARs and improve dose conformity to PTV, at the expense of a moderate increase in mean dose to the femoral heads.


Conclusion:
The study for prostate radiotherapy demonstrated that a choice of MEPAs for VMAT has the potential to minimize doses to OARs and improve dose conformity to PTV, at the expense of a moderate increase in mean dose to the femoral heads. conform to the target volume. Subsequently, intensity modulated radiation therapy (IMRT) allowed modulation of fluence across the geometrically shaped field by using multiple radiation beams of nonuniform intensities. Currently, IMRT is widely practiced in clinics owing to its dosimetric advantages such as superior target dose conformity and better OARs sparing. 1 During the last decade, volumetric modulated arc therapy (VMAT) using modulated arcs is gaining popularity due to its improved efficiency compared to IMRT. VMAT involves the simultaneous rotational movement between the linear accelerator along with varying dose rate, gantry speed, and the shaping of multileaf collimator (MLC) leaves to produce modulated fluence while the beam is on. It has been reported by a number of studies that VMAT results in improved delivery efficiency than IMRT for various types of cancer. [2][3][4][5][6][7] A comprehensive meta-analysis on preferred technique in prostate treatment has shown that, in addition to improvement in the delivery efficiency, VMAT also protects OARs better than IMRT for prostate cancer. 8 Both IMRT and VMAT utilize inverse planning algorithms for optimization of dose to target and OARs. A clinically available optimization software optimizes fluence map for each beam angle to achieve dosevolume objectives. However, it does not optimize for couch angle or photon energy. The selection of these parameters depends on the tumor location and the experience of a treatment planner. The preference on selection of photon beam energy for deep seated targets varies due to various energy-related dosimetric consequences. For instance, use of low energy photon beams (≤6 MV) generates narrow penumbra, which results in tighter dose distribution around the target.
However, for deep seated targets, it may result in a higher surface dose. Higher energy photon beams, on the other hand, increase forward scattering of electrons and photons, resulting in a low skin dose, but may result in undesirable dose to the patient from secondary neutrons (especially for 18 MV). A number of previous studies for prostate cancer reported dosimetric benefits of using a higher energy photon beam over 6 MV photon beam. [9][10][11][12] Only a handful of studies, however, have compared dosimetric results of mixed energy (both low and high MV) IMRT plans with a single energy IMRT for deep seated targets. 12,13 While Park et al. 12 performed a sequential optimization of photon beam energy (i.e., generation of 6 MV fluence maps followed by 15 MV fluence maps) using a commercial treatment-planning software, McGeachy et al. 13 performed simultaneous optimization of photon beam energy and fluence maps using an external optimizer. Nonetheless, both studies showed that mixed energy IMRT improved overall quality of the treatment plans including better sparing of OARs.
To our knowledge, for VMAT, only one study has investigated the dosimetric influence of mixed energy VMAT approach for prostate cancer. 14 Pokharel compared the mixed energy full arcs VMAT plans (a composite of 6 MV primary plan and 16 MV boost plan) with a single-energy full arcs VMAT plans of either low or high energy.
Pokharel reported mixed energy VMAT plans to be superior over a single-energy VMAT plans in better sparing of OARs while maintaining dose conformity to the target. Since the current commercial VMAT optimizers are not capable of optimizing a single plan with more than one energy, a mixed energy VMAT plan can only be created by combining two or more individual plans. 15 In this work, we created mixed energy partial arcs (MEPAs) plans by manually merging a 6 MV partial arcs plan and a 15 MV partial arcs plan. To our knowledge, the investigation on the dosimetric impacts of MEPAs on VMAT plans for prostate has not been reported in the literature. The aim of this work, therefore, was to further explore the scope of using two mixed energy VMAT techniques for prostate cancer by: • evaluating the additive effects of photon energy and dose weight-

2.A | Patient selection
A cohort of 25 patients with intermediate risk of prostate cancer who underwent radiation therapy was randomly selected for Part 1 of this study. A subset of 15 patients was randomly selected for the Part 2 of this study. For both studies, mean and standard deviation of planning measurements such as anterior-posterior separation, lateral separation, planning target volume (PTV), bladder, rectum, and femoral head volumes are summarized in Table 1. Figure 1 illustrates the steps taken in generating MEPAs plans and their comparisons with single energy plans in each part of the study.

2.B | CT simulation and contouring
Computed tomography (CT) scanning and simulations were performed using Philips Brilliance Big Bore Scanner (Philips Medical, Cambridge, MA) with patients in a supine position and by following the standard CT scan protocol. The thickness of each CT image in axial dimension was 1.5 mm. The contouring of prostate, left femur, right femur, bladder, and rectum was performed by a radiation oncologist on the axial slices of the CT using the Varian Eclipse™ treatment planning system version 13.7 (Varian Medical Systems, Palo Alto, CA). The OARs included bladder, rectum, left, and right femur. The OAR volumes were contoured according to the radiation therapy oncology group (RTOG-0815) protocol. 16 The prostate was defined as a clinical target volume from which the PTV was generated by adding a 5 mm margin in all directions. Mean PTV volume was 86 ± 25 cc.

2.C | Treatment planning and optimization
In both parts of this study, the total prescription dose (PD) was 79.2 Gy in 44 fractions, with a daily dose of 180 cGy. The goal of treatment plan was to cover 95% of the PTV volume by at-least 95% of the PD with no more than 2% of the PTV receiving 107%.
The dosimetric constraints were originally derived based on the quantitative analysis of normal tissue effects (QUANTEC) requirement for prostate cancer. 17 For OARs, the goal was to meet the clinically acceptable dose-volume requirements as shown in Table 2.

2.C.1. | Treatment Plans
For each of the 25 patients in the Part 1 of the study, three volumetric modulated arc plans were generated using the RapidArc™ module in Eclipse™: (a) 6 MV plan using a SA, (b) 15 MV plan using a SA, (c) composite plan using 6 MV anterior-posterior partial arcs,  In both parts, the collimator angle was set to 90°for all plans as it is considered to be a good choice for better OARs sparing in prostate cancer VMAT. 18 The isocenter was placed at the center of mass of the PTV for all the plans.

2.C.3. | Optimization parameters
In Part 1 of this study, two separate single-energy single-arc (a 6 MV-SA plan and a 15 MV-SA) plans were generated by setting the optimization objectives, dose volume constraints and priority weighting factors as illustrated in (Table 3). For MEPAs, the following steps were followed:   Table 4. MEPAs (6/15 MV 1:1 weighted) plans were generated by following the aforementioned steps 2 and 3, but with an equal dose weighting.
In both parts of this study, the beam arrangement The QUANTEC based dose-volume restrictions for OARs including femoral heads, rectum, and bladder.
Femoral heads V50 < 5% represents no more than 5% of either femoral heads should receive a dose of 50 Gy or more. D max = Maximum Dose.
F I G 2 . Arc start and stop angles for a Volumetric Arc Therapy (VMAT) mixed energy partial arcs plan using partial arcs in Eclipse treatment planning system. combination with the falloff value of 0.05 cm −1 . The NTO distance from the target border, start dose, and end dose were 1 cm, 105%, and 60%, respectively. No normalization was required in both studies to achieve dosimetric goals of the treatment.

2.D | Dosimetric parameters
The dose volume histograms (DVH) were generated for each plan in Eclipse for dosimetric evaluation and comparison. The dose calculation was performed with the anisotropic analytical algorithm (AAA where TV T,ref . represents the volume of the target volume covered by the 95% of the isodose, V ref represents the total volume receiving 95% of the isodose (V ref was determined by converting isodose to structure feature in Eclipse), V T represents PTV volume. This conformity assessment in Eq. (1) accounts for both target coverage (the first brackets) and the proximity of isodose line to the target (the second brackets). A CN value closer to 1 is considered a perfectly conformal plan.
Similarly, the mean and maximum dose, and hotspot determined by D 2% (dose received by 2% of PTV) were recorded for each case.
To evaluate the dose homogeneity within the PTV, the homogeneity index (HI) was defined as per ICRU83 by taking a ratio of difference of D 2% (dose delivered to 2% of the PTV) and D 98% (dose delivered to 98% of the PTV), and dose delivered to 50% of the PTV. 21 The plan is considered homogeneous if the value of HI is close to zero.

2.E | Statistical analysis
In Part 1, the dosimetric parameters of MEPAs (6/15 MV 1:2 weighted) plans were statistically compared with the dosimetric parameters of 6 and 15 MV-SA using a two-tailed paired-sample t-test.
In addition, the 95% confidence interval is included for each P-value. weighted) plans are shown in Table 6. The average differences, D 6MV avg andD 15MV avg , for dosimetric parameters of the PTV, bladder, rectum, and as well as number of Monitor Units (MU), CI, and HI are shown in Table 7.  (Table 6).

3.B.5 | Monitor units
The total number of monitor units for MEPAs (6/15 MV 1:1 weighted) plans was higher than that of 15 MV-PA plans (480 vs 442 MUs; P = 0.04; Tables 8 and 9) with an average negative difference of 9% (Table 10), but lower than that of 6 MV-PA plans (480 vs 553 MUs; P < 0.0005; Tables 8 and 9) with an average positive difference of 13% (Table 10).   With an exception of degraded HI and lower MUs, the results of Part 1 of this study are in agreement with a previous study, 14 which compared dosimetric quality of single-energy partial-arc (30°-165°a nd 195°-330°) VMAT plans with that of a single-energy full-arc (0°-359°) VMAT plans for prostate and demonstrated that partial arcs technique results in lower doses to the bladder and rectum but at an expense of higher doses to femoral heads. 14 28 It should be noted that the V 70Gy ranged from 6% to 7% in Part 1 (Table 5) and 8% to 11% in Part 2 of this study ( which involves the risk of Grade 3 toxicity as a late response. 29 However, this was mainly due to not including maximum bladder dose constraints during optimization for any of the three techniques.

3.B.6 | Dose distribution
This was because it is considered a strict constraintrequired to be achieved by every single voxel of a structure, which, in turn, would require us to change the optimization parameters and optimize the plans individually. Instead, the goal was to optimize all the plans with a fixed optimization setup to highlight superiority among different techniques. In terms of prostate motion, a greater prostate motion has been reported to occur in anterior and posterior direction than lateral direction. 33 Furthermore, it has been demonstrated that intrafraction prostate motion from breathing is a major cause of prostate positional variation. 34 Although lower MUs would reduce the total treatment time resulting in lower probability of such organ motion, the total treatment time for MEPAs technique, regardless of the lower MUs, may not be reduced significantly as two different energies need to be moded up at the console for each treatment fraction.
Historically, patient separation in anterior posterior direction greater than 20 cm were considered as a threshold for using higher photon energy, 35 the mean AP separation in our study was~23 cm.
The rationale behind using the lowest clinical range (6 MV) to the highest clinical range (15 MV) was to exploit the maximum difference in dose deposition. Both MEPAs techniques in this study involved 15 MV, which raises a question of additional dose deposited by photo-neutrons produced in the linac head. This may be of some concern for MEPA (6/15 MV 1:2 weighted) technique as 2/3 of the PD is delivered by 15 MV beam. One study on the measurement of photo-neutron dose at isocenter from an 18 MV linac showed that the total neutron equivalent dose is two to three orders of magnitude smaller than the photon dose delivered to the patient. 36 Nonetheless the amount of neutron dose in the vicinity of the patient should not be neglected, which is one of the limitations of this study. Therefore, prior to clinically employing MEPA with 15 MV and higher, additional risks of secondary cancers due to photo-neutrons should be considered. Furthermore, mixed energies VMAT involving higher energy would not be recommended for patients with pacemakers as it can result in the device malfunction. 37 Since the neutron production for higher energy (>10 MV) in FFF mode is reduced as much as 70%, 29 similar mixed energy technique for flattening filter free (FFF) modality would be an interesting topic for future investigation, though clinical use of FFF modality is currently limited to ≤10 MV.
Another limitation of our work is the same set of optimization parameters including priority weighting factors used for all the patients in Part 1 and 2 of this study. Our rationale behind maintaining same parameter set was to ensure that the differences were only due to energy and dose weighting selection in Part 1, and energy Nevertheless, once established, MEPAs can easily be implemented for post optimization stages (i.e, patient specific QA) as the patient specific QA for MEPAs plans can be performed similarly to that of a single-energy VMAT plans. This study was based on comparisons of TPS generated dosimetric outcomes. Any quality assurance of these plans was not considered as it was beyond the scope of this work.
Finally, the radiobiological impact of any of the techniques used in this study was not investigated.
The TPS used in this study (RapidArc™, Eclipse, Palo Alto, CA, USA) does not allow optimization of a single plan with two different energies. Therefore, a composite plan was generated by summing a lower energy and a higher energy plan. Beside the TPS used in this study, the RayStation™ (Raysearch Laboratories, Stockholm, Sweden) and the Monaco™ (Elekta, Stockholm, Sweden) are two major treatment planning systems that are currently being used to optimize VMAT treatment plans. However, to our knowledge, no current treatment planning system, including the one used in this study, allows simultaneous optimization of two different energies. The current study, thus, involved the manual selection of dose weighting per energy to achieve the desire dosimetric outcome. An algorithm that simultaneously optimizes for both energies is necessary as it will generate a plan with an optimal proportion of PD dedicated to each energy, which, in turn, will further improve the quality of a mixed energy VMAT plan. While it was beyond the scope of this work to investigate the most suitable TPS for MEPAs technique, it would be interesting to investigate MEPAs on RayStation™, which utilizes multicriteria optimization where the user navigates through many pareto optimal plans to arrive at a plan with desired dosimetric tradeoffs. However, the dosimetric comparisons between two plans may not be suitable for RayStation™ as due to selection of best possible tradeoff between different dose-volume objectives of various structures, the parameters may not remain same in the two plans.

| CONCLUSION S
This study investigated the potential scope of using MEPAs VMAT technique to treat prostate cancer compared to single-energy VMAT techniques. In Part 1 of this study, MEPAs (6/15 MV 1:2 weighted) plans were found to be superior in sparing bladder and rectum, but resulting in slightly reduced target homogeneity compared to either 6 and 15 MV-SA plans. In Part 2 of this study, the impact of multiple energies alone was investigated by equally weighting both 6 and 15 MV in MEPAs (6/15 MV 1:1 weighted) and comparing with single-energy partial arcs (6 and 15 MV-PA).
MEPAs (6/15 MV 1:1 weighted) plans resulted in improved target dose conformity and, lower doses to bladder and rectum compared to 6 and 15 MV-PA. In both parts, however, mixed energy VMAT plans increased doses to femoral heads compared to single-energy VMAT plans.

CONFLI CT OF INTEREST
None.