Worldwide, intensity modulated radiation therapy (IMRT) and volumetric modulated arc therapy (VMAT) are routinely used in clinical practice. These techniques have improved patient outcomes compared with three-dimensional (3D) conformal radiotherapy   .
The multi-leaf collimator (MLC), gantry position, and/or dose rate vary continuously during IMRT and VMAT delivery. Such complicated beam delivery techniques lead to dosimetric differences between measurements and calculations; therefore, patient-specific quality assurance (QA) is required to validate the calculated dose distribution  . Patient-specific QA procedures differ from traditional approaches using an ionization chamber and films to two-dimensional (2D) planar or 3D volumetric dosimeters. One of the advantages of 3D volumetric dosimetry over 2D planar dosimetry is that volumetric dose distributions can be assessed with a single measurement.
Dose calculation algorithms have evolved dramatically from measurement- based to model-based, linear Boltzmann transport equation (LBTE)-based, and Monte Carlo (MC)-based methods. Model-based algorithms, such as superposition/convolution algorithms, anisotropic analytic algorithms (AAA), and collapsed cone convolution/superposition (CCCS), provide more accurate results in low-density areas than measurement-based algorithms. The model-based calculation algorithm computes transport and dose deposition using radiological and density scaling deposited by the absorbed dose in water. LBTE- and MC-based dose calculation algorithms, such as Acuros XB (AXB) and X-ray voxel Monte Carlo (XVMC), respectively, are more accurate in heterogeneous regions than model-based dose calculation algorithms  -  . As a feature of the AXB and XVMC, these algorithms can calculate different dose reporting modes in water [dose-to-water (Dw)] and in medium [dose-to-medium (Dm)] through the stopping power ratio of water-to-medium    .
Recently, ArcCHECK (Sun Nuclear Corp., Melbourne, FL), which is one of the 3D volumetric dosimeters, has been clinically used as a patient-specific QA tool. Several studies on ArcCHECK have been reported with a focus on angular, directional, and field size dependencies, and response to scattering and leakage radiations in a fixed arc  -  . In these reports, patient-specific QA using ArcCHECK was also performed with AAA or CCCS algorithms  -  . Eaton et al. reported the comparison of measured dose distributions and calculated ones with several dose calculation algorithms, including MC-based method, by use of ArcCHECK. However, the report makes no mention of dose reporting modes in MC-based method  . The dose distribution between Dw and Dm was different in the “water-unlike” tissues with densities, due to the difference in stopping power ratio and fluence  . Taken together, a comparison between measured and calculated dose distributions in Dw and Dm using ArcCHECK were not fully elucidated. Before clinical implementation of ArcCHECK for IMRT and VMAT plans, qualified medical physicists should familiarize themselves with its characteristics against dose calculation algorithms.
The purpose of this study was to compare measured dose distributions and calculated ones in Dw and Dm in simple, patient-specific IMRT and VMAT plans using ArcCHECK with the fixed phantom density recommended by the vendor.
2. Materials and Methods
2.1. ArcCHECK Device
ArcCHECK is a cylindrical polymethylmethacrylate (PMMA) phantom with a 3D array of 1386 diode detectors with a 1-cm detector spacing. ArcCHECK has an outer diameter of 26.6 cm and an inner-hole diameter of 15.1 cm, with the curved plane of diodes at a distance of 10.4 cm from the center. The detectors spiral down the cylinder length to increase the spatial sampling rate and reduce the detector overlap from the beam’s eye view. The active detector size is 0.8 mm × 0.8 mm. ArcCHECK measures the dose data in 50-ms intervals, saves all the measurement data as a function of time, and measures both relative and absolute doses.
2.2. Absolute Dose Calibration
A water tank (WP1D Water Phantom; IBA Dosimetry GmbH, Germany) was used to measure the absolute dose at a calibration depth of 3.3 cm according to the guidelines provided by Sun Nuclear Corporation (SNC)  . The physical distance from the top surface of the ArcCHECK to the diode detectors was 2.85 cm, which corresponds to a water-equivalent depth of 3.3 cm, as described above.
Before measuring the dose at the calibration depth of 3.3 cm in water, the dose per monitor unit (D/MU) was measured using the ionization chamber (PTW30013; PTW Co., Freiburg, Germany) annually calibrated by the Japan calibration service system. The chamber was irradiated perpendicular to the beam axis with a radiation field size of 10 × 10 cm2 at a depth of 10 cm in water. The center of the ionization chamber was at a source-axis-distance of 100 cm. The dose delivered at the given depth was calculated by multiplying the delivered monitor unit, D/MU, and the tissue maximum ratio. Thereafter, the doses were measured at a depth of 3.3 cm and a source-surface distance of 86.3 cm in water, according to SNC instructions  . The effective point of the measurement of the ionization chamber was located at the measuring depth. The absolute dose was corrected by D/MU and a displacement correction factor.
2.3. Absolute Dose Calibration
Measurements were performed on four treatment machines, including a Clinac iX (Varian Medical Systems, Palo Alto, CA), a TrueBeam STx (Varian), a Vero (Mitsubishi Heavy Industries [MHI], Ltd., Tokyo, Japan) at Kyoto University Hospital, and a TrueBeam (Varian) at Osaka Red Cross Hospital. A 120-leaf Millennium MLC was mounted on the Clinac iX and TrueBeam.
Treatment plans were designed on Eclipse (version 11; Varian) for Varian treatment machines, and iPLAN RT Dose (version 4.5.3; Brainlab AG, Feldkirchen, Germany). AXB (version 11.0.31), with a 2.5-mm grid resolution for Eclipse, and XVMC (version 4.5.3), with a 2.0-mm grid resolution and 2% dose variance for iPLAN, were used for dose calculation. Table couch structures were used in the calculations. The treatment planning information is summarized in Table 1.
2.4. ArcCHECK Phantom-Mass-Density
The computed tomography (CT) dataset of the ArcCHECK was provided from SNC for dose calculations. The user manual recommended that the ArcCHECK phantom-mass-density should be assigned a density close to that of the PMMA material for the treatment planning system (TPS) dose calculation  . The recommended density value of 1.18 g/cm3 for LBTE- and MC-based dose calculation algorithms (e.g., AXB and XVMC), was assigned to the ArcCHECK on CT images. The density value of 1.18 g/cm3 for AXB and XVMC was the combination of cartilage and bone density in TPS.
Table 1. Treatment planning information.
Abbreviations: MLC, multi-leaf collimators; HD, high-definition; AXB, Acuros XB; XVMC, X-ray Voxel Monte Carlo; FFF, flattening filter free.
2.5. Dosimetric Evaluation for Simple Plans
A total of 45 simple plans, including the 1-field anterior-posterior direction (1F), the 3-field anterior and left-right direction (3F), the 4-field anterior-posterior and left-right direction (4F), a half-arc delivery (half-arc) from 270˚ to 90˚ gantry angle under clockwise rotation, and a full-arc delivery (full-arc) from 181˚ to 179˚ gantry angle under clockwise rotation, were assessed. A field size of 10 × 10 cm2 was used for all simple plans. The dose distributions were assessed quantitatively by gamma analysis with a 3% normalized global dose error and a 3-mm DTA threshold (γ3%G/3mm) criterion using SNC software (ver. 6.6) for all plans.
2.6. Dosimetric Evaluation for Patient-Specific Plans
A total of 96 patients were assessed retrospectively in this study. Of them, QA for 76 plans was conducted at Kyoto University Hospital and 20 at Osaka Red Cross Hospital. Eighty-five patients were irradiated with VMAT by Varian machines; static IMRT by Vero was used for the remaining 11 patients. This study was approved by institutional review board. A summary of treatment machines, energy, delivery technique, and representative disease site is given in Table 2.
Table 2. Summary of treatment machines, energy, and delivery technique.
Abbreviations: KUHP, Kyoto University Hospital; ORC, Osaka Red Cross Hospital; VMAT, volumetric modulated arc therapy; IMRT, intensity modulated radiotherapy; FFF, flattening filter free.
Gamma analysis was used to evaluate patient-specific QA. The gamma index at γ3%G/3mm was assessed in Dw and Dm. As subgroup analyses, dose distributions due to differences in energies, delivery techniques, and treatment machines were also assessed in Dw and Dm.
The paired Wilcoxon signed-rank test explored the change in γ3%G/3mm between Dw and Dm. Data were statistically analyzed using JMP Pro 11 software (SAS Institute, Inc., Cary, NC, USA). The level of significance for all tests was set at a p-value of 0.05.
3.1. Absolute Dose Calibration
The differences between the measured dose at a depth of 2.85 cm in ArcCHECK and the calculated dose at a depth of 3.3 cm in the water phantom were within ±1.0% for any combination in Dw and Dm.
3.2. Dosimetric Evaluation for Simple Plans
A summary of the γ3%G/3mm for simple plans is presented in Figure 1. The median values of γ3%G/3mm for all simple plans, including 1F, 3F, 4F, half-arc, and full-arc, in Dw and Dm were 98.1% (range, 75.2% - 100%) and 95.5% (range, 23.7% - 100%), respectively. The median pass rates of γ3%G/3mm for all simple plans in Dw were significantly higher than those in Dm (p < 0.01).
3.3. Dosimetric Evaluation for Patient-Specific Plans
The pass rates of γ3%G/3mm in Dw were significantly higher than those in Dm.
Figure 1. Boxplot of γ3%G/3mm in dose-to-water (Dw) and dose-to-medium (Dm) for simple plans. The γ3%G/3mm in Dw was significantly higher than those in Dm (p < 0.01).
Figure 2 shows a boxplot of γ3%G/3mm for patient-specific plans. The median pass rates of γ3%G/3mm in Dw and Dm were 98.6% (range, 90.1% - 100%) and 90.5% (range, 38.5% - 97.2%), respectively. Figure 3 shows boxplots of γ3%G/ 3mm under various energies, delivery techniques, and treatment machines. The results also showed that all γ3%G/3mm in Dw were significantly higher than those in Dm (p < 0.01). Figure 4 shows an example of dose profiles of patient-specific QA with the 10FFF-MV X-ray from the TrueBeam STx. The maximum dose difference between measurement and calculation was −0.8% and 3.5% in Dw and Dm, respectively. The doses calculated in Dw were 4.2% systematically higher than those in Dm.
The present study is the first reported to assess the comparison between measured dose distributions and calculated ones in Dw and Dm using ArcCHECK for various plans. We found that the median pass rates of γ3%G/3mm for all simple plans in Dw were significantly higher than those in Dm (Figure 1). In addition, the dose distribution calculated in Dw agreed with the measured one to a high level of significance, compared with those in Dm (Figure 2). AAPM TG 119 reported that the average percentage of γ3%G/3mm for IMRT was 96.3% ± 4.4% (mean ± 2σ [σ, standard deviation])  . In addition, several studies have reported pass rates of around 95% with γ3%G/3mm using ArcCHECK  -   . In this study, the median pass rate of γ3%G/3mm was 98.6% in Dw, which is supported by previous reports.
Worldwide, the traceability of QA dosimetry is established in water; therefore, dose calibration protocols for treatment machines and treatment planning
Figure 2. Boxplot of γ3%G/3mm in dose-to-water (Dw) and dose-to-medium (Dm) for patient-specific plans. The γ3%G/3mm in Dw was significantly higher than those in Dm (p < 0.01).
Figure 3. Boxplots of γ3%G/3mm under various (a) energies, (b) delivery techniques, and (c) treatment machines in dose-to-water (Dw) and dose-to-medium (Dm). The legend for the box plots is described in Figure 2. The results also showed that all γ3%G/3mm in Dw were significantly higher than those in Dm (p < 0.01).
Figure 4. Dose profiles of patient-specific QA with 10FFF-MV X-ray from TrueBeam STx. (a) dose-to-water (Dw) and (b) dose-to-medium (Dm). Dots show measurement values. The yellow, red, and blue dots indicate agreement, higher, and lower doses compared with the calculated dose distribution in Dw and Dm using γ3%G/3mm criteria, respectively. The dose of the representative point in box information shows the measured dose and calculated dose. The doses calculated in Dw were 4.2% systematically higher than those in Dm.
systems should be water-based. The SNC user manual recommends that the ArcCHECK dose calibration be performed in water 25. In this study, the differences between the measured dose at a depth of 2.85 cm in ArcCHECK and the calculated dose at a depth of 3.3 cm in the water phantom were within ±1.0%. Therefore, the assigned absolute dose to ArcCHECK was clinically acceptable from the viewpoint of traceability.
The conversion of Dm to Dw or vice versa requires the application of Bragg- Gray cavity theory; , where is the unrestricted water-to-medium mass collision stopping power averaged over the energy spectra of primary electrons at the point of interest    . The term refers to the energy loss by electrons per unit path length of a material, which indicates that the amount of scattered photons and secondary electrons is dependent on . When dose distributions are calculated in Dw, the behavior of the energy loss by electrons per unit path length for PMMA might be identical to those for AAA and CCCS  . Therefore, the γ3%G/3mm of our results calculated in Dw were consistent with those calculated in AAA and CCCS in the previous studies  -  .
The possibility, at first provided by Monte Carlo simulation, of reporting the dose as Dm, has still become a subject of controversy in medical physics community   . For tissues with densities around 1.00 g/cm3, the dose difference between Dw and Dm for megavoltage photon beams is generally small (1% - 2%)  . The clinical impact of switching from Dm to Dw is not expected to be significant, mainly because most tissues of interest in radiotherapy are similar to water   . Actually, the dose distributions are routinely calculated in Dm in our institution. Therefore, even though the pass rates of γ3%G/3mm is low in Dm of the ArcCHECK, it does not mean that quality of treatment plans in the patients calculated in Dm is poor. When dose distributions are calculated in Dm for patient-specific QA for IMRT and VMAT plans, the density value of 1.00 g/cm3 should be assigned to the ArcCHECK on CT images. Note that, however, this density assignment is not recommended by SNC  .
The cause of lower γ3%G/3mm in Dm may be due to , which varies with phantom-mass-density. According to Siebers et al., the of water-to-PMMA was 1.035, which was almost constant, even at different energies  . Therefore, the tendencies of γ3%G/3mm between Dw and Dm were similar under various energies [Figure 3(a)], and systematic declines in absolute dose were observed in Dm due to (Figure 4). Similar tendencies were also seen under various delivery techniques and treatment machines [Figure 3(b) and Figure 3(c)]. From our results and the rationale discussed in this study, our approach may be clinically reasonable.
Disease heterogeneity, included in this study, and characteristics of γ3%G/3mm would be causes of large variations between the results in Dw and those in Dm. As shown in the results of the 10FFF-MV X-ray in Figure 3(a) and Vero in Figure 3(c), the variations were almost the same between Dw and Dm due to disease homogeneity. However, larger variations were observed for other results, which would be derived from the characteristics of γ3%G/3mm. The concept of the gamma evaluation method combines the dose difference criterion with a DTA criterion for each point of interest  . If a global dose difference of
This study demonstrated that measured dose distributions and calculated ones in Dw agreed well with each other, and showed a greater agreement than those in Dm for a range of energies, delivery techniques, and treatment machines from different institutions. From the viewpoint of the rationale of dosimetry, Dw shows better agreement with measured dose distribution when using the fixed phantom density recommended by the vendor.
We acknowledge financial support by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (Grant 25253078).