The local enhancement of radiation dose from photons of MeV energies
obtained by introducing materials of high atomic number into the
treatment region
Ahmad Alkhatib, Yoichi Watanabe,a兲 and John H. Broadhurstb兲
Department of Therapeutic Radiology, University of Minnesota, Minneapolis, Minneapolis 55455
共Received 8 July 2008; revised 20 May 2009; accepted for publication 15 June 2009;
published 2 July 2009兲
With the advent of therapeutic radiation treatment machines with photon end point energies of
several MeV, a new channel is available to transfer the photon energy to biological material,
namely, pair production. This process has a photon threshold energy of 1.02 MeV. The probability
of pair production, which depends on the square of the atomic number 共Z兲 of the interacting
material, increases markedly as the photon energy is further increased. As the goal of treatment
planning in radiation therapy is to locally maximize the absorbed dose in abnormal cells and
minimize the dose in surrounding normal cells, in this study the authors measured the dose enhancement which could be expected if a high-Z material such as gold was present adjacent to tumor
sites during irradiation. The authors used photon beams produced by electron accelerators with
energies ranging from 6 to 25 MV. They chose either gold or lead foils as high-Z materials, the
measurements being repeated using the same geometry but replacing the high-Z materials with a
low-Z material 共aluminum兲. The comparison of the experimental results using low- and high-Z
materials verified the theoretical prediction of the expected dose enhancement. The effect of finite
range of the electron-positron pairs was also studied by varying the spacing between two foils
placed parallel or orthogonal to the incident photon beam. Using an 18 MV photon beam, the
authors observed a maximum dose enhancement of 44%. They intend therefore to proceed from
these phantom studies to animal measurements. © 2009 American Association of Physicists in
Medicine. 关DOI: 10.1118/1.3168556兴
Key words: pair production, MV photon beam, dose enhancement, high-Z
I. INTRODUCTION
The goal of radiation therapy planning is to maximize the
absorbed dose in abnormal or malignant cells and minimize
it elsewhere. Our aim was to increase the absorbed dose by
inserting a material with a high atomic number 共Z兲 locally in
the region of the malignant cells. In vivo this would probably
be achieved by injecting submicron gold particles. The injection of gold particles to enhance dose has already been
proven to be viable using an animal model together with low
energy photon beams 共250 kVp兲;1,2 however, with orthovoltage x rays the dose enhancement was due mainly to photoelectric interactions. At higher photon energies, the primary
modes of energy transfer to the biological material are
Compton scattering and pair production. As the photon energy is further increased 共above 5 MeV兲 and if high-Z materials 共such as gold兲 are present, pair production becomes
dominant 共as seen in Fig. 1兲. By using foils of high-Z material, this study examines the effects of geometry 共the thickness of the foils, the gap width between the foils, and the
orientation of the foils relative to the incident photon beam
direction兲 together with the effects of both the photon beam
energy and the atomic number of the foils on the local enhancement of absorbed dose.
II. THEORY
For megavoltage photon beams, the two competing processes for transferring energy from incident photons to tissue
3543
Med. Phys. 36 „8…, August 2009
are Compton scattering and pair production. These processes
can produce high energy electrons but with different probabilities. Compton scattering is photon-atomic electron interaction, which therefore is proportional to Z, whereas pair
production occurs primarily by interaction with the atomic
nuclei and is proportional to Z2.
Assuming that the incoming photon has much higher energy than the threshold energy for pair production 共i.e., 1.02
MeV兲, the electron-positron pair produced will have large
kinetic energy and will initially interact with the surrounding
tissue as relativistic particles, i.e., their linear energy transfer
will be, to first order, independent of their kinetic energy.
Unlike Compton processes, a pair production process produces two charged particles originating from the same location and no scattered photon, so the local linear energy transfer will be considerably larger than the energy transferred by
Compton interaction. Furthermore, the electrons and positrons are emitted in the forward direction with respect to the
direction of the incident photon beam. Therefore, the increase in linear energy transfer will be concentrated downstream of the pair production event. As the particles lose
energy and become nonrelativistic, the normal pattern of
electron-electron and electron-nuclear scatterings will take
place; however, overall there is still the doubling of the linear
energy transfer due to the presence of two particles in the
same region. Once the positron has reached epithermal ener-
0094-2405/2009/36„8…/3543/6/$25.00
© 2009 Am. Assoc. Phys. Med.
3543
3544
Alkhatib, Watanabe, and Broadhurst: Dose enhancement by high-Z material for photon radiotherapy
3544
FIG. 1. Pair production probabilities for water and gold as a function of
photon energy. The figure was drawn based on the photon interaction data
provided by Hubbell 共Ref. 4兲.
gies, it will interact with an atomic electron, creating two
gamma rays each of 0.511 MeV energy. These photons are
unlikely to interact locally, and therefore will contribute little
to the local dose.
Figure 2 is the photon energy spectrum of a 25 MV photon beam generated by an SL25 linear accelerator 共Elekta,
Stockholm, Sweden兲.3 We calculated the relative numbers of
electron-positron pairs with varying total kinetic energy for
water and gold by convolving the pair production
probability4 with the photon energy spectrum given in Fig. 2,
the results being shown in Fig. 3. Since the area under the
gold curve is 84 times that of water, the total number of
electron-positron pairs produced for the 25 MV photon beam
is 84-fold larger when gold replaces water. In addition, the
curves show that the electron-positron pairs from the high-Z
FIG. 2. Photon energy spectrum of a 25 MV photon beam from an Elekta
SL25 linear accelerator.
Medical Physics, Vol. 36, No. 8, August 2009
FIG. 3. Relative number of electron-positron pairs for water and gold as a
function of the total kinetic energy of the electron-positron pair. Note that
the data for water are multiplied by 10 for presentation purpose.
material 共gold兲 will on average have a somewhat lower total
kinetic energy than those pairs produced from water 共i.e., 8.6
MeV vs 9.0 MeV兲. Note that the total kinetic energy of an
electron-positron pair from pair production is divided randomly between the electron and the positron. This implies
that the electrons and positrons deposit their energy in a
smaller volume surrounding the high-Z material than electrons produced by both pair production and Compton events
in water because with lower energy these particles will have
a shorter range.
We computed the kinetic energy carried by secondary
electrons 共and positrons兲 produced through photoelectric interactions, Compton scattering processes, and pair productions when photon beams produced by various accelerator
energies, i.e., 6, 10, 18, and 25 MV, interact with gold atoms.
For simplicity we made the following three assumptions.
First the energy of photoelectrons is the same as the incident
photon energy. Second the energy of the Compton electrons
is the maximum possible energy of electrons emitted when
photons are scattered backward with respect to the initial
photon flight direction. Third the total energy of an electron
and a positron is the incident photon energy minus twice the
electron mass. We used the published photon energy spectra3
and the photon interaction data available at NIST.4 Table I
shows the fractions of energy carried by secondary electrons
共and positrons兲 from three different photon interaction processes. When the beam energy increases above 18 MV, the
energy carried by electron-position pairs exceeds the energies of photoelectrons and Compton electrons. Hence, one
can conclude that secondary charged particles 共i.e., electrons/
3545
Alkhatib, Watanabe, and Broadhurst: Dose enhancement by high-Z material for photon radiotherapy
3545
TABLE I. The fraction of energies carried by electrons from photoelectric,
Compton, and pair production interactions for photon beams with various
energies. The values are percentage ratios of the electron energy transferred
from the incident photon and the initial photon energy.
Photon beam energy Photoelectric Compton scattering Pair production
共MV兲
共%兲
共%兲
共%兲
6
10
18
25
32.1
18.5
8.3
6.2
58.3
54.4
42.7
37.1
9.5
27.1
49.1
56.7
positrons兲 are produced mainly by pair production for high
energy photon beams such as 18 and 25 MV beams.
III. METHODS AND MATERIALS
All measurements were made using phantoms assembled
from acrylic slabs, between which both high-Z materials and
dose measurement devices were inserted. For this work, the
high-Z materials were gold 共Z = 79兲 or lead 共Z = 82兲 foils of
different thicknesses. Note that gold is desirable for biological applications; however, we also used lead foils because of
cost considerations. The thickness of 1 ⫻ 1 cm2 square foils
varied from 0.3 to 1.0 mm. To further verify that the dose
enhancement was due to pair production in the lead or gold
foils, the experimental measurements were repeated using
the same geometry but with aluminum foils 共Z = 13兲.
The two-dimensional 共2D兲 distributions of absorbed dose
were obtained using Kodak EDR2 films 共Eastman Kodak
Co., Rochester, NY兲 placed between the sheets of the acrylic
phantoms. The film analyses were performed using the
RIT113 software version 4 共RIT Inc. Colorado Springs, CO兲.
As there was some concern about the unavoidable introduction of a higher-Z material in the silver halide of the
Kodak films, spot measurements were also made using
tissue equivalent thermoluminescent dosimeters 共TLDs兲
共3 ⫻ 3 mm2 square size Harshaw TLD-100 chips from
Thermo Scientific, Franklin, MA兲.
The irradiating photon beams were provided by Varian
2300CD 共Varian Medical Systems, Palo Alto, CA兲 and Elekta Synergy linear accelerators 共Elekta, Stockholm, Sweden兲.
The phantom assembly was placed at 100 cm source-tosurface distance 共SSD兲. To mimic normal therapeutic treatment conditions, a dose of 200 cGy at the depth of dose
maximum, dmax, was delivered at a dose rate of 400 MU/min
over a 10⫻ 10 cm2 field size.
Two different geometries were used to test the thesis of
dose enhancement, which takes advantage of increased pair
production in high-Z materials with high energy photon
beams. In all measurements the enhancement volume was
chosen to be in a region downstream of the maximum dose,
or dmax. In this chosen region many of the lower energy
photons would already have been absorbed, thus making an
enhancement mainly due to pair production. The first geometry 共called perpendicular configuration兲 was to place a foil
共or the first foil兲 upstream over the volume, in which enhancement was expected, with the foil surface orthogonal to
Medical Physics, Vol. 36, No. 8, August 2009
FIG. 4. Illustration of geometrical arrangement for foil surfaces placed orthogonal to the photon beam direction 共perpendicular configuration兲. The
distance between the 1 ⫻ 1 cm2 foils was varied from 3 to 9 mm. The foil at
upstream was placed at 5 cm depth in solid phantom.
the beam direction. A second foil was placed parallel to the
first foil below the volume as shown in Fig. 4. The purpose
of the second foil was to backscatter electrons and positrons
which had passed through the enhancement volume without
depositing most of their kinetic energy. A single foil configuration was considered before through experimental measurements and Monte Carlo simulations.5–7 Our initial measurements indicated that, without the second foil, no observable
dose enhancement would be obtained. Furthermore, this
single foil arrangement does not mimic an attainable configuration for obtaining dose enhancement in clinical settings. Thus, the single foil geometry will not be further discussed in this article. A second geometry 共called parallel
configuration兲 as shown in Fig. 5 more closely simulated a
high-Z elemental distribution that could be realized in patient
treatment. In this case foils were placed parallel to the beam
axis on two sides of the enhancement volume.
To confirm our measurements 共of the perpendicular con-
FIG. 5. Illustration of geometrical arrangement for foil surfaces placed parallel to the photon beam direction 共parallel configuration兲. 1 ⫻ 1 cm2 square
foils were placed in solid phantom. The depth of the foils was 13.5 cm. The
interfoil spacing between the foils was variable 共less than 10 mm兲.
3546
Alkhatib, Watanabe, and Broadhurst: Dose enhancement by high-Z material for photon radiotherapy
FIG. 6. Absorbed dose comparison between gold foils and aluminum foils.
A 25 MV photon beam from a Varian 2300CD was used. The interfoil
spacing between two foils was 4 mm. Figure 4 shows an illustration of the
geometry 共perpendicular configuration兲 for this measurement.
figuration兲, we did Monte Carlo simulations using the MCNP
code.8 For simplicity we approximated the geometry using a cylindrical model with the incident beam being along
the axis of the cylinder. We used the energy spectrum of the
25 MV photon beam as shown in Fig. 2. A 10 cm diameter
photon beam entered at the top surface of a 40 cm diameter
cylindrical volume filled with water. Gold foils were modeled as thin disks of 1 cm diameter. The absorbed dose in the
region between the two foils was calculated on a 1 cm diameter surface. The contributions by electrons 共or positrons兲
produced in the gold foils upstream and downstream were
differentiated by using the Cell Flagging Card 共CFn card兲
available with the MCNP code. This option enabled us to trace
back the locations where the secondary electrons 共or positrons兲 passed through before contributing to the dose at the
dose calculation points.
3546
FIG. 7. Photon energy dependence of dose enhancement. 6, 10, and 18 MV
photon beams from an Elekta Synergy were used. Lead foils 共1 mm thickness at the upstream position and 0.5 mm thickness at the downstream
position兲 were used. The geometry was the same as that used for Fig. 6.
4C
IV. RESULTS
The first measurements were performed using the geometry of Fig. 4 共perpendicular configuration兲. A 0.9 mm thick
gold foil 共or the first foil兲 was used for the upstream position
at a depth of 5 cm. The thickness of the second gold foil at
the downstream position was 0.3 mm. The distance between
the two parallel foils was 4.0 mm. A 25 MV photon beam
from a Varian 2300CD linear accelerator was used. An additional measurement was made under the same condition using aluminum foils 共0.9 mm thick foil for the upstream position and 0.3 mm thick foil for the downstream position兲 to
verify the expected Z dependence of the dose enhancement.
Figure 6 shows the dose enhancement obtained from the use
of gold foils and the lack of measurable enhancement obtained when gold was replaced by aluminum in the same
geometry. The amount of dose enhancement we obtained
was 20% for the gold foils.
To eliminate the possibility of dose enhancement as a result of the high-Z component in the Kodak EDR2 films, we
measured the doses at the center of the dose enhancement
region between the foils and at 2.5 cm off the centerline
using TLD chips with the same configuration as that used for
Medical Physics, Vol. 36, No. 8, August 2009
the result shown in Fig. 6. The ratio of the measured doses at
the center and off-centerline positions was 1.24, or a 24%
dose enhancement downstream from the first gold foil. The
similar dose enhancement obtained using TLD chips confirms the validity of the dose enhancement observed with the
film-based measurements.
Figure 7 presents the energy dependence of the dose enhancement for the perpendicular configuration depicted in
Fig. 4 using two lead foils of 1 and 0.5 mm thicknesses. The
depth of the first lead foil was set to be 5 cm below the
phantom surface. The three different photon energy beams
used 共6, 10, and 18 MV兲 were from an Elekta Synergy accelerator. The figure shows that the dose downstream of the
first foil 共1 mm thick lead兲 on the beam axis in a volume
between the foils decreased for 6 and 10 MV photon beams,
whereas the result of the 18 MV photon beam indicates a
dose enhancement occurring both near the beam axis and
regions along the foil edges.
Figure 8 shows the effect of changing the interfoil gap
FIG. 8. Absorbed dose comparison between two gold foils with different
separations 共4 and 7 mm兲 using an 18 MV photon beam from an Elekta
Synergy. The perpendicular configuration as depicted in Fig. 4 was used.
3547
Alkhatib, Watanabe, and Broadhurst: Dose enhancement by high-Z material for photon radiotherapy
3547
FIG. 9. 2D dose distribution taken with a radiographic film, which was placed at the middle of two 1 mm thick lead foils separated by 4 mm in solid phantom
共parallel configuration as depicted in Fig. 5兲. An 18 MV photon beam from an Elekta Synergy was used.
width between two gold foils on the dose enhancement for
an 18 MV photon beam. The dose enhancement increased
with the decreasing distance between the two foils; a dose
enhancement of 37% being observed with 4 mm spacing.
The physical mechanisms of the energy and the interfoil gap
width dependence of the dose enhancement are discussed in
Sec. V.
Dose was also measured using the parallel configuration,
in which foils were arranged parallel to the beam direction as
illustrated in Fig. 5. Two 1 mm thick and 1 ⫻ 1 cm2 square
lead foils were placed at a depth of 13.5 cm from the phantom surface with an interfoil spacing between the foils of 4
mm. An EDR2 film was placed centrally between the two
foils oriented parallel to the 18 MV photon beam direction. A
color wash plot of the 2D dose distribution shown in Fig. 9
clearly demonstrates the dose enhancement obtained in a
volume between the foils. Figure 10 shows the dose profile
along the transverse plane taken across the enhancement re-
FIG. 10. Dose profile plotted along the transverse direction through the dose
enhancement region at 13.5 cm depth, or along line A-A⬘ as shown in Fig. 9.
Medical Physics, Vol. 36, No. 8, August 2009
gion shown in Fig. 9. Note that in this case the enhancement
was 44% and the peak dose in the enhancement area exceeded the dose at dmax.
V. DISCUSSION
The energy dependence of dose enhancement seen in Fig.
7 is due to the energy dependence of three physical processes
共namely, photoelectric, Compton, and pair production兲 in
high-Z materials. The dose in the downstream region
strongly depends on the secondary electron transport processes inside and near the high-Z material.9 For low photon
energies 共i.e., a 6 MV photon beam兲 the high-Z material
causes increases in both the photoelectric interactions and
Compton scattering. Photoelectrons have a low energy and
do not travel very far out from the high-Z material. Most
Compton electrons are ejected in the directions away from
the incident photon direction. Hence the dose in the region
downstream to the first foil actually decreases. As shown in
the Sec. II, for higher photon energies 共i.e., an 18 MV photon
beam兲 the pair production process in high-Z material efficiently converts the photon energy to the energy of secondary electrons/positrons 共an average energy of about 4.3
MeV兲, which travel a short distance 共about 2 MeV/1 cm兲
before transferring all their kinetic energy in the solid phantom. This leads to the observed downstream dose enhancement.
The effect of interfoil gap width on the dose enhancement
presented in Fig. 8 can be explained as follows. As previously stated, for 18 MV or higher photon energy beams, pair
production is the primary photon interaction process in a
high-Z material 共i.e., gold foils兲. The total dose between two
foils is the sum of the dose due to the electrons/positrons
produced in the first 共upstream兲 foil and the dose due to the
electrons/positrons backscattered into the same region by the
second 共downstream兲 foil. A Monte Carlo simulation of the
material and geometry configuration of the measurements in
3548
Alkhatib, Watanabe, and Broadhurst: Dose enhancement by high-Z material for photon radiotherapy
fact showed that the backscattered electrons contributed 18%
of the total dose at points between the two foils. Since the net
increase of the dose by replacing water with gold foils was
26%, the backscattered electron contribution to the dose enhancement was large. As the separation between the foils is
increased, slower electrons/positrons from the pair production process do not reach the downstream foil. They therefore do not backscatter and again deposit energy in the same
volume between the foils. The above argument is also supported by a measurement we made using a single gold foil,
in which no measurable dose enhancement was observed
downstream of the foil.
Our initial results indicate that dose enhancement by pair
production due to high-Z material placed on the upstream
side was to a large extent negated by the loss of the lower
energy photons, both through photoelectric absorption and
scattering out of the initial beam direction due to increased
Compton interactions in the high-Z material upstream of the
volume of interest. However, when the geometry of the
phantom was modified so that high-Z foils were introduced
around the area of interest but parallel to the incident photon
beam direction, the photons did not have to traverse the upstream foil to reach the area of interest, while Compton scattered photons from both the regions inside and outside the
area of interest could pass through the surrounding foils, interact, and produce either electron-positron pairs for local
dose enhancement or more electrons by a second Compton
process. This explains why a larger dose enhancement is observed with the parallel configuration than that with the perpendicular configuration of foils relative to the incident photon beam.
For the current study we used thin parallel metallic foils,
the geometrical arrangements being essentially two dimensional as seen in Figs. 4 and 5. Both foil arrangements
showed dose enhancement. A further increase in the dose
enhancement could be expected if the gold is arranged in a
three-dimensional 共3D兲 geometry so that the increased production of secondary electrons/positrons more effectively
contributes to the dose between the gold. An optimization of
the geometry, i.e., the shape/size of gold 共foils, cylinders, or
micro-/nanoparticles兲 and the gap distance among the gold
materials, may well lead to even larger dose enhancement in
the 3D configuration. As an example, further phantom measurements can be made with the high-Z material formed in a
cylinder surrounding the region of interest.7 When progress-
Medical Physics, Vol. 36, No. 8, August 2009
3548
ing to animal and eventually human studies, thin foils would
be replaced by submicron gold particles suspended in a gel10
or gold nanoparticles injected at many sites to simulate the
optimum geometry. As stated, depending on the biological
geometry, larger enhancements than reported in this paper
may be possible with a 3D distribution of gold microparticles
suspended in a gel carrier.
VI. CONCLUSIONS
These phantom based studies have shown that local dose
enhancement of as much as 40% can be achieved by the
introduction of high-Z materials close to the region of interest with 18 MV or higher energy photon beams. Detailed
work is in order to study the interactions more carefully and
gain much better understanding of dose distributions in all
the type of geometries that were suggested here using measurements and more detailed Monte Carlo simulations. Simultaneously, animal experiments will be undertaken to ensure that the enhancement can be sustained under
biologically meaningful conditions.
a兲
Author to whom correspondence should be addressed. Electronic mail:
watan016@umn.edu; Telephone: 612-626-6708; Fax: 612-626-7060.
Permanent address: Department of Physics, University of Minnesota,
Minneapolis, MN 55455.
1
J. F. Hainfeld, D. N. Slatkin, and H. M. Smilowitz, “The use of gold
nanoparticles to enhance radiotherapy in mice,” Phys. Med. Biol. 49,
N309–N315 共2004兲.
2
J. F. Hainfeld et al., “Radiotherapy enhancement with gold nanoparticles,” J. Pharm. Pharmacol. 60, 977–985 共2008兲.
3
D. Sheikh-Bagheri and D. W. Rogers, “Monte Carlo calculation of nine
megavoltage photon beam spectra using the BEAM code,” Med. Phys.
29, 391–402 共2002兲.
4
J. H. Hubbell, “Photon mass attenuation and energy-absorption coefficients from 1 keV to 20 MeV,” J. Appl. Radiat. Isot. 33, 1269–1290
共1982兲.
5
I. J. Das, F. M. Khan, and B. J. Gerbi, “Interface dose perturbation as a
measure of megavoltage photon beam energy,” Med. Phys. 15, 78–81
共1988兲.
6
B. Ciesielski et al., “Dose enhancement in buildup region by lead, aluminum, and lucite absorbers for 15 MV photon beam,” Med. Phys. 16,
609–613 共1989兲.
7
X. A. Li et al., “Dose enhancement by a thin foil of high-Z material: A
Monte Carlo study,” Med. Phys. 26, 1245–1251 共1999兲.
8
J. F. Briesmeister, MCNP-A general Monte Carlo N-particle transport code,
Version 4C, Los Alamos National Laboratory, Los Alamos, NM, 2000.
9
B. L. Werner et al., “Dose perturbations at interfaces in photon beams,”
Med. Phys. 14, 585–595 共1987兲.
10
O. Holte et al., “Preparation of a radionuclide/gel formulation for localised radiotherapy to a wide range of organs and tissues,” Pharmazie 61,
420–424 共2006兲.
b兲