Dosimetric dependence on the collimator angle in prostate volumetric modulated arc therapy

Purpose: The purpose of this study is to investigate the dose-volume variations of planning target volume (PTV) and organs-at-risk (OARs) in prostate volumetric modulated arc therapy (VMAT) when varying collimator angle. The collimator has the largest impact and is worth considering, so, its awareness is essential for a planner to produce an optimal prostate VMAT plan in a reasonable time frame. Methods: Single-arc VMAT plans at different collimator angles (0o, 15o, 30o, 45o, 60o, 75o and were created systematically using a Harold heterogeneous pelvis phantom. The conformity index (CI), homogeneity index (HI), gradient index (GI), machine monitor units (MUs), dose-volume histogram and mean and maximum dose of the PTV were calculated and analyzed. On the other hand, the dose-volume histogram and mean and maximum doses of the OARs such as the bladder, rectum and femoral heads for different collimator angles were determined from the plans. Results: There was no significant difference, based on the planned dose-volume evaluation criteria, found in the VMAT optimizations for all studied collimator angles. A higher CI (0.53) and lower HI (0.064) were found in the 45o collimator angle. In addition, the 15o angle provided a lower value of HI similar to the 45o collimator angle. Collimator angles of 75o and 90o were found to be good rectum sparing, and collimator angles of 75o and 30o were found to be good for sparing of right and left femur, respectively. The PTV dose coverage for each plan was comparatively independent of the collimator angle. Conclusion: Our study indicates that the dosimetric results provide support and guidance to allow the clinical radiation physicists to make careful decisions in implementing suitable collimator angles to improve the PTV coverage and OARs sparing in prostate VMAT.


Introduction
Volumetric modulated arc therapy (VMAT) has become a standard delivery option in the field of prostate radiotherapy, due to its shortened delivery time and the smaller monitor units (MUs), as compare to step-and-shoot intensity modulated radiotherapy (IMRT). [1][2][3][4][5][6] Patient dosimetry between prostate VMAT and IMRT has been extensively studied, which reveals that prostate VMAT can produce comparable or even improved target coverage and normal tissue (bladder, rectum and femoral heads) sparing. [7][8][9][10][11] VMAT encloses more dose delivery parameters such as dynamic multileaf collimator movement, dose rate, and gantry speed with single or multiple photon arcs in the treatment, [12][13][14][15] which requires a more powerful machine, patient quality assurance procedures, dose calculation algorithm, and dosimetric evaluation for the treatment. [16][17][18][19] Until the availability of the Elekta linear accelerator VMAT in 2008 20 , the only commercially available treatment planning system (TPS) was ERGO++ (3D Line Medical Systems/ Elekta Ltd, Crawley, UK), which needed an initial definition of sub-arcs and had manual version of the multileaf collimator (MLC) before automatic weight optimization and was not considered a full inverse planning system. 11, 21 ,22 In December 2009, two manufacturers introduced a new system of VMAT delivery that employed a VMAT treatment planning tool, implemented in Oncentra with Master-plan v3.3 (Nucletron BV, Veenendal, The Netherlands) with VMAT application on a Synergy linac (Elekta Ltd, Crawley, UK). Initially, the Synergy linac was used for a limited number of patients. 8,23 RapidArc (Varian Medical Systems, Palo Alto, CA) is a VMAT technique delivering radiation dose over one or several continuous arcs with the simultaneous adjustment of dose rate, gantry rotation speed, and multi-leaf collimator (MLC) field aperture. RapidArc has gained enormous interest because of its potential in delivering quality dose distribution with significantly shortened treatment time and lower number of MU. Several recent studies have reported the use of arc-based radiation dose delivery methods in prostate cancer. 5,7,11,[24][25][26] Multileaf collimators (MLC) are the best tool for beam shaping, and an important way to minimize the absorbed dose to healthy tissue and critical organs. They have moveable leaves arranged in pairs that can block a certain part of beam. Owing to its ability to control leaf position and with a large number of controlled leaves, it can be used to shape any desired field. 27 Its manufacturers have established the necessary mechanisms for precision, control and reliability, together with reduction of leakage and transmission of radiation between and through the leaves. Moreover, it provides precise dose delivery to any part or the treated volume, accurately. 28 Otto has stated 3 and later on approved 29 that a 45 o collimator angle is feasible dosimetrically in most cases. While, Bortfield et al. 29 found that the superiority of the above collimator angle (45 o ) was ambiguous. Furthermore, Bortfield and Webb did their work with a 0 o collimator angle for a 2D model. 30 Treutwein et al. 31 concluded that the approximation was still effective for 4° gantry spacing and same passing rates were found for IMRT. The work of Feygelman et al. 32 33 was used for this study. Computed tomography (CT) images (2 mm slice thickness and slice interval) were taken from the Toshiba scanner (Aquilion ONE TSX-301A; Toshiba medical systems, USA) containing 512 × 512 pixels in each slice. The Harold phantom was irradiated by a 120 kV photon beam with 300 mA current perpendicular to the phantom surface. After the CT simulation, digital imaging and communication in medicine (DICOM) CT images were transferred to the Pinnacle treatment planning system (TPS) for contouring and planning preparation.
The rectum, bladder, PTV, and femoral heads were contoured on the TPS. The whole prostate was assigned as gross tumor volume (GTV). The PTV was drawn by expanding 1 cm around the CTV in all directions uniformly except in the posterior direction, where an expansion of 0.7 mm was performed for a total contoured volume of 85.89 cm 3 . The bladder, rectum, and femoral heads have contoured volumes of 59.83 cm 3 , 36.26 cm 3 and 166 cm 3 , respectively.

VMAT plan and treatment delivery
For planning the data, a Synergy S ® linear accelerator with energy of 6 MV, equipped with beam modulator head, an iViewGT electronic portal imaging device, and on board cone beam CT XVI was used for VMAT delivery. There were no moveable jaws and the maximum field size was 16 cm × 21 cm. Maximum variable dose rate for each VMAT plan was 600MU/min and the gantry was rotated from 180 to 179.9 in the clockwise direction with 91 control points.
Smart-arc prostate VMAT plans were generated on Pinnacle (Philips, Version 9.2.0, Fitchburg, WI, U.S.A) with AC-QSim 3TM and were optimized with the direct machine parameter optimization (DMPO) algorithm. The isocenter was positioned at the center of the CTV and plans were set up in 39 fractions for 78 Gy minimum doses to the CTV. All calculations were performed using adaptive convolve (AC) having a calculation grid spacing of 0.25 cm. In order to make fair comparisons, no modification was done throughout the optimization to the dose-volume constraints and weighting.

Dosimetric evaluation
The dosimetric comparison was carried out using the following parameters such as D99%, D95%, D5%, maximum dose (Dmax), mean dose (Dmean), Conformity Index (CI), Homogeneity Index (HI), Gradient index (GI) and MUs for the PTV for collimator angle as shown in Table 1.
By definition, RTOG CI (98) is the volume of the target receiving > 98% of the prescribe dose divided by the volume of the PTV which has optimal value of 1. HI is defined as the dose received by 5% of the PTV minus the dose received by 95% of the PTV divided by the mean dose (its optimal value is 0) as shown in Equation (1 (1) GI is defined as the ratio of volume covered by at least a given percentage of the prescription dose. 35 Mathematically, GI in this study is expressed in (2) as: where, V50 is the volume covered by the at least 50% of the prescription dose. A value closer to unity embodies a faster dose fall-off in normal tissue, which may indicate a lower dose to critical structure.

Dose-volume histogram (DVH) evaluation
Dose-volume histogram plots were used to provide quantitative comparisons among the VMAT plans using the different collimator angles. Considerable attention should be placed on ensuring an unbiased comparison for successive computation of numerous indices. The DVHs data for each collimator angle was gathered from Pinnacle with a bin size of 0.01 Gy. PTV and organ specific individual DVHs for each collimator angles were calculated.

Results
This study has been carried out on a Harold phantom and clinically acceptable VMAT plans satisfying a minimum of 99% prescribed coverage to PTV were achieved. Mean doses for all collimator angles were found between 75.96 (Gy) and 76.42 (Gy). The values of CI for all collimator angles are summarized in Table 1 revealing that a 45 o collimator angle is closer to unity than any other studied collimator angles. A collimator angle of 0 o requires fewer MUs while 75 o and 90 o collimator angles require the most MUs. The highest HI values were established for a 60 o collimator angle whereas we found lower values for 45 o and 15 o angles. It was found that a 30 o collimator angle showed as lower GI value of GI that was closer to unity while higher values were found at 0 o collimator angle. Figure 1 showes Dose distribution at collimator 90.   Average accumulated DVHs of the PTV, rectum, bladder and femoral heads are shown in Figures 2-4, which are planned using VMAT with different collimator angles. The planning dose objectives of the rectum and bladder agree well with the prescribed dose; their mean, maximum, D30% and D50% doses are shown in Table 2. V14% and V38% were chosen since they have been used as physics quality assurance evaluation criteria at the Princess Margaret cancer center. V30% and V38% were calculated for rectum as well as for bladder and are shown in Table 2. The dose to the femoral heads was found to be within the acceptable range; their mean, maximum, D5%, V14% and V22% are calculated and shown in Table 2.

Dose-volume indices
An investigation of the collimator angles reveals that a 45°c ollimator angle has a 0.3% higher CI, 0.14% lower HI and 0.02% lower requires MUs than all other studied collimator angles. According to Bortifield 29 a 45 o collimator angle is preferred to 0 o collimator angle. He also clarified the hypothesis that the leaves of the MLC in a parallel opposed beam move in and orthogonal direction and consequently these beams are not terminated. Additionally, Otto 36 explained that only a single leaf pair can be used to modulate the intensity within a CT slice without collimator rotation and secondly that an 8% lower MU requirement can be found using a 45 o collimator angle verses 0 o angle. This also explains the fact that with a 45° collimator angle, one can irradiate the right and left side of the PTV as well add spare the rectum and bladder in a fashion that is not possible with a 0 o angle. In our investigation, the number of MUs required are (0.02%) lower using a collimator angle of 45° than when using a collimator angle of 90 o . Fogliata et al. 37

Dose-volume criteria, maximum and mean dose
Mean dose-volume criteria, maximum and mean dose are the important parameters for plan evaluation. Table 1 shows the dosimetric results of the PTV and Table 2 summarizes the mean dose-volume criteria of the bladder, rectum and femoral heads calculated by the treatment planning system. In this study the dose-volume evaluation criteria for the prostate VMAT plan are: D99% of PTV ≥ 74.1Gy, D30% of rectum and bladder ≤ 70Gy, D50% of rectum and bladder ≤ 53 Gy, D5% of femoral heads ≤ 53Gy. For the mean D30% and D50% of the rectum and bladder, all the collimator angles satisfy the corresponding dose-volume criteria. The mean D50% and D30% of bladder are found to be lower for the 60 o collimator angle (on average 0.03% and 0.02%) than other studied collimator angles. However, the 90 o collimator angle had a higher D50% and D30% for the rectum (on average 0.02% and 0.01%) than other studied collimator angles. For the left and right femoral head, the 90 o collimator angle had a mean D5% which was on average 0.23% and 0.9% higher more than the other collimator angles, respectively. For percentage bladder and rectum volume receiving at least the given dose, lower V30%, V38%, values were found using collimator angles of 90 o and 60 o , respectively. The percentage of the right and left femur volume receiving at least the given dose was lower for the V14Gy, V22Gy criteria at collimator angles of 75 o and 30 o , respectively. Figure 2 shows the average DVH of the PTV for all collimator angles planned using the VMAT technique. The dose range in Figure 2 begins at 70 Gy rather than 0 Gy to focus on the drop-off region of the curve. No noticable difference has been found using all studied collimator angles as seen in Figure 2. It is obvious in Figure 3(a) that the percentage of volume receiving chosen doses (e.g. V30Gy and V38Gy) are constantly lower for a 75 o collimator angle. This shows that collimator angle of 75 o is good for bladder sparing with V38Gy value is 40.51. It is apparent in the Figure 3( 40 reported that double arc technique could produce a better plan with improved PTV coverage and reduced treatment time compared to intensity modulated radiation therapy. They found that though the single arc technique resulted in a higher rectal dose, the technique had higher efficiency than the double arc. For a busy treatment unit demanding high patient throughput, single arc technique could be an acceptable option for simple prostate cases. However, for complex cases involving lymph doses, more than one single full arc may be required. It is worthwhile to study the collimator angle effect on different photon arc techniques in prostate VMAT. This is the future work in this study.

Conclusion
This work explores the impact of different collimator angles on a dosimetric scoring function. Collimator angle selection could play vital role in improving the quality of treatment plans. It is concluded from the results that the dose variations with the change of collimator angle are significant. VMAT plans with said collimator angles do not play a substantial role in PTV coverage but for more accuracy, a 45 o collimator angle provides superior PTV dose distribution than all other studied collimator angles as shown by a higher value of CI, lower value of HI and 1.4% higher value of MUs. It was observed that a 75 o collimator angle appropriate for sparing of rectum and right femur. In our investigation, 90 o and 30 o collimator angles showed the highest sparing of the rectum and left femur, respectively. The results of our study set the groundwork for guiding the collimator angle selection with regards to PTV dose distribution and sparing of OARs in prostate VMAT planning. This work also can be extended to other treatment sites using VMAT.