Calculation of Photoneutron Contamination of Varian Linac in ICRU Soft-Tissue Phantom Using MCNPX Code.

PURPOSE: The aim of this research was to calculate the fluence, dose equivalent (DE), and kerma of thermal, epithermal and fast photoneutrons separately, within ICRU soft-tissue-equivalent phantom in the radiotherapy treatment room, using MCNPX Monte Carlo code.
MATERIALS AND METHODS: For this purpose, 18 MV Varian Linac 2100 C/D machine was simulated and desired quantities were calculated on the central axis and transverse directions at different depths.
RESULTS: Maximum fluence, DE and kerma of total photoneutrons on central axis of the phantom were 43.8 n.cm-2.Gy-1, 0.26, and 3.62 mGy.Gy-1, at depths 2, 0.1, 0.1 cm, respectively. At any depth, average of fluence, DE and kerma in the outer area of the field were less than the inner area and in general were about 72%, 52%, and 45%, respectively.
CONCLUSION: According to this research, within the phantom; variation of fluence, DE and kerma in transverse direction were mild, and along the central axis at shallow area were sharp. DE of fast photoneutrons at shallow and deep areas were one order of magnitude greater than thermal photoneutrons. Copyright:

INTRODUCTION

Radiation therapy is an established way for treating cancers. The main challenge in cancer treatment using radiotherapy (RT) is delivering high doses to the tumor volume without harming the healthy tissues.

Unfortunately, despite applying accurate and conform field shape, some unwanted radiation reaches to normal tissues around the tumor and cause damage. This unwanted radiation emerges from scattered photons from photon treatment beam, gamma rays due to produced radionuclides in the treatment room, secondary gamma rays due to inelastic interactions, electron contamination of the treatment beam (i.e. in photon mode) as well as produced neutrons from electronuclear and photonuclear interactions which are named electroneutron and photoneutron.

Interactions between high-energy treatment beam and nuclei of constituent minerals of the medical linear accelerator (Linac), beam collimation system, couch, patient's body, air, and walls of treatment room can produce photoneutrons. Since the threshold energy of (γ, n) reaction for constituent minerals of the head of Linac, such as lead, tungsten, copper, and iron, is generally in the range of 6.74–11.20 MeV.[1]

Since the quality factor of neutrons is about 2–20 times more than photons,[2] they have a substantially higher biological effectiveness than photons; therefore, even a small number of neutrons can lead to a nonnegligible effective dose to both patients, in the form of non-target and out-of-field dose, and staff due to activation of in-room materials.[3] The estimation of photoneutron contamination in RT has been studied by several researchers in various experimental and simulation methods.[45678910] Bezak et al.[11] and Bezak et al.[12] measured the total dose equivalent (DE) in Rando and water equivalent phantoms, using thermoluminescent dosimeter (TLD) and estimated the risk of secondary cancer in organs of Rando phantom in treatment of prostate. Sohrabi and Hakimi measured the dose of thermal and epithermal photoneutrons using a self-made experimental method within a polyethylene phantom.[13] Bagheri et al.[14] and Bagheri et al.[15] measured the dose of thermal photoneutrons in the treatment of breast cancer within the breast Rando phantom using TLD. Results of these experiments mostly were inconsistent with each other and often with low precision. It seems applying Monte Carlo codes such as MCNPX is more reliable. Barquero et al. calculated the effects of total photoneutrons on various organs using MCNP code in a computational phantom.[16] Many others calculated the spectra of photoneutrons and DE due to photoneutrons in tissue[17181921] and some of them[1822] calculated the DE of fast neutrons in voxel-based phantoms. Calculating the effects of each category of photoneutrons along the beam axis, in water equivalent and water phantoms were conducted by many other researchers.[523222425] Effects of the compensator, pelvic prosthesis, circular cones, grid, and dental restorations were also investigated, too. Investigating the effect of the compensator on photoneutron production was performed using the Monte Carlo MCNPX code and SSNTD CR-39 dosimeter[26] and calculating the effects of the grid on photoneutrons production was conducted using MIRD phantom.[42728]

As it is well-known, absorption is the predominant mechanism for losing the energy of thermal neutrons in soft tissue through reactions of 14N (n, p) 14c and 1H (n, γ) 2H. Hence, to consider the role of nitrogen and similarity to the tissue of the body, we used ICRU soft tissue equivalent in simulations. Most of the researchers have studied photoneutrons distribution in the air of the treatment room and only few works studied photoneutrons within the phantom. In addition, previous researches often have reported photoneutrons contaminations in whole or focused to the proportion of fast photoneutrons. In this research, we determined the proportion of thermal (0.001–0.5 eV), epithermal (0.5 eV–0.1 MeV), and fast (0.1–20 MeV) photoneutrons in axial and transverse directions at 159 points of the phantom.

MATERIALS AND METHODS

Simulation

A typical treatment room[28] with walls, ceiling (thickness of 1.7 m) and floor (thickness of 1 m) from concrete (with density of 2.35 g/cm3 and weight fractions of H 0.6%, O 49.8%, Na 1.7%, Mg 0.3%, Al 4.6%, Si 31.5%, S 0.1%, K 1.9%, Ca 8.3% and Fe 1.2%) simulated [Figure 1].[2728] The MCNPX Monte Carlo code, Version 2.6, was applied for simulating the 18 MV Varian Linac 2100 C/D machine [Figure 2]. The head of the Linac including all effective components therein containing the target (W), primary collimator (W), vacuum window (Be), flattening filter (Fe and Ta), ionization chamber (Cu and Kapton), secondary collimator (W and Pb)), mirror (Mylar), Jaws (W), and upper circle (Fe) were simulated.

Figure 1

Simulated treatment room[28]

Figure 2

ICRU phantom and different components of the head of 18 MV Varian Linac 2100 C/D. These components were simulated using Visual Editor V22. Dimensions of each component were shown within parentheses

The source was an electron beam incident on a target made of tungsten and copper and produce Bremsstrahlung photons. Both spatial distribution and energy distribution of the electron beams were considered gaussian, with FWHM of 0.14 cm and FWHM of 1.2 MeV respectively, and the mean energy of electrons set to 18.3 MeV to have the most conformity between measurements and calculations for percentage depth dose and dose profile

Verification of simulation

At first, percentage depth dose and dose profile of the photon beam (set at energy 18 MV, field size 10 cm × 10 cm and source-surface distance [ssd] =100 cm) were measured by a 0.6 cc Farmer ionizing chamber (PTW Freiburg, Germany) at an IBA Blue Phantom (IBA dosimetry Schwarzenbruck, Germany) with dimensions of 50 cm × 50 cm × 50 cm and constitution of water. For validating the accuracy of simulation, the mentioned parameters were used exactly by simulation. Percentage depth dose (PDD) and relative dose Calculated using *F8 tally and 2 × 109 histories. Energy Cut-off cards were used with values of 0.7 and 0.01 MeV for electron and photon, respectively. For calculating the PDD, voxels' dimensions were set to be 2 cm × 2 cm × 0.2 cm along the central axis (Z-axis) from 0 to 29 cm. For calculating the dose profile, the voxels in the transverse axis (X-axis) had dimensions of 0.4 × 2 × 0.4 cm3 [Figure 3].

Figure 3

Side view of simulated water phantom. The voxels along Z axis used for calculation of percentage depth dose and the voxels along X axis used for dose profile

Calculation of photoneutron contamination

For calculating photoneutron contamination, ICRU soft-tissue-equivalent phantom (with weight fractions of 10.1% H, 11.1% C, 2.6% N, 76.2% O) with dimensions of 100 cm × 50 cm × 30 cm was simulated [Figures 2 and 4]. In all calculations, the Linac was considered to be in photon mode with 18 MV energy, 10 cm × 10 cm field size and SSD = 100 cm. In these calculations, 2 × 109 electron histories were traced and the relative statistical uncertainties at the majority of points was <10%. To reduce run time, the electron and photon energy Cut-offs were set to be 7 MeV. In Z direction, thickness of the cells were as follows: First cell 0.2 cm, from depth 0.2 cm to 10 cm, 0.5 cm and after that 1.5 cm [Figure 4]. Scoring cells considered along the beam axis with dimensions of 1 × 1 × (0.2, 0.5, 1.5) cm3 and at transverse directions 1 × 2 × (0.2, 0.5, 1.5) cm3.

Figure 4

Location of scoring cells in ICRU phantom (side view). The voxels in Z direction were along the central axis of the beam

The mode of the simulation was electron-photon-neutron with cross-section libraries of MCPLIB04, EL03, and ENDEF/B-V2 for photon, electron, and neutron, respectively. In the production of photoneutrons KAERI01u, LA150u, and CNDC01u libraries were also used. For electron, photon, and neutron, the default physics cards were used unless in the PHYS: N card the maximum energy of neutrons set to 40 MeV for variance reduction and the fourth entry at PHYS: P card set to 1 to reduce the relative statistical uncertainty and allow photoneutrons production interactions.

Neutron source strength (Qn), was calculated based on McGinley and Landry method.[10] First, a sphere, with center at the target and radius of 100 cm, was considered, then the number of passed neutrons was calculated using F1 tally on the surface of the sphere, with closed collimator jaws. The code computes tallies for one electron of the source. For comparing the results with findings of others, output was derived in terms of 1 Gy of photon dose in the water phantom at “depth of maximum absorbed dose” (dmax). Hence, the number of required electrons for delivering 1 Gy of photon dose at dmax was derived and multiplied with the outputs of the code. Accordingly, the neutron source strength obtained in terms of n/Gy.

The fluence of photoneutrons was calculated using F4 tally in terms of neutrons/cm2/electron and converted to n/cm2/Gy unit.

DE obtained by F4 tally and fluence-to-DE coefficients. These coefficients derived from NCRP No. 38 and exerted using “dose energy” and “dose function” cards. Unit of this quantity converted to mSv/Gy.

Kerma acquired using F6 tally in terms of MeV/g/electron, which changed to mGy/Gy.

RESULTS

In Figures 5, 6 and 7 results of measurements and calculations for PDD and dose profile were compared with each other. Uncertainty of calculated doses was often <1%. For evaluating the agreement between calculated and measured results, the gamma index values were derived for criteria of 2 mm for “distance to agreement” and 2% for “dose difference.” In all points, the values of the gamma index were less than unity. Gamma index <1 means the existence of an acceptable agreement between simulation and measurement. Hence, the accuracy of the simulation was verified.

Figure 5

(a) Dose profiles in water phantom were determined by measurement and calculation at depth 15 cm, (b) gamma index values for comparing measurement with and calculation

Figure 6

(a) Percentage depth doses in water phantom were determined by measurement and calculation, (b) Gamma index values for comparing the differences between measurement and with calculation

Figure 7

(a) Dose profiles in water were determined by measurement and calculation at depth of 4 cm, (b) gamma index values for comparing measurement and with calculation.

The number of required electrons for delivering 1 Gy of photon dose in the water phantom at dmax, was calculated. This number was depicted by Ne in Table 1. An important quantity, which demonstrates the ability of each Linac for producing photoneutrons, was the neutron source strength. The maximum fluence of photoneutrons on the beam axis (Фmax) with related and the corresponding depths was were shown in this table. As it is obvious from Table 1, our results were in agreement with published data and validate the simulation for photoneutron calculations as well

Table 1

Required electrons for delivering 1 Gy of photon dose (Ne), neutron source strength (Qn), maximum fluence of photoneutrons on the beam axis (Фmax) and depth of maximum fluence of photoneutrons (dФmax) for Varian 2100 C/D Linac operated at 18 MV in a number of studies

Authors	Ne (105electrons/Gy)	Qn (1012n/Gy)	Фmax (10-8n cm-2/electrons)	dФmax (cm)
Mesbahi et al.[29]	_	1.3	_	_
Martínez-Ovalle et al.[18]	_	_	3.25	2.03
Chegeni et al.[28]	1.26	1.38	_	_
This work	1.28	1.37	3.42	2

Figure 8 shows the fluence of thermal, epithermal, fast and total photoneutrons along the central axis as well as transverse direction at depths of 0.1, 1, 2, 10, and 20 cm of the phantom. Results of Kry et al.[19] and Martínez-Ovalle et al.[18] (i. e., with the same type of Linac, energy, phantom, field size, and SSD) for total photoneutrons were depicted on Figure 8a and shows agreement between our calculations and these data.

Figure 8

Fluence (a), dose equivalent (b), and kerma (c) of photoneutrons along central axis of the 18 MV photon treatment beam and 10 cm × 10 cm radiation field in the ICRU phantom. For comparison the results of Martínez-Ovalle et al.[18] and Kry et al.[19] were also depicted (a)

Figure 9 shows the fluence of thermal, epithermal, fast and total photoneutrons in transvers direction at depths of 0.1, 1, 2, 10 and 20 cm of the phantom

Figure 9

Fluence of thermal (a), epithermal (b), fast (c) and total (d) photoneutrons at transvers direction of the 18 MV photon treatment beam for 10 cm × 10 cm radiation field at different depths in the ICRU phantom

Figure 10 shows DE of thermal, epithermal, fast, and total photoneutrons in transverse direction at depths of 0.1, 1, 2, 10, and 20 cm of the phantom. Calculations of Kry et al.[19] (i. e., with the same type of Linac, energy, phantom, field size, and SSD) for total photoneutrons are depicted in Figure 10d that were consistent with this research

Figure 10

Dose equivalent of thermal (a), epithermal (b), fast (c) and total (d) photoneutrons at transvers direction of the 18 MV photon treatment beam and 10 cm × 10 cm radiation field at different depths in the ICRU phantom. For comparison the results of Kry et al.[19] in the same situation were also depicted (d)

Figure 11 shows kerma of thermal, epithermal, fast, and total photoneutrons in transverse direction at depths of 0.1, 1, 2, 10, and 20 cm of the phantom

Figure 11

Kerma of thermal (a), epithermal (b), fast (c) and total (d) photoneutrons at transvers direction of the 18 MV photon treatment beam and 10 cm × 10 cm radiation field at different depths in the ICRU phantom.

DISCUSSION

Figure 8 shows fluence, DE and kerma of photoneutrons on the central axis of the treatment photon beam. In Figure 8a, the findings of Kry et al.[19] and Martínez-Ovalle et al.[18] are also displayed. Since, in the study of Kry et al.,[19] the parameters were in terms of an arbitrary unit, Martínez-Ovalle et al.[18] multiplied them by 1.28 to fit the peaks of the curves. After depth of 15 cm Martínez-Ovalle et al.[18] 's curve shows a little difference with both Kry et al.[19] 's study and ours. Ignoring the role of walls, was the mean reason for continuing decline after the depth of 15 cm in this graph. It has been shown, if we ignore the walls of the treatment room, the fluence of thermal photoneutrons at isocenter will estimated less than the actual amount.[22] Therefore, the total fluence will be underestimated too. Martínez-Ovalle et al.,[18] have attributed this effect to the existence of couch and backscattering of photoneutrons from it, that noted in Kry et al.[19] 's simulation, however; we think it is because of ignoring the walls, since; we did not consider the couch and observed the same behavior. According to the Figure 8 effect of walls is negligible. Fluence of total photoneutrons on the central axis reached to its maximum value at depth 2 cm and then declined rapidly. Fast photoneutrons after entering the phantom, due to elastic interaction with Hydrogen nuclei and also elastic and inelastic interactions with other nuclei, gradually lose their energy and the number of thermal photoneutrons increase up to the depth of 4 cm then decrease rapidly because their capture cross-section will increase [Figure 12].[30] Due to the large number of Hydrogen nuclei relative to Nitrogen nuclei,[31] these nuclei have dominant role in the absorption of neutrons. From Figures 8 and 9, it could be observed that the fluence of thermal photoneutrons at depths more than about 1 cm was greater than epithermal and fast photoneutrons in all points of the phantom. DE and kerma had their maximum at the surface layer of the phantom (i.e. depth 0.1 cm) and rapidly reduce until the depth of 15 cm and after that remain nearly constant. Therefore, we can divide the thickness of the phantom into two areas: Shallow area (i.e. depths <15 cm) and deep area (i.e. depths more than 15 cm). In Table 2 average of fluence, DE and kerma of total photoneutrons at deep area were compared with their maximums. From data of Figure 8b, it could be found out that from the surface of the phantom to the depth of 5 cm the DE of fast photoneutrons was one order of magnitude greater than values related to thermal photoneutrons and at depths more than 5 cm their order of magnitudes were the same. From Figure 8c we can result that from surface of the phantom to depth of 5 cm the kerma of fast photoneutrons was two order of magnitude greater than values related to thermal photoneutrons and at depths more than 5 cm this ratio was one order of magnitude.

Figure 12

Thermal neutron capture cross section with hydrogen nucleus

Table 2

Fluence, dose equivalent and kerma of total photoneutrons on central axis of the treatment beam in the phantom

Dosimetry quantity	dmax (cm)	Maximum	ADA	(ADA/maximum) × 100
Fluence	2	43.8 n/cm2/Gy	3.89 n/cm2/Gy	9%
DE	0.1	0.26 mSv/Gy	0.004 mSv/Gy	2%
Kerma	0.1	3.62 mGy/Gy	0.08 mGy/Gy	2%

dmax: Depth of maximum quantity, ADA: Average in deep area. DE: Dose equivalent

Figure 9 shows the Fluence of thermal, epithermal, fast, and total photoneutrons at transverse direction. Attention to these graphs indicates that fluence within the photon beam and lateral distances more than 10 cm for each group of photoneutrons, at all depths, were almost invariable and gradually decrease between these two areas. So, we can consider the phantom in the lateral direction in three areas: Inner area of photon treatment beam (i.e. transverse distances less than 5 cm), photoneutron's penumbra area (i.e. transverse distances between 5 cm and 10 cm) and outer area (i.e. transverse distances more than 10 cm). In Table 3, as a sample, average of total photoneutrons at the inner area and outer area in different depths were compared. The average fluence at the outer area was less than the inner area and in the whole was about 72%. This comparison could be carried out for thermal, epithermal, and fast photoneutrons separately, too. At each depth, fluence of thermal, epithermal, and fast photoneutrons had the same order of magnitude.

Table 3

Average fluence of total photoneutrons in inner and outer areas at different depths and corresponding relative values

Depth (cm)	AIA (n/cm2/Gy)	AOA (n/cm2/Gy)	(AOA/AIA) × 100
0.1	28.7	21.1	74%
1	37.4	26.1	70%
2	40.3	26.9	67%
10	15.2	10.0	66%
20	2.8	2.3	82%
			Average at all depths 72%

AIA: Average in inner area, AOA: Average in outer area

Figure 10 shows DE of thermal, epithermal, fast, and total photoneutrons at transverse direction. In Table 4, as a sample, the average DE of total photoneutrons at the inner area and outer area in different depths have been compared with each other. The average of DE at the outer area was less than the inner area, especially for fast photoneutrons, and in whole, was about 52% of its inner value. At shallow area, DE of fast photoneutrons was one order of magnitude larger than thermal and epithermal photoneutrons, but in deep area, all three categories had the same order of magnitude

Table 4

Average DE of total photoneutrons in inner and outer areas at different depths and corresponding relative values

Depth (cm)	AIA (mSv/Gy)	AOA (mSv/Gy)	(AOA/AIA) × 100
0.1	3.23	1.69	52%
1	2.85	1.48	52%
2	2.43	1.19	49%
10	0.41	0.20	47%
20	0.059	0.036	60%
			Average at all depths 52%

AIA: Average in inner area, AOA: Average in outer area

Figure 11 shows kerma of thermal, epithermal, fast, and total photoneutrons at transverse direction. In Table 5, as a sample, the average kerma of total photoneutrons in inner and outer areas at different depths have been compared with each other. Average of kerma in the outer area was less than the inner area, especially for fast photoneutrons, and in whole was about 45% of its inner value. At shallow area, kerma of fast photoneutrons was two order of magnitude larger than thermal photoneutrons

Table 5

Average kerma of total photoneutrons in inner and outer areas at different depths and corresponding relative values

Depth (cm)	AIA (mGy/Gy)	AOA (mGy/Gy)	(AOA/AIA) × 100
0.1	0.23	0.12	51%
1	0.20	0.10	50%
2	0.17	0.08	46%
10	0.023	0.009	38%
20	0.003	0.001	42%
			Average at all depths 45%

AIA: Average in inner area, AOA: Average in outer area

CONCLUSION

In ICRU soft-tissue phantom, variation of fluence, DE and kerma in transverse direction were mild, and along the central axis at the shallow area were sharp. At any depth, average of fluence, DE and kerma in the outer area of the field were less than the inner area and in general were about 72%, 52%, and 45%, respectively. Fluence of thermal photoneutrons at all points, except at depths <1 cm, was more than other categories of photoneutrons and their order of magnitudes were the same. DE of fast photoneutrons at shallow area was one order of magnitude greater than thermal photoneutrons, and for deep area had the same order. At shallow area, kerma of fast photoneutrons was two order of magnitude larger than thermal photoneutrons.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.