https://orcid.org/0000-0002-1885-4747
Figures
Abstract
Introduction
This study presents an empirical method to model the high-energy photon beam percent depth dose (PDD) curve by using the home-generated buildup function and tail function (buildup-tail function) in radiation therapy. The modeling parameters n and μ of buildup-tail function can be used to characterize the Collimator Scatter Factor (Sc) either in a square field or in the different individual upper jaw and lower jaw setting separately for individual monitor unit check.
Methods and materials
The PDD curves for four high-energy photon beams were modeled by the buildup and tail function in this study. The buildup function was a quadratic function in the form of with the main parameter of d (depth in water) and n, while the tail function was in the form of e−μd and was composed by an exponential function with the main parameter of d and μ. The PDD was the product of buildup and tail function, PDD =
. The PDD of four-photon energies was characterized by the buildup-tail function by adjusting the parameters n and μ. The Sc of 6 MV and 10 MV can then be expressed simply by the modeling parameters n and μ.
Results
The main parameters n increases in buildup-tail function when photon energy increased. The physical meaning of the parameter n expresses the beam hardening of photon energy in PDD. The fitting results of parameters n in the buildup function are 0.17, 0.208, 0.495, 1.2 of four-photon energies, 4 MV, 6 MV, 10 MV, 18 MV, respectively. The parameter μ can be treated as attenuation coefficient in tail function and decreases when photon energy increased. The fitting results of parameters μ in the tail function are 0.065, 0.0515, 0.0458, 0.0422 of four-photon energies, 4 MV, 6 MV, 10 MV, 18 MV, respectively. The values of n and μ obtained from the fitted buildup-tail function were applied into an analytical formula of Sc = nE(S)0.63μE to get the collimator to scatter factor Sc for 6 and 10 MV photon beam, while nE, μE, S denotes n, μ at photon energy E of field size S, respectively. The calculated Sc were compared with the measured data and showed agreement at different field sizes to within ±1.5%.
Conclusions
We proposed a model incorporating a two-parameter formula which can improve the fitting accuracy to be better than 1.5% maximum error for describing the PDD in different photon energies used in clinical setting. This model can be used to parameterize the Sc factors for some clinical requirements. The modeling parameters n and μ can be used to predict the Sc in either square field or individual jaws opening asymmetrically for treatment monitor unit double-check in dose calculation. The technique developed in this study can also be used for systematic or random errors in the QA program, thus improves the clinical dose computation accuracy for patient treatment.
Citation: Li X-J, Ye Y-C, Zhang Y-S, Wu J-M (2022) Empirical modeling of the percent depth dose for megavoltage photon beams. PLoS ONE 17(1): e0261042. https://doi.org/10.1371/journal.pone.0261042
Editor: Suhairul Hashim, Universiti Teknologi Malaysia - Main Campus Skudai: Universiti Teknologi Malaysia, MALAYSIA
Received: September 14, 2021; Accepted: November 22, 2021; Published: January 6, 2022
Copyright: © 2022 Li et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and supporting information.
Funding: This work was supported by the Funding: Key R&D plan of Science and Technology Program of Gansu Province, China. (19YF3FH001).
Competing interests: The authors have declared that no competing interests exist.
I. Introduction
The measurement of x-ray dose at the central axis in radiation oncology is usually tabulated and used for the clinical dose calculation. Percentage depth dose (PDD) and tissue-phantom ratio (TPR) are dominated by the scattering effect of depth and field size [1]. They consist mainly in two parts, primary fluence (adjusted for inverse square, beam hardening effects) and another part, which represents the effects of attenuation. PDD and the other quantities such as TPR and field size factor (Scp) are typically measured for simple square-shaped fields for each therapy machine and modality [2, 3].
The collection of clinical data for the implantation of treatment planning system for dose accuracy calculation is time consuming and repetitive. The achievement of this study provides a lot of help in double check of the measurement results to reduce the time of measurements, and the confidence to use interpolation. It would be helpful to know the smallest number of measurements needed to characterize the high-energy x-ray beam [4]. Several protocal provide a general framework and describes a large number of tests and procedures that should be considered by the users of RTPSs. However, the workload for the implementation of the recommendations from those documents are enormous and requires far more personnel and instrumentation resources than is available in most facilities, particularly within smaller hospitals. These hospitals are not always able to perform complete characterization, algorithm validation and software testing of dose calculation algorithms used in RTPS [5]. These may include the Sc factors, also known as head scatter factors, which account for the variation in beam output with field size from changes in direct and indirect radiation from the head of the linear accelerator. The term Scp contains both the collimator and phantom scatter that is defined as the ratio of dose for the field of interest to that of a reference field for the same delivered monitor units measured under full scatter conditions in a large water tank at the reference depth [6].
In this study, we proposed a simple mathematic equation to model the PDD by using the buildup-tail function. The quantity of interest is the parameters extracted from the well modeled PDD buildup-tail function for characterizing the Sc either in a square field or in the different individual upper jaw and lower jaw setting separately for specific patient treatment monitor unit calculation.
II. Materials and methods
II. A Experiment design and steps
The experiment was conducted in the following steps:
- The PDD for four high-energy photon energies of 4 MV, 6 MV, 10 MV, 18 MV were modeled in this study. The measurements data were 6 MV, 10 MV (Elekta infinity, Stockholm, Sweden, and Varian trueBeam, Palo Alto, Ca), while 4 MV and 18 MV were published data [7, 8].
- Measurement of photon beam PDD was conducted by two linear accelerators with two-photon energies of 6 MV and 10 MV at SSD = 100cm with different field sizes. The quantity percentage depth dose is defined as the quotient, expressed as a percentage, of the absorbed dose at any depth d to the absorbed dose at a reference depth (usually at the depth of dose maximum) along the central axis of the beam.
- The Sc for radiation beams of two-photon energies of 6 MV and 10 MV were measured.
- High energy photon PDD curves were modeled by the buildup and tail function generated in this study.
- The PDD of four-photon energies were modeled by the buildup-tail function by adjusting the parameters n and μ to get the best fitting.
- The Sc of 6 MV and 10 MV can then be expressed simply by the modeling parameters n and μ.
The details of each step are described in the following sections.
II. B Percent depth dose measurement
PDDs were acquired with a PTW MP3-T water phantom (PTW ionization Freiburg Gmbh) at WuWei Heavy Ion Center, Wuwei Cancer Hospital, Gansu, China (WHICH). PDDs were measured with a PTW Semiflex parallel-plate ionization chamber (PTW ionization Freiburg Gmbh, type 31010, volume 0.125 cm3). For the acquisition of PDDs the chamber position was automatically corrected to the effective point of measurement. In both photon energies water phantom measurements were performed at 100 cm source-to-surface distance (SSD) for square field sizes of 5 x 5, 10 x 10, 15 x 15, 20 x 20, and 40 x 40 cm2 with a step size of 0.1 cm. PDDs were normalized to 100% at dmax depth. Since the parallel-plate chamber has a small plate separation and it is explicit that the point of measurement is the front surface of the cavity. The depth curve measured by parallel-plate chamber was then compared with the PDD curve measured by films.
II. C The comparison of depth dose curve converted via parallel-plate chamber ionization curve with GAF EBT 3film
We used Gafchromic EBT3 films (Ashland Specialty Ingredients GP, NJ USA; Lot # 04022001) for the depth dose curves measurement in for comparing the PDD measured by plane-parallel ion chamber. The film processing and dose profile measurements followed the international protocols [9]. A pre-exposure technique was used for the calibration curve derivation [10]. This was performed by giving each film an initial dose of 2 Gy with a photon energy of 6 MV to homogenize the film polymer density in Gafchromic film experiment. The Hurter-Driffield calibration curve (H-D curve) [11] and the PDD films data scanned with a red filter [12] were analyzed by the VeriSoft imaging procession software (PTW-Freiburg, Germany, VeriSoft version 3.1) for further comparing with the depth dose curves measured by the parallel-plate chamber.
Absolute output and machine quality assurance were performed before conducting the measurements of percent ionization depth by the parallel-plate chamber as well as PDD curve by Gafchromic films.
II.D Measurement of Sc
The measurement of Sc for Elekta infinity and Varian trueBeam with the high-energy photon energy of 6 and 10 MV were conducted using a PTW Semiflex chamber, type31010 (PTW, Freiburg), coupled to either a PTW UNIDOS or a Scanditronix-Wellhofer Dose1 (IBADosimetry, Schwarzenbruck) electrometer. The chamber and phantom fulfill the suitability criteria for this type of measurement according to the recommendation by the AAPM TG-74 report. The chamber was placed in an aluminum(ρ = 2.7 gcm3) mini phantom with 3.9 cm of the material above the chamber, a measurement depth beyond dmax equivalent to a depth of 10 cm water for sufficiently avoiding contaminant electrons. The phantom and chamber axis were vertically aligned to the beam central axis, and the chamber reference point was set at 100 cm source-to-chamber distance. All measurements were normalized to the reference field reading of 10 x10 cm2. Sc values were measured for a selection of square fields from 3 x 3 to 40 x 40 cm2 with the chamber in mini-phantom.
II. E percent depth dose numerical equation
There are two home-generated numerical equations for describing the PDD curves of high energy photon beam, buildup function, and tail function. The buildup function was a quadratic function in the form of with two main parameters of d (depth in water) and n, while the tail function was in the form of e−μd and was composed by an exponential function with main parameters of d and μ. The modeled PDD was the product of buildup and tail function to be, PDDb-t =
.
PDDb-t is described separately as buildup function and tail function in the following,
where d is the depth in water along the central axis in unit cm, n is a beam hardening factor with unitless, a scalar.
where d is the depth in water along the central axis in unit cm, μ is a linear attenuation coefficient factor in unit cm-1 for adjusting the slope of the tail. The tail function in the form of e−μd was composed of an exponential function with main parameters of d and μ.
The empirical function of percentage depth dose is the combination of these two functions, denoted as PDDb-t (the abbreviation of PDDbuildup-tail)
Eq 1
The PDD of four photon energies were modeled by the buildup-tail function by adjusting the parameters n and μ to get the best fitting.
All PDD of four high-energy photon beams with different field sizes at SSD = 100cm were adjusted by the main parameters of n and μ to get the best fitting.
III. Results
III. A. The best fitting of percent depth dose was conducted by empirical function in four-photon energies
The dose variations of PDD of high energy photon beams measured by films and by ion chamber were less than 0.5%.
The PDD with different energies adopted in this study was already measured by the water phantom during commissioning and was spot checked in this experiment. By adjusting the main parameters of n and μ, the best fitting for four-photon PDD curves in every energy at the field sizes of 10 cm x 10 cm were listed in Fig 1A–1D. The comparison of four high-energy photon measured and modeled PDD with two published and two facility data is listed in Table 1.
The best fitting of photon PDD curve at energy 4, 6, 10 and 18 MV from a–d), representatively. The measurements data were 6 MV (b) and 10 MV (c) while 4 MV (a) and 18 MV (d) were published data.
Fig 1A–1D represents the fitting results of PDD curves of photon energy from 4 to 18 MV, representatively. Table 2 lists the best fitting parameters n and μ for four-photon energies.
III. B. The Sc expressed by the parameters n and μ generated in the empirical buildup-tail function of photon energy 6 and 10 MV
The Sc of photon energy 6 and 10 MV can be expressed in a certain acceptable deviation within 0.8% by the parameters n and μ as a function of field size in empirical buildup-tail function by the equation:
Eq 2
nE and μE denote the parameters n and μ in empirical buildup-tail function at photon energy E, while FS stand for field size of interest.
Fig 2A and 2B show the best fitting of the Sc by the Sc,E equation.
The Sc of photon energy 6 MV (a) and 10 MV (b) can be expressed in a certain acceptable deviation within 0.8% by using the parameters n and μ modeled in empirical Buildup-tail function by the equation of Sc,E = nE · (FS)0.63μE, nE and μE denote the parameters n and μ in empirical Buildup-tail function at photon energy E, FS denotes the field size.
The measured Sc of two-photon energies can be characterized by the parameters n and μ generated in the buildup-tail function and the comparison between characterized and measured was listed in Table 3.
Table 3 shows the measured Sc in a range of 0.9709 to 1.046 for Varian 6 MV flattened photon beam and 0.9739 to 1.0408 for 10 MV flattened photon beam at square field sizes from 4 cm x 4 cm to 40 cm x 40 cm. The deviation of Sc was characterized by Eq 2 and the measurements were between 0.8% to -0.2% for 6 MV, while for 10 MV were within 0.1%.
IV. Discussion
To let readers better understand the origin of the home-generated Buildup-tail function. The author describes the derivation of this function in the following.
The buildup-tail model originated from the proportion function y(x) = . When x increases from -∞ to 0, the curve of y is located in the region of the (-,-) quadrant. When x goes from 0 to +∞, the curve of y is located in the region of the (+, +) quadrant. Let
be
, then the curve of y(x) falls in the (-,+) and the (+,+) quadrants. The curve of y(x) =
has a left and right tail of the dose-profile-shape pattern. Let y(x) =
=
. When x = 0, y(x) becomes infinite which does not happen in real dose-profiles. Therefore, we insert n into y(x) to be: f(x) =
, where n>0., let tail(x) =
, where x is the depth in water in unit of cm, n is a scalar of spread factor in real number.
On other hand, the function tail(x) = demonstrates an ascending values tail(x)with an increasing depth of x in water. When introduces an exponential function eμx to tail(x), the combination becomes PDDb-t function
, namely, PDDb-t =
, where d is the depth in water in the unit of cm, n>0 and is a harden factor of real number scalar. When x = 0, n plays an important role to avoid primary(x) become infinite, while μ is the linear attenuation factor to fine turn the growth of the
value.
Finally, the buildup-tail model can be expressed as follows:
The PDD can be fitted by using the buildup-tail modeling by adjusting the main parameters of n and μ in all photon energy for the standard PDD curves in Fig 1A–1D. The random variations of modeled PDD with measured PDD had a maximum deviation within 1.5% as shown in Table 1.
The parameters n and μ represent the photon beam hardening factor of the buildup function and the beam penetration ability of the tail function, respectively. The more photon energy, the more n and less μ it has for the PDD curve fitting as shown in Table 2.
Fig 2 show the parameter n and μ in describing the measured Sc in the energy of 6 and 10 MV photon beam in Fig 2A and 2B, respectively. The Sc of photon energy 6 and 10 MV can be expressed with a certain deviation within 0.8% by using the parameters n and μ generated in empirical buildup-tail function by Eq 2,
Where nE and μE denote the parameters n and μ as a function of field size, FS, in empirical buildup-tail function at photon energy E.
A high energy photon beam usually has a high penetration ability, namely, has a small attenuation coefficient μ and also has a large n to own a less surface dose and the deeper dmax. The combination of a small μ and a large n, large μ, and small n can characterize a high and low energy photon beam percentage depth dose, respectively (Table 2).
The concept of in-air output ratio Sc was introduced to characterize how the incident photon fluence per monitor unit varies with collimator settings, and the author believe the Sc of a larger photon energy is more sensitive in accordance with the field size, namely, the larger collimator setting, the more Sc it is. This might be owing to the strong side scatter of the higher energy of photons. As we can see in Fig 2B, the Sc of 10 MV increases gradually when the field sizes open widely, on the contrary, Fig 2A, the Sc of 6 MV becomes slightly saturation when the field sizes open widely. This phenomenon causes the curvature of Sc of photon energy 10 MV to be easy fitted by the parameters n and μ as a function of field size than in a photon energy of 6 MV.
Table 3 shows the deviation of characterized and measured Sc of Varian photon energy at 6 and 10 MV.
Since parameter n represents photon beam hardening factor in buildup function, in other words, the larger n (n = 4.95) the high beam quality, therefore you can see the larger n, the less surface dose, and the deeper dmax (to compare n = 4.95 and n = 0.0495 in Fig 3).
The parameter μ represents the attenuation coefficient of the photon beam in the medium. The larger μ (μ = 0.458) the more attenuation, thus, the steeper curves and to have a shortened range (to compare μ = 0.458 and μ = 0.00458).
On the other hand, μ represents the attenuation coefficient, meanwhile, tail function represents the beam penetration ability of high energy photon beam.
As shown in Fig 3, the larger μ (μ = 0.458) the more attenuation when photon penetrating in the medium, therefore, you can see the larger μ, the steeper curves and to have a shortened range (to compare μ = 0.458 and μ = 0.00458).
It can be seen that the change of Sc at square field size with a range of 0.951 to 1.048 for the Elekta 6 MV flattened beam. It is very interesting that a weaker variation can be seen when the upper jaw and lower jaw collimator setting was reversed. For example, the Sc of upper jaw x lower jaw setting at 10 cm x 15 cm and 15 cm x 10 cm was 1.0049, 1.0104 in Table 4, respectively.
The Sc of Elekta 6 MV photon beam with different upper and low jaw setting from 4 cm x 4 cm to 40 cm x 40 cm was listed in Table 4. The orthogonal equal upper and lower jaw setting (square field) from the upper left to the lower right of 4 cm x 4 cm to 40 cm x 40 cm follow the Eq 2 while Sc can be calculated for the different upper jaw and lower jaw setting separately following the equation below,
Eq 3
The Sc of different upper and lower jaw settings can be calculated separately by using Eq 3. In Table 5a, the Sc can be calculated sequentially by using Eq 3 for various lower settings with upper jaw fixed at 4 cm, 5 cm, 20 cm, 30 cm, and 40 cm, while in Table 5b, the Sc was calculated sequentially for various upper setting with lower jaw fixed at 4 cm, 5 cm, 20 cm, 30 cm, and 40 cm. The Sc calculation by separating the individual upper or lower jaws setting does a lot of favor in individual treatment monitor unit double-checking at jaws opened asymmetrically [13].
b. shows the Sc was calculated sequentially for various upper settings with lower jaw fixed at 4 cm, 5 cm, 20 cm, 30 cm, and 40 cm.
V. Conclusions
The dosimetric quantities Sc, even the quantity Sp by dividing Scp with Sc, and hence required for planning system measurement, or a monitor unit check methodology, were easily and accurately parameterized for our flattened beams using a simple mathematical expression in this study. The data reproduced may be used as expectation values for comparison when commissioning similar beams, as there are scant published data on Sc of all photon beams from these accelerator types.
In this study, we presented a study with an empirical method to model the PDD curve for high energy photon beam by using the buildup and tail function in radiation therapy. The modeling parameters n and μ can also be used to predict the Sc in either square field or with jaws opening asymmetrically for individual treatment monitor unit double check in the patient’s dose calculation.
The achievement of this study provides a lot of help in double check of the measurement results to reduce the time of measurements, and the confidence to use interpolation. It would be helpful to know the smallest number of measurements needed to characterize the high-energy x-ray beams.
Acknowledgments
The author appreciates Professor Yan-Cheng Ye for his great contribution to this study. Professor Yan-Cheng Ye is juxtaposed with first author (co-first authors with equal contribution). Professor Yan-Shan Zhang is juxtaposed with correspondence author (co-corresponding authors with equal contribution).
References
- 1.
Khan FM. Dose distribution and scatter analysis. In Khan FM, ed The Physics of Radiation Therapy. 4th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2010: 140–157.
- 2. Zhu T.C, Ahnesjo A., Lam K.L., et al. In-air output ratio, Sc, for megavoltage photon beams. Med. Phys. 2009: 36(11):5261–90. pmid:19994536
- 3. Cashmore J. The characterization of unflattened photon beams from a 6 MV linear accelerator Phys. Med. Biol., 2008: 53(7), 1933–1946.
- 4. Commissioning of Radiotherapy Treatment Planning Systems: Testing for Typical External Beam Treatment Techniques Report of the Coordinated Research Project (CRP) on Development of Procedures for Quality Assurance of Dosimetry Calculations in Radiotherapy. IAEA-TECDOC-1583, 2008.
- 5. Gibbons J., Antolak J, Followill D.et al. Monitor unit calculations for external photon and electron beams: Report of the AAPM Therapy Physics Committee Task GroupNo.71, Med. Phys. 41(3);2014. pmid:24593704
- 6. Wang Ken Kang-Hsin, Zhu Timothy C. Modeling scatter-to-primary dose ratio for megavoltage photon beams. Med Phys, 2010 37, 5270. pmid:21089761
- 7. Kase Kenneth R. and Svensson G. K., Head scatter data for several linear accelerators (4–18 MV); Med Phys; 1986, 13, 530. pmid:3736512
- 8. Fátima Padilla-Cabal Mailyn Pérez-Liva, Lara Elier, et al. Monte Carlo calculations of an Elekta Precise SL-25 photon beam model, Journal of Radiotherapy in Practice. 2015; 14, 311–322.
- 9. Niroomand-Rad A., et al. Radiochromic film dosimetry: An update to TG-55 –Report of AAPM Task Group 235. Med. Phys. 2020, 47(12):5986–6025. pmid:32990328
- 10. García-Garduño OA, Lárraga-Gutiérrez JM, Rodríguez-Villafuerte M. Effect of correction methods of radiochromic EBT2 films on the accuracy of IMRT QA. App. Radi. Isot. 2016;107:121–126. pmid:26492322
- 11. Lewis D., et al. An efficient protocol for radiochromic film dosimetry combining calibration and measurement in a single scan. Med Phys, 2012, 39(10): 6339–6350. pmid:23039670
- 12. García-Garduño O.A., Lárraga-Gutiérrez J.M., Rodríguez-Villafuerte M. Effect of correction methods of radiochromic EBT2 films on the accuracy of IMRT QA. App. Radi. Isot. 2016, 107,121–126. pmid:26492322
- 13. Dutreix A, Bjarngard B, Bridier A, et al. Monitor unit calculation for high energy photon beams. Leuven: Estro Booklet No. 3; 1997.