ORIGINAL ARTICLE Year : 2019  Volume : 44  Issue : 4  Page : 246253 Experimental determination of radial dose function and anisotropy function of GammaMed Plus ^{192}Ir highdoserate brachytherapy source in a bounded water phantom and its comparison with egs_brachy Monte Carlo simulation Rekha Reddy Buchapudi^{1}, Ravikumar Manickam^{2}, Varatharaj Chandaraj^{3}, ^{1} Department of Radiation Physics, Kidwai Memorial Institute of Oncology, Bengaluru, Karnataka, India ^{2} Department of Radiation Physics, Kidwai Memorial Institute of Oncology; Department of Radiotherapy, Sri Shankara Cancer Hospital and Research Centre, Bengaluru, Karnataka, India ^{3} Department of Radiation Physics, Kidwai Memorial Institute of Oncology, Bengaluru, Karnataka, India; Department of Physics, Carleton Laboratory for Radiotherapy Physics, Carleton University, Ottawa, ON, Canada Correspondence Address: Objective: The aim of the present study is to experimentally measure the radial dose function g(r) and anisotropy function F(r,θ) of GammaMed Plus ^{192}Ir highdoserate source in a bounded water phantom using thermoluminescent dosimeter (TLD) and film dosimetry and compare the obtained results with egs_brachy Monte Carlo (MC)calculated values for the same geometry. Materials and Methods: The recently developed egs_brachy is a fast Electron Gamma Shower National Research Council of Canada MC application which is intended for brachytherapy applications. The dosimetric dataset recommended by Task Group 43 update (TG43U1) is calculated using egs_brachy for an unbounded phantom. Subsequently, radial dose function g(r) and anisotropy function F(r,θ) are measured experimentally in a bounded water phantom using TLD100 and Gafchromic EBT2 film. Results: The TG43U1 dosimetric parameters were determined using the egs_brachy MC calculation and compared with published data which are found to be in good agreement within 2%. The experimentally measured g(r) and F(r,θ) and its egs_brachy MC codecalculated values for a bounded phantom geometry are found to be good in agreement within the acceptable experimental uncertainties of 3%. Conclusion: Our experimental phantom size represents the average patient width of 30 cm; hence, results are closer to scattering conditions in clinical situations. The experimentally measured g(r) and F(r,θ) and egs_brachy MC calculations for bounded geometry are well in agreement within experimental uncertainties. Further, the confidence level of our comparative study is enhanced by validating the egs_brachy MC code for the unbounded phantom with respect to consensus data.
Introduction Although a miniaturized highspecific activity 60 Co source is available on remote after loading equipment, a 192 Ir source is being still widely used for highdoserate (HDR) brachytherapy treatment, mainly addressing gynecological lesions. Any radioactive source used in HDR brachytherapy for clinical practice needs a substantial amount of dosimetric data, as recommended by the American Association of Physicists in Medicine, TG43U1.[1] Experimental measurement of such data may result in large uncertainties because of the rapid fall of the dose at distances near the source, and this limitation can be overcome by accurate Monte Carlo (MC) simulations.[2] There have been many studies comparing these dosimetric data using either experimental or MC studies.[3],[4],[5],[6],[7] However, most of these studies have compared their results with MC calculations, which do not resemble the exact geometry conditions used in an experimental setup. It may not be possible to compare the MC calculations of unbounded geometry with bounded experimental results or MCcalculated values because of the differences in scattering conditions. This would result in dose differences of >10% occurring near the periphery of the bounded phantom, as mentioned by Granero et al.[8] and Venselaar et al.[9] The aim of the present study is to compare the experimentally measured radial dose function g(r) and anisotropy function F(r,θ) in a bounded water phantom using thermoluminescent dosimeter (TLD100 rods) and EBT2 Gafchromic film with egs_brachy MC code calculation by simulating the similar experimental conditions. Before this comparison, the efficacy of egs_brachy MC code was validated by calculating the dose rate constant (Λunb), along–away dose rate data, radial dose function (g(r) unb), and anisotropy function (F(r,θ) unb) around the GammaMed (GM) Plus 192 Ir source in an unbounded (unb) liquid water phantom in comparison with published consensus data.[10] This was found to be the routine validation procedure for any new MC code calculation. The dataset calculated in the present work can be considered an input for radiotherapy treatment planning systems (TPSs) or for their quality control. Materials and Methods Monte Carlo simulation details egs_brachy Monte Carlo code Because there were not many fast MC codes available specifically for brachytherapy simulations, the Electron Gamma Shower National Research Council of Canada (EGSnrc) user code called BrachyDose [11] was developed to address this need. It uses a multigeometry package by Yegin.[12] In addition, a C++based EGSnrc library called egs ++ was introduced,[13] which led to the development of egs_brachy for modeling the particle sources and geometry specifically for brachytherapy applications. The main features of egs_brachy are that it includes a comprehensive library of brachytherapy source geometries, enhanced simulation efficiency, calculation of collision kerma using the track length estimator, phasespace sources, efficient radiation transport and geometry modeling, particle recycling, and variance reduction techniques for electronic brachytherapy. A publication by Chamberland et al.[14] provides a general overview of the code, complete discussion of all egs_brachy features, details on egs_brachy benchmarking, and characterization of the simulation efficiency of egs_brachy MC code. Modeling of the GammaMed Plus 192 Ir highdoserate source The GM Plus 192 Ir source (Mallinckrodt Medical B. V., Petten, The Netherlands) is one of the HDR brachytherapy sources commonly used for the management of most malignancies. The GM Plus 192 Ir source consists of a 3.50mmlong 192 Ir core with a diameter of 0.70 mm, enclosed in a 0.90mmdiameter and 4.52mmlength AISI 316 L stainless steel capsule (density of 7.8 g/cm 3). The Ir192 source emits a wide spectrum of relatively low energies, mostly in the range of 201–884 keV with an average value of 360 keV. A total of 6.0 cm of stainless steel cable is included in this simulation. The geometric design and material of the GM Plus source details are taken from a published study.[15] The schematic egs_brachy modeled source is shown in [Figure 1].{Figure 1} egs_brachy calculation for unbounded phantom The dosimetric dataset was calculated for GM Plus 192 Ir as recommended by TG43U1 for an unbounded phantom similar to the approach used by Taylor and Rogers.[16],[17] A cylindrical phantom of 80 cm in length and 40 cm in radius filled with liquid water having a density of 0.998 g cm − 3 was modeled for MC simulations. The scoring region, voxel sizes, and other parameters for calculations are as described by Chamberland et al.[14] All TG43U1 parameters for unbounded phantom calculations are denoted as superscript to the respective parameters, such as Λunb,g(r) unb, and F(r,θ)unb. For MC calculations, only the photon part of the 192 Ir source spectrum is included, as the dose contribution from the electron is negligible, because it is stopped by the stainless steel encapsulation around the source. The cutoff energy for the photon calculation is up to 1 keV. The photoelectric absorption, Rayleigh scattering, fluorescent emission of characteristic Xrays, and bound Compton scattering are modeled in calculations of egs_brachy MC code. The XCOM database [18] and Livermore Evaluated Atomic Data Library [19] for the photon cross sections and atomic transitions, respectively, are included and used the EGSnrc user code “g” for the massenergy absorption coefficients. The simulations are done up to 4 × 109 histories to get 1 σ statistical uncertainties (Type A) of 2% or less. Experimental dosimeters and phantom design Thermoluminescent dosimeters A fresh batch of TLD100 square rods (LiF:Mg, Ti) with dimensions of 1 mm × 1 mm × 6 mm was used. The annealing procedure before each experiment called the “prereadout” method was performed as described by Booth et al.[20] The Harshaw Bicron TLD reader (Model 3500) and the Thermolyne Furnace (Model 47900) were used for analyzing the TLD response and for annealing purposes, respectively. The thermoluminescent output was measured in nanocoulombs by integrating the area under the glow curve for a temperature of 270°C. The whole batch of TLD was irradiated using 60 Co γrays from a Theratron 780E telecobalt unit to deliver a dose of 2 Gy, and the relative responses, the elemental correction factors, were determined. This procedure was repeated five times, and the TLD rods that showed variation above 2% were discarded. The output of the each TLD was corrected with respective calibration factors to obtain a precision on the order of ± 1% (1 σ, Type A), and the TLD rods showed a linear response up to 10 Gy. Gafchromic EBT2 film The Gafchromic EBT2 films (ISP Technologies) are highly sensitive, having high spatial resolution, and they were used in the dose range of 0.01–40 Gy. For the purpose of calibration, the film was cut into 4 cm × 4 cm samples and marked at the left corner to reproduce the orientation. These sample films were irradiated for the dose range of 0.1–40 Gy in 60 Co γrays from a Theratron 780E telecobalt unit. The scanning of the irradiated film after 24 h of exposure was carried out on an EPSON Dual Lens Perfection V700 desktop scanner. The filmscanning protocol was adopted from the published literature.[21] PTW Verisoft version 6.0.1 (PTWFreiburg, Germany) software was used for analyzing scanned films in tag image file format. The pixel values of the irradiated and unirradiated films were used to obtain the optical density and converted into dose as described in the literature.[22] Experimental water phantom and slab inserts A precisely machined 30 cm × 30 cm × 30 cm water phantom with polymethyl methacrylate (PMMA) walls with thicknesses of 1 cm was indigenously fabricated for the experimental measurement. For measuring radial dose function g(r), a PMMA slab insert with dimensions of 30 cm × 30 cm × 1 cm was carefully machined for housing TLD rods and an MRI compatible tube, inside which the source was driven as shown in [Figure 2]a. With this TLD arrangement in the slab, eight measurements were simultaneously performed at each distance. This pattern for the TLD locations was selected to minimize the interference of anyone TLD with the absorbed dose measured by the other TLD rods. The slab containing the TLDs was inserted horizontally in a water phantom and located at the center of the phantom, surrounded by a water medium. For the film measurements, Gafchromic EBT2 film was attached to the PMMA slab to hold the film rigidly in the water medium and was inserted into the water phantom. The irradiation conditions were kept the same as those for the TLDs. Another PMMA slab insert of 30 cm × 30 cm × 1 cm was fabricated for measuring anisotropy functions, as shown in [Figure 2]b. This PMMA slab containing the TLDs was placed vertically to be at the center of the water phantom. The measurement conditions for EBT2 film were kept similar to those of the TLDs. The design of the PMMA slab inserts was taken from Meigooni et al.[23] The positional accuracy of the source in the phantom with respect to the TLD and film was verified before each measurement, using square samples of the EBT2 films.{Figure 2} Measurement techniques The measurements were carried out using Varian GM Plus HDR unit (Varian Medical Systems, USA), and the dwell position and dwell times were planned with BrachyVision TPS. The dwell times for irradiating TLD100 rods to a dose of 3 Gy at all measurement distances from 1 cm to 10 cm varied from 23.8 s to 856.8 s, respectively, for a nominal activity of 370 GBq (10 Ci) source strength. For the EBT2 film measurements, dwell time was 82.9 s for a dose of 8 Gy at 1 cm. Thermoluminescent dosimeter100 For experimental measurements with TLDs, the g(r) and F(r,θ) are denoted as g(r) TLD and F(r,θ) TLD. The g(r) TLD and F(r,θ) TLD were measured at the distances and polar angles, as shown in [Figure 2]. An average reading with reproducibility of better than 1% from three consecutive measurements with TLD at each point was considered. As mentioned by Thomason and Higgins,[24] volume correction factors for the finite size of the TLDs are calculated as 1.028 for 1cm distance and 1.0 for distances beyond 1 cm and applied at respective radial distances. All the TLD100 rods were exposed to doses <3 Gy. EBT2 Gafchromic film The measured g(r) and F(r, θ) of EBT2 films are denoted as g(r) film and F(r,θ) film. The measurement distances and polar angles for g(r) film and F(r, θ) film were similar to the TLD measurement for unirradiated and irradiated films. The experimental setup to measure radial dose function g(r) and anisotropy function F(r,θ) using the fabricated water phantom with respective PMMA slabs is shown in [Figure 3]. The irradiated films were scanned using “face up” protocol at 150 DPI and 48bit color depth.{Figure 3} egs_brachy calculation of g(r) and F(r,θ) for bounded phantom For the bounded (bou) phantom, egs_brachy calculation g(r) and F(r,θ) are denoted as g(r) bou and F(r,θ) bou. A liquid water phantom of dimension 30 cm × 30 cm × 30 cm was simulated to reproduce the experimental phantom geometry in the MC calculations. The calculation parameters were the same as mentioned in the earlier section on egs_brachy calculation for the unbounded phantom. It is relevant to compare the experimentally measured values in a bounded geometry against egs_brachy MCcalculated values of a similar bounded geometry. Results The obtained dose rate constant Λunb for the GM plus 192 Ir source using egs_brachy MC calculations was 1.110 ± 0.005 cGyh − 1 U − 1 and deviates with consensus data by 0.6%. The g(r) unb results calculated from a radial distance of 0.2–20 cm with corresponding calculated uncertainties are shown in [Table 1]. [Figure 4] shows a comparison with consensus data [10] and the results reported by Taylor et al.[17] The maximum variation was observed as 0.8% at 0.2 cm distance with consensus data and 0.6% at 8 cm distance with Taylor et al.,[17] and the approximate uncertainty in our calculation was 0.3%. [Table 2] summarizes the calculated F(r,θ) unb, and it is found to be in agreement with consensus data within 2%. The overall uncertainties (Type A) were 1.5%, 0.5%, and 1.5% for θ < 5, 7≥ θ ≤170, and 175≥ θ ≤179, respectively, for the radial distance of 0.25–20 cm. The along–away dose rate values are shown in [Table 3] which shows a variation of <2% in comparison with consensus data. The dose rate along z = 0 cm and away y = 0.2 cm is 22.8 cGy h − 1 U − 1, and the corresponding value calculated by Ballester et al.[15] is 23.2 cGy h − 1 U − 1. In addition to the consensus data, the TG43U1 dataset is found to be in good agreement, with a variation of <2%, with Taylor et al.[17] for the same source.{Table 1}{Figure 4}{Table 2}{Table 3} [Table 4] shows the experimentally measured g(r) TLD and g(r) film for radial distances from 1 to 10 cm in comparison with egs_brachy MCcalculated g(r) bou values from 0.2 to 10 cm. A maximum variation of 2.8% was found between g(r) TLD and g(r) bou, and the range of variation was from 0.7% to 2.8%. In a similar comparison between g(r) film and g(r) bou, the maximum variation was found to be 2.1%, and the range was from 0.08% to 2.1%. The approximate uncertainties (Type A) in the measurements were 1% for film, 1.5% for TLD100, and 0.5% for egs_brachy calculation.{Table 4} [Figure 5] shows the comparison between experimentally measured F(r,θ) TLD and F(r,θ) film with egs_brachy MCcalculated values F(r, θ) bou for the radial distances of 1, 5, and 10 cm. The maximum variation of 2.5% was found between F(r,θ) TLD in comparison with egs_brachy, and most of the variations were well below 2%. Similarly, comparing F(r,θ) film with F(r,θ) bou, most of the values are within 1.5% variation, and the maximum variation was 1.7%. The approximate uncertainties (Type A) in the measurements were 1% for film, 1.5% for TLD100, and 1.5% for egs_brachy calculation.{Figure 5} Discussion In HDR brachytherapy dosimetry, it is a routine methodology to verify the MC calculations with experimental methods by any possible dosimeters, such as TLD and film, in regions where the experimental uncertainties are minimum. On successful validation, the MCcalculated dosimetric parameters can be used as input to the clinical dosimetry through TPSs. Williamson [25] compared the MC calculations for a 192 Ir source assuming an unbounded liquid water medium with the experimentally measured data in a 20 cm × 20 cm × 20 cm 3 phantom reported in the literature and found that up to 5 cm from the source showed good agreement but varied from 5% to 10% at larger distances. Richter et al.[26] compared the dosimetric parameters of 60 Co and 192 Ir sources in HDR brachytherapy using MC calculation in spherical phantoms of two different radii of 15 and 50 cm to find the influence of phantom size on the dose at larger distances from the source. Their results show that radial dose function is influenced by the phantom size. The consensus data [10] for the GM plus 192 Ir HDR source refer to Ballester et al.,[15] in which GEANT3 MC code was used for simulating an unbounded cylindrical water phantom of 40cm diameter and 40cm length. The TG43U1 dataset consists of dose rate constant, radial dose function, anisotropy function, and twodimensional alongandaway dose rate table. In the present study, the TG43U1 dosimetric dataset was calculated for a cylindrical liquid water phantom of 80cm length and 40cm radius using egs_brachy MC code and compared with consensus data for its validation. In a literature survey, numerous publications were found on TLDs used for brachytherapy dose measurements including the dosimetric dataset based on TG43U1.[27],[28],[29] However, there are many conflicting results reported in the literature based on TLD measurements in a phantom. The reasons are (i) energy dependence of TLD when calibrating TLDs in a 60 Co photon beam and measuring in a 192 Ir photon spectrum and (ii) the depthdependent response of TLD from the source. Mobit et al.[30] found that, even for 60 Co gamma energies, the TLD100 rods behave like a large cavity rather than a small cavity, and the MCcalculated average dose ratio of water to LiF for 60 Co gamma rays is not more than 0.8%. Karaiskos et al.[31] also performed weighted photon spectra calculations for LiF using MC code, and their findings are in agreement with Mobit et al.[30] They also found that the variation in TLD response with the shift in the 192 Ir spectrum toward lower energies up to a depth of 15 cm is within 3%, which is within the error for the experimental setup. Das et al.[27] did not apply either an energy correction factor or a depth correction factor when calibrating TLDs using a 6MV beam for the measurements in a 192 Ir beam and justified their methodology based on Pradhan and Quast [32] They acknowledged that there is an overresponse of TLDs at depth in a 192 Ir phantom, and this correction was found to be within 3% up to 10 cm from the source, which is negligible when measurement uncertainties are considered. As mentioned by Arjomandy et al.,[33] EBT 2 Gafchromic film shows a weak energy dependence for clinical useful energies. Based on these studies, herein, no correction factors for both energy and depth dependence were considered in the experimental work. Subsequent to the validation by egs_brachy calculation, the radial dose function g(r) and anisotropy function F(r,θ) of the GM Plus 192 Ir source using TLD100 and EBT2 Gafchromic film in an indigenously fabricated water phantom were measured and compared with the egs_brachy calculation. To reduce the uncertainty in MC calculation and to get accurate calculation results, the geometry of the phantom was mimicked to be similar to the experimental setup as a bounded water phantom with PMMA wall material. From the experimental results of g(r) and F(r,θ), the observed variation with egs_brachy code calculation was found to be reasonably well within the acceptable experimental uncertainties of 3%. Conclusion The experimentally measured parameters and their comparison with egs_brachy MC calculations for bounded geometry are well within the experimental uncertainties. There are no published values in literature for the source type studied with a bounded water phantom using two different dosimeters in comparison with MC calculation for the same geometry. Further, the confidence level of the comparative study was enhanced by validating egs_brachy MC code for an unbounded phantom with respect to consensus data. The experimental phantom size represents the average patient width of 30 cm; hence, the results are closer to scattering conditions in clinical situations. Acknowledgment The authors would like to thank Dr. D.W.O. Rogers and Dr. Marc J P Chamberland, CLRP, Carleton University, Ottawa, Canada, for their valuable inputs related with egs_brachy MC code. The authors are also grateful to Mr. T. VijayaReddy, KMIO, Mr. Sundarraj Pillai, Kalyani Radiotherapy Specialty India (P) Ltd, and Mr. K. Kanakavel, PTW Dosimetry India (P) Ltd, for their support related with the experimental work. This work has been supported by a UICC International Cancer Technology Transfer Fellowship. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest. References


