Home Search Collections Journals About Contact us My IOPscience

Energy spectra, angular spread, fluence profiles and dose distributions of 6 and 18 MV photon

beams: results of Monte Carlo simulations for a Varian 2100EX accelerator

This content has been downloaded from IOPscience. Please scroll down to see the full text.

2002 Phys. Med. Biol. 47 1025

(http://iopscience.iop.org/0031-9155/47/7/303)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 134.99.128.41
This content was downloaded on 19/11/2013 at 08:01

Please note that terms and conditions apply.

INSTITUTE OF PHYSICS PUBLISHING PHYSICS IN MEDICINE AND BIOLOGY
Phys. Med. Biol. 47 (2002) 1025–1046 PII: S0031-9155(02)28915-0

Energy spectra, angular spread, fluence profiles and

dose distributions of 6 and 18 MV photon beams:
results of Monte Carlo simulations for a Varian
2100EX accelerator

George X Ding
Medical Physics, Fraser Valley Cancer Center, British Columbia Cancer Agency, 13750,
96th Avenue, Surrey, BC, Canada V3V 1Z2

E-mail: gding@bccancer.bc.ca

Received 19 September 2001, in final form 6 December 2001

Published 12 March 2002
Online at stacks.iop.org/PMB/47/1025

Abstract
The purpose of this study is to provide detailed characteristics of incident
photon beams for different field sizes and beam energies. This information is
critical to the future development of accurate treatment planning systems. It
also enhances our knowledge of radiotherapy photon beams. The EGS4 Monte
Carlo code, BEAM, has been used to simulate 6 and 18 MV photon beams from
a Varian Clinac-2100EX accelerator. A simulated realistic beam is stored in a
phase space data file, which contains details of each particle’s complete history
including where it has been and where it has interacted. The phase space files
are analysed to obtain energy spectra, angular distribution, fluence profile and
mean energy profiles at the phantom surface for particles separated according to
their charge and history. The accuracy of a simulated beam is validated by the
excellent agreement between the Monte Carlo calculated and measured dose
distributions. Measured depth–dose curves are obtained from depth–ionization
curves by accounting for newly introduced chamber fluence corrections and
the stopping-power ratios for realistic beams. The study presents calculated
depth–dose components from different particles as well as calculated surface
dose and contribution from different particles to surface dose across the field.
It is shown that the increase of surface dose with the increase of the field size
is mainly due to the increase of incident contaminant charged particles. At
6 MV, the incident charged particles contribute 7% to 21% of maximum
dose at the surface when the field size increases from 10 × 10 to 40 ×
40 cm2. At 18 MV, their contributions are up to 11% and 29% of maximum
dose at the surface for 10 × 10 cm2 and 40 × 40 cm2 fields respectively.
However, the fluence of these incident charged particles is less than 1% of
incident photon fluence in all cases.
(Some figures in this article are in colour only in the electronic version)
0031-9155/02/071025+22$30.00 © 2002 IOP Publishing Ltd Printed in the UK 1025
1026 G X Ding

1. Introduction

BEAM is a general purpose Monte Carlo user code for simulation of radiotherapy beams
from treatment units (Rogers et al 1995). It is based on the EGS4 code system. It has been
benchmarked against measurements for electron and photon beams from accelerators as well as
beams from Co-60 units (Ding 1995, Mora et al 1999, Rogers et al 1995, Sheikh-Bagheri et al
2000). The original studies mainly focused on electron beams (Ding et al 1995, 1996, 1997,
Zhang et al 1998, 1999). Using the BEAM code, the characteristics of the electron beams
in commercial medical accelerators have been extensively studied (Ding and Rogers 1995,
Ma et al 1997). However, similar detailed information for photon beams is not easily
available although the BEAM code has been widely used to study photon beams from medical
accelerators (Deng et al 2000, Faddegon et al 1999, Keall et al 2000, Liu et al 1997a, 1997b,
Schach von Wittenau et al 1999, 2000, Siebers et al 1999, van der Zee and Welleweerd 1999).
The energy and angular distributions of photons and contaminant charged particles
(electrons and positrons) from medical accelerators are the most important characteristics
of radiotherapy photon beams. Knowledge of clinical beams is essential for dosimetry and
the development of a new accurate treatment planning system. Experimentally it is difficult
to obtain detailed information because of various limitations in the clinical environment and
detectors. A number of experimental investigations of photon energy spectra have been
reported (Huang et al 1982, 1983, Jessen 1973, Levy et al 1974, 1976, Nath and Schulz 1976,
Swindell 1983). Some of the techniques used in the literature employ reconstruction methods
to obtain the energy spectra from measured narrow-beam transmission data. Others deduce
energy spectra from the spectroscopy of Compton scattered photons. One of the major
advantages of the Monte Carlo technique is that it allows detailed information about each
particle’s history to be known. Monte Carlo simulation can be used to obtain the information
that cannot be measured experimentally. Originally, Mohan et al (1985) had calculated a
photon’s energy spectra and angular distributions using the EGS3 Monte Carlo code. The
charged particles in the photon beams were not studied. Using the BEAM code van der Zee
and Welleweerd (1999) investigated some characteristics of a 10 MV photon beam from
an ELEKTA SL linac. Deng et al (2000) also studied photon beam characterization and
modelling for treatment planning of 4 to 15 MV beams from Varian Clinac 2100C and
2300CD accelerators.
This study aims to provide more comprehensive information for radiotherapy photon
beams including incident photons as well as contaminating electrons and positrons in a
radiotherapy beam for different field sizes and beam energies. This information enhances
our knowledge of radiotherapy photon beams. It also serves as a benchmark to demonstrate
the accuracy of the Monte Carlo technique in simulating the radiotherapy photon beams. In
addition, it provides detailed information on the Monte Carlo computing speed required to
simulate an incident beam and to calculate a dose distribution on current computers.

2. Methods and materials

2.1. Monte Carlo calculation

The EGS4 user code BEAM was used to simulate 6 and 18 MV radiotherapy photon beams
emerging from an accelerator. The position, energy, angle, charge and weight of scored
particles were stored in a phase space file. The variable LATCH (Rogers et al 1995) was used
to record each particle’s complete history of where a particle has been or where a particle
has interacted in the beam simulation. The stored phase space files were used repeatedly for
Energy spectra, angular spread, fluence profiles and dose distributions of photon beams 1027

Table 1. BEAM simulation benchmark data when bremsstrahlung splitting technique is not
employed. The simulation is performed on a Pentium III with 1 GHz processor. The total
particle’s entries include photons, electrons and positrons. Each particle’s entry has a weight of 1.

Phase space
Energy Field size Histories/h file size (kb) Total entries Photon entries
(MV) (cm2) (×106) (per 1000 history) (per 1000 history) (per 1000 history)

6 10 × 10 7.55 1.05 × 10−3 2.063 2.055

6 40 × 40 7.40 5.78 × 10−2 37.49 37.36
18 10 × 10 4.56 0.243 8.669 8.559
18 40 × 40 4.53 4.32 153.4 153.0

Table 2. BEAM simulation benchmark data when the selective bremsstrahlung splitting (SBS)
technique is employed. The parameters used with SBS are: Nmin = 10, Nmax = 100. The simulation
is performed on a Pentium III with 1 GHz processor. Each particle’s entry has weights between
0.01 and 0.1.
Beam
nominal Phase space Total entries Photon entries
energy Field size Histories/h file size (kb) (per 1000 history) (per 1000 history)
(MV) (cm2) (×106) (per 1000 history) (×102) (×102)
6 10 × 10 2.96 3.07 1.095 1.091
6 40 × 40 2.09 74.5 26.61 26.51
18 10 × 10 0.538 17.9 6.392 6.333
18 40 × 40 0.271 378 135.0 133.8

analysing beam or as input to the EGS4 user code DOSXYZ to calculate the dose distributions
in a water phantom. This investigation presents simulated 6 and 18 MV beams from a Varian
Clinac-2100EX linear accelerator.
In the incident beam simulation, the EGS4 options and settings used to generate the
phase space files were: AE = ECUT = 0.700 MeV, AP = PCUT = 0.010 MeV, no photon
interaction forcing and no Rayleigh scattering. The electron range rejection technique was used
to increase the simulation speed. The value for ESAVE = 2 MeV was used since it provided a
factor of 2–3 increase in speed and ignored only 0.1% of photons (Sheikh-Bagheri et al 2000)
reaching the phantom surface, produced due to bremsstrahlung anywhere in the accelerator
head except in the target. Range rejection was turned off in the target to provide the most
accurate model for bremsstrahlung production. A variance reduction technique called selective
bremsstrahlung splitting (SBS) (Rogers et al 2000) was also used in the beam simulation to
increase the speed. The parameters used with SBS were: Nmin = 10, Nmax = 100. A Russian
roulette of secondary electrons is not employed. It is worth mentioning that simulations
were performed initially with both range rejection and SBS turned off. The results of the
simulations using both methods were the same within statistics. However, the simulation speed
with SBS turned on was much faster (almost two orders of magnitude) as shown in tables 1
and 2.
The CPU used for the simulation was a Pentium III with 1 GHz processors. For 6 MV
beams, a total of 5 × 108 and 2 × 107 electron histories were simulated for the 10 × 10 cm2 and
40 × 40 cm2 fields respectively. While at 18 MV, a total of 1 × 108 and 5 × 106 electron histories
are simulated for 10 × 10 cm2 and 40 × 40 cm2 fields respectively. The detailed benchmark data
of BEAM simulation when the bremsstrahlung splitting technique was off and on are shown in
1028 G X Ding

table 1 and 2 respectively. The dose calculation using the DOSXYZ code ran at 1.5 × 107 and
1.1 × 107 histories per h for 6 and 18 MV respectively. The dose calculation speed depends
on the requirement of voxel size and its statistical uncertainty, which was less than 1% in this
study. In this study the incident beam simulation took about 100–200 h for a 10 × 10 cm2
field and 10–20 h for a 40 × 40 cm2 field at 6 or 18 MV. The dose calculation using
DOSXYZ code took about 50–100 h for each field. The size of a phase space file was about
1–2 GB.
The starting incident electron energy and radial spread were both adjusted to produce the
best match between Monte Carlo calculated and measured dose distributions of the largest
field size for a given beam energy. These were the only two adjustable parameters in the
beam simulation. The final incident electrons had an energy distribution (which was taken
as a Gaussian with a FWHM of 1 MeV) and the distribution was centred at 6.02 MeV and
18.00 MeV for 6 MV and 18 MV beam respectively. The electron beam radial intensity
distribution was also taken as Gaussian with the FWHM of 1.2 mm and 1.5 mm for 6 MV and
18 MV beam respectively.
It is worth mentioning that dose profiles of large fields were very sensitive to the size of
the radial spread of incident electrons. While the central axis depth–dose curve only depended
on the incident electron energy, finding the best choice for the incident electron on target was
time consuming. Fortunately, it was only required to be done once.
When the Monte Carlo calculated dose distributions were in agreement with measured
dose distributions, the phase space files were then analysed to obtain the incident particles’
energy spectra, angular distributions, fluence profiles and mean energy profiles at the phantom
surface.

2.2. Experimental measurement

The photon beams were generated by a Varian Clinac-2100EX linear accelerator. All
measurements were made using a WELLHOFER WP700 scanner with an IC-10 ionization
chamber. The scanning system has a position accuracy of
0.5 mm and a reproducibility
of
0.2 mm for scanning. The IC-10 cylindrical chamber has an inner diameter of 6 mm.
The effective point of measurement for the ion chamber is taken to be 0.6rinner (where rinner is
the radius of the chamber cavity) upstream of the centre of the chamber, consistent with the
AAPM TG-51 protocol (Almond et al 1999) or IAEA code of practice (Andreo et al 2000).
The water phantom has a dimension of 48 × 48 × 48 cm3. Due to the limited width of
the phantom size, the scans of beam profiles at large depths (20 and 35 cm) for the largest
field size (40 × 40 cm2) are obtained by combining three partial overlapping scans. Two of
these partial overlapping scans are obtained by offseting the water phantom by 10 to 15 cm
in opposite directions. All measurements were made in water. A reference chamber,
which is mounted on the accelerator head, is used to correct beam output variations during
scanning.

2.3. Chamber fluence correction factors

Here a new concept of applying cylindrical chamber fluence correction to photon beams
was introduced. The measured depth–ionization curves were corrected for chamber fluence
correction factor, Pfl. In the electron beam for cylindrical chambers, values of Pfl were given
in the AAPM TG-21 protocol (TG-21 1983) as a function of chamber radius and mean energy
Ē z at the point of measurement. Strictly speaking the tabulated data in TG-21 were for the
point of measurement at dmax where most previous measurements were done (Almond et al
Energy spectra, angular spread, fluence profiles and dose distributions of photon beams 1029

1994, Johansson et al 1977, Van der Plaetsen et al 1994). However, if one assumes that the
values of Pfl at dmax in TG-21 still apply to another depth (Ding and Yu 2001, Huq et al
1997, Reft and Kuchnir 1999, Rogers 1998) away from dmax, the data of Pfl given in TG-21
(TG-21 1983) could be used to correct the component of depth ionization contributed by the
contaminant electrons. In this study the following polynomial fitted curve (Ding and Yu 2001)
was used
Pfl (R50 , z) = a + bĒ z + c(Ē z )2 − d(Ē z )3 + e(Ē z )4 (1)
where the polynomial fitting parameters for a chamber with diameter of 6 mm were: a =
0.95202, b = 1.3725 × 10−3, c = 3.7981 × 10−4, d = 2.8342 × 10−5, and e = 5.6462 ×
10−7, and Ē z was the calculated mean electron energy at depth z in water from the incident
contaminant electrons. Equation (1) was assumed to apply to positrons as well. Although this
was a rough approximation the end result was not sensitive because the ionization contributed
by contaminant positrons was much smaller compared to electrons.
The amount of actual corrections to the depth–ionization curves depended on the amount
of ionization contributed by the contaminant charged particles. This correction factor applied
to the component of the depth ionization that is contributed by the contaminant charged
particles. In the past, this correction has never been made because the exact amount of dose
contributed by the contaminant electrons is hard to determine without a full Monte Carlo
simulation.

2.4. Water-to-air stopping-power ratio corrections

Measured depth ionizations are converted to depth–dose by accounting for the Spencer–Attix
water-to-air stopping-power ratios, (L̄/ρ)water
air , which are calculated using EGSnrc (Kawrakow
2000) code SPRRZnrc (Rogers et al 2001). The variation of the stopping-power ratios with
depth is about 0.5% at 6 MV and 1% at 18 MV from surface to depth of 40 cm in water.
Calculated stopping-power ratios, (L̄/ρ)water
air , presented in figure 1 are generally in agreement
with calculations made by Andreo and Nahum (1985). The differences are due to the realistic
incident beams in the present study, which include field size dependent as well as contaminant
charged particles. The incident beams used in their calculations (Andreo and Nahum 1985)
were mono-energetic, plane parallel, broad photon beams.

2.5. Determination of the measured depth–dose curve from a measured depth–ionization

curve
Assuming that pdi(z) was a percentage-depth ionization curve measured using a cylindrical
chamber and that pcc(z) was the ionization component contributed by the charged particles,
the measured depth–dose curve (mdd) was determined by

water
L̄
mdd(z) = [pdi(z) − (1 − Pfl (z)) pcc(z)] (z) (2)
ρ air
where Pfl(z) was the chamber fluence correction factors as a function of depth expressed by
equation (1). The ionization component contributed by the charged particles (pcc) was obtained
from the Monte Carlo calculation. The Monte Carlo calculated dose component contributed
by the charged particles was used as the ionization component since the differences between
these two had negligible effect on the final corrections. The measured percentage depth–dose
(%dd) curve was obtained by re-normalizing the mdd curve.
1030 G X Ding

1.125

1.120 6 MV, 40x40

1.115

stopping power ratios

1.110

1.105 6 MV, 10x10

1.100

1.095
18 MV, 40x40
1.090

1.085

1.080
18 MV, 10x10
1.075
0 5 10 15 20 25 30 35 40

Depth in water /cm

Figure 1. Calculated water-to-air stopping-power ratios as a function of depth for realistic photon
beams from a Varian Clinac-2100EX accelerator calculated using EGSnrc code SPRRZnrc. The
SSD is 100 cm for the incident beams.

100
6 MV beam, 10x10 cm 2 field
90

80 Measured
Calculated (total)
70
Photons
%dd

60 Primary photons
Electrons
50

40

30

20

10

0
0 5 10 15 20 25 30 35 40
Depth in water/cm
(a)

Figure 2(a). Comparison of measured and Monte Carlo calculated central axis depth–dose curves
along with calculated components from incident photons, primary photons, electrons and positrons
for a 6 MV photon beam from a Varian Clinac-2100EX accelerator, field size = 10 × 10 cm2 and
SSD = 100 cm. The x- and y-axes are offset from zero for clarity.

3. Results and discussion

3.1. Comparison between measured and Monte Carlo calculated dose distributions

Comparisons of depth–dose curves are shown in figures 2(a)–(d) for 10 × 10 cm2 and 40 ×
40 cm2 fields at 6 and 18 MV respectively. The Monte Carlo calculation also breaks the depth–
dose curve into various components. The dose contribution from incident photons is referred
to as photons. Photons, which have not interacted with any part of the accelerator except at
the target, are referred to as primary photons. The contribution from incident contaminant
Energy spectra, angular spread, fluence profiles and dose distributions of photon beams 1031

100
6 MV beam, 40 x 40 cm 2 field
90
Measured
80
Calculated (total)
70 Photons
Primary photons
%dd 60
Electrons
50

40

30

20

10

0
0 5 10 15 20 25 30 35 40
Depth in water/cm
(b)

Figure 2(b). Comparison of measured and Monte Carlo calculated central axis depth–dose curves
along with dose contributions from incident photons, primary photons, electrons and positrons for
a 40 × 40 cm2 field at 6 MV (SSD = 100 cm).

100
18 MV, 10 x 10 cm 2 field
90
Measured
80 Calculated (total)
Photons
70 Primary photons
Electrons
60
Positrons
%dd

50

40

30

20

10

0
0 5 10 15 20 25 30 35 40
Depth in water/cm
(c)

Figure 2(c). Comparison of measured and Monte Carlo calculated central axis depth–dose curves
along with dose contributions from incident photons, primary photons, electrons and positrons for
a 10 × 10 cm2 field at 18 MV (SSD = 100 cm).

electrons is referred to as electrons. The contribution from incident contaminant positrons is

referred to as positrons.
All measured depth–dose curves were derived from the measured depth–ionization
curves by accounting for chamber fluence correction factors and stopping-power ratios using
equation (2). Each measured depth–dose curve was normalized at the depth of maximum dose
at the beam central axis. The Monte Carlo calculated depth–dose curve was scaled so that it
1032 G X Ding

100

90
18 MV, 40 x 40 cm2 field

80

70

%dd 60

50
Measured
40 Calculated (total)
Photons
30 Primary photons
Electrons
20
Positrons
10

0
0 5 10 15 20 25 30 35 40
Depth in water/cm
(d)

Figure 2(d). Comparison of measured and Monte Carlo calculated central axis depth–dose curves
along with dose contributions from incident photons, primary photons, electrons and positrons for
a 40 × 40 cm2 field at 18 MV (SSD = 100 cm).

had the same value at a depth 10 as that of the corresponding measured data. All profile data
were normalized and scaled in the same manner.
It can be seen that the charged particle contamination contributes significantly to the
surface dose. The increase of surface dose with the increase of the field size is mainly due
to the increase of these contaminant charged particles. At the largest field size the electron
contamination is almost tripled compared to the 10 × 10 cm2 field. At 6 MV the contaminant
charged particles contribute 7% and 21% of the maximum dose at the surface for 10 × 10 cm2
and 40 × 40 cm2 fields respectively. Beyond a few millimetres their contributions are
negligible. However, at 18 MV, their contributions are 4% and 6% even at dmax (=3 cm)
for 10 × 10 cm2 and 40 × 40 cm2 fields respectively. At the surface their contributions are
up to 11% and 29% of maximum dose for 10 × 10 cm2 and 40 × 40 cm2 fields as shown in
figures 2(c) and 2(d) respectively. The electron contamination is more than tripled compared
to the 10 × 10 cm2 field. It is interesting to see that the contaminant charged particles
contribute more dose than the incident photons at the surface as shown in figure 2(d). The
results here generally agree with the experimental results by Butson et al 1996a, 1996b, 1997a,
1997b.
Figures 2(e)–(f) present Monte Carlo calculated surface dose and contribution from
different particles to surface dose across the fields for 10 × 10 cm2 and 40 × 40 cm2 fields at
6 and 18 MV respectively. It is seen that the contaminant charged particles play a significant
role in making the surface dose flat across the field, particularly at 18 MV.
Figures 2(g)–(h) show the comparisons between measured and the Monte Carlo calculated
dose profiles and contributions from primary photons at various depths for the 40 × 40 cm2
field at 6 and 18 MV respectively. The calculated dose profiles match the measurements
at all depths. It is seen that the dose profiles from all particles and from primary photons
have a very similar shape beyond the depth where contaminant charged particles can reach.
The dose contribution from primary photons has almost the same value between a depth
of 3 and 5 cm at the central axis while their contribution differs between a depth of 3
Energy spectra, angular spread, fluence profiles and dose distributions of photon beams 1033

70 Surface dose 70
6 MV 40 x 40 field

60 60

relative dose
50 10x10 field 50

40 40
Total
Photons
30 Primary photons 30
Electrons
Positrons
20 20

10 10

0 0
-30 -20 -10 0 10 20 30

X-axis/cm
(e)

Figure 2(e). Monte Carlo calculated results showing the total surface dose and contribution from
different particles to surface dose across the field for 10 × 10 cm2 and 40 × 40 cm2 fields at 6 MV.
The calculated dose is the average dose between 0 and 0.25 cm depth in water.

50 50
Surface dose
18MV 40 x 40 field
Total
relative dose

40 Photons 40
Primary photons
Electrons
Positrons
30 30
10 x 10 field

20 20

10 10

0 0
-30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30

X-axis/cm
(f)

Figure 2(f). Monte Carlo calculated results showing the total surface dose and contribution from
different particles to surface dose across 10 × 10 cm2 and 40 × 40 cm2 fields at 18 MV. The
calculated dose is the average dose between 0 and 0.25 cm depth in water.

and 5 cm away from the central axis. This is caused by different incident photon spectra
between the central axis and away from the central axis. The mean energy of the incident
photons at phantom surface decreases away from the central axis. More details will be shown
later.
The agreements between Monte Carlo calculated and measured dose distributions are
excellent in all cases except for the 18 MV at the 40 × 40 cm2 field in the build-up region,
where the discrepancies are significant as seen in figure 2(d).
1034 G X Ding

110 110
6 MV
100 100
40 x 40 field
90 90

relative dose
80 80

70 70

60 Measured 60
Calculated (all)
Primary photons
50 50

40 40

30 30

20 20

10 10

0 0
-30 -20 -10 0 10 20 30

X-axis/cm
(g)

Figure 2(g). Comparison of measured and Monte Carlo calculated dose profiles and contribution
from primary photons for a 40 × 40 cm2 field at 6 MV, SSD = 100 cm. The profiles are at depths
of 1.5, 5, 10, 20 and 35 cm.

110 110
Primary photons
18 MV depth = 3 cm
100 100
40 x 40 field
relative dose

90 90

80 80

70 70

60 60

50 50

40 40

30 30

20 Measured 20
Calculated (all)
10 Primary photons 10

0 0
-30 -20 -10 0 10 20 30

X-axis/cm
(h)

Figure 2(h). Comparison of measured and Monte Carlo calculated dose profiles and contribution
from primary photons for the 40 × 40 cm2 field at 18 MV, SSD = 100 cm. The profiles are at
depths of 3, 5, 10, 20 and 35 cm.

In a recent publication Hartmann Siantar et al (2001) hypothesized that their similar

discrepancy was caused by a source of electrons in the accelerator head that was not fully
accounted for in the treatment head simulation. However, a new study (Ding 2001) has
demonstrated that these electrons are not responsible. This implies that something present
in the high-energy photon beams but not modelled in the EGS calculations is responsible,
possibly neutrons? d’Errico et al (1998) measured depth–dose curves of neutrons in the
photon beams, which are different from a photon’s depth–dose curves. Nath et al (1993) have
Energy spectra, angular spread, fluence profiles and dose distributions of photon beams 1035

1x10-5

counts/MeV/per incident e-
Photons 6 MV, 10 x 10 cm2 field
9x10-6

8x10-6
Primary photons
7x10-6

6x10-6
Electrons x 0.3 x 10-2
5x10-6

4x10-6
Positrons x 0.3 x10 -3
3x10-6

2x10-6

1x10-6

0
0 1 2 3 4 5 6 7

Energy/MeV
(a)

Figure 3(a). The energy spectra of photons, primary photons, electrons and positrons at the
phantom surface and inside the 10 × 10 cm2 field of a 6 MV beam from a Varian Clinac-2100EX
accelerator.

1.8x10-5
counts/MeV/per incident e-

6 MV, 40 x 40 cm2 field

1.6x10-5 Photons
1.4x10-5
Primary photons
1.2x10-5

1.0x10-5
Electrons x10-2
8.0x10-6

6.0x10-6 Positron x10-3

-6
4.0x10

2.0x10-6

0.0
0 1 2 3 4 5 6 7
Energy/MeV
(b)

Figure 3(b). The energy spectra of photons, primary photons, electrons and positrons at the
phantom surface and inside the 40 × 40 cm2 field of a 6 MV beam (SSD = 100 cm).

shown that the neutron dose in high-energy photon beams increases by more than a factor
of 2 for a field changing from the 10 × 10 to 30 × 30 cm2 field. However, the amount of
neutron dose in a high-energy photon beam is too small to explain the discrepancies. Further
investigation is needed to find the true cause of the discrepancy.

3.2. Energy spectra

Figures 3(a)–(d) present energy spectra of photons, primary photons, electrons and positrons
at the phantom surface for the 10 × 10 cm2 and 40 × 40 cm2 fields for the 6 and 18 MV
beams respectively. Only particles that are inside the field are counted. It is seen that the
majority of photons are primary photons, which have very similar spectra as all photons.
1036 G X Ding

1.4x10-5

counts/MeV/per incident e-
Photons
18 MV, 10 x 10 cm2 field
-5
1.2x10 Primary photons

1.0x10-5

8.0x10-6 Electrons x 10-2

6.0x10-6

Positrons x 10-2
4.0x10-6

2.0x10-6

0.0
0 5 10 15
Energy/MeV
(c)

Figure 3(c). The energy spectra of photons, primary photons, electrons and positrons at the
phantom surface and inside the 10 × 10 cm2 field of a 18 MV beam from a Varian Clinac-2100EX
accelerator CL2100EX (SSD = 100 cm).
counts/per MeV/per incident e-

3.0x10-5
Photons 18 MV, 40 x 40 cm 2 field
2.5x10-5
Primary photons
-5
2.0x10

Electrons x 10-2
1.5x10-5

1.0x10-5
Positrons x 10-2

5.0x10-6

0.0
0 5 10 15

Energy/MeV
(d)

Figure 3(d). The energy spectra of photons, primary photons, electrons and positrons at the
phantom surface and inside the 40 × 40 cm2 field of an 18 MV beam (SSD = 100 cm).

Energy spectra of incident photons peak at 0.5 MeV and 0.3 MeV for the 10 × 10 cm2 and
40 × 40 cm2 fields at 6 MV respectively, while at 18 MV, energy spectra of incident photons
peak at 1.5 MeV and 0.5 MeV for the 10 × 10 cm2 and 40 × 40 cm2 fields respectively.
The counts per unit energy and per incident electron history at the target for photons are 2–3
magnitudes higher than those of contaminant charged particles.
The spectra of contaminant electrons peak at low energy while the spectra of positrons
peak at the average energy.
Energy spectra, angular spread, fluence profiles and dose distributions of photon beams 1037

100
6 MV, 10 x 10 cm 2 field

relative counts /per degree

90 Photons
80

70 Electrons x 0.2 x 10-3

60

50 Positrons x 0.2 x 10-4

40

30

20

10

0
0 5 10 15 20 25 30
angle /degree
(a)

Figure 4(a). The angular distributions of photons, electrons and positrons at the phantom surface
and inside the 10 × 10 cm2 field of a 6 MV beam from a Varian Clinac-2100EX accelerator.

100
6 MV, 40x40 cm 2 field
relative counts /per degree

90 Photons
80

70

60 Electrons x 10 -3
50

40

30 Positrons x 10 -4

20

10

0
0 10 20 30 40 50 60
angle /degree
(b)

Figure 4(b). The angular distributions of photons, electrons and positrons at the phantom surface
and inside the 40 × 40 cm2 field of a 6 MV (SSD = 100 cm).

3.3. Angular distributions

Figures 4(a)–(d) present the angular distributions of photons, electrons and positrons at a
phantom surface for the 10 × 10 cm2 and 40 × 40 cm2 fields for the 6 and 18 MV beams
respectively. Only particles, which are inside the fields, are counted. The angular distributions
of incident photons are similar to a point source while the contaminant electrons show a wide
angular spread. The wide angular spread of contaminant charged particles in the 10 × 10 cm2
field at 6 MV reflects the fact that a lot of them are created or scattered in the air gap (about
43 cm) between the accelerator head and phantom surface.
1038 G X Ding

100
18 MV, 10x10 cm 2 field

relative counts /per degree

90 Photons
80

70 Electrons x 10-3
60

50

40

30 Positrons x 10-3
20

10

0
0 5 10 15 20
angle /degree
(c)

Figure 4(c). The angular distributions of photons, electrons and positrons at the phantom surface
and inside the 10 × 10 cm2 field of a 18 MV beam (SSD = 100 cm).
relative counts /per degree

100
Photons 18 MV, 40x40 cm 2 field
90

80

70

60

50
Electrons x 0.5 x 10-2
40

30

20 Positrons x 0.5 x 10-2

10

0
0 10 20 30 40 50 60
angle /degree
(d)

Figure 4(d). The angular distributions of photons, electrons and positrons at the phantom surface
and inside the 40 × 40 cm2 field of a 18 MV beam (SSD = 100 cm).

3.4. Fluence profiles

Figures 5(a)–(d) show particles’ fluence profiles of photons, primary photons, electrons and
positrons at the phantom surface for the 10 × 10 cm2 and 40 × 40 cm2 fields at 6 and 18 MV
beams respectively (SSD = 100 cm).
For a 10 × 10 cm2 field, the photon’s fluence remains relatively constant until a sharp
decrease at the field edge. For a 40 × 40 cm2 field, the photon’s fluence profile increases up
to 30% away from central axis. The primary photons dominate the photon fluence.
For contaminant electron fluence there is no sharp decrease outside the field. For a
10 × 10 cm2 field at 6 MV, it is interesting to see that the majority (70%) of the contaminant
Energy spectra, angular spread, fluence profiles and dose distributions of photon beams 1039

-2
2.5x10-5

planar fluence/incident e/cm

6 MV, 10 x 10 cm 2 field
Photons

-
2.0x10-5

1.5x10-5 Photons (primary only) Electrons x 0.2 x 10 -2

Electrons (created in air) x 0.2 x 10-2

1.0x10-5

Positrons x 0.2 x 10 -3
5.0x10-6

0.0
0 5 10 15

off-axis distance/cm
(a)

Figure 5(a). The planar fluence profiles of photons, primary photons, electrons and positrons at
the phantom surface for a 10 × 10 cm2 field at 6 MV from a Varian Clinac-2100EX accelerator
(SSD = 100 cm).

6 MV, 40 x 40 cm 2 field
2

Photons
planar fluence/incident e/cm

2.5x10 -5
-

2.0x10 -5
Primary photons

Electrons x 0.5 x 10 -2
1.5x10 -5

1.0x10 -5
-2
Electrons (created in air) x 0.5x 10

5.0x10 -6
Positrons x 10 -3

0.0
0 5 10 15 20 25 30
off-axis distance/cm
(b)

Figure 5(b). The planar fluence profiles of photons, primary photons, electrons and positrons at
the phantom surface of a 40 × 40 cm2 field at 6 MV (SSD = 100 cm).

electrons are created in the air gap (about 43 cm) between accelerator head and phantom
surface. However, for the 40 × 40 cm2 field at 6 MV, the majority of the contaminant electrons
(60%) are created in the accelerator head and only 40% are created in the air gap. At 18 MV,
the majority (>75%) of contaminant electrons are created in the accelerator head for field size
of 10 × 10 cm2 or 40 × 40 cm2.
At 6 MV, even at the field centre the fluence of the contaminant electrons is only about
0.1% and 0.4% of photon fluence for 10 × 10 cm2 and 40 × 40 cm2 fields respectively. For
positrons, their fluence is a magnitude smaller compared to the electrons.
At 18 MV near the central axis, the contaminant electron fluence is about 0.3% and 0.9%
of the photon’s fluence for the 10 × 10 cm2 and 40 × 40 cm2 fields respectively, while for
1040 G X Ding

18 MV, 10 x 10 cm 2 field

2
planar fluence/incident e/cm
8.0x10 -5 Photons

-
7.0x10 -5

6.0x10 -5
Primary photons Electrons x 0.5 x 10-2
5.0x10 -5

4.0x10 -5
Electrons (created in air) x 0.2 x 10-2
3.0x10 -5

2.0x10 -5
Positrons x 0.5 x 10-2

1.0x10 -5

0.0
0 5 10 15
off-axis distance/cm
(c)

Figure 5(c). The planar fluence profiles of photons, primary photons, electrons and positrons at
the phantom surface for a 10 × 10 cm2 field at 18 MV (SSD = 100 cm).

1.0x10 -4
18 MV, 40 x40 cm 2 field Photons
planar fluence/incident e-/cm 2

9.0x10 -5
Primary photons
8.0x10 -5

7.0x10 -5
Electronsx 10 -2
6.0x10 -5

5.0x10 -5

4.0x10 -5 Electrons (created in air) x 0.5x 10 -2

-5
3.0x10

2.0x10 -5

1.0x10 -5

Positrons x 10-2
0.0
0 5 10 15 20 25 30
off-axis distance/cm
(d)

Figure 5(d). The planar fluence profiles of photons, primary photons, electrons and positrons at
the phantom surface for a 40 × 40 cm2 field at 18 MV (SSD = 100 cm).

positrons it is only about 0.07% and 0.2% of photon fluence for 10 × 10 cm2 and 40 × 40 cm2
fields respectively. It is worth mentioning that this relatively small fluence of contaminant
charged particles contributes up to 29% of maximum dose at the surface shown in figure 2(f).

3.5. Mean energy distributions

Figures 6(a)–(d) present mean energy profiles of photons, electrons and positrons at a phantom
surface for the 10 × 10 cm2 and 40 × 40 cm2 fields at 6 and 18 MV beams respectively
(SSD = 100 cm).
At 6 (or 18) MV for the 10 × 10 cm2 field, the mean energy of incident photons is
relatively flat inside the field and decreases slightly towards the field edge until a sharp drop
Energy spectra, angular spread, fluence profiles and dose distributions of photon beams 1041

2.0
6 MV, 10 x10 cm 2 field
1.9

mean energy/MeV
Positrons
1.8
Photons
1.7

1.6

1.5

1.4
Electrons
1.3

1.2

0 5 10
off-axis distance/cm
(a)

Figure 6(a). The mean energies of incident photons, electrons and positrons at the phantom
surface are shown as a function of off-axis distance for a 10 × 10 cm2 field at 6 MV from a Varian
Clinac-2100EX (SSD = 100 cm).

2.0
1.9 6 MV, 40x40 cm 2 field
1.8 Positrons
mean energy/MeV

1.7
1.6
1.5
1.4
1.3
1.2
1.1 Photons Electrons
1.0
0.9
0.8
0.7
0 5 10 15 20 25 30
off-axis distance/cm
(b)

Figure 6(b). The mean energies of incident photons, electrons and positrons at the phantom surface
are shown as a function of off-axis distance for a 40 × 40 cm2 field at 6 MV (SSD = 100 cm).

at the field border. It is interesting to see that both mean incident energies are higher than
that of incident photons at 18 MV. It is worth mentioning that the rapid increase in mean
energy outside the treatment field for the 6 MV 10 × 10 cm2 field is real. It is similar
to the hardening effect of a wedged beam, where the photons that pass the wedge have
higher average energy. Here the photons, which have gone through the thick collimators,
have much higher mean energy. Figure 6(e) shows the mean energy profiles of primary
photons versus all photons as a function of off-axis distance for the 6 MV 10 × 10 cm2 field.
As the distance increases from the field edge, the portion of primary photons in the total
1042 G X Ding

7 18 MV, 10 x10 cm 2 field

mean energy/MeV
6
Positrons

Photons
3 Electrons

0 5 10
off-axis distance/cm
(c)

Figure 6(c). The mean energies of incident photons, electrons and positrons at the phantom surface
are shown as a function of off-axis distance for a 10 × 10 cm2 field at 18 MV (SSD = 100 cm).

Positrons 18 MV, 40x40 cm 2 field

mean energy /MeV

3
Electrons
2
Photons

0 5 10 15 20 25 30
off-axis distance/cm
(d)

Figure 6(d). The mean energies of incident photons, electrons and positrons at the phantom surface
are shown as a function of off-axis distance for a 40 × 40 cm2 field at 18 MV (SSD = 100 cm).

photon fluence increases rapidly. Therefore the average mean energy of all photons increases
rapidly.
For the 40 × 40 cm2 field at 6 MV, the mean energy of incident photons gradually
decreases away from the central axis from 1.7 MeV at the centre to 1.2 MeV at the field
edge; then there is a sharp drop outside the field. While at 18 MV, the mean energy of
photons decreases away from central axis from 4.5 MeV at the centre to 3 MeV at the field
edge. The average photon’s energy inside the field is less than 1/3 of the nominal beam
energy.
Energy spectra, angular spread, fluence profiles and dose distributions of photon beams 1043

3.5

6 MV, 10 x10 cm 2 field

mean energy/MeV
3.0

2.5 Primary photons

2.0

Photons
1.5

0 5 10
off-axis distance/cm
(e)

Figure 6(e). The mean energy of all photons versus primary photons at the phantom surface as
a function of off-axis distance for a 10 × 10 cm2 field at 6 MV from a Varian Clinac-2100EX
(SSD = 100 cm).

The distribution of mean energy of incident contaminant electrons or positrons shows a

slow decrease away from central axis. The mean energy of incident positrons is generally
higher than that of electrons or photons.

4. Conclusions

This investigation presents incident particles’ energy spectra, angular distributions, fluence
profiles and mean energy profiles at the phantom surface. Because of the extensive use of
the LATCH feature (Rogers et al 1995), which allows us to separate particles according to
their history, detailed information is obtained for realistic incident particles including photons,
primary photons, contaminant electrons and positrons, and electrons created in the air between
the accelerator head and the phantom surface. The data presented here are for 6 and 18 MV
radiotherapy photon beams from a Varian Clinac-2100EX accelerator at field sizes of 10 ×
10 cm2 and 40 × 40 cm2. The characteristics of the incident photon beams presented are based
on the Monte Carlo simulation using EGS4 user code BEAM. This information enhances our
knowledge on many aspects of radiotherapy photon beams.
The study has presented the calculated depth–dose components from different particles
as well as calculated surface dose and contribution from different particles to surface dose
across the field. It is shown that the majority of the incident photons are primary photons. The
shape of the depth–dose curves and dose profiles is dominated by primary photons beyond
the build-up depth. The primary photons can be easily modelled by a point source. The
information of beam characteristics is very useful to the future development of an accurate
treatment planning system. It is also useful to improve photon beam dosimetry and to design
new accelerators.
The calculated and measured dose distributions in water phantoms are in remarkable
agreement in all cases except for 18 MV at a very large field in the dose build-up region.
The unresolved discrepancy may indicate the limitation in the EGS modelling as well as the
limitation of measurements at the build-up region in the high-energy photon beams. In any
case the real cause of the discrepancy needs further investigation.
1044 G X Ding

The calculated water-to-air stopping-power ratios of realistic photon beams show

negligible variation as a function of depth. The variation in depth is less than 0.5% between
0.5 cm and 40 cm in water. At 18 MV there is about 1% increase in stopping-power ratios
between the depth of 0.1 cm and 0.5 cm for a 40 × 40 cm2 field.
It must be emphasized that the newly introduced chamber fluence corrections strongly
depend on the amount of the ionization contributed by the contaminant charged particles
in the measured depth–ionization curve. This correction is negligible when the component
contributed by the contaminant charged particles is small. The correction varies with depth
because both the electron fluence correction factor and the ionization component from the
contaminant charged particles vary with depth. The effect of this correction is about 1%
reduction to the measured depth–ionization curve at the build-up region for the largest open
field at 18 MV. The effect of the correction reduces the values of the measured depth–
ionization curve in the build-up region, which improves the agreement between the Monte
Carlo calculation and the measurement slightly.
In the build-up region, the doses contributed by contaminant charged particles are
significant and increase with increase in field size. It is shown that increase of surface
dose with increase of the field size is mainly due to increase of incident contaminant charged
particles. At 6 MV, the incident charged particles contribute 7% to 21% of maximum dose
at the surface when the field size increases from 10 × 10 to 40 × 40 cm2. At 18 MV, their
contributions are up to 11% and 29% of the maximum dose at the surface for 10 × 10 cm2
and 40 × 40 cm2 fields respectively.
The majority of contaminant charged particles originate from the accelerator head. The
number of contaminant electrons generated in the air gap between accelerator head and
phantom surface (about 43 cm for SSD = 100 cm) are also significant and vary with incident
beam energy and field size. It has been shown from the Monte Carlo calculated surface
dose that unlike doses from photons, the doses from contaminant charged particles do not
reduce sharply outside the treatment field. A clinical implication is that a shielding just above
the patient skin may be necessary to avoid unwanted radiation to a critical organ even when
the organ is outside the treatment field. A 1 or 2 mm thick lead sheet is adequate to shield
these contaminant electrons.
To simulate an incident beam takes 10–200 h CPU times on a 1 GHz processor depending
on the field size simulated. This is done with the Monte Carlo variance reduction technique
(SBS) (Rogers et al 2000, Sheikh-Bagheri et al 2000). Without using SBS, it will be about
100 times slower. To calculate the dose distribution takes 50–100 h in this study. Although
the calculation times are for this study, it gives an indication of the feasibility of using full
Monte Carlo simulation for treatment planning on a current computer. This also indicates the
real need to develop new techniques to speed up the Monte Carlo simulation.

Acknowledgments

The author gratefully acknowledges Varian Medical Systems for providing the information on
the Varian Clinac linear accelerators. The author also wishes to thank Dr Cheryl Duzenli for
her support in the project.

References

Almond P R, Attix F H, Humphries L J, Kubo H, Nath R, Goetsch S and Rogers D W 1994 The calibration and use of
plane-parallel ionization chambers for dosimetry of electron beams: an extension of the 1983 AAPM protocol
report of AAPM Radiation Therapy Committee Task Group No 39 Med. Phys. 21 1251–60
Energy spectra, angular spread, fluence profiles and dose distributions of photon beams 1045

Almond P R, Biggs P J, Coursey B M, Hanson W F, Huq M S, Nath R and Rogers D W 1999 AAPM’s TG-51 protocol
for clinical reference dosimetry of high-energy photon and electron beams Med. Phys. 26 1847–70
Andreo P, Burns D T, Hohlfeld K, Huq M S, Kanai T, Laitano F, Smyth V G and Vynckier S 2000 Absorbed dose
determination in external beam radiotherapy: an international code of practice for dosimetry based on standards
for absorbed dose to water Volume 398 of Technical Report Series (IAEA, Vienna: International Atomic Energy
Agency)
Andreo P and Nahum A E 1985 Stopping-power ratio for a photon spectrum as a weighted sum of the values for
monoenergetic photon beams Phys. Med. Biol. 30 1055–65
Butson M J, Mathur J N and Metcalfe P E 1997a Skin dose from radiotherapy x-ray beams: the influence of energy
Australas. Radiol. 41 148–50
Butson M J, Rozenfeld A, Mathur J N, Carolan M, Wong T P and Metcalfe P E 1996a A new radiotherapy surface
dose detector: the MOSFET Med. Phys. 23 655–8
Butson M J, Wong T P, Law A, Law M, Mathur J N and Metcalfe P E 1996b Magnetic repulsion of linear accelerator
contaminates Med. Phys. 23 953–5
Butson M J, Yu P, Kan M, Carolan M, Young E, Mathur J N and Metcalfe P E 1997b Skin dose reduction by a
clinically viable magnetic deflector Australas. Phys. Eng. Sci. Med. 20 107–11
Deng J, Jiang S B, Kapur A, Li J, Pawlicki T and Ma C M 2000 Photon beam characterization and modelling for
Monte Carlo treatment planning Phys. Med. Biol. 45 411–27
d’Errico F, Nath R, Tana L, Curzio G and Alberts W G 1998 In-phantom dosimetry and spectrometry of photoneutrons
from an 18 MV linear accelerator Med. Phys. 25 1717–24
Ding G X 1995 An investigation of radiotherapy electron beams using Monte Carlo techniques PhD Thesis Department
of Physics, Carleton University, Ottawa Carleton Institute for Physics, Ottawa, Canada
Ding G X 2001 Dose discrepancies between the Monte Carlo calculation and measurement for high-energy photon
beam at buildup region Int. Workshop on Recent Developments in Accurate Radiation Dosimetry (Montreal,
Canada, 10–13 Oct. 2001) (abstract)
Ding G X and Rogers D W 1995 Energy spectra, angular spread, and dose distributions of electron beams from
various accelerators used in radiotherapy National Research Council of Canada, Report No PIRS-0439, Ottawa
(http://www.irs.inms.nrc.ca/inms/irs/papers/PIRS439/pirs439.html)
Ding G X, Rogers D W, Cygler J E and Mackie T R 1997 Electron fluence correction factors for conversion of dose
in plastic to dose in water Med. Phys. 24 161–76
Ding G X, Rogers D W and Mackie T R 1995 Calculation of stopping-power ratios using realistic clinical electron
beams Med. Phys. 22 489–501
Ding G X, Rogers D W and Mackie T R 1996 Mean energy, energy-range relationships and depth-scaling factors for
clinical electron beams Med. Phys. 23 361–76
Ding G X and Yu C W 2001 Determination of percentage depth–dose curves for electron beams using different types
of detectors Med. Phys. 28 298–302
Faddegon B A, O’Brien P and Mason D L 1999 The flatness of Siemens linear accelerator x-ray fields Med. Phys. 26
220–8
Hartmann Siantar C L et al 2001 Description and dosimetric verification of the PEREGRINE Monte Carlo dose
calculation system for photon beams incident on a water phantom Med. Phys. 28 1322–37
Huang P H, Kase K R and Bjarngard B E 1982 Simulation studies of 4 MV x-ray spectral reconstruction by numerical
analysis of transmission data Med. Phys. 9 695–702
Huang P H, Kase K R and Bjarngard B E 1983 Reconstruction of 4 MV bremsstrahlung spectra from measured
transmission data Med. Phys. 10 778–85
Huq M S, Yue N and Suntharalingam N 1997 Experimental determination of fluence correction factors at depths
beyond dmax for a Farmer type cylindrical ionization chamber in clinical electron beams Med. Phys. 24 1609–13
Jessen K A 1973 Measurements of primary spectra from a kilocurie 60Co unit and a 6 MeV linear accelerator Acta
Radiol. Ther. Phys. Biol. 12 561–8
Johansson K A, Mattsson L O, Lindborg L and Svensson H 1977 Absorbed-dose determination with ionization
chambers in electron and photon beams having energies between 1 and 50 MeV IAEA Symp. Proc. (Vienna:
IAEA-SM-222/35) pp 243–70
Kawrakow I 2000 Accurate condensed history Monte Carlo simulation of electron transport: II. Application to ion
chamber response simulations Med. Phys. 27 499–513
Keall P J, Siebers J V, Jeraj R and Mohan R 2000 The effect of dose calculation uncertainty on the evaluation of
radiotherapy plans Med. Phys. 27 478–84
Levy L B, Waggener R G, McDavid W D and Payne W H 1974 Experimental and calculated bremsstrahlung spectra
from a 25 MeV linear accelerator and a 19 MeV betatron Med. Phys. 1 62–7
Levy L B, Waggener R G and Wright A E 1976 Measurement of primary bremsstrahlung spectrum from an 8 MeV
linear accelerator Med. Phys. 3 173–5
1046 G X Ding

Liu H H, Mackie T R and McCullough E C 1997a Calculating dose and output factors for wedged photon radiotherapy
fields using a convolution/superposition method Med. Phys. 24 1714–28
Liu H H, Mackie T R and McCullough E C 1997b Calculating output factors for photon beam radiotherapy using a
convolution/superposition method based on a dual source photon beam model Med. Phys. 24 1975–85
Ma C M, Faddegon B A, Rogers D W and Mackie T R 1997 Accurate characterization of Monte Carlo calculated
electron beams for radiotherapy Med. Phys. 24 401–16
Mohan R, Chui C and Lidofsky L 1985 Energy and angular distributions of photons from medical linear accelerators
Med. Phys. 12 592–7
Mora G M, Maio A and Rogers D W 1999 Monte Carlo simulation of a typical 60Co therapy source Med. Phys. 26
2494–502
Nath R, Meigooni A S, King C R, Smolen S and d’Errico F 1993 Superheated drop detector for determination of
neutron dose equivalent to patients undergoing high-energy x-ray and electron radiotherapy Med. Phys. 20
781–7
Nath R and Schulz R J 1976 Determination of high-energy x-ray spectra by photoactivation Med. Phys. 3 133–41
Reft C S and Kuchnir F T 1999 Measured overall perturbation factors at depths greater than dmax for ionization
chambers in electron beams Med. Phys. 26 208–13
Rogers D W 1998 A new approach to electron-beam reference dosimetry Med. Phys. 25 310–20
Rogers D W, Faddegon B A, Ding G X, Ma C M, We J and Mackie T R 1995 BEAM: a Monte Carlo code to simulate
radiotherapy treatment units Med. Phys. 22 503–24
Rogers D W, Kawrakow I, Seuntjens J P and Walters B R 2001 NRC user codes for EGSnrc National Research
Council of Canada Report No PIRS-702
Rogers D W, Ma C, Ding G X, Walters B R, Sheikh-Bagheri D and Zhang G G 2000 BEAM users manual National
Research Council of Canada, Ottawa Report PIRS-0509(a)revE
Schach von Wittenau A E, Bergstrom P M Jr and Cox L J 2000 Patient-dependent beam-modifier physics in Monte
Carlo photon dose calculations Med. Phys. 27 935–47
Schach von Wittenau A E, Cox L J, Bergstrom P M, Jr Chandler W P, Hartmann Siantar C L and Mohan R 1999
Correlated histogram representation of Monte Carlo derived medical accelerator photon-output phase space
Med. Phys. 26 1196–211
Sheikh-Bagheri D, Rogers D W, Ross C K and Seuntjens J P 2000 Comparison of measured and Monte Carlo
calculated dose distributions from the NRC linac (in process citation) Med. Phys. 27 2256–66
Siebers J V, Keall P J, Libby B and Mohan R 1999 Comparison of EGS4 and MCNP4b Monte Carlo codes for
generation of photon phase space distributions for a Varian 2100C Phys. Med. Biol. 44 3009–26
Swindell W 1983 A 4 MV CT scanner for radiation therapy: spectral properties of the therapy beam Med. Phys. 10
347–51
TG-21 1983 A protocol for the determination of absorbed dose from high-energy photon and electron beams
Med. Phys. 10 741–71
Van der Plaetsen A, Seuntjens J, Thierens H and Vynckier S 1994 Verification of absorbed doses determined with
thimble and parallel-plate ionization chambers in clinical electron beams using ferrous sulphate dosimetry Med.
Phys. 21 37–44
van der Zee W and Welleweerd J 1999 Calculating photon beam characteristics with Monte Carlo techniques Med.
Phys. 26 1883–92
Zhang G G, Rogers D W, Cygler J E and Mackie T R 1998 Effects of changes in stopping-power ratios with field
size on electron beam relative output factors Med. Phys. 25 1711–6
Zhang G G, Rogers D W, Cygler J E and Mackie T R 1999 Monte Carlo investigation of electron beam output factors
versus size of square cutout Med. Phys. 26 743–50