Electron contamination for 6 MV photon beams from an Elekta linac: Monte Carlo simulation

This study investigates electron contamination for 6 MV photon beams from an Elekta linear accelerator (linac) using Monte Carlo simulation. The linac head was simulated by the BEAMnrc code and the absorbed dose in a phantom was calculated using the DOSXYZnrc code. The parameters of the initial electron beams on the target, such as mean energy and radial intensity distribution, were determined by matching the calculated dose distributions with the measured dose (at 10 x 10 cm2 field size and 90 cm source-skin distance). The central axis depth-dose curves of electron contamination were calculated for various field sizes from 5 x 5 cm2 to 40 x 40 cm2. We investigated the components that generated the electron contamination for a field size of 10 x 10 cm2. The optimal initial electron beam energy was 6.3 MeV with a full-width half maximum (FWHM) of the radial intensity distribution of 1.0 mm. These parameters were found to be in good agreement with the measured data. Electron contamination increased as the field size increased. At a depth of 1.0 mm and field sizes of 5 x 5, 10 x 10, 20 x 20, 30 x 30, and 40 x 40 cm2, the doses from electron contamination were 3.71, 5.19, 14.39, 18.97 and 20.89 %, respectively. Electron contamination decreased with increased depth. At a depth of 15 mm, the electron contamination was about 1 %. It was mainly generated in the air column between the linac head and the phantom (3.65 %), the mirror (0.99 %), and the flattening filter (0.59 %) (for a depth of 1.0 mm and field size of 10 x 10 cm2).


Introduction
Radiotherapy is an essential modality in cancer treatment. Radiotherapy employs ionizing radiation and its interaction with tissue can activate cell death, either directly or indirectly [1][2][3][4]. The radiosensitivity of most cancer cells is higher than healthy cells. Thus, ionizing radiation can kill cancer cells while maintaining as many healthy cells as possible. The ionizing radiation is carefully planned at the specific direction, energy, field size and intensity to achieve a maximum and homogeneous dose in the cancer tumor, while surrounding healthy tissue receives the minimum possible dose [3,4].
In general, the radiation beam used in radiotherapy comprises either electrons or photons. Electron beams are produced by a linear accelerator (linac) machine, while photon radiation can be generated by either a Cobalt-60 (Co-60) machine or by a linac machine [5][6][7][8][9]. Electron beams are usually used for tumors located at or near the surface of the skin (superficial tumors), while photon beams are used for tumors located at a greater depth [3].
Photon beams from radiotherapy units (Co-60 or linac), in fact, are no longer pure photon beams, but a mixture of photons and a small amount of electrons produced by the photon beams [10]. This effect is known as electron contamination [4]. The electron contamination has been studied by numerous authors [11][12][13][14][15][16]. Various techniques were used to determine the amount of electron contamination [7], including direct measurement using electron filter plates and sweeping magnets to remove electrons, analytical methods, and Monte Carlo simulation. However, no studies have been conducted on the parts and components that produce the electron contamination. The electron contaminations increase skin and subcutaneous tissue doses, however it does not have an impact to the clinical target volume (CTV) [17][18][19][20]. It is important to note that it is impossible to remove all electron contaminations. If the characteristics of the electron contaminations and the components producing it are well-understood, then the electron contaminations can be minimized [4,16]. In this study, the sources of electron contaminations are investigated using Monte Carlo simulation.

Methodology 2.1. Linac head and phantom
We investigated electron contamination by Monte Carlo simulation using the EGSnrc code. The EGSnrc code consisted of the BEAMnrc for linac head simulation [21] and the DOSXYZnrc for calculation of absorbed dose in the phantom [22]. For all steps of the simulations, an electron cut-off energy (AE) of 0.521 MeV and photon cut-off energy (AP) of 0.01 MeV were chosen. The number of initial electrons on the target to produce x-ray beams was taken to be 3 x 10 8 and the skin-source distance (SSD) was 90 cm.
All components of the linac head were specified exactly from an Elekta SL Series Linac Physics Manual. The components of the linac head were target, target block, primary collimator, flattening filter, monitor chamber, mirror, and jaws ( Fig. 1). The phantom was simulated using the DOSXYZnrc code. The phantom was made from water, and had dimensions of 40 cm x 40 cm x 40 cm. The dose zones for percentage depth dose (PDD) were determined along the central axis. Determination of the build-up dose was performed in the same way as the PDD determination, except it was focused on the region between the surface and the maximum depth dose (dm) and used smaller voxels. The dose zones for the lateral profile were perpendicular to the central axis. It was measured at a depth of 1.5 cm. To determine the optimum model, the energies of initial electrons were varied from 5.7 to 6.6 MeV. The full-width half maximums (FWHMs) of radial intensity distribution of incident electrons were varied from 0.5 to 2 mm. After an optimum model was achieved, i.e. in the good agreement with the measured data [23], the model was used for the next simulation.

Dose of electron contaminations
The dose of electron contamination was determined by investigating electrons from the phase space file (PSF) produced by the BEAMnrc code. The source of electron contamination was traced by the LATCH (program code to trace particle's history). Number LATCH for Target, Target Block, Primary Collimator, Flattening Filter, Monitor Chamber, Mirror, Jaws and the air between linac head and phantom, were 1, 2, 3, 4, 5, 6, 7, and 8, respectively.

Results and Discussions 3.1. Standard model determination
Before calculation of electron contamination, it is important to accurately determine a standard model of the initial electron beams, in terms of energy and radial distribution of incident electrons. The energy and the FWHM of their radial intensity distribution of incident electrons were varied [24]. The resulting maximum depth dose, PDD, and the dose profile were compared with those obtained from direct measurements. Fig. 2(a) shows that the maximum depth dose (dm) increased as the energy of incident electron increased. The maximum depth dose is 1.5 cm for electron energies of 5.7 and 6.0 MeV, 1.6 cm for an energy of 6.3 MeV, and 1.7 cm for an energy of 6.6 MeV. The measurement of maximum depth dose was at a depth of 1.5 cm. Figure 2 There are two factors to determine standard model of FWHM, i.e. PDD and dose profile. Figure 3 shows that the PDDs and dose profiles from measurement and simulation for various FWHMs are very similar. However, a FWHM of 1.0 mm is in best agreement with measurement data. In addition, Elekta also recommends this value [25]. This value is recommended to ensure that the X-ray source is small enough to reduce the effect of penumbra.  Figure 4(a) shows the doses at the build-up region for a field size of 10 x 10 cm 2 . The photon doses increase with increased depth. Doses in build-up region increase due to secondary charged particles (electrons and positrons) that are released in the phantom by photon interactions (i.e. photoelectric effect, Compton effect, and pair production). The charge particles have relatively long range and then deposit their energy kinetic in the phantom. Figure 4(b) shows that the doses of electron contamination decreased with depth. Decreasing electron contamination dose is due to a decreasing energy of electron contaminations in their tracks. It is shown from Figure 4(b) that the doses of electron contaminations are about 1.0% at a depth 1.5 cm (i.e. depth of dose maximum (dm)). An increase of field size causes an increase of electron contamination at the surface or beneath the surface of the phantom [16]. At a depth of 1.0 mm, the doses of electron contaminations were 3.71, 5.19, 14.39, 18.97, and 20.89 % for field sizes of 5 x 5, 10 x 10, 20 x 20, 30 x 30, and 40 x 40 cm 2 , respectively. These findings are in agreement with previous studies [3,24]. It is shows that the energy of electron contaminations at field sizes from 5 x 5 to 40 x 40 cm 2 were similar, but the number of contaminating electrons varies with field size.

Electron contamination
In Monte Carlo simulation, the probability of every interaction can be followed and the source of electron contamination can be traced. The electron contamination generated from the various components of the linac head can be found using Figure 5. The electron contamination is mainly generated in the air column between the linac head and the phantom, mirror and flattening filter. The other parts of the linac head do not contribute much. The fluctuation of percent doses at a field sizes of 30 x 30 and 40 x 40 cm 2 due to the random nature of small number of electron contaminations. If the simulation used greater number of particles, then greater number electron contaminations will be produced and the fluctuation of electron contaminations will be reduced. However, it needs additional computation time.
The electron contamination generated in the air column between the linac head and the phantom was very significant [26], due to the large volume of the air column and its position near the water phantom. The electron contamination generated in the mirror is also significant because all photon beams interact with the mirror, and there are no materials between mirror and phantom except the air. Although the mirror is very thin (consisting of mylar with a thickness of 0.3 mm), it provides a significant contribution to the electron contamination.
The contribution from flattening filter is smaller than that from the mirror, because most of its generated electrons are absorbed within the flattening filter itself.
The monitor chamber contributes insignificant electron contamination, because it is very thin and made from mylar film, plates and air. The primary collimator and jaws also give a small contribution, because the interaction with the photon beams occur only at the edge of the beam [17]. And the contribution of the primary collimator is insignificant because there are many components that absorb electron contamination generated within it.

Conclusions
The dose due to electron contamination and their sources have been investigated. It is important to determine a standard model before performing Monte Carlo simulation. The standard initial electron beam parameters for the model were 6.3 MeV energy and 1.0 mm FWHM of the radial intensity distribution. These values were found to be in best agreement with the measured data. The electron contamination increased as the field size increased, and the dose from electron contamination dropped off rapidly with depth. At a depth of 1.5 cm, the contribution of electron contaminations is about 1%. These contaminations are generated mainly in the air column between the linac head and the phantom, as well as in the mirror and the flattening filter.