Journal of Cancer Research and Therapeutics

: 2013  |  Volume : 9  |  Issue : 3  |  Page : 402--409

Dosimetric comparison between three dimensional treatment planning system, Monte Carlo simulation and gel dosimetry in nasopharynx phantom for high dose rate brachytherapy

Zeynab Fazli1, Mahdi Sadeghi2, MH Zahmatkesh3, Seied Rabei Mahdavi4, Claudio Tenreiro5,  
1 Department of Nuclear Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran
2 Agricultural, Medical and Industrial Research School, Nuclear Science and Technology Research Institute, Karaj, Iran
3 Department of Medical Physics, Novin Medical Radiation Institute, Tehran, Iran
4 Department of Medical Physics, Faculty of Medicine, Tehran University of Medical Sciences, Tehran, Iran
5 Department of Energy Science, Sung Kyun Kwan University, Suwon, Korea

Correspondence Address:
Mahdi Sadeghi
Agricultural, Medical and Industrial Research School, Nuclear Science and Technology Research Institute, P.O. Box 31485-498, Karaj


Purpose: For the treatment of nasopharnx carcinoma (NPC) using brachytherapy methods and high-energy photon sources are common techniques. In the common three dimensional (3D) treatments planning, all of the computed tomography images are assumed homogeneous. This study presents the results of Monte Carlo calculations for non-homogeneous nasopharynx phantom, MAGICA normoxic gel dosimetry and 3D treatment planning system (TPS). Materials and Methods: The head phantom was designed with Plexiglas cylinder, head bone, and nasopharynx brachytherapy silicon applicator. For the simulations, version 5 of the Monte Carlo N-particle transport code (MCNP5) was used. 3D treatment planning was performed in Flexiplan software. A normoxic radiosensitive polymer gel was fabricated under normal atmospheric conditions and poured into test tubes (for calibration curve) and the head phantom. In addition, the head phantom was irradiated with Flexitron afterloader brachytherapy machine with 192 Ir source. To obtain calibration curves, 11 dosimeters were irradiated with dose range of 0-2000 cGy. Evaluations of dosimeters were performed on 1.5T scanner. Results: Two-dimensional iso-dose in coronal plan at distances of z = +0.3, –0.3 cm was calculated. There was a good accordance between 3D TPS and MCNP5 simulation and differences in various distances were between 2.4% and 6.1%. There was a predictable accordance between MAGICA gel dosimetry and MCNP5 simulation and differences in various distances were between 5.7% and 7.4%. Moreover, there was an acceptable accordance between MAGICA gel dosimetry and MCNP5 data and differences in various distances were between 5.2% and 9.4%. Conclusion: The sources of differences in this comparison are divided to calculations variation and practical errors that was added in experimental dosimetry. The result of quality assurance of nasopharynx high dose rate brachytherapy is consistent with international standards.

How to cite this article:
Fazli Z, Sadeghi M, Zahmatkesh M H, Mahdavi SR, Tenreiro C. Dosimetric comparison between three dimensional treatment planning system, Monte Carlo simulation and gel dosimetry in nasopharynx phantom for high dose rate brachytherapy.J Can Res Ther 2013;9:402-409

How to cite this URL:
Fazli Z, Sadeghi M, Zahmatkesh M H, Mahdavi SR, Tenreiro C. Dosimetric comparison between three dimensional treatment planning system, Monte Carlo simulation and gel dosimetry in nasopharynx phantom for high dose rate brachytherapy. J Can Res Ther [serial online] 2013 [cited 2021 Oct 21 ];9:402-409
Available from:

Full Text


In the nasopharynx as in other head and neck sites, carcinomas are the most common type of malignancy. The management of nasopharyngeal carcinoma (NPC) has been of great interest to radiation oncologists. Surgery has had a limited role, and radiation is necessary for almost of the patients. In the case nasopharynx cancer we treat the primary cancer by external radiotherapy, after completion of the external radiotherapy part of the treatment; a booster dose is applied by brachytherapy. The local control rates achieved by radiation suggest that NPC may be more radiosensitive than carcinomas arising from other head and neck sites. Nevertheless, management remains a challenge because high doses are needed for an efficient treatment and the tumors are often adjacent to critical neural structures. Therefore, irradiation of critical neural could lead to devastating side-effects. One of the advancements in the management of NPC is the use of high dose rate (HDR) brachytherapy. 192 Ir sources are the most commonly used sources for interstitial brachytherapy. [1] HDR brachytherapy using 192 Ir remote after-loaders has gained acceptance in recent years because the irradiation times are short enough to allow treatment on an out-patient basis. HDR brachytherapy is a highly extended practice in clinical brachytherapy today. It should be mentioned that the quality dose rate distribution datasets of the HDR sources used in a clinical treatment are required.

Little dosimetric practice has been emphasized, especially for the treatment planning. Hence, new question arise: Even though current treatment planning in HDR brachytherapy can calculate radiation dose distribution in three dimensional (3D), it is rather primitive compared with teletherapy. For example, no inhomogeneity correction for the patient's tissues is taken into account in brachytherapy treatment plans. This may take a risk of significant under dosage to breast tumors, owing to the lack of the scattering medium, the attenuation by the contrast medium and the low density of lung. Similar dose errors should also occur in NPC treatments because the anatomical location of nasopharynx is surrounded by air cavity and bone structure. Therefore, a major concern is whether dosimetry surveys should be included in the system commissioning to indicate such intrinsic system limitations.

Therefore, to our knowledge, the present study is the first on the nasopharynx to investigate the effect of air cavity and bone structure on dose distribution. The Monte Carlo (MC) method was selected to perform this study because it is one of the most widely used methods in dosimetric studies of brachytherapy source. In the present study, the accuracy of the algorithm used by the treatment planning system (TPS), Flexiplan, designed by developed by Sonotech GmbH, Germany was tested.

The perplexity of assessing the final complex delivered dose in modern radiotherapy modalities has raised a substantial need and motivation toward finding a comprehensive dosimetry method, which should have the capability of accurate multi-dimensional dose verification and quality assurance (QA) of the whole treatment. [2] Up to now, gel dosimetry systems are the only true 3D dosimeters. [3] No other conventional dosimeter is capable of fulfilling the requirements of a comprehensive 3D measurement of a dose distribution with sharp gradients and irregular shape. [2]

The laborious manufacturing process of conventional polymer gels has been significantly simplified by the introduction of a new generation of polymer gels that can be fabricated under normal atmospheric conditions, [4] for which they are called normoxic.

Gel dosimetry has been examined as a clinical dosimeter since the 1950s. During the last two decades, however, the number of investigators in this field has increased rapidly, and the body of knowledge regarding gel dosimetry has expanded considerably. Gel dosimetry is still in it's a research phase, and the introduction of this tool into clinical use is proceeding slowly. [5]

In this study, a new type of gel dosimeter with acronym of MAGICA was used. This type of gel dosimeter was manufactured by adding agarose to the ingredient of MAGIC gel dosimeter. [4] MAGICA gel dosimeter was manufactured in Novin Radiation Medicine Institute (I.R. Iran) in 2004. [6]

 Materials and Methods

Head phantom design and treatment planning simulation

To accomplish the gel dosimetry, the hollow phantom was designed with Plexiglas cylinder that head bone was embedded within phantom. Therefore, the head phantom in this study was included a Plexiglas cylinder with 20 cm inner diameter and 12 cm height, head bone and brachytherapy silicon applicator for nasopharynx cancer.

In this study, computed tomography (CT) images were used in MC based dose calculation for brachytherapy treatment planning; with this study treatment planning for nasopharynx carcinoma was planned with the presence of the skull bone with more accurately than 3D TPS. For this propose, nasopharynx applicator has been placed inside the head phantom. Then, CT-scan was performed for the head phantom [Figure 1]. It should be noted that the CT scan procedure was performed for 48 slices with 0.2 cm thickness.{Figure 1}

To identify clinical target volume (CTV) and patient target volume (PTV) the whole of the nasopharynx is considered to CTV and PTV is determined by CTV with 0.5 cm margin. With this plan, Flexipaln software was determined 70 dwell position for each applicator (140 dwell position totally).

Preparation of polymer gel dosimeter

A MAGICA polymer gel dosimeter was prepared. All chemicals (gelatin, ascorbic acid [AA], CuSO 4 .5H 2 O, hydroquinone and methacrylic acid [MAA]) were provided by Sigma Aldridge and Flucka with experimental grade. High Performance Liquid Chromatography (HPLC) water was obtained from Novin Medical Radiation Institute in Tehran. The preparation of the gel was carried out in a similar procedure as described by Fong et al. with a slight difference due to the presence of agarose in MAGICA formulation. First, water was divided into 5 flasks of varying sizes, ready for dissolving each substance. Gelatin was added in to about 60% of the total HPLC de-ionized water. Two electrical heating plates provided with magnetic stirring and thermostatic controls were used to heat the solutions. Gelatin was allowed to swell for about an hour and then the solution was stirred and heated to about 50°C until a clear solution was obtained, ensuring all gelatin powder has been dissolved. When the temperature of gelatin solution reached near 40°C, agarose was added to about 30% of warm water which had been heated up to 50°C. Agarose solution was stirred and heated to about 90°C at which agarose was thoroughly dissolved. At this time gelatin solution should have reached near 50°C. Both solutions were allowed to cool. The gelatin solution was larger in volume compared to the agarose solution; thus, agarose solution cools faster in spite of its higher temperature. However, the cooling rates can be adjusted with respect to each other by proper adjustment of the heating plates. When both solutions cooled to an equal temperature about 47°C, agarose solution was added to the gelatin solution and stirring continued. Stirring never stopped before the end of fabrication. At 45°C, Hydroquinone, which had been solved in about 5% of water, was added to the mixture. The remaining 5% of water were divided into two portions and in each portion AA and Copper (II) sulphate were dissolved after being weighed. These two chemicals, which together play the role of oxygen scavenger, were added to the mixture when temperature declined to about 37°C. MAA was added at the same temperature. The amount of MAA for all gel fabrications was 9% of the total weight of gel except in one experiment in which more MAA was used; Composition of 1000 g MAGICA gel dosimeter is represented in [Table 1]. The gel was then decanted into test tubes and poured into the head phantom and left in a typical refrigerator at about 4°C to set [Figure 2]. Gel phantom and calibration tubes were not irradiated in the first 24 h after being manufactured. All irradiations were performed after that period. [4],[6] {Figure 2}{Table 1}

Irradiation of polymer gel dosimeter

To obtain calibration curves, 11 dosimeters were irradiated with doses in the range of 0-2000 cGy (0, 200, 400, 600, 800, 1000, 1200, 1400, 1600, 1800, 2000 cGy) for 6 MeV photon beam (24 h) after gel preparation. To ensure homogeneous irradiation of the dosimeter, the gel phantom was fixed in the center of a water-filled cylindrical shape container. Perpendicular field source-to-surface (SSD = 100 cm, field size = 10 × 40 cm 2 ) was used for polymer gel irradiation. All exposures were performed on the same date by the same repetition rate of 200 cGy/min.

The head phantom was irradiated with Flexitron afterloader brachytherapy machine. 192 Ir brachytherapy source had stopped in 140 dwell positions that were determined with Flexipaln software. The total exposure time was 14 min.

Magnetic resonance imaging (MRI) evaluation of polymer gel dosimeter and preparing an R2 map

Evaluations of dosimeters were performed on a Siemens Symphony Germany, 1.5T scanner in the body coil 1 day after irradiation. All samples of the polymer gel dosimeter were left inside the MRI room for a sufficiently long time (4 h) to become temperature equilibrated with the room temperature. A multiecho sequence with 16 equidistant echoes was used for the evaluation of irradiated polymer gel dosimeters. The parameters of the sequence were shown in [Table 2].{Table 2}

Container was centered in the body coil. To ensure that the obtained R2 values were not influenced by possible temperature gradients in the gel, phantoms were left in MRI room 24 h before scanning.

Since, the gel temperature during imaging increased up to 3°C, a little motion artifact would exist in MRI. [7] 24 h after irradiation, it was sure that the polymerization mechanism was completed. R2 (=1/T2) maps were computed using the modified radiotherapy gel dosimetry image processing software coded in MATLAB.

MCNP simulation

Brachytherapy dose distributions were simulated using the Monte Carlo N-particle transport code (MCNP) published by the Los Alamos National Laboratory. [8] The MCPLIB04 photon cross-section library was applied using data from Evaluated Nuclear Data File [ENDF/B] [9],[10] The results from the MCNP5 calculations contain numerous flexible tallies : s0 urface current and flux, volume flux (track length), point or ring detectors, particle heating, fission heating, pulse height tally for energy or charge deposition, mesh tallies, and radiography tallies. Particle fluence and cell-heating tallies (F4 and F6, respectively) were employed to calculate kerma and absorbed dose in this study. The 192 Ir photon spectrum used in these simulations was obtained from NuDat database. [11] For the calculations, the stainless steel characteristic X-ray production was suppressed with δ = 2 keV (where δ is the energy cutoff). [12] The spherical water phantom was modeled with a 30 cm diameter with an atomic ratio of 2:1 for H:O and ρ = 0.998 g/cm 3 . The line source placed in the center of the phantom for the calculation of all dosimetric parameters at radial distances of r = 0.5, 1, 2, 3, 4, 5, 7, 8, 9, and 10 cm, away from the source [Figure 3]. According to the TG43-U1 recommendation, radial dose function, g(r), and dose rate constant, L, are the most important and general dosimetric parameters for source validation and bench marking. To evaluate the simulation accuracy in this study, the Λ and g(r), 2D anisotropy function F(r,θ) values were calculated for 192 Ir HDR source (type Ir2.A85-2) from Sadeghi et al. study. [13] {Figure 3}

To obtain the dose-rate distribution produced by the 192 Ir source, its geometry was modeled as follow: The nasopharynx applicator was placed inside the head phantom and then CT scan was done for 50 slices with 0.2 cm diameter. These images have been used for MCNP5 simulation [Figure 3]. 84456 detector voxels (0.3 × 0.3 × 0.2 cm 3 ) were defined in this phantom for MC simulation. The simulation was performed up to 2.7 × 10 6 histories. With this number of histories, the average statistical uncertainty is lower than 4%. For this area, 140 dwell position was selected and this coordination was selected as a point source position in MCNP5 [Figure 4]. The 192 Ir photon spectrum used in these simulations was obtained from [Table 3]. [14] {Figure 4}{Table 3}

192 Ir has 26 photon beams, but in this study to reduce the variance and compatibilities of MC simulation, photon beams with a probability of higher than 3% is used (9 photon) [Table 3].


R2 map calculation

The spin-spin relaxation rate R2 = 1/T2 is related to the absorbed dose, which was delivered to a gel phantom. R2 map in each image was calculated with the selection of regions of interest with an equal number of pixels and the mean values of R2 and standard deviation of R2 in those regions were calculated. In this study, the number of pixels for each calibration tubes was 364 and the Regions of Interest (ROIS) were selected in the center of the calibration tubes. [Figure 5] illustrate the R2 = (1/T2) as a function of absorbed dose by imaging in water. Standard deviations in figures are also shown.{Figure 5}

MCNP5 bench marking

The MCNP simulation method in this work was benchmarked with the 192 Ir source. The comparison of MCNP-calculated value of Λ, with the previously published data for the source, demonstrates the accuracy of our simulation method. The calculated dose rate constant for the ideal orientation was found to be 1.108 ± 0.032 cGyU−1 /h compared to the value, 1.113 ± 0.033 cGyU−1 /h, measured by Sadeghi et al. [13]

The calculated line source radial dose function for the ideal orientation of the 192 Ir source in water in this work and those determined from calculated by Sadeghi et al. are presented in [Figure 6]. [Figure 6] shows a comparison of the radial dose function of 192 Ir line source and Sadeghi et al. work. Furthermore, anisotropy function F(r,θ) results are shown in [Table 4]. According to the figure, there was an excellent agreement between the calculated value for this study and the previous work.{Figure 6}{Table 4}


Comparison between iso-doses in 3D TPS, MCNP5 simulation and gel dosimetry

This section presents the results of the MCNP5 and Flexiplan simulations and MAGICA gel dosimetry performed in the head phantom with the nasopharynx applicator fully loaded with 192 Ir sources. [Figure 7] presents MCNP two dimensional (2D) iso-dose in coronal plan at distances of z = +0.3, −0.3 cm. In all distances the doses are normalized to the maximum dose.{Figure 7}

When treatment planning was performed, Flexiplan software creates a DICOM file, that in this file all the information about iso-doses in each cross section was available. This file was extracted and proceeds by MATLAB software. Typically, a 2D and 3D iso-dose in 2 coronal cross-sections were drawn. [Figure 8] shows the iso-doses in coronal plan at distances of z = 0.3, −0.3 cm.{Figure 8}

R2 map in head phantom images were calculated with MATLAB software. As before all doses are normalized to the maximum dose. [Figure 9] represents calculated R2 map in MATLAB software in coronal plan at distances of z = 0.3, −0.3 cm.{Figure 9}

Comparison between iso-doses in 3D TPS and MCNP5 simulation

As can be seen in [Figure 7] and [Figure 8] and [Table 5], there is good accordance between 3D TPS and MCNP5 simulation, differences in various distances are between 2.4% and 6.1%. In Flexiplan TPS, all the simulations were carried out in a water phantom while in MCNP5 simulation skull bone and air cavities were considered. As the Iridium maximum and average energy is 0.828 MeV and 0.370 MeV respectively, most of the gamma photons move from Compton to photoelectric area and in this area, photoelectric cross section is related to Z 4 . [14] Therefore, non-homogeneity is become more important in photoelectric. In minimum doses, deference between the skeleton and water Mass energy absorption coefficient became larger, and the differences in minimum doses is cause by this differences. Another reason for these differences is the variety of energy photon beam. As you know, 192 Ir source has more than 20 photons with different energy (0.061-1.378 MeV). However, TPS use just average energy for dose calculation.{Table 5}

Comparison between iso-doses in MAGICA gel dosimetry and mcnp5 simulation

As can be seen in [Figure 7] and [Figure 9] and [Table 6], there is predictable accordance between MAGICA gel dosimetry and MCNP5 simulation and differences in various distances are between 5.7% and 7.4%.{Table 6}

One of the reasons for these differences could be the simplifications in MCNP5 simulation. For variance reduction in MCNP5 simulation, some simplifications were performed. However, the most important reason for these differences could be practical errors in experimental dosimetry. Temperature dependence in normoxic dosimetric gel is better than the others. Additional, MAGICA gel has a better resistance to temperature increases, but, in HDR brachytherapy system, due to the high dose gradients close to source, increase the temperature is uncontrollable.

Comparison between isodoses in MAGICA Gel dosimetry and 3D TPS

As can be seen in [Figure 8] and [Figure 9] and [Table 7], there is acceptable accordance between MAGICA gel dosimetry and MCNP5 simulation and differences in various distances are between 5.2% and 9.4%.{Table 7}

Sources of differences in this comparison are divided to first calculations variation and next practical errors that was added in experimental dosimetry. This comparison was much closer to reality, because in real HDR brachytherapy treatments all of these errors can occur together.


Two major advantages of polymer gel dosimeters were their ability to determine integrated 3D dose distribution, as well as their ability to form in different shapes. [15] In fact, polymer gel dosimeters were monomers, which distributed in a gelling matrix. Ionizing radiations convert these monomers to polymers via distinguished mechanism. [16] The polymerization degree is dependent on the absorbed dose in gel dosimeter. After polymerization, magnetic properties of polymer surrounding protons are changed. [15] These changes could be exhibited by MRI. Artifacts have always been major problems in medical imaging, and gel dosimeter MRI are not immune from these problems. [7],[17]

A major problem in MRI read out method, especially when dealing with small vials (or other gel container) is the serious reduction of contrast in edges. Since in gel dosimetry spatial resolution has a great importance, one should be careful in using image processing filters. In this study, contrast enhancement of MAGICA gel dosimeters with MRI in water environment. Contrast and noise of R2 (=1/T2) in gel MRI images in water and optimum condition for MAGICA gel dosimeters MR imaging was obtained for the best contrast and resolution.

In order to compare 3D TPS and MCNP5 simulation iso-dose curves for nasopharynx HDR brachytherapy, non-homogeneous phantom with CT images was simulated in MCNP5. For bench marking, the dosimetric parameters including, dose rate constant, Λ and radial dose function, g(r) of the 192 Ir source have been calculated by using the MCNP5 MC code in the three geometric orientations. These calculations were performed following the American Association of Physicists in Medicine (AAPM) TG-43UI task group recommendations and compared with the measured data of Sadeghi et al. [13]

According to AAPM TG-43UI task group recommendations, the errors less than 10% for brachytherapy is acceptable. Therefore, in this study, the result of QA of nasopharynx HDR brachytherapy was consistent with international standards.

In general, TPS that is used in hospitals perform the good simulation with the best approximations, but in brachytherapy that patient absorption dose is high in just one or two treatment session(s), these errors may cause the serious problem in planning (receive the insufficient dose to cancerous tissue or receive excessive dose to normal tissues and sensitive organs). MC simulation performs better simulation, because non homogeneous and photon energy spectrum is simulated in MCNP5. However, the biggest disadvantage of this system is long running time to obtain high accuracy. With computers enhancement, use of better systems with accurate estimation will be possible.


1Lamberts SW. Intraoperative HDR Brachytherapy: Present and Future. Noorwegen: Isodose Control B.V Press; 2007.
2Oldham M, Siewerdsen JH, Kumar S, Wong J, Jaffray DA. Optical-CT gel-dosimetry I: Basic investigations. Med Phys 2003;30:623-34.
3Podgorsak EB. Radiation Oncology Physics: A Handbook for Teachers and Students. Austria: International Atomic Energy Agency (IAEA); 2005.
4Fong PM, Keil DC, Does MD, Gore JC. Polymer gels for magnetic resonance imaging of radiation dose distributions at normal room atmosphere. Phys Med Biol 2001;46:3105-13.
5Ibbott GS. Application of gel dosimetry. J Phys 2004;3:58-77.
6Zahmatkesh MH, Kousari R, Akhlaghpour S, Bagheri SA. MRI gel dosimetry with methacrylic acid, ascorbic acid, hydroquinon and copper in agarose (MAGICA) gel. In: Preliminary Proceedings of DOSGE. Third International Conference on Radiotherapy Gel Dosimetry; Sep 13-16; Ghent, Belgium, (Australian Journal of Basic and Applied Sciences) 2004.
7De Deene Y, De Wagter C. Artefacts in multi-echo T2 imaging for high-precision gel dosimetry: III. Effects of temperature drift during scanning. Phys Med Biol 2001;46:2697-711.
8Monte Carlo Team. MCNP-5 General Monte Carlo N-Particle Transport Code-Version 5. Los Alamos National Laboratory. Available from: [Last reviewed 2008].
9Cross Section Evaluation Working Group. ENDF/B-VI summary documentation (ENDF-201). Brookhaven National Laboratory Report no. BNL.NCS-17541. 8 th ed. New York .USA: National Nuclear Data Center; 2000.
10NuD at 2.0 National Nuclear Data Center nuclear data from NuDat, a web-based database maintained by the National Nuclear Data Center Brookhaven National Laboratory, Upton, NY, 2004. Available from: [Last accessed on 2013].
11Nath R, Anderson LL, Luxton G, Weaver KA, Williamson JF, Meigooni AS. Dosimetry of interstitial brachytherapy sources: Recommendations of the AAPM Radiation Therapy Committee Task Group No. 43. American Association of Physicists in Medicine. Med Phys 1995;22:209-34.
12Granero D, Pérez-Calatayud J, Ballester F. Monte Carlo study of the dose rate distributions for the Ir2.A85-2 and Ir2.A85-1 Ir-192 afterloading sources. Med Phys 2008;35:1280-7.
13Sadeghi M, Taghdiri F, Saidi P. Dosimetric characteristics of the ¹ 9²Ir high-dose-rate afterloading brachytherapy source. Jpn J Radiol 2011;29:324-9.
14Cember H. Introduction to Health Physics. Chicago, U.S.A: Medical Physics Press; 2001.
15De Deene Y. Essential characteristics of polymer gel dosimeters. J Phys 2004;3:34-57.
16De Deene Y, Hurley C, Venning A, Vergote K, Mather M, Healy BJ, et al. A basic study of some normoxic polymer gel dosimeters. Phys Med Biol 2002;47:3441-63.
17De Deene Y, De Wagter C, De Neve W, Achten E. Artefacts in multi-echo T2 imaging for high-precision gel dosimetry: I. Analysis and compensation of eddy currents. Phys Med Biol 2000;45:1807-23.