- Email: [email protected]
Surface and build-up region dose analysis for clinical radiotherapy photon beams
ARTICLE IN PRESS
Applied Radiation and Isotopes 66 (2008) 1438–1442 www.elsevier.com/locate/apradiso
Surface and build-up region dose analysis for clinical radiotherapy photon beams E. Ishmael Parsai, Diana Shvydka, David Pearson, M. Gopalakrishnan, John J. Feldmeier Department of Radiation Oncology, University of Toledo Health Science Campus, 3000 Arlington Avenue, Toledo, OH 43614, USA Received 8 June 2007; received in revised form 15 February 2008; accepted 20 February 2008
Abstract The standard clinical approach of dose measurement using a Farmer type fixed plane parallel cylindrical ionization chamber produces erroneous results in the build-up region. We studied dose distribution in this region using a Monte Carlo simulation technique and compared our results with data measured using extrapolation, parallel plate, and cylindrical farmer type ionization chambers for 6 and 10 MV photon beams from two different accelerators. The extrapolation chamber data agreed favorably with the Monte Carlo results, suggesting that dose at the skin surface and a few mm beneath is significantly lower than conventionally accepted values. r 2008 Elsevier Ltd. All rights reserved. keywords: Build-up region; Surface dose; Extrapolation chamber; Monte Carlo simulations
1. Introduction As the energy of the ionizing radiation increases, the penetrating power of the secondary charged particle as well as of primary particle increases, leading to a deeper position of the maximum dose point. Even though the dose maximum is achieved at a larger depth, the dose near the patient skin surface is not negligible and needs to be accounted for. The primary dose contribution of contaminant electrons to the build region originates mostly from the primary photon interaction with components of accelerator head, such as the secondary jaws and beam modifiers, as well as the air column between the treatment head and patient surface. The secondary contribution arises from the back scattering of the photons inside the patient body. In clinical settings, the percentage depth dose values are typically measured using computerized scanning systems in a water phantom. This method, however, is susceptible to erroneous results as the dose registered in the first few mm below water surface is prone to high degree of uncertainty. This could be attributed to the rapid change of dose gradient, the lack of charge particle equilibrium and also Corresponding author. Tel.: +1 419 383 5113; fax: +1 419 383 6313.
E-mail address: [email protected] (E. Ishmael Parsai). 0969-8043/$ - see front matter r 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.apradiso.2008.02.089
the nature and sensitivity of the ionization chamber being used. It was established that the instrument of choice for measurements made in this region of high-dose gradient is the extrapolation chamber (Nilsson and Montelius, 1986). As this particular chamber is cumbersome to use and is not widely available, a common method of measuring the percentage depth dose (PDD) in the region above the dose maximum of high-energy photon beams is to employ fixed plane parallel plate ionization chambers (Lawrence, 1973). A comparison of the extrapolation and fixed plane parallel plate chamber shows an increase of charge collection for parallel plate chamber, which is attributed to increased inscattering of the electrons from the sidewalls into the active volume of the chamber (Manson et al., 1975; Velkley et al., 1975). It was found that the parallel plate chamber overestimated the PDD values by 10–40% depending upon the energy of the photons. A method for correcting for the over-response of the parallel plate chamber was proposed by Velkley et al. (1975) and further modified by Gerbi and Khan (1987, 1990) to account for the effect of separation of the guard ring width with respect to the central electrode. The Monte Carlo (MC) simulation technique has been demonstrated to be an accurate method for dose calculations in radiotherapy. Rogers et al. (1985, 1995) have shown that OMEGA BEAM, a general purpose MC
ARTICLE IN PRESS E. Ishmael Parsai et al. / Applied Radiation and Isotopes 66 (2008) 1438–1442
modeling code, can be used to study the beam characteristics of a linear accelerator. The MC calculations have been bench-marked by various investigators (Ding et al., 1996; Van der Zee and Welleweerd, 1999; Libby et al., 1999a, b; Deng et al., 2000; Sheikh-Bagheri, 2000), validating the accuracy of the simulations with respect to the various variance reduction techniques and coding geometry. Even though MC has become a standard tool for the verification of radiation dosimetry measurements, there is still a disagreement between measurements and simulation results in the build-up region, where MC shows significantly lower dose (Rogers and Bielajew, 1985; Ding, 2002; Abdel-Rahman et al., 2005). The purpose of this study is to investigate the PDD variation in the build-up region for 6 and 10 MV photon beams from a Varian 1800 Clinac and an Elekta SL 25 linear accelerator using MC simulations and to compare the simulation results with the data acquired using an extrapolation chamber, parallel plate chamber and a conventional 0.125 cm3 cylindrical chamber. We also compare a subset of our results for a 6 MV beam with the data recently reported by Abdel-Rahman 2005 for Varian 2300 Clinac. 2. Materials and methods 2.1. Extrapolation chamber measurements An extrapolation chamber (Far West Technology) was used as a ‘‘golden’’ standard against which all the measured and simulated depth dose data were compared. The central electrode as well as the guard ring of the chamber was made of A-150 Tissue Equivalent Plastic; the chamber window was made of 6.9 mg/cm2 polyethylene. The saturation voltage of 50 V/mm was determined by varying the voltage across the chamber for a fixed dose and plotting measured charge vs. voltage. We employed this voltage as a parameter for data collection, ensuring that it remained constant with the change of plate separation and that small fluctuations in applied voltage would not affect the electrometer readings. The measurement of the collected charge was done in a solid water phantom for specific depths and field sizes at an SSD of 100 cm. For each depth the ionization charge was measured at the positive and negative bias voltage and the average taken. Readings were taken for each depth using 4, 3, 2 and 1 turn of the scribe, which varied plate separation from 0.65 to 4.65 mm. A linear regression equation was used to fit the measured PDD, which was extrapolated to the surface to determine the surface dose. The doses at successive depths were measured by adding thin layers of solid phantom material over the entrance foil, while maintaining 100 SSD at the top of the phantom.
1439
depths. The chamber consists of a cylindrical body made of Perspex, a window made of polyethylene coated with a graphite layer (2.3 mg/cm2), and a collecting electrode made of Perspex coated with a graphite layer. The chamber measuring volume has a diameter 6 mm and a height 2 mm, while the collecting electrode has an effective diameter of 5.4 mm. The separation between the electrodes is 2 mm. The chamber is provided with an add-on protective cap of 1 mm to be used for measurements in water. The surface dose measurement at d ¼ 0 was done without the protective cap by placing the chamber carefully at the water surface. The successive dose measurements inside the phantom were performed with the protective cap in place and positioning the chamber at respective depths. Interpretation of the data acquired with a parallel plate chamber requires special consideration. As previously shown by Velkley et al. (1975), parallel plate chambers over-respond at the phantom surface down to a depth of about 20% of dmax. The difference is greater for the lowenergy photons due to the contribution of electrons scattered from the sidewalls of the chamber and collected in the chamber active volume. As the beam energy increases the scattered electrons are more forward peaked and are less likely to add to the dose in the chamber active volume. Velkley et al. (1975) proposed an empirical formula for correction of PDD measured with parallel plate chamber for the region above dmax. It has been shown by Gerbi and Khan (1990) that the above-mentioned formalism either under- or over-compensates the surface dose measurements for chambers at different depths. The Velkley method was extended to include the effects of collector edge to sidewall distance to represent the chamber response in a more accurate way. The new corrected PDD equation was given by P0 ðdÞ ¼ PðdÞ xðE; 0Þl exp a d=d max (1) where xðE; 0Þ ¼ 27:19 32:59IR þ ð1:666 þ 1:982IRÞC Here P0 (d) is the corrected PDD, P(d) is the uncorrected PDD, E is the maximum energy of the photon spectrum, d and dmax are the depths, l is the plate separation in mm, x(E,0) is the over response in percent per mm (%/mm) of plate separation at the surface of the phantom for energy E, IR is the ionization ratio measured at 10 and 20 cm for 10 10 field size at a constant chamber distance of 100 cm, C is the collector edge to side wall distance in mm and a ¼ 5.5 is an empirically determined constant of proportionality. We used Eq. (1) to obtain the corrected readings for our Markus parallel plate chamber. 2.3. Cylindrical chamber measurements
2.2. Parallel plate chamber measurements A Markus chamber of type 23343 from PTW-Freiburg was used to measure the PDD for various field sizes and
For comparison with other techniques we collected PDD data in the build-up region with a setup typically used in the clinic environment for radiation dosimetry and quality
ARTICLE IN PRESS 1440
E. Ishmael Parsai et al. / Applied Radiation and Isotopes 66 (2008) 1438–1442
assurance measurements. A three-dimensional computercontrolled scanning water phantom system (Wellhofer, Scanditronix) equipped with waterproof 0.125 cm3 cylindrical ion chambers for both scanning and reference chamber was employed for the measurements. 2.4. Monte Carlo simulations 6 and 10 MV photon beams from a Varian Clinac 1800 and an Elekta SL 25 were simulated using the BEAM/EGS4 code system (Rogers et al., 2001). The incident electron energy was determined to be 6.4 MeV for a nominal energy of 6 MV and 10.85 MeV for the nominal 10 MV for the Varian machine. The MC beam parameters were adjusted until a close match was obtained for the computed data against the measured ones for PDD and beam profiles for various field sizes and different depths. The computed depth of maximum dose dmax for the Varian 6 and 10 MV photon beams were found to be 1.5 and 2.35 cm in water, respectively. For the Elekta, 6 and 10 MV beams were found to be produced from incident electron energies of 6.30 and 9.50 MeV, respectively. The depth of maximum dose was found, by MC in the case of the Elekta, to be at 1.6 and 2.35 cm for the 6 and 10 MV beams, respectively. All dmax values were determined from PDD data using a 10 10 field size. The values of electron cutoff energy (ECUT) and photon cutoff energy (PCUT) used in the simulation were 0.7 and 0.01 MeV, respectively. Various variance reduction techniques, such as bremsstrahlung splitting and photon forcing, were employed to shorten simulation times while achieving good statistics. Both 6 and 10 MV simulations were split into three parts. First, we obtained a phase-space file at a plane after the flattening filter and before the jaws. The phase space file obtained in this way was independent of field size and was then used as an input to the second, field-defining, segment of the accelerator. The second stage of simulations resulted in a set of output phase space files obtained at an SSD of 100 cm for the various field sizes. These were used as an input for the third part, simulating a 30 30 30 cm3 water phantom, with a central axis voxel size ranging from 1 1 0.05 to 1 1 0.1 cm3 along x, y and z directions, respectively. The phantom was modeled using the DOSXYZ code (Ma et al., 2001). The program STATDOSE was then used to analyze the output of DOSXYZ and obtain the PDD curves. The PDD curves obtained here were then compared to those obtained with the various ionization chamber measurements. For all simulations, sufficient histories were run to ensure that the error for each voxel size was less than 1%. 3. Results and discussion Our measurements indicated that the uncorrected parallel plate ionization chamber data over-estimated the dose at the surface by 30–40% for 6 MV and by 20–30% for 10 MV photons for a 10 10 field. As mentioned
before, we used Eq. (1) to correct the parallel plate chamber readings. All presented PDD data were obtained by normalization of the reading at a given depth to the values at dmax. Tables 1 and 2 show summarized comparisons between the readings of the percent depth dose for the extrapolation chamber and the parallel plate chamber (for both uncorrected and corrected) at the surface of the water phantom. The % error column values indicate the percent difference between the extrapolation and corrected parallel plate ionization chamber measurements for each field size. These measurements show the following trends: (a) As the energy of the photon beam increases, the difference between the extrapolation chamber and the uncorrected parallel plate readings decreases. (b) The surface dose increases with field size. (c) The surface dose decreases with an increase of energy due to the generation of forward peaked high-energy electrons. (d) With the appropriate correction factors proposed by Gerbi and Khan (1990), a fixed parallel plate chamber could be employed to measure dose accurately in the region above the depth of maximum dose. MC simulations were run for field sizes ranging from 5 5 to 25 25 for both 6 and 10 MV photon beams to compare with the measured data. The calculated PDDs were normalized to dmax. Figs. 1 and 2 show the PDD along the central axis plotted for extrapolation, fixed parallel plate, and 0.125 cm3 chambers compared to MC Table 1 Summary of the differences in corrected and uncorrected PDD values measured using the parallel plate chamber compared with the extrapolation chamber for 6 MV photons at the surface of the phantom Field size
Extrapolation readings
Uncorrected parallel plate readings
Corrected parallel plate readings
Error %
55 10 10 15 15 25 25
10.53 16.04 21.47 31.45
20.77 26.64 32.81 43.67
10.28 16.14 21.9 32.17
2.4 0.6 2.0 2.3
Table 2 Summary of the differences in corrected and uncorrected PDD values measured using the parallel plate chamber compared with the extrapolation chamber for 10 MV photons at the surface of the phantom Field size
Extrapolation readings
Uncorrected parallel plate readings
Corrected parallel plate readings
Error %
55 10 10 15 15 25 25
6.99 12.72 18.71 29.32
13.18 19.41 26.10 37.81
6.77 12.88 19.25 30.03
3.2 1.3 2.9 2.4
ARTICLE IN PRESS E. Ishmael Parsai et al. / Applied Radiation and Isotopes 66 (2008) 1438–1442
1441
Table 3 Calculated and measured doses at depth of 0.5 mm in the water phantom for 6 MV photons Field size
Extrapolation readings
Monte Carlo simulation
Error %
55 10 10 15 15 25 25
28.32 33.34 38.02 46.64
27.5 34.60 36.95 48.20
2.8 3.6 2.6 3.2
Table 4 Calculated and measured doses at depth of 0.5 mm in the water phantom for 10 MV photons
Fig. 1. Percentage depth dose in the build-up region measured by extrapolation, parallel plate and ionization chambers in water phantom, compared with MC simulation results for 6 MV photon beam, 10 cm 10 cm field size at 100 cm SSD.
Field size
Extrapolation readings
Monte Carlo simulation
Error %
55 10 10 15 15 25 25
17.91 24.39 29.7 40.21
17.4 23.8 29.09 42
2.8 2.4 2.1 4.0
values and in close agreement with MC simulations. We can compare a small subset of our results to those available from Abdel-Rahman et al. (2005) for a Varian 2300 Clinac beam, particularly, the PDD values for 6 MV beam, for a 10 10 field size, very near the surface and at a depth of 1 mm. We obtained somewhat lower figures for both the measured and simulated PDD at 1 mm depth. At the same time, PDD values very near the surface seem to match closer, although our extrapolation chamber result of 16.04 is higher than the previously reported value of 12.8, as well as closer to the result of the MC simulation. In addition, cylindrical chamber measurements of Abdel-Rahman et al. (2005) show the same over-estimation as in the present work. 4. Conclusions Fig. 2. Percentage depth dose in the build-up region measured by extrapolation, parallel plate and ionization chambers in water phantom, compared with MC simulation results for 10 MV photon beam, 10 cm 10 cm field size at 100 cm SSD.
simulation results. The data shown here are for a 10 10 field size at 100 SSD for both 6 and 10 MV photons. Tables 3 and 4 summarize the differences between calculated and measured doses at a depth of 0.5 mm in the water phantom for 6 and 10 MV photons, respectively. The error % column is calculated to be the percent difference between the extrapolation chamber reading and MC simulation result for each field size. The maximum deviation is within 3–4% at 0.5 mm depth in the phantom, decreasing to below 3% further from the surface. To the best of our knowledge, we have presented the first study showing the surface dose for different field sizes being significantly lower than the conventionally accepted
Good agreement of the data points was observed for the PDD measured using extrapolation chamber, corrected fixed parallel plate chamber readings and MC simulations. As the depth of maximum dose is reached, the difference in response of both chambers becomes very small due to establishment of charge particle equilibrium. On the contrary, the data acquired with cylindrical chamber over-responds by more than 50% for depth d ¼ 0 for both 6 and 10 MV photons for a 10 10 field at 100 SSD. This over-response is probably due to the fact that the chamber collects more low-energy electrons while it is positioned half inside and half outside the water surface during scanning. Studying the response of the chamber to lowenergy electrons can provide further insight to this hypothesis. The dose at 0.5 mm into the phantom showed a maximum deviation of 3.6% and 4% for both 6 and 10 MV photons, respectively. At the surface, however, the
ARTICLE IN PRESS 1442
E. Ishmael Parsai et al. / Applied Radiation and Isotopes 66 (2008) 1438–1442
measured dose using the extrapolation chamber and the corrected fixed parallel plate chamber showed maximum deviation of 2.4% for 6 MV, and 3.2% for the 10 MV photons. It can also be noted that as a depth of 1.25 cm is reached in the phantom, there is a close correlation of the measured charge collected by all the chambers. Measured PDD from a fixed parallel plate chamber or extrapolation chamber should be used for the data in the build-up region as all other available chamber types tend to over-estimate the values by a very high percentage (450%). We believe that the results of this study could be used as PDD data in the build-up region in a treatment planning system (TPS) to ensure accurate calculations of dose delivered to subcutaneous skin and superficial tissue regions. In the future we plan to extend the measurements to beam modifiers, oblique beams and electron beams. References Abdel-Rahman, W., Seuntjens, JP., Verhaegen, F., 2005. Validation of Monte Carlo calculated doses for megavoltage photon beams. Med. Phys. 32, 286–298. Deng, Jun., Jiang, S.B., Kapur, A., Pawlicki, T., Ma, C.M., 2000. Photon beam characterization and modelling for Monte Carlo treatment planning. Phys. Med. Biol. 45, 411–427. Ding, G.X., 2002. Dose discrepancies between Monte Carlo calculations and measurements in the build-up region for high-energy photon beam. Med. Phys. 29, 2459–2463. Ding, G.X., Rogers, D.W.O., Mackie, T.R., 1996. Mean energy, energy–range relationships and depth-scaling factors for clinical electron beams. Med. Phys. 23, 361–376.
Gerbi, B.J., Khan, F.M., 1987. The polarity effect of commercially available plane–parallel ionization chambers. Med. Phys. 14, 210–215. Gerbi, B.J., Khan, F.M., 1990. Measurement of dose in the buildup region using fixed-separation plane–parallel ionization chambers. Med. Phys. 17, 210–215. Lawrence, G., 1973. Relative surface doses from supervoltage radiation. Radiology 109, 437–442. Libby, B., Siebers, J., Mohan, R., 1999a. Validation of Monte Carlo generated phase-space descriptions of medical linear accelerators. Med. Phys. 26, 1476–1483. Libby, B., Siebers, J., Mohan, R., 1999b. Validation of Monte Carlo generated phase-space descriptions of medical linear accelerators. Med. Phys. 26, 1476–1483. Ma, C.-M., Rogers, D.W.O., Walters, B., 2001. DOSXYZnrc Users Manual. National Research Council Report PIRS-0509(B)revF, Ottawa. Manson, D.J., Velkley, D.E., Purdy, J.A., Oliver, G.D., 1975. Measurements of surface dose using buildup curves obtained with an extrapolation chamber. Radiology 115, 473–474. Nilsson, B., Montelius, A., 1986. Fluence perturbation in photon beams under non-equilibrium conditions. Med. Phys. 13, 191–195. Rogers, D.W.O., Bielajew, A.F., 1985. Calculated buildup curves for photons with energies up to 60Co. Med. Phys. 12, 738–744. Rogers, D.W.O., Faddegon, B.A., Ding, G.X., Ma, C.-M., We, J., Mackie, T.R., 1995. BEAM: A Monte Carlo code to simulate radiotherapy treatment units. Med. Phys. 22, 503–524. Rogers, D.W.O., Ma, C.-M., Walters, B., Ding, G.X., Sheikh-Bagheri, D., Zhang, G., 2001. BEAMnrc Users Manual. National Research Council Report PIRS-0509(A)revF, Ottawa. Sheikh-Bagheri, D., 2000. Comparison of measured and Monte Carlo calculated dose distributions from the NRC linac. Med. Phys. 27, 2256–2266. Van der Zee, W., Welleweerd, J., 1999. Calculating photon beam characteristics with Monte Carlo techniques. Med. Phys. 26, 1883–1892. Velkley, D.E., Mansion, D.J., Purdy, J.A., Oliver, G.D., 1975. Buildup region of megavoltage photon radiation sources. Med. Phys. 2, 14–19.
Applied Radiation and Isotopes 66 (2008) 1438–1442 www.elsevier.com/locate/apradiso
Surface and build-up region dose analysis for clinical radiotherapy photon beams E. Ishmael Parsai, Diana Shvydka, David Pearson, M. Gopalakrishnan, John J. Feldmeier Department of Radiation Oncology, University of Toledo Health Science Campus, 3000 Arlington Avenue, Toledo, OH 43614, USA Received 8 June 2007; received in revised form 15 February 2008; accepted 20 February 2008
Abstract The standard clinical approach of dose measurement using a Farmer type fixed plane parallel cylindrical ionization chamber produces erroneous results in the build-up region. We studied dose distribution in this region using a Monte Carlo simulation technique and compared our results with data measured using extrapolation, parallel plate, and cylindrical farmer type ionization chambers for 6 and 10 MV photon beams from two different accelerators. The extrapolation chamber data agreed favorably with the Monte Carlo results, suggesting that dose at the skin surface and a few mm beneath is significantly lower than conventionally accepted values. r 2008 Elsevier Ltd. All rights reserved. keywords: Build-up region; Surface dose; Extrapolation chamber; Monte Carlo simulations
1. Introduction As the energy of the ionizing radiation increases, the penetrating power of the secondary charged particle as well as of primary particle increases, leading to a deeper position of the maximum dose point. Even though the dose maximum is achieved at a larger depth, the dose near the patient skin surface is not negligible and needs to be accounted for. The primary dose contribution of contaminant electrons to the build region originates mostly from the primary photon interaction with components of accelerator head, such as the secondary jaws and beam modifiers, as well as the air column between the treatment head and patient surface. The secondary contribution arises from the back scattering of the photons inside the patient body. In clinical settings, the percentage depth dose values are typically measured using computerized scanning systems in a water phantom. This method, however, is susceptible to erroneous results as the dose registered in the first few mm below water surface is prone to high degree of uncertainty. This could be attributed to the rapid change of dose gradient, the lack of charge particle equilibrium and also Corresponding author. Tel.: +1 419 383 5113; fax: +1 419 383 6313.
E-mail address: [email protected] (E. Ishmael Parsai). 0969-8043/$ - see front matter r 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.apradiso.2008.02.089
the nature and sensitivity of the ionization chamber being used. It was established that the instrument of choice for measurements made in this region of high-dose gradient is the extrapolation chamber (Nilsson and Montelius, 1986). As this particular chamber is cumbersome to use and is not widely available, a common method of measuring the percentage depth dose (PDD) in the region above the dose maximum of high-energy photon beams is to employ fixed plane parallel plate ionization chambers (Lawrence, 1973). A comparison of the extrapolation and fixed plane parallel plate chamber shows an increase of charge collection for parallel plate chamber, which is attributed to increased inscattering of the electrons from the sidewalls into the active volume of the chamber (Manson et al., 1975; Velkley et al., 1975). It was found that the parallel plate chamber overestimated the PDD values by 10–40% depending upon the energy of the photons. A method for correcting for the over-response of the parallel plate chamber was proposed by Velkley et al. (1975) and further modified by Gerbi and Khan (1987, 1990) to account for the effect of separation of the guard ring width with respect to the central electrode. The Monte Carlo (MC) simulation technique has been demonstrated to be an accurate method for dose calculations in radiotherapy. Rogers et al. (1985, 1995) have shown that OMEGA BEAM, a general purpose MC
ARTICLE IN PRESS E. Ishmael Parsai et al. / Applied Radiation and Isotopes 66 (2008) 1438–1442
modeling code, can be used to study the beam characteristics of a linear accelerator. The MC calculations have been bench-marked by various investigators (Ding et al., 1996; Van der Zee and Welleweerd, 1999; Libby et al., 1999a, b; Deng et al., 2000; Sheikh-Bagheri, 2000), validating the accuracy of the simulations with respect to the various variance reduction techniques and coding geometry. Even though MC has become a standard tool for the verification of radiation dosimetry measurements, there is still a disagreement between measurements and simulation results in the build-up region, where MC shows significantly lower dose (Rogers and Bielajew, 1985; Ding, 2002; Abdel-Rahman et al., 2005). The purpose of this study is to investigate the PDD variation in the build-up region for 6 and 10 MV photon beams from a Varian 1800 Clinac and an Elekta SL 25 linear accelerator using MC simulations and to compare the simulation results with the data acquired using an extrapolation chamber, parallel plate chamber and a conventional 0.125 cm3 cylindrical chamber. We also compare a subset of our results for a 6 MV beam with the data recently reported by Abdel-Rahman 2005 for Varian 2300 Clinac. 2. Materials and methods 2.1. Extrapolation chamber measurements An extrapolation chamber (Far West Technology) was used as a ‘‘golden’’ standard against which all the measured and simulated depth dose data were compared. The central electrode as well as the guard ring of the chamber was made of A-150 Tissue Equivalent Plastic; the chamber window was made of 6.9 mg/cm2 polyethylene. The saturation voltage of 50 V/mm was determined by varying the voltage across the chamber for a fixed dose and plotting measured charge vs. voltage. We employed this voltage as a parameter for data collection, ensuring that it remained constant with the change of plate separation and that small fluctuations in applied voltage would not affect the electrometer readings. The measurement of the collected charge was done in a solid water phantom for specific depths and field sizes at an SSD of 100 cm. For each depth the ionization charge was measured at the positive and negative bias voltage and the average taken. Readings were taken for each depth using 4, 3, 2 and 1 turn of the scribe, which varied plate separation from 0.65 to 4.65 mm. A linear regression equation was used to fit the measured PDD, which was extrapolated to the surface to determine the surface dose. The doses at successive depths were measured by adding thin layers of solid phantom material over the entrance foil, while maintaining 100 SSD at the top of the phantom.
1439
depths. The chamber consists of a cylindrical body made of Perspex, a window made of polyethylene coated with a graphite layer (2.3 mg/cm2), and a collecting electrode made of Perspex coated with a graphite layer. The chamber measuring volume has a diameter 6 mm and a height 2 mm, while the collecting electrode has an effective diameter of 5.4 mm. The separation between the electrodes is 2 mm. The chamber is provided with an add-on protective cap of 1 mm to be used for measurements in water. The surface dose measurement at d ¼ 0 was done without the protective cap by placing the chamber carefully at the water surface. The successive dose measurements inside the phantom were performed with the protective cap in place and positioning the chamber at respective depths. Interpretation of the data acquired with a parallel plate chamber requires special consideration. As previously shown by Velkley et al. (1975), parallel plate chambers over-respond at the phantom surface down to a depth of about 20% of dmax. The difference is greater for the lowenergy photons due to the contribution of electrons scattered from the sidewalls of the chamber and collected in the chamber active volume. As the beam energy increases the scattered electrons are more forward peaked and are less likely to add to the dose in the chamber active volume. Velkley et al. (1975) proposed an empirical formula for correction of PDD measured with parallel plate chamber for the region above dmax. It has been shown by Gerbi and Khan (1990) that the above-mentioned formalism either under- or over-compensates the surface dose measurements for chambers at different depths. The Velkley method was extended to include the effects of collector edge to sidewall distance to represent the chamber response in a more accurate way. The new corrected PDD equation was given by P0 ðdÞ ¼ PðdÞ xðE; 0Þl exp a d=d max (1) where xðE; 0Þ ¼ 27:19 32:59IR þ ð1:666 þ 1:982IRÞC Here P0 (d) is the corrected PDD, P(d) is the uncorrected PDD, E is the maximum energy of the photon spectrum, d and dmax are the depths, l is the plate separation in mm, x(E,0) is the over response in percent per mm (%/mm) of plate separation at the surface of the phantom for energy E, IR is the ionization ratio measured at 10 and 20 cm for 10 10 field size at a constant chamber distance of 100 cm, C is the collector edge to side wall distance in mm and a ¼ 5.5 is an empirically determined constant of proportionality. We used Eq. (1) to obtain the corrected readings for our Markus parallel plate chamber. 2.3. Cylindrical chamber measurements
2.2. Parallel plate chamber measurements A Markus chamber of type 23343 from PTW-Freiburg was used to measure the PDD for various field sizes and
For comparison with other techniques we collected PDD data in the build-up region with a setup typically used in the clinic environment for radiation dosimetry and quality
ARTICLE IN PRESS 1440
E. Ishmael Parsai et al. / Applied Radiation and Isotopes 66 (2008) 1438–1442
assurance measurements. A three-dimensional computercontrolled scanning water phantom system (Wellhofer, Scanditronix) equipped with waterproof 0.125 cm3 cylindrical ion chambers for both scanning and reference chamber was employed for the measurements. 2.4. Monte Carlo simulations 6 and 10 MV photon beams from a Varian Clinac 1800 and an Elekta SL 25 were simulated using the BEAM/EGS4 code system (Rogers et al., 2001). The incident electron energy was determined to be 6.4 MeV for a nominal energy of 6 MV and 10.85 MeV for the nominal 10 MV for the Varian machine. The MC beam parameters were adjusted until a close match was obtained for the computed data against the measured ones for PDD and beam profiles for various field sizes and different depths. The computed depth of maximum dose dmax for the Varian 6 and 10 MV photon beams were found to be 1.5 and 2.35 cm in water, respectively. For the Elekta, 6 and 10 MV beams were found to be produced from incident electron energies of 6.30 and 9.50 MeV, respectively. The depth of maximum dose was found, by MC in the case of the Elekta, to be at 1.6 and 2.35 cm for the 6 and 10 MV beams, respectively. All dmax values were determined from PDD data using a 10 10 field size. The values of electron cutoff energy (ECUT) and photon cutoff energy (PCUT) used in the simulation were 0.7 and 0.01 MeV, respectively. Various variance reduction techniques, such as bremsstrahlung splitting and photon forcing, were employed to shorten simulation times while achieving good statistics. Both 6 and 10 MV simulations were split into three parts. First, we obtained a phase-space file at a plane after the flattening filter and before the jaws. The phase space file obtained in this way was independent of field size and was then used as an input to the second, field-defining, segment of the accelerator. The second stage of simulations resulted in a set of output phase space files obtained at an SSD of 100 cm for the various field sizes. These were used as an input for the third part, simulating a 30 30 30 cm3 water phantom, with a central axis voxel size ranging from 1 1 0.05 to 1 1 0.1 cm3 along x, y and z directions, respectively. The phantom was modeled using the DOSXYZ code (Ma et al., 2001). The program STATDOSE was then used to analyze the output of DOSXYZ and obtain the PDD curves. The PDD curves obtained here were then compared to those obtained with the various ionization chamber measurements. For all simulations, sufficient histories were run to ensure that the error for each voxel size was less than 1%. 3. Results and discussion Our measurements indicated that the uncorrected parallel plate ionization chamber data over-estimated the dose at the surface by 30–40% for 6 MV and by 20–30% for 10 MV photons for a 10 10 field. As mentioned
before, we used Eq. (1) to correct the parallel plate chamber readings. All presented PDD data were obtained by normalization of the reading at a given depth to the values at dmax. Tables 1 and 2 show summarized comparisons between the readings of the percent depth dose for the extrapolation chamber and the parallel plate chamber (for both uncorrected and corrected) at the surface of the water phantom. The % error column values indicate the percent difference between the extrapolation and corrected parallel plate ionization chamber measurements for each field size. These measurements show the following trends: (a) As the energy of the photon beam increases, the difference between the extrapolation chamber and the uncorrected parallel plate readings decreases. (b) The surface dose increases with field size. (c) The surface dose decreases with an increase of energy due to the generation of forward peaked high-energy electrons. (d) With the appropriate correction factors proposed by Gerbi and Khan (1990), a fixed parallel plate chamber could be employed to measure dose accurately in the region above the depth of maximum dose. MC simulations were run for field sizes ranging from 5 5 to 25 25 for both 6 and 10 MV photon beams to compare with the measured data. The calculated PDDs were normalized to dmax. Figs. 1 and 2 show the PDD along the central axis plotted for extrapolation, fixed parallel plate, and 0.125 cm3 chambers compared to MC Table 1 Summary of the differences in corrected and uncorrected PDD values measured using the parallel plate chamber compared with the extrapolation chamber for 6 MV photons at the surface of the phantom Field size
Extrapolation readings
Uncorrected parallel plate readings
Corrected parallel plate readings
Error %
55 10 10 15 15 25 25
10.53 16.04 21.47 31.45
20.77 26.64 32.81 43.67
10.28 16.14 21.9 32.17
2.4 0.6 2.0 2.3
Table 2 Summary of the differences in corrected and uncorrected PDD values measured using the parallel plate chamber compared with the extrapolation chamber for 10 MV photons at the surface of the phantom Field size
Extrapolation readings
Uncorrected parallel plate readings
Corrected parallel plate readings
Error %
55 10 10 15 15 25 25
6.99 12.72 18.71 29.32
13.18 19.41 26.10 37.81
6.77 12.88 19.25 30.03
3.2 1.3 2.9 2.4
ARTICLE IN PRESS E. Ishmael Parsai et al. / Applied Radiation and Isotopes 66 (2008) 1438–1442
1441
Table 3 Calculated and measured doses at depth of 0.5 mm in the water phantom for 6 MV photons Field size
Extrapolation readings
Monte Carlo simulation
Error %
55 10 10 15 15 25 25
28.32 33.34 38.02 46.64
27.5 34.60 36.95 48.20
2.8 3.6 2.6 3.2
Table 4 Calculated and measured doses at depth of 0.5 mm in the water phantom for 10 MV photons
Fig. 1. Percentage depth dose in the build-up region measured by extrapolation, parallel plate and ionization chambers in water phantom, compared with MC simulation results for 6 MV photon beam, 10 cm 10 cm field size at 100 cm SSD.
Field size
Extrapolation readings
Monte Carlo simulation
Error %
55 10 10 15 15 25 25
17.91 24.39 29.7 40.21
17.4 23.8 29.09 42
2.8 2.4 2.1 4.0
values and in close agreement with MC simulations. We can compare a small subset of our results to those available from Abdel-Rahman et al. (2005) for a Varian 2300 Clinac beam, particularly, the PDD values for 6 MV beam, for a 10 10 field size, very near the surface and at a depth of 1 mm. We obtained somewhat lower figures for both the measured and simulated PDD at 1 mm depth. At the same time, PDD values very near the surface seem to match closer, although our extrapolation chamber result of 16.04 is higher than the previously reported value of 12.8, as well as closer to the result of the MC simulation. In addition, cylindrical chamber measurements of Abdel-Rahman et al. (2005) show the same over-estimation as in the present work. 4. Conclusions Fig. 2. Percentage depth dose in the build-up region measured by extrapolation, parallel plate and ionization chambers in water phantom, compared with MC simulation results for 10 MV photon beam, 10 cm 10 cm field size at 100 cm SSD.
simulation results. The data shown here are for a 10 10 field size at 100 SSD for both 6 and 10 MV photons. Tables 3 and 4 summarize the differences between calculated and measured doses at a depth of 0.5 mm in the water phantom for 6 and 10 MV photons, respectively. The error % column is calculated to be the percent difference between the extrapolation chamber reading and MC simulation result for each field size. The maximum deviation is within 3–4% at 0.5 mm depth in the phantom, decreasing to below 3% further from the surface. To the best of our knowledge, we have presented the first study showing the surface dose for different field sizes being significantly lower than the conventionally accepted
Good agreement of the data points was observed for the PDD measured using extrapolation chamber, corrected fixed parallel plate chamber readings and MC simulations. As the depth of maximum dose is reached, the difference in response of both chambers becomes very small due to establishment of charge particle equilibrium. On the contrary, the data acquired with cylindrical chamber over-responds by more than 50% for depth d ¼ 0 for both 6 and 10 MV photons for a 10 10 field at 100 SSD. This over-response is probably due to the fact that the chamber collects more low-energy electrons while it is positioned half inside and half outside the water surface during scanning. Studying the response of the chamber to lowenergy electrons can provide further insight to this hypothesis. The dose at 0.5 mm into the phantom showed a maximum deviation of 3.6% and 4% for both 6 and 10 MV photons, respectively. At the surface, however, the
ARTICLE IN PRESS 1442
E. Ishmael Parsai et al. / Applied Radiation and Isotopes 66 (2008) 1438–1442
measured dose using the extrapolation chamber and the corrected fixed parallel plate chamber showed maximum deviation of 2.4% for 6 MV, and 3.2% for the 10 MV photons. It can also be noted that as a depth of 1.25 cm is reached in the phantom, there is a close correlation of the measured charge collected by all the chambers. Measured PDD from a fixed parallel plate chamber or extrapolation chamber should be used for the data in the build-up region as all other available chamber types tend to over-estimate the values by a very high percentage (450%). We believe that the results of this study could be used as PDD data in the build-up region in a treatment planning system (TPS) to ensure accurate calculations of dose delivered to subcutaneous skin and superficial tissue regions. In the future we plan to extend the measurements to beam modifiers, oblique beams and electron beams. References Abdel-Rahman, W., Seuntjens, JP., Verhaegen, F., 2005. Validation of Monte Carlo calculated doses for megavoltage photon beams. Med. Phys. 32, 286–298. Deng, Jun., Jiang, S.B., Kapur, A., Pawlicki, T., Ma, C.M., 2000. Photon beam characterization and modelling for Monte Carlo treatment planning. Phys. Med. Biol. 45, 411–427. Ding, G.X., 2002. Dose discrepancies between Monte Carlo calculations and measurements in the build-up region for high-energy photon beam. Med. Phys. 29, 2459–2463. Ding, G.X., Rogers, D.W.O., Mackie, T.R., 1996. Mean energy, energy–range relationships and depth-scaling factors for clinical electron beams. Med. Phys. 23, 361–376.
Gerbi, B.J., Khan, F.M., 1987. The polarity effect of commercially available plane–parallel ionization chambers. Med. Phys. 14, 210–215. Gerbi, B.J., Khan, F.M., 1990. Measurement of dose in the buildup region using fixed-separation plane–parallel ionization chambers. Med. Phys. 17, 210–215. Lawrence, G., 1973. Relative surface doses from supervoltage radiation. Radiology 109, 437–442. Libby, B., Siebers, J., Mohan, R., 1999a. Validation of Monte Carlo generated phase-space descriptions of medical linear accelerators. Med. Phys. 26, 1476–1483. Libby, B., Siebers, J., Mohan, R., 1999b. Validation of Monte Carlo generated phase-space descriptions of medical linear accelerators. Med. Phys. 26, 1476–1483. Ma, C.-M., Rogers, D.W.O., Walters, B., 2001. DOSXYZnrc Users Manual. National Research Council Report PIRS-0509(B)revF, Ottawa. Manson, D.J., Velkley, D.E., Purdy, J.A., Oliver, G.D., 1975. Measurements of surface dose using buildup curves obtained with an extrapolation chamber. Radiology 115, 473–474. Nilsson, B., Montelius, A., 1986. Fluence perturbation in photon beams under non-equilibrium conditions. Med. Phys. 13, 191–195. Rogers, D.W.O., Bielajew, A.F., 1985. Calculated buildup curves for photons with energies up to 60Co. Med. Phys. 12, 738–744. Rogers, D.W.O., Faddegon, B.A., Ding, G.X., Ma, C.-M., We, J., Mackie, T.R., 1995. BEAM: A Monte Carlo code to simulate radiotherapy treatment units. Med. Phys. 22, 503–524. Rogers, D.W.O., Ma, C.-M., Walters, B., Ding, G.X., Sheikh-Bagheri, D., Zhang, G., 2001. BEAMnrc Users Manual. National Research Council Report PIRS-0509(A)revF, Ottawa. Sheikh-Bagheri, D., 2000. Comparison of measured and Monte Carlo calculated dose distributions from the NRC linac. Med. Phys. 27, 2256–2266. Van der Zee, W., Welleweerd, J., 1999. Calculating photon beam characteristics with Monte Carlo techniques. Med. Phys. 26, 1883–1892. Velkley, D.E., Mansion, D.J., Purdy, J.A., Oliver, G.D., 1975. Buildup region of megavoltage photon radiation sources. Med. Phys. 2, 14–19.