The ArcCHECK diode array for dosimetric verification ...

Home

Search

Collections

Journals

About

Contact us

My IOPscience

The ArcCHECK diode array for dosimetric verification of HybridArc

This content has been downloaded from IOPscience. Please scroll down to see the full text. 2011 Phys. Med. Biol. 56 5411 (http://iopscience.iop.org/0031-9155/56/16/021) View the table of contents for this issue, or go to the journal homepage for more

Download details: IP Address: 132.199.169.193 This content was downloaded on 16/03/2016 at 13:46

Please note that terms and conditions apply.

IOP PUBLISHING

PHYSICS IN MEDICINE AND BIOLOGY

Phys. Med. Biol. 56 (2011) 5411–5428

doi:10.1088/0031-9155/56/16/021

The ArcCHECK diode array for dosimetric verification of HybridArc A L Petoukhova, J van Egmond, M G C Eenink, R G J Wiggenraad and J P C van Santvoort Radiotherapy Centre West, PO Box 432, NL-2501 CK, The Hague, The Netherlands E-mail: [email protected]

Received 22 December 2010, in final form 7 July 2011 Published 29 July 2011 Online at stacks.iop.org/PMB/56/5411 Abstract The aim of this work is to evaluate dosimetric accuracy of a new treatment modality, HybridArc, in iPlan RT Dose 4.5 (BrainLAB, Feldkirchen, Germany) using a four-dimensional diode array (ArcCHECK, Sun Nuclear Corporation, Melbourne, USA). HybridArc is able to enhance dynamic conformal arcs with inversely planned elements. HybridArc plans for various sites (intracranial and extracranial) were constructed and after that these plans were recalculated for the ArcCHECK diode array with Monte Carlo (MC) and Pencil Beam (PB) dose algorithms in iPlan RT Dose. All measurements of these HybridArc plans were performed with 6 MV photon beams of a Novalis accelerator (BrainLAB, Feldkirchen, Germany) using the ArcCHECK device without and with an insert containing an ionization chamber. Comparison of the absolute dose distributions measured and calculated in iPlan RT Dose with the MC algorithm at the cylinder of the ArcCHECK diode array for HybridArc plans gives good agreement, even for the 2% dose difference and 2 mm distance-to-agreement criteria. The PB calculations significantly differ from the ArcCHECK measurements so that the MC algorithm is found to be superior to the PB algorithm in the calculation of the HybridArc plans. One of the drawbacks of the PB calculations in iPlan RT Dose is the too large arc step size of 10◦ . Use of a finer angular resolution may improve the PB results significantly. (Some figures in this article are in colour only in the electronic version)

1. Introduction Interest in rotational radiotherapy has grown significantly in the last few years. Different techniques such as dynamic arc therapy, intensity-modulated arc therapy (IMAT), helical Tomotherapy and volumetric arc therapy (VMAT) have been developed. The IMAT technique 0031-9155/11/165411+18$33.00

© 2011 Institute of Physics and Engineering in Medicine

Printed in the UK

5411

5412

A L Petoukhova et al

(Yu 1995, Ma et al 2001) is a treatment modality which combines beam rotation and dynamic multileaf collimation and for which the dose rate and the gantry speed do not change during the delivery. The HybridArc technique (BrainLAB 2010) is an IMAT solution proposed by BrainLAB for accelerators that can neither change dose rate nor gantry speed during the delivery of arcs. HybridArc is a combination of dynamic conformal arcs (DCA) and intensity-modulated radiotherapy (IMRT) treatments. One or more IMRT beams are added to each DCA at the same table position to further optimize the dose distribution while minimizing the additional setup effort. There are a number of methods to evaluate dosimetric performance of rotational therapy, such as gel dosimetry (Vergote et al 2004), electronic portal imaging device (EPID) with a commercial portal dose image prediction tool (Iori et al 2010) or based on back projection (Mans et al 2010) and film dosimetry (radiographic or radiochromic (Pallotta et al 2007)). Gel and film dosimetry are very time consuming, whereas portal dosimetry is by far the most time efficient for daily patient quality assurance (QA). Unfortunately, the Novalis accelerator does not have any EPID. IMAT dosimetry is commonly verified by using cylindrical phantoms with ionization chambers and radiographic films inside. Film can be shaped into cylindrical (Zeidan et al 2006) and spiral paths (Paliwal et al 2000) in order to sample the 3D dose distribution better than that can be done with a plane, but it is still time consuming. Alternatively, various two-dimensional detector arrays can be used for dosimetric verification of rotational therapy. For example, Han et al (2010) applied a MatriXX ion chamber array (IBA Dosimetry GmbH, Schwarzenbruck, Germany) with proper corrections, whereas Van Esch et al (2007) used a Seven29 ion chamber array (PTW, Freiburg, Germany) in combination with the Octavius phantom to compensate for the directional dependence of the 2D array. Jursinic et al (2010) demonstrated that the modified MapCHECK can reduce the angular dependence from ±20% for the original MapCHECK to ±2%. Existing 2D detector arrays are limited by their planar design and are not ideal for pre-treatment QA with arc delivery because of missing of a big part of the lateral beams and an uncertain calibration for these fields. The Delta4 (Scandidos, Uppsala, Sweden) is a 3D diode array consisting of diode matrices in two orthogonal planes inserted in a cylindrical acrylic phantom. The gantry angle is independently recorded by an inclinometer attached to the gantry or accelerator head. This allows the device to identify which control point of a dynamic arc is being delivered. The diode array was evaluated by Bedford et al (2009) for IMRT and VMAT verification. Bedford concluded that the Delta4 is a complex device and requires careful testing before it can be introduced into clinical practice. Although the system has detectors in only two planes, it provides a novel interpolation algorithm that is capable of estimating doses at points where no detectors are present (Sadagopan et al 2009). The ArcCHECK, a cylindrical diode array, was developed by Sun Nuclear Corporation to meet the need of rotational therapy QA. It is well suited for QA of VMAT (L´etourneau et al 2009) and IMRT (Eenink et al 2010). It allows measurements at arbitrary gantry angles with sufficient spatial resolution: even for the Novalis accelerator with leaf widths as small as 3 mm (Eenink et al 2010). Suitability of the ArcCHECK diode array for QA of HybridArc (a combination of DCA and IMRT) plans is expected to be based on these references. The aim of this work is to evaluate dosimetric accuracy of HybridArc for various treatment sites. This is performed by comparing the plans with experimentally determined dose values using ArcCHECK and those calculated using Monte Carlo (MC) and Pencil Beam (PB) dose algorithms in iPlan RT Dose.

The ArcCHECK diode array for dosimetric verification of HybridArc

5413

Figure 1. Example of a HybridArc plan with one arc (shown in khaki with light khaki and brown denoting the beginning and the end of the arc, respectively) and five IMRT beams (displayed in yellow).

2. Material and methods 2.1. HybridArc treatment modality HybridArc plans for various sites (intracranial and extracranial) were made using the HybridArc option in iPlan RT Dose 4.5. With this modality, a combination of the DCA and IMRT techniques can be produced by manually selecting the ratio between these two components. During automatic preplanning, a treatment group can be placed with dynamic arcs and IMRT beams distributed per arc (see figure 1). Weighting of each arc is automatically adjusted using arc length. Table angles for arcs were optimized manually. For optimization of the aperture of the dynamic arcs and IMRT beams, inverse planning was used. The HybridArc option gives tools for parameter adjustment and plan optimization during inverse planning. There are four optimization options that give different priorities for competing objectives for planning target volume (PTV) and organs at risk (OARs): one can select either PTV dose uniformity only, or OAR sparing low priority, or OAR sparing medium priority, or OAR sparing high priority.

5414

A L Petoukhova et al Table 1. Characteristics of the patients, the tumours and the dose prescriptions.

Tumor site

Diagnosis

Dose prescription

Thorax Spine

NSCLC Spinal tumor

Head

Meningioma

PTV: 24 × 2.75 Gy prescribed to the 100% isodose PTV: 10 × 3 Gy prescribed to the 100% isodose Boost: 10 × 4.8 Gy prescribed to the 100% isodose PTV: 28 × 1.8 Gy prescribed to the 80% isodose

2.2. HybridArc plans for various treatment sites and dose calculation As a starting point a HybridArc plan was made for a spine patient with one arc without any IMRT beams. This case was used to test the performance of the ArcCHECK. For dosimetric evaluation of the HybridArc option, three patients with the PTV in the vicinity of the OARs, who might have had a clinical advantage of using this treatment modality, were chosen. These patients were diagnosed with meningioma, spinal tumor and non-small cell lung carcinoma (NSCLC), respectively, and were treated before with an IMRT technique in our department. A summary of patient characteristics and dose prescriptions is given in table 1. For the lung patient, the HybridArc plan consisted of one arc with five IMRT beams with a DCA to IMRT ratio of 60% to 40%. The same DCA to IMRT ratio was used for all treatment sites described in this work. For the patient with a spinal tumor, two coplanar arcs with two IMRT beams per arc were chosen to spare the pharynx. Because of the irregular shape of meningioma, three non-coplanar arcs with two IMRT beams per arc were needed to get an acceptable coverage of the PTV and to fulfill the OAR constraints. For inverse planning, the objectives of the original IMRT treatment plans for each patient were used. CT scans of each patient were taken on an AcQSim single-slice wide bore CT scanner (Philips Medical Systems Ltd, Stevenage, UK) using a slice thickness of 2 mm for the intracranial patient and 3 mm for the extracranial patients. These data were used in iPlan RT Dose for three-dimensional dose calculation with a PB algorithm with path-length correction (BrainLAB 2008). In the present version of iPlan RT Dose, inverse planning can be performed only with the PB algorithm.

2.3. ArcCHECK device The ArcCHECK device is a cylindrical acrylic phantom containing an array of 1386 diodes in a helical configuration with 1 cm inter-diode spacing, 1 cm pitch and 21 cm diameter (ArcCHECKTM User’s Guide (Sun Nuclear Corporation 2009)). Active detector size of each diode is 0.8 × 0.8 mm2. The diodes have the same design as those used in the MapCheck 1175 (Sun Nuclear, Melbourne, USA), which showed negligible field size and source-to-surface distance (SSD) dependence (Li et al 2009). The ArcCHECK reads out the accumulated diode charge, acquiring frames with 50 ms updates. After each measurement, the individual frames are corrected for diode leakage current and angular dependence (Yan et al 2010). In our study, the angular dependence was not used because of the small field size (9.8 × 9.8 cm2) of the Novalis accelerator. The processed frames are then summed and saved to disk. The MapCHECK software can be used to import a DICOM RT Dose file for comparison to an ArcCHECK measured file. The import filter extracts a cylindrical dose plane from the imported 3D volume for 2D dose comparison with the ArcCHECK diodes.

The ArcCHECK diode array for dosimetric verification of HybridArc

(a)

5415

(b)

Figure 2. ArcCHECK with (a) and without (b) the optional cavity insert for measuring dose at the isocenter.

The ArcCHECK measures the entry dose in front of the isocenter and the exit dose behind the isocenter at two effective depths for every angle. This ensures that errors that are present in the isocenter will also be seen by the ArcCHECK. We used the ArcCHECK with and without a cavity plug with detector insert to measure the dose at the isocenter as shown in figure 2. The ArcCHECK cradle supports the device during calibration and measurements (see figure 2). Marks on the outer surface of the cylinder were used to position the device in the isocenter. In this study, the axis of the cylindrical phantom was aligned with the gantry rotation axis. 2.4. Dose calculation and verification with ArcCHECK For the ArcCHECK with and without the cavity insert, megavoltage computed tomography (MVCT) images and a CT-density (electron density) conversion file from MVCT were provided by Sun Nuclear. MVCT images were used in order to avoid streak artifacts from high atomic number components (Jursinic et al 2010). The HybridArc plans were recalculated with the MC and PB algorithms in iPlan RT Dose for the ArcCHECK with and without the insert. The MC algorithm in iPlan RT Dose is based on XVMC (x-ray voxel Monte Carlo) code developed by Kawrakow et al (1996) and Fippel (1999). The XVMC algorithm consists of three main components (BrainLAB 2008): a virtual energy fluence model (Fippel et al 2003), full MC geometry simulation (Fippel 2004) and patient dose computation (Kawrakow et al 1996, Fippel 1999, Kawrakow and Fippel 2000). The MC algorithm in iPlan RT Dose has been validated by two groups (Fragoso et al 2010, Petoukhova et al 2010). MC calculations were performed with full MLC geometry simulation (accuracy optimized model) with a spatial resolution of 2 mm and a variance of 1% following Chetty et al (2006). The option ‘dose to medium’ instead of ‘dose to water’ was used. PB calculations were performed with a spatial resolution of 2 mm and an arc step calculation of 10◦ . In the present version of iPlan RT Dose, this parameter cannot be changed. Monitor units (MUs) from the original patient plan were used. Because of intrinsic variations in the diode sensitivity, an array calibration is performed by Sun Nuclear (prior to shipment) to obtain a uniform detector response on the beam central axis, irrespective of the gantry angle (Yan et al 2010). For absolute calibration, the dose was calculated at the upper two diodes for a 9.8 × 9.8 cm2 field in the geometry presented in figure 3. The water-equivalent depth between the outer surface of the ArcCHECK and the

5416

A L Petoukhova et al

Figure 3. Dose distribution calculated in iPlan RT Dose at the ArcCHECK device with the insert and an ionization chamber in the isocenter (top panel) and without the insert (bottom panel) for a 9.8 × 9.8 cm2 field (gantry = 0◦ , collimator = 90◦ and SSD = 866 cm). These calculations were performed with the MC algorithm.

The ArcCHECK diode array for dosimetric verification of HybridArc

5417

detector junction is 2.85 cm. To check the dose at this particular depth, a water phantom was set up in iPlan RT Dose such that the detector position is the same as the detector position in ArcCHECK and has the same amount of buildup. It is important to note that the diode array of the ArcCHECK is surrounded by a thin layer of air. Since the PB simulations are not able to properly account for the loss of electronic equilibrium, their result is systematically higher by about 4% for the ArcCHECK with and without the insert. For comparison purposes we have, therefore, used different absolute calibration factors for MC and PB. Although the absolute calibration factors for the ArcCHECK with and without the insert were the same (within 0.2%) for the same algorithm, we performed different calibration for these two cases with each calculation algorithm. It is important to note that the calculated dose at the isocenter of the ArcCHECK with the insert was the same (within 0.8%) for MC and PB. For the MC calculation, in the isocenter the ‘dose to water’ option gave the same result as the ‘dose to medium’ option. On each measurement day, the calibration was checked by measurement of the reference field. Only deviations smaller than 0.5% were accepted. Otherwise a new calibration factor was applied. Dose measurements at the isocenter of the ArcCHECK device with the insert were performed with an NE2571 ionization chamber. For absolute calibration of this chamber, the same reference field was used as described before. The absolute dose in the reference point given a certain number of MU is determined according to the Dutch dosimetry protocol (NCS report 2 1986) using a NE2571 ionization chamber (NE Technology, Berkshire, UK) whose calibration is traceable to an absolute dosimetry standard. The produced non-clinical plans were irradiated at a Novalis accelerator (BrainLAB, Feldkirchen, Germany) with the BrainLab m3 micro-MLC (multileaf collimator) with 6 MV photon energy. 2.5. Evaluation To compare dose distributions between calculated and measured doses, a gamma analysis was performed with 2% dose difference (of the maximum dose) and 2 mm distance-to-agreement (DTA) criteria according to Van Dyk (Van Dyk et al 1993, Low and Dempsey 2003). These criteria were chosen tighter than usual (3%/3mm) for IMRT commissioning (Ezzell et al 2009) because of the stereotactic applications of the Novalis. The gamma analysis was evaluated in terms of the number of diodes which satisfied specified tolerances of dose difference between calculations and measurements relative to the maximum value on the calculated dose map (normalization point) and DTA criterion. Only the diodes with the dose values larger than 10% of the maximum value on the dose map were included in the analysis and their number is denoted as N0. A threshold of 10% was proposed by the AAPM Task Group 119 (Ezzell et al 2009) for the gamma evaluation with 3%/3 mm criteria. The number of them failing the 2% global dose difference and 2 mm DTA criteria is denoted as n2 further in the text. Moreover, the pass rate γ 2 = 1 − n2/N0 is calculated. Additionally, similar results of the gamma analysis for 3% global dose difference and 3 mm DTA criteria (n3, γ 3) were calculated. 3. Results 3.1. Reference field Reference measurements with the ArcCHECK device were carried out to check the dose calibration with a 9.8 × 9.8 cm2 field in the geometry as illustrated in figure 3 for the ArcCHECK with (top) and without the insert (bottom). The results of the measurements in

5418

A L Petoukhova et al

Figure 4. Comparison of the absolute dose cross-profiles measured (dots) and calculated in iPlan RT Dose with the Monte Carlo algorithm (solid lines) in the isocenter plane of the ArcCHECK diode array with (top panel) and without (middle panel) the insert for a 9.8 × 9.8 cm2 field is given. The bottom panel presents the comparison of the absolute dose cross-profiles measured and calculated in iPlan RT Dose with the Pencil Beam algorithm in the isocenter plane of the ArcCHECK diode array with the insert. The red and blue points display the failed diodes with the dose above and below the 2% dose difference criterion, respectively.

comparison with the calculations in iPlan with MC are presented in figure 4 for the diodes lying on the isocenter plane, i.e. the plane normal to the cylinder axis and passing through the isocenter. In this case with the insert, 99.1% of diodes (N0 = 223, n2 = 2) passed 2% dose difference relative to the normalization point and 2 mm DTA criteria. Pass rate γ 3 = 100% and n3 = 0. The normalization used in this case was 2.50 Gy. For the ArcCHECK without the insert, pass rates of 98.7.0% (N0 = 230, n2 = 3) and 100.0% (n3 = 0) were calculated for the γ 2 and γ 3 criteria, respectively, with a normalization of 2.51 Gy. For the ArcCHECK with the insert, results of the comparison of the measurements with the PB dose calculations for the reference field are presented in figure 4 (bottom panel). In this case, the pass rates of 86.5% (N0 = 222, n2 = 30) and 96.8% (n3 = 7) were found for the γ 2 and γ 3 criteria, respectively, with a normalization of 2.53 Gy.

3.2. One DCA without IMRT beams For benchmark purposes, a HybridArc plan was made for the spine patient with one arc without any IMRT beams. In figure 5, an example of MapCHECK software is presented for this case. The results for the arc measured and calculated with MC are presented in figure 6 for the ArcCHECK with (top panel) and without (middle panel) the insert and table 2. Comparison of the ArcCHECK (with insert) measurements to the PB calculations is shown in figure 6 (bottom panel) and table 3. There are more failing points for the PB dose calculation than the MC. Most of these points are red, which indicates that the measured dose is higher than the PB calculated dose. Moreover, the failing red points are located along the dose calculation steps that are the result of the 10◦ resolution. Ionization chamber measurements at the isocenter of the ArcCHECK device with insert are presented in table 4 for a single arc.

The ArcCHECK diode array for dosimetric verification of HybridArc

5419

Figure 5. Screen shot of the MapCHECK software for a plan with only one arc is given. Comparison of the absolute dose distribution measured (set 1, top-left panel) and calculated in iPlan RT Dose with the Monte Carlo algorithm (set 2, top-right panel) for the cylindrical ArcCHECK diode array with the insert for this plan is presented. The green point shows the normalization point. In the middle panel, the results of the gamma analysis are presented. The red and blue points display the failed diodes with the dose above and below the γ 2 criterion, respectively. In the bottom panel the calculated dose profile (solid line) at the level of the green line in the middle panel is compared to the ArcCHECK measurement symbols; the color legend is the same as in the middle graph. Any diode can be selected to view the calculated and measured dose values at that position.

Table 2. Pass rates and mean gamma values obtained by comparing MC calculations with ArcCHECK measurements with and without the insert.

HybridArc plan ArcCHECK

Test plan Thorax Spine Head With Without With Without With Without With Without insert insert insert insert insert insert insert insert

Pass rate γ 2 Mean γ 2 Pass rate γ 3 Mean γ 3 Number diodes N0 Normalization dose

95.9% 0.44 99.1% 0.32 651 1.40 Gy

97.5% 0.38 99.5% 0.27 646 1.63 Gy

96.2% 0.41 98.8% 0.29 419 2.19 Gy

96.8% 0.39 100.0% 0.28 506 2.15 Gy

93.9% 0.45 98.9% 0.34 544 1.60 Gy

94.6% 0.44 98.7% 0.34 554 1.87 Gy

94.5% 0.39 99.5% 0.27 584 0.74 Gy

91.8% 0.42 99.3% 0.31 704 0.73 Gy

5420

A L Petoukhova et al

Figure 6. Dose map calculated in iPlan RT Dose with the Monte Carlo algorithm for the cylindrical ArcCHECK diode array with (top panel) and without the insert (middle panel) for a single arc is shown. The bottom panel presents dose map calculated in iPlan RT Dose with the Pencil Beam algorithm for the cylindrical ArcCHECK diode array with the insert. The green point shows the normalization point. The coordinates are given in centimeters.

3.3. HybridArc plans For the lung patient, the measurements of a HybridArc plan with one arc and five IMRT beams show very good agreement with the MC dose calculations even in the high-dose gradient (see table 2 and figure 7, top panel and middle panel for the ArcCHECK with the insert and without insert, respectively). Analysis of the measurements in comparison to the PB dose calculations for the lung plan is presented in table 3. In table 4, the results of the ionization chamber measurements for this HybridArc plan are given. For the spine patient, the measurements of a HybridArc plan with two coplanar arcs and two IMRT beams per arc still give good results in comparison to the MC dose calculations in

The ArcCHECK diode array for dosimetric verification of HybridArc

5421

Table 3. Pass rates and mean gamma values obtained by comparing PB calculations with ArcCHECK measurements with and without the insert.

Test plan

Thorax

Spine

Head

HybridArc plan ArcCHECK

With insert

Without With insert insert

Without With insert insert

Without With insert insert

Without insert

Pass rate γ 2 Mean γ 2 Pass rate γ 3 Mean γ 3 Number diodes N0 Normalization dose

66.4% 0.84 83.4% 0.60 649 1.42 Gy

61.5% 0.83 87.1% 0.60 649 1.66 Gy

72.9% 0.71 89.1% 0.51 516 2.29Gy

67.9% 0.83 82.5% 0.60 567 1.93 Gy

75.8% 0.67 91.5% 0.48 731 0.73 Gy

62.1% 0.86 81.4% 0.61 404 2.29 Gy

65.8% 0.90 81.1% 0.67 512 1.61 Gy

77.8% 0.69 92.5% 0.50 563 0.72 Gy

Table 4. Ionization chamber measurements at the isocenter of the ArcCHECK device with insert for the various HybridArc plans and deviations from the measurements for the MC and PB calculations.

Dose

Test plan MC

PB

Thorax Dose MC PB

Dose

Spine MC

PB

Dose

Head MC

PB

3.55 Gy −2.1% −1.3% 2.22 Gy 3.0% 1.2% 2.69 Gy 3.7% 5.8% 1.79 Gy 0.8% 1.6%

iPlan RT Dose (see table 2 and figure 8, top panel and middle panel for the ArcCHECK with the insert and without insert, respectively). Comparison of the ArcCHECK with the insert measurements to the PB calculations for the same plan is shown in figure 8 (bottom panel) and table 3. For this simultaneous integrated boost plan, more failing points for the PB dose calculation than for the MC were found. Most of these points are red and positioned along the dose calculation steps. Moreover, the dose profiles calculated with PB display much more oscillations than both the measured points and the MC results (compare top and bottom panels in figure 8). See table 4 for the ionization chamber measurements for this spine patient. For the meningioma patient, the measurements of a HybridArc plan with three noncoplanar arcs and six IMRT beams show very good results in comparison to the MC dose calculations in iPlan RT Dose (see table 2 and figure 9, top panel and middle panel for the ArcCHECK with the insert and without insert, respectively). For the ArcCHECK with insert, an analysis of the measurements in comparison with the PB dose calculations for the meningioma plan is presented in table 3. The results of the ionization chamber measurements for the meningioma patient are also shown in table 4. 4. Discussion Rotational therapy is a topic currently receiving significant attention (Bedford and Warrington 2009, Tang et al 2010, Yu and Tang 2011). The HybridArc modality in iPlan is one of the recent developments in this field. One of the advantages of HybridArc would be the inverse planning, which makes planning faster and more targeted than forward planning. Another feature is that before optimizing the IMRT beams, the DCA aperture is optimized according to the defined IMRT constraints to further improve the dose distribution. In this way HybridArc could combine the advantages of the DCA and IMRT treatments.

5422

A L Petoukhova et al

Figure 7. Comparison of the absolute dose cross-profiles measured (dots) and calculated in iPlan RT Dose with the Monte Carlo algorithm (solid lines) in the isocenter plane of the ArcCHECK diode array with (top panel) and without the insert (bottom panel) for a lung HybridArc plan with one arc and five IMRT beams is given. The red and blue points display the failed diodes with the dose above and below the 2% dose difference criterion, respectively.

To our knowledge, HybridArc treatment modality has not been discussed in the literature so far. In this work, dosimetric verification of HybridArc was investigated to be able to safely apply this treatment modality in clinical practice. This publication is not meant to give a detailed description of the applied HybridArc planning techniques. Detailed comparison of HybridArc with DCA and IMRT is also beyond the scope of this work. 4.1. Comparison between MC and PB calculations of HybridArc plans For the reference field of 9.8 × 9.8 cm2, the γ 2 pass rates were 99.1% and 86.5% for the MC and the PB calculations, respectively. Probably the reason for worse results for the PB

The ArcCHECK diode array for dosimetric verification of HybridArc

5423

Figure 8. Comparison of the absolute dose cross-profiles measured in the isocenter plane (dots) and calculated in iPlan RT Dose with the Monte Carlo algorithm (solid lines) for the ArcCHECK with (top panel) and without the insert (middle panel) for a spine HybridArc plan with two arcs and four IMRT beams is shown. The bottom panel presents the comparison of the absolute dose cross-profiles measured and calculated in iPlan RT Dose with the Pencil Beam algorithm for the ArcCHECK with the insert. The red and blue points display the failed diodes with the dose above and below the 2% dose difference criterion, respectively.

5424

A L Petoukhova et al

Figure 9. Dose map calculated in iPlan RT Dose with the Monte Carlo algorithm for the cylindrical ArcCHECK diode array with (top panel) and without the insert (bottom panel) for a meningioma HybridArc plan with three arcs and six IMRT beams is given. The green point shows the normalization point. The red and blue points display the failed diodes with the dose above and below the 2% dose difference criterion, respectively. The coordinates are given in centimeters.

calculations is that the PB model in iPlan RT dose does not have enough possibilities to properly define the shoulders of the beam profile, where the disagreement between the PB calculations and measurements was recorded. For the HybridArc plans studied here, good agreement between the measurements and the MC calculations at the cylindrical ArcCHECK diode array is observed (see table 2). It is slightly better than our clinical ArcCHECK results for QA IMRT at the Novalis accelerator calculated with PB: for 42 patients, an average pass rate of 94.9 ± 3.0 and mean gamma of 0.47 ± 0.06 for γ 3. The agreement is worse for the PB calculations of the HybridArc plans (see table 3), probably because of the 10◦ discretization in dose calculation. To the best of our knowledge, this article is the first report of the application of the commercial version of the ArcCHECK diode array. Similar studies were performed by L´etourneau et al (2009) and Yan et al (2010) using a prototype of the ArcCHECK, which had much fewer diodes. Moreover, different gamma criteria were used to analyze the results. Therefore, the quantitative comparison between the results is rather difficult. Our results for the MC calculations (see table 3 for the pass rate γ 3) are better or comparable with the results by L´etourneau (see figure 3(b) for 3%/3 mm of L´etourneau et al) for the clinical VMAT plans calculated using 180 control points in PINNACLE3 (version 8.1×, Philips Medical Systems, Madison, USA). To analyze the difference in dose calculation between MC and PB, a comparison of the MC and PB calculations for the ArcCHECK array with the insert is performed for the HybridArc plan. The results are presented in figure 10. One can see that deviations beyond the 2% dose difference criterion are observed not only in the low-dose region, but also within

The ArcCHECK diode array for dosimetric verification of HybridArc

5425

Figure 10. Dose map calculated in iPlan RT Dose with the Monte Carlo algorithm for the HybridArc plan with one arc and five IMRT beams at the central axial plane (top panel) and the cylindrical plane (bottom panel) of the ArcCHECK with insert. The red and blue areas display the failed points (difference > 2% of the normalization dose) with the Pencil Beam dose above and below the Monte Carlo result, respectively. Normalization doses are 2.86 Gy (top panel) and 3.03 Gy (bottom panel). The coordinates are given in centimeters.

the target, where the dose is high, the two calculation algorithms show significant differences. On the other hand, the majority of the points displaying large relative differences are at larger distances from the target, i.e. in the low-dose region. Therefore, we think that the PB algorithm can still be suitable for dose calculation for most of stereotactic patients, even with the 10◦ discretization in dose calculation. In the case of inhomogeneities such as lung or air cavities, the MC algorithm is superior to the PB algorithm (Reynaert et al 2007, Fragoso et al 2010, Petoukhova et al 2010) because the MC algorithm is able to properly calculate lateral electron transport. Moreover, according to Das et al (2008), the MC algorithm better treats the loss of

5426

A L Petoukhova et al

the electronic equilibrium for small fields (