Radiological properties of a waxgypsum compensator material

F. C. P. du Plessisa and C. A. Willemse

Medical Physics Department, Faculty of Health Sciences, University of the Free State, P.O. Box 339,
Bloemfontein, 9300 South Africa
Received 27 October 2004; revised 25 February 2005; accepted for publication 1 March 2005;
published 12 April 2005
In this paper the radiological properties of a compensator material consisting of wax and gypsum is
presented. Effective attenuation coefficients EACs have been determined from transmission mea-
surements with an ion chamber in a Perspex phantom. Measurements were made at 80 and 100 cm
source-to-skin distance SSD for beam energies of 6, 8, and 15 MV, for field sizes ranging from
narrow beam geometries up to 40 40 cm2, and at measurement depths of maximum dose build-up,
5 and 10 cm. A parametrization equation could be constructed to predict the EAC values within 4%
uncertainty as a function of field size and depth of measurement. The EAC dependence on off-axis
position was also quantified at each beam energy and SSD. It was found that the compensator
material reduced the required thickness for compensation by 26% at 8 MV when compared to pure
paraffin wax for a 10 10 cm2 field. Relative surface ionization RSI measurements have been
made to quantify the effect of scattered electrons from the waxgypsum compensator. Results
indicated that for 80 cm SSD the RSI would exceed 50% for fields larger than 15 15 cm2. At 100
cm SSD the RSI values were below 50% for all field sizes used. 2005 American Association of
Physicists in Medicine. DOI: 10.1118/1.1896455

Key words: effective attenuation, compensator, SSD, waxgypsum, relative surface dose

I. INTRODUCTION Apart from attenuation, compensators introduce scatter

Compensators play an important role in radiation dose
modulation. It can be used to modulate the dose to compen-
sate for missing tissue, e.g., to obtain a uniform dose at a
certain reference depth in the patient,1 or as a beam intensity
modulation device where a uniform dose is sought in a plan-
ning target volume.2
Diverse materials are used for compensator manufactur-
ing that include wax,3 gypsum,4 lead,5 aluminum,1,6 brass,1,7
stainless steel granulates,8 gypsum/steel,9 tungsten powder
and silicon,2 cement,10 and tissue equivalent plastic.11
The shape of compensators can be realized with methods
such as thickness ratios,6,11,12 the use of effective attenuation
coefficients,1318 and the use of narrow beam linear attenua-
tion coefficients with the inclusion of scatter radiation.13 In
missing tissue compensation methods the relevant contour
information of the patient can be obtained with numerous
methods.1926 Compensator manufacturing methods vary
from cutting out self-adhesive lead strips to build the
compensator27 to making a negative cast with a computer-
ized milling machine after which the compensator material is
poured into8 and allowed to set.3,4,9,10
The density or attenuation properties of compensator ma-
terials also have its advantages and disadvantages.4,8,9 If the
density of the material is high, e.g., lead or low melting point
alloys, then a small error in the required thickness can have a
large error on the modulated dose. On the other hand, low
density compensator materials such as wax would be the
least sensitive to such errors, but would sometimes require
impractical thicknesses to reduce the dose to a required per-
centage level.
and beam hardening into the heterogeneous radiation beam,
causing alterations in its effective energy as a function of
field size and depth of measurement.2831 The beam flatten-
ing filter in the radiation head causes relative beam softening
at locations displaced from the beam central axis, therefore,
also changing the beam characteristics as well as the attenu-
ation properties of a compensator material.32,33
Effective attenuation coefficients EACs can take ac-
count of the attenuation properties of compensator materials
with changing beam characteristics. These can be determined
with a variety of methods that include the use of
tissue-maximum-ratios15 or can be calculated from transmis-
sion ratio data using sector integration methods28 or can be
measured with ion chambers30 in large water phantoms,10,11
polystyrene or Styrofoam phantoms,4,8,12 in air using
build-up caps,9 in mini-phantoms,31 or with parallel plate ion
chamber measurements.17 These EACs are used to determine
the thickness of the compensator.9,20
The attenuation properties of wax can be enhanced by the
introduction of suitable materials to give a uniform compo-
sition of the resulting material. In this study the radiological
characteristics of a compensator material made from a wax
gypsum mixture are presented. These are expressed in terms
of an EAC obtained from transmission measurements for
x-ray beam energies of 6, 8, and 15 MV for different field
sizes ranging from narrow beams up to 40 40 cm2 fields.
The effect of the flattening filter in the radiation head on
off-axis EAC values were also investigated for a 5 5 cm2
field located at different off-axis positions. The effect on the
relative surface ionization RSI34,35 due to the presence of
the compensator medium was also investigated.

1246 Med. Phys. 32 5, May 2005 0094-2405/2005/325/1246/10/$22.50 2005 Am. Assoc. Phys. Med. 1246
1247 F. C. P. du Plessis and C. A. Willemse: Radiological properties of a waxgypsum compensator material
II. METHODS AND MATERIALS
A. Compensator manufacturing
The following criteria influenced the design of the com-
pensator material:
a The material has to be readily available and inexpen-
sive.
b The resulting material must have attenuation properties
that would allow sufficient tissue thickness compensa-
tion with the maximum thickness 9 cm of compensa-
tor material, based on the thickness of the polystyrene
blocks commonly used as compensator molds.
c The attenuation properties of the compensator material
must be stable over time and must not be influenced by
factors such as humidity.
d The resulting material should be easy to handle and to
mill during compensator manufacturing.
It was found that a suitable compensator material could be
produced from a mixture of paraffin wax and gypsum pow-
der Kaldent Dental Lab Plaster, Bombay, India. The op-
timal mixing ratio was two mass units of wax to three mass
units of gypsum powder. The wax was melted and the gyp-
sum was added to it while it was stirred at 78 C. For every
70 g of compensator material 1 ml of mineral oil was added.
This enhanced miscibility and lowered the mixture viscosity
so that it could be poured more easily into a compensator
mould. It also eliminated any visible coagulation. A set of
moulds were manufactured to produce 28 28 cm2 plates of
the compensator material with thicknesses of 1, 2, and 5 cm.
The waxgypsum mixture was poured into the moulds while
they were being vibrated to eliminate the formation of gas
bubbles. After the mixture had set, the material was ma-
chined to the desired thickness. The physical density of the
compensator material was measured as 1.48 0.01 g / cm3. Its
effective atomic number was determined to be 10.56 and the
mass-electron density Z / A = 0.533 if the atomic composi-
tion is taken to be 40% dental wax 60 % gypsum and SSD unless otherwise stated. The bottom surface of the com-
the weight composition of dental wax is given by the per-
centage weight ratio: H : C : O = 13: 83: 4 and for gypsum:
Ca: S : O : H = 23.3: 18.6: 55.8: 2.3. A sample of the pure
gypsum was mixed with water with a water/gypsum mixing
ratio of 0.58, poured into a mould, and allowed to set under
vibration to minimize gas bubble formation, and allowed to
dry. The physical density was determined as 1.29
0.01 g / cm3.
The influence of sedimentation of the gypsum powder in
the wax on the density was also investigated. A 20 cm col-
umn of the mixture was poured under vibration and allowed
to set. Test samples were taken at the top and bottom of this
column. The density difference between these test samples
was within the measurement error.
B. Determination of the EAC of the compensator
material
To calculate the necessary thickness of a compensator ma-
terial for desired dose distribution inside patient/phantom, its
Medical Physics, Vol. 32, No. 5, May 2005
1247
EAC values must be known. The EAC takes into account the
spectral changes and field size of the radiation beam, as well
as the scatter properties of the measurement phantom. EAC
data can be obtained from transmission measurements using
Eq. 1, under the assumption that the attenuation curve is a
simple exponential:
DSt = DS0eefft , 1a
where eff is the EAC obtained for a compensator material of
thickness t, DS0 is the ion chamber reading detector sig-
nal without any attenuator in the beam, and DSt represents
the signal from the ion chamber when an attenuator slab of
thickness t is present. The transmission, Tt = DSt / DS0,
is taken as the ratio of the detector reading with and without
any compensator material in the beam and is a ratio of ion-
ization measurements. The EAC for the compensator mate-
rial can be obtained from the transmission values through the
following relation:
lnTt
eff = . 1b
t
Alternatively, Eq. 1b can also be used to determine the
thickness, t, of compensator required to yield a given trans-
mission, T, if the EAC is known.
C. Central axis transmission measurements
Transmission measurements were carried out with a
Farmer type Nuclear Enterprises 2571 0.6 cc ion chamber
and a Nuclear Enterprises IONEX Dosemaster electrometer
in a Perspex phantom with dimensions large enough to attain
full lateral electron equilibrium conditions. Measurements
were taken on a Philips SL25 linear accelerator for 8 and 15
MV beam energies and on a Philips SL75/5 machine at 6
MV at a source-to-skin distance SSD of 100 and 80 cm.
Square fields of side lengths 5, 10, 15, 20, 30, and 40 cm
were used. In this study all field sizes are defined at 100 cm
pensator slabs was located at a distance of 67 cm from the
target 33 cm from the phantom surface and was placed on
an aluminum shadow-tray, mounted on the radiation head.
The slab thickness, t, was varied from 0 to 10 cm in 1 cm
steps. The measurements were taken at water equivalent
depths of dmax, 5.5 and 11 cm. The depth of maximum dose
build-up was 1.5, 2.0, and 2.5 cm for 6, 8, and 15 MV,
respectively. For all the open beam measurements, i.e., no
compensator material in the beam the shadow tray was left
in place.
The narrow beam EAC was also determined in a Perspex
phantom. A set of transmission measurements were taken
with the detector placed at a distance of 2 m from the target
and a field size set to cover a 5 5 cm2 field area at the
position of the detector. This was repeated for all beam en-
ergies and measurement depths.
The influence of water absorption on the attenuation prop-
erties of the compensator material was determined by im-
mersing the slabs of compensator material in water for 24 h.
Transmission measurements were taken for a field size of
1248 F. C. P. du Plessis and C. A. Willemse: Radiological properties of a waxgypsum compensator material 1248

10 10 cm2 at 100 cm SSD and compared with the corre-

sponding measurements before the slab immersion. It was
found that the difference in transmission was within experi-
mental uncertainties of 1%.

D. Off-axis transmission measurements

The beam flattening filter is designed to scatter photons
emerging from the bremsstrahlung target in order to make
the fluence more uniform over the field area. The filters are
cone shaped and thick in the center, becoming thinner with
increasing off-axis distance. Thus, the x-ray beam is more
filtered hardened in the central axis area and relatively
more soft in the off-axis beam area. In order to investigate FIG. 1. Transmission data and fitted exponential curves for the waxgypsum
this effect on the EAC, transmission measurements were compensator material at 6 MV for the indicated field sizes. Depth of mea-
taken for a 5 5 cm2 field with the field center located at surement was 1.5 cm at an SSD of 100 cm.
off-axis positions of 5, 10, and 15 cm at a SSD of 100 and 80
cm for all beam energies. The measurement depth was at
maximum build-up dose dmax. nitude of the percentage error is of the order of 1% which is
smaller than the size of the symbols in the figure. The EAC
decreases as the field size increases since the transmission
E. Relative surface ionization
for a fixed absorber thickness increases with field size, as
The RSI is defined as the ratio of the corrected ionization shown in Fig. 1, due to increasing scatter from the compen-
at the surface to that at a certain reference depth, chosen in sator. The EAC values at 80 cm SSD are significantly lower
this study to be dmax for each respective beam energy. It was at larger field sizes than the corresponding EAC values at
measured with a PTW 30-329 Markus parallel plate cham- 100 cm SSD. This is due to more of the compensator-
ber and a Baldwin Farmer electrometer. Measurements were induced scatter reaching the ion chamber for the shorter
taken in a Perspex phantom at dmax the reference depth and SSD. The compensator retractal distance at 100 cm SSD is
on the surface for SSDs of 80 and 100 cm, and field sizes of 33 cm and at 80 cm it is only 13 cm. Thus at 80 cm SSD
5 5, 10 10, 20 20, and 30 30 cm2 at 100 cm SSD. relatively more large angle scattered photons originating in
Care was taken that the ion chamber fitted tightly into the the compensator material are detected by the chamber. An
phantom with its front surface flush with that of the phantom
surface for surface measurements. For dmax measurements
additional phantom slabs were added while the SSD was
kept at a constant value. Chamber polarity effects and over
response corrections of the chamber were made according to
the method of Gerbi et al.34 The influence of the aluminum
shadow tray and the addition of 1 cm and 10 cm slabs of
waxgypsum material on the surface ionization relative to an
open beam were investigated.
All the data in this study were fitted using a least-squares
residual sum of squares based regression procedure in Mi-
crosoft Excel.

III. RESULTS AND DISCUSSION

Figure 1 shows a set of transmission curves at 6 MV for
various field sizes measured at dmax = 1.5 cm in the measure-
ment phantom for a SSD of 100 cm. The EAC for each field
size was determined by fitting an exponential function Eq.
1a, using a linear least-squares method, to the natural loga-
rithm of the measured data. The correlation coefficient for
each fit was better than 0.999. The percentage transmission
for 10 cm thickness ranged from 50% at 6 MV to 60% at 15 FIG. 2. a The EAC values for 6 MV. In the graph the legends indicate the
MV not shown for the narrow beam. For the 40 40 cm2 measurement depth in the phantom cm and in brackets the SSD value
field the corresponding values are 60% and 70%. The mea- cm. Zero field size data correspond to the narrow beam EAC data. The
solid lines indicate the calculated EAC values using Eqs. 2 and 3. b
surement error was of the order of the size of the symbols on
EAC vs depth at 6 MV and 100 cm SSD for a narrow beam and square fields
the curves. Figure 2a shows the calculated EAC values for with S = 5, 10, and 40 cm. The solid lines indicate the calculated EAC values
6 MV as a function of field size S at both SSDs. The mag- using Eqs. 2 and 3.

Medical Physics, Vol. 32, No. 5, May 2005

1249 F. C. P. du Plessis and C. A. Willemse: Radiological properties of a waxgypsum compensator material 1249

TABLE I. Fitting parameters for Eq. 2 for 100 and 80 cm SSD.

Energy a1 108 a2 107 a3 107 a4 105 a5 105 a6 105 SSD

MeV cm5 cm4 cm4 cm3 cm3 cm2 cm

6 1.88 5.8 7.45 3.5 1.26 3.3 100

0.576 5.3 3.54 2.1 4.25 6.3 80
8 1.96 4.2 11.5 2.7 0.928 7.8 100
3.69 6.0 17.3 3.0 0.462 28 80
15 0.623 3.3 7.29 1.8 3.06 11 100
0.602 3.3 7.29 1.4 2.88 35 80

important consequence of this is that, e.g., if IMRT was to be
done with this compensator material using an isocentric tech-
nique, then the change of EAC values with SSD must be
taken into account. This would also enhance the surface dose
as is discussed later.
The EAC for a fixed field size is found to decrease with
increasing depth due to increased in-phantom scatter, as
shown in Fig. 2b for 6 MV at 100 cm SSD. But it was
found that this trend is reversed for fields larger than 18
18 cm2. This effect can also be observed in Fig. 2a for
the 100 cm SSD case. Similar effects not shown here were
observed for 8 and 15 MV for field sizes of 20 20 cm2 and
25 25 cm2, respectively. Thus as the beam energy increases
so thus the field size where the EAC is virtually invariant
with depth. A linear fit of the field side length S, versus
energy yields an equation of the form, S = 1.308E 17.795,
with a correlation better than 0.994
The data in Figs. 2a and 2b, were fitted by
effd,S = 0d + a1d + a2S3 + a3d + a4S2
+ a5d + a6S,
where 0d is the measured narrow beam EAC at each
depth. S is the side length of the square field, and d the
measurement depth in the phantom expressed in terms of
the equivalent depth in water through the relation:
depthwater = 1.11 depthPerspex. The terms in Eq. 2 indicate
a linear dependence with depth, d, for a constant square field
of side length, S, and is confirmed in Fig. 2b. Further, 0d
was parametrized as a function of depth with
0d = 0dmax + b1d2 + b2d,
where b1 and b2 are fitting parameters and 0dmax is the
narrow beam EAC measured in the phantom at an equivalent
depth of dmax in water.
The fitting parameters in Eq. 2 are shown in Table I for
80 and 100 cm SSD and the fitting parameters for Eq. 3 in
TABLE II. Fitting parameters for Eq. 3.
Table II. The uncertainties for fitting parameters ai in Table
I were found to be between 1% and 3% for a1, a3, and a5.
For parameters a2, a4, and a6 the uncertainty varied between
6% and 9%. For the fitting parameters in Table II the uncer-
tainty was within 1%. The percentage difference between
EAC values calculated with Eqs. 2 and 3 and measure-
ments is summarized in Table III. From Table III it can be
seen that Eqs. 2 and 3 reproduce the measured EAC data
within 3.6% for field sizes 20 cm and within 4.7% for
larger fields. This would lead to an error in the dose modu-
lation of the order of 2% if an EAC of 0.07 cm1 and
a maximum thickness of 9 cm of compensator material is
assumed.
A. Dependence of EAC on off-axis position
Figure 3 shows the variation of the EAC for a 5 5 cm2
field at 6, 8, and 15 MV as a function of off-axis position as
measured at dmax. The EACs were calculated from transmis-
sion data using the effective oblique thickness of the com-
2 pensator slabs in the direction of the central axis of the off-
axis fields. For both SSDs the EAC values increase as a
function of off-axis position. This increase is related to the
diminishing thickness of the flattening filter with off-axis dis-
tance, leading to a relative beam softening toward the edges
of the flattening filter. The EAC values at 100 cm SSD is
larger than those at 80 cm SSD, due to fewer compensator
scattered photons reaching the measurement phantom and
enhancing the dose at the point of measurement.
It was found that the variation of the EAC as a function of
3 off-axis distance x at both SSD values could be modeled
with an accuracy of 1.5% with an equation of the form:
TABLE III. Percentage difference between the measured EAC values and the
calculated EAC values using Eqs. 2 and 3 at 100 and 80 cm SSD.
Percentage deviation
Energy SSD
MeV S 20 cm S 25 cm cm

6 1.4 0.9 100

Energy 0dmax b1 106 b2 104
3.1 4.7 80
MeV cm1 cm3 cm2
8 2.7 4.6 100
6 0.0685 6.84 2.36 3.6 4.0 80
8 0.0603 31.0 4.15 15 2.1 2.9 100
15 0.0503 1.07 2.20 1.7 2.4 80

Medical Physics, Vol. 32, No. 5, May 2005

1250 F. C. P. du Plessis and C. A. Willemse: Radiological properties of a waxgypsum compensator material 1250

FIG. 4. The EAC values for the waxgypsum mixture as a function of beam
energy taken at dmax for a 10 10 cm2 and a 40 40 cm2 field size at a SSD
of 100 cm.

FIG. 3. The variation of the EAC for a 5 5 cm2 field at 6, 8, and 15 MV as

a function of off-axis positions of 0, 5, 10, and 15 cm, at dmax. The legends to relatively more beam filtration in the center as compared
indicate the SSD value while the values in brackets indicate the beam energy
in MV units.
to the off-axis regions, when compared to the filter shape at
8 MV and is also about 1.5 times thicker. At 15 MV the beam
quality per unit radius is thus expected to change more rap-
EACx = c1x3 + c2x2 + c3x + effdmax,5. 4 idly than for 8 MV. This could explain why the percentage
difference in the EACs on the CAX and 15 cm off-axis is of
The parameter effdmax , 5 is the EAC calculated with Eq. the order of 11% at 15 MV compared to 9% at 8 MV. From
2 for d = dmax and S = 5 cm at the respective SSD. Table IV the literature36 the Compton interaction coefficient is also on
shows the fitting parameters for Eq. 4 at 80 and 100 cm average three to seven times larger than the pair production
SSD. The uncertainties in the values for the parameters ci coefficient at energies up to 12 MeV for stainless steel. At
were less than 1%. higher energies, pair production increases and is on average
For the 100 cm SSD cases in Fig. 3 the EAC change with 1.3 times larger than the Compton coefficient up to 15 MeV.
4.4% at 6 MV, 9.2% at 8 MV, and 11.4% at 15 MV with Thus for the 15 MV case, pair production could soften the
respect to the central axis EAC values. In terms of dose beam, but the relative amount of these interactions is out-
calculations this could lead to dose discrepancies of up to 3% weighed by the Compton scatter events at lower energies.
at off-axis locations of 15 cm at 6 MV. At 8 and 15 MV the This would most likely cause overall beam hardening in a
difference would be of the order of up to 5%. For example: If polychromatic 15 MV beam. It should be bear in mind that
the compensator thickness is taken as t = 9 cm and the EAC the beam quality depends on the spectral distribution of the
at 6 MV on the CAX at 100 cm SSD is taken as 0 bremsstrahlung photons emerging through the target, which
= 0.068 cm1, then the fraction by which the dose is altered will result in a higher beam quality on the CAX,37 also on
at an off-axis location of 15 cm is given by D / D0 the shape of the flattening filter and its material composition.
= expt* 0, where = 0.071 g / cm1 or 1.044*0, Figure 4 shows the variation of the EAC values, measured
which is the EAC value at the 15 cm off-axis position. This at dmax for the waxgypsum mixture as a function of beam
gives D / D0 = 0.973 or a 3% difference. energy.
The difference in the off-axis EACs compared to CAX The results show that the EAC values decrease as the
measured EACs can be ascribed to a difference in beam beam energy increases, but levels off due to increasing pair
quality at these locations; it is well known that the beam production in the waxgypsum material. This trend is con-
quality degrades at off-axis locations in the beam. Stainless sistent with data produced by Arora et al.4 for a gypsum
steel flattening filters are used at the three beam energies. attenuator.
The filter for 15 MV is sharply peaked in the center, leading Figure 5 shows how the EAC depends on the SSD for the
6, 8, and 15 MV cases over the range 80120 cm for a 10
10 cm2 field, measured at dmax.
TABLE IV. Fitting parameters for Eq. 4 for a field with S = 5 cm at dmax. The EAC values increase slowly at a given beam energy
Energy c1 106 c2 106 c3 105 effdmax,5 SSD
for an increase of SSD. Less scattered photons from the com-
MeV cm4 cm3 cm2 cm1 cm pensator would reach the detector at longer SSDs. If the
thickness of the attenuation material is increased, it would
6 0.216 0.183 26.7 0.0676 100 increase the fraction of photons that are scattered. As a result
0.655 7.71 42.4 0.0648 80
the detector signal will decrease more rapidly with an in-
8 3.11 72.1 0.526 0.0593 100
crease in attenuation material at long SSDs compared to
3.24 0.796 0.141 0.0576 80
15 2.16 54.8 5.30 0.0491 100
shorter SSDs. This would result in larger EAC values. The
3.06 75.5 7.20 0.0478 80 EAC values shown in Fig. 5 could be predicted to within 2%
with the empirical relation:

Medical Physics, Vol. 32, No. 5, May 2005

1251 F. C. P. du Plessis and C. A. Willemse: Radiological properties of a waxgypsum compensator material
FIG. 5. EAC as a function of SSD for the 6, 8, and 15 MV cases for 10
10 cm2 field at dmax.
EACE,z = 0.000 28*z*E0.3547 + 0.088E0.3162 5
as a function of beam energy E and SSD z. Here, a linear
dependence of the EAC vs SSD was assumed. Equation 5
was only validated over the SSD interval in Fig. 5. Figure 6
shows the transmission values for the waxgypsum mixture
compared with that of paraffin wax and pure gypsum for
beam energy of 8 MV at dmax.
The results in Fig. 6 indicate that the addition of gypsum
to the wax decreases the transmission by 8% at 8 MV at both
SSDs. The addition of gypsum in effect reduces the thickness
of the compensator by 26% compared to wax if a certain
transmission level is desired, e.g., 7.5 cm of the wax
gypsum material can produce the same transmission as 10
cm of paraffin wax. Due to the low density of the gypsum
1.29 g / cm3 its transmission characteristics are similar to
that of paraffin wax at both SSDs. This type of gypsum has a
high degree of porosity probably due to its mixing ratio of
0.58 to water. The addition of wax to the pure gypsum re-
places these pores to yield a compensator material that is
more dense than its constituents. Gypsum used by others4
has a density on the order of 2 g / cm3 and will reduce the
transmission of x rays more efficiently than the gypsum used
in this study to enhance the attenuation of pure paraffin wax.
The EAC values were determined from relative ionization
measurements. If the EACs are related to relative dose then
the change in beam quality should be assessed. Evaluation of
FIG. 6. Transmission values for waxgypsum wx-gp and paraffin wax
wx for a beam energy of 8 MV and a field size of 10 10 cm2 at dmax. The
quantity in brackets indicates the SSD in cm units.
Medical Physics, Vol. 32, No. 5, May 2005
1251
FIG. 7. The effect of compensator-induced scatter on the detector signal for
eight cases for a 6 MV beam, as measured at 100 cm SSD. Each measure-
ment is labeled according to the case under consideration, with the
compensator setup shown directly below for cases 13 and directly above
for cases 48.
the transmission measurements indicated a very good corre-
lation better than 0.999 between the measured data and
pure exponentials fitted to these data, e.g., Eq. 1a. This
indicated that a constant value for eff was sufficient to
model transmission through the compensator and no need to
adjust it for compensator thickness was necessary. This indi-
cated that the effective energy of the beam was virtually
invariant over the range of compensator thicknesses used.
The reduction in the absorbed dose for a thickness of com-
pensator material would be directly proportional to the trans-
mission as measured and expressed in terms of relative ion-
ization. Furthermore for the 0.6 cc Farmer type ionization
chamber a small change in the beam quality, indicated by a
change in eff as a function of thickness, would not alter its
response noticeably in the megavoltage energy range.
B. Scatter influence of compensator shape on CAX
ionization measurements
The EACs in this paper were derived from ionization
measurements of the beam transmitted through uniform
thicknesses of homogenous waxgypsum slabs. An impor-
tant aspect that should be addressed is how the off-axis com-
pensator shape influences these ionization measurements.
Compensator-induced scatter and varying in-phantom scat-
tering conditions could influence the ionization measure-
ments on the central axis CAX. These effects would be
more enhanced at lower beam energies and large field sizes.
Figure 7 shows CAX ionization measurements for eight
cases for different compensator shapes measured at 6 MV in
a 40 40 cm2 beam at a depth of 1.5 cm in a Perspex phan-
tom at a SSD of 100 cm. In case 1, the ionization measure-
ment was taken with the shadow tray only, referred to as
open beam. In case 2, a 5 cm slab was positioned at an
off-axis distance of 3 cm from the CAX shown as the ver-
tical line through each figure at 100 cm SSD. Its measure-
ment is labeled with 2. For case 3 the same thickness was
used with the off-axis position set at 6 cm as shown in the
figure, with its measurement labeled by 3. From these
measurements it is evident that they vary with respect to the
open beam due to the presence of the 5 cm compensator slab
at the mentioned off-axis positions. This can be ascribed to
1252 F. C. P. du Plessis and C. A. Willemse: Radiological properties of a waxgypsum compensator material
induced scatter from the compensator that enhances the de-
tector signal and decreases slightly as the off-axis position is
increased. This decrease is however within a standard error
of 0.3% as shown on all measurements. The in-phantom
scatter is such that about 35%43% of the phantom would
receive attenuated primary radiation the open beam receives
100% primary radiation. At these locations the phantom re-
ceives about 75% transmitted primary x rays. The combina-
tion of compensator scatter and changing in-phantom scatter
leads to an increase in the detector signal on the order of 3%.
In cases 48 the baseline thickness of the compensator
was fixed at 2 cm, with case 4 showing the ionization mea-
surement on the CAX with the 2 cm slab on the shadow tray.
Case 5 shows the ionization measurement for the configura-
tion shown where an extra 3 cm of material was added at 3
cm off-axis position. In case 6 the thickness was increased to
a total of 7 cm. The thicker compensating medium gives rise
to slightly higher ionization values. Here the detector signal
increase at a constant off-axis position is on the order of
1.5%. In case 7 the same setup is used as in case 5 with the
offset now 6 cm. Here the ionization is approximately the
same as in case 5, experimental uncertainties taken in con-
sideration. So as in cases 13 the off-axis position of the
compensator did not cause a significant variation in the ion-
ization between shifts of 3 and 6 cm.
Case 8 presents a more complicated case where an addi-
tional 2 cm slab is added at an off-axis distance of 3 cm to
the 2 cm base slab. Here it can be seen that the addition of
compensating medium enhance the ionization and compares
well with the value for case 6. It thus seems that the addition
of compensating material at off-axis positions influences the
ionization values between 1.5% and 3%, depending on the
thickness of the material on the CAX. Variations between the
3 and 6 cm off-axis positions did not change the ionization
significantly and might be due to scatter from the peripheri-
cal regions of the compensator that enhances the detector
signal that is not sensitive for these off-axis off-sets. Similar
studies made by others,38 using a low melting point alloy that
covered half the radiation beam with measurements at 2 cm
off-axis, showed that the relative increase in dose due to
scatter from of the attenuator was on the order of 2.5% at an
optimal attenuator thickness of 2 cm for a 6 MV Varian
Clinac beam. This thickness is roughly equal to one mean-
free path, 1 / eff. In our case this will correspond to a thick-
ness of 18.8 cm for eff = 0.053 at 6 MV in agreement with
an increase in the detector reading 1 and 2, as well as 4, 5,
and 6 where the off-axis thicknesses ranged from 2 to 7 cm.
C. Relative surface ionization
The introduction of low atomic number materials such as
wax can have a significant effect on the surface dose on a
patient and also depends on the distance from the absorber to
the surface ASD.35 The effect of the waxgypsum compen-
sator material on the relative surface ionization was investi-
gated at ASDs of 33 cm SSD= 100 cm and 13 cm SSD
= 80 cm.
Medical Physics, Vol. 32, No. 5, May 2005
1252
FIG. 8. The effect of the open beam - - -, shadow tray , tray 1 cm
waxgypsum , tray 10 cm waxgypsum , tray 1 cm wax ,
tray 10 cm wax , tray 1 cm gypsum and tray 10 cm gypsum
material on the RSI expressed as a percentage relative surface ionization
vs field size S.
In Fig. 8 the RSI is shown as a function of field size S
for a SSD of 100 cm and beam energy of 8 MV. Eight cases
are presented to show the effect of field size for an open
beam, the tray, the tray 1 cm waxgypsum, and the tray
10 cm waxgypsum and the same thicknesses for pure
wax and gypsum on the RSI. The errors associated with the
measurements correspond to the size of the symbols.
The RSI increases as the field size is increased due to
more scatter that reaches the parallel plate ion chamber. The
RSI enhancement from the shadow tray is of the order of
3%, while the addition of waxgypsum enhances the RSI by
about 7% relative to the open beam case. The RSI values for
wax corresponded to that of waxgypsum but for gypsum
alone it is less since its electron density is lower than that of
wax. Here the RSI is enhanced by 4.5% relative to the open
beam RSI.
It was found that the RSI depended linearly on the field
side length S for all studied beam energies and at both SSDs.
The parametrization equation was of the form:
RSIS = S 5 + RSI5, 6
where parameter is listed in Table V. The uncertainties in
the values for were within 2.5% for almost all the cases.
The RSIS for an equivalent field of side length, S, is ex-
pressed in terms of the RSI for the 5 5 cm2 field size. This
equation is of the same form as that used by Arora et al. in
their characterization of gypsum attenuators.4
In their study they calculated the values for a gypsum
material at 4 and 15 MV at SSDs of 80 and 100 cm. At 15
MV and 100 cm SSD our values as measured on a Philips
SL 25 machine for the open beam was 1.02, compared to
their value of 0.92. For their Lexan tray was 1.26 and for
our aluminum shadow tray it was 1.25. The addition of gyp-
sum to the tray led to slightly lower values than in our case
at 15 MV where wax also forms part of the compensator
material.
Figure 9 shows the plots top of the RSI5 as used in Eq.
6, and bottom the RSI30 as a function of beam energy.
As expected the RSI5 values are higher at 80 cm SSD due
to more scatter contribution to the parallel plate ion chamber
1253 F. C. P. du Plessis and C. A. Willemse: Radiological properties of a waxgypsum compensator material
TABLE V. Fitting parameter for each beam modifier accessory, energy, and SSD. Parameter is expressed in
units of cm1.
values at 80 cm SSD
Energy MeV 6 8
Open 1.26 1.12
Tray 1.69 1.57
Tray 1 cm a 2.36 2.16
Tray 10 cm a 3.08 2.26
a
Data for waxgypsum compensator material.
signal than at 100 cm SSD. The RSI5 decreases slowly
with an increase in beam energy. This might be related to
relatively less photon interactions in the materials in the
beam, resulting in a lower RSI. At 80 cm SSD the RSI5
decreases relatively rapidly from 6 to 8 MV for the tray
waxgypsum materials. This indicates that the RSI is en-
hanced at 6 MV due to large angle scattered electrons from
the waxgypsum material that increases the RSI. This effect
is not as prominent at 100 cm SSD.
For the 30 30 cm2 case the dose enhancement due to the
waxgypsum material is of the order of 30% relative to the
open beam case at 80 cm SSD compared to 8% for the cor-
responding case at 100 cm SSD. For this field size the effect
of beam energy is such that the RSI30 values level-off in
comparison with the slow decrease for the RSI5 case. It
seems that the increased number of scattered photons and
electrons produced by the beam modifiers in field size of
FIG. 9. The percentage relative surface ionization RSI as a function of
beam energy. Solid lines and symbols correspond to the 100 cm SSD case.
Broken lines and symbols correspond to the 80 cm SSD case. The top figure
represents the RSI for a field size of 5 5 cm2, and the bottom figure, a
30 30 cm2 field.
Medical Physics, Vol. 32, No. 5, May 2005
1253
values at 100 cm SSD
15 6 8 15
1.02 1.01 0.93 1.14
1.28 1.04 1.07 1.24
1.37 1.18 1.36 1.45
1.46 1.12 1.41 1.57
30 30 cm2 balances the decrease in the number of photon
interactions at higher beam energies, resulting in virtually
constant RSI30 values for 8 and 15 MV. The most impor-
tant result is, however, that the RSI30 is enhanced dramati-
cally due to the presence of waxgypsum material at a SSD
of 80 cm, this enhancement is also virtually identical at
thicknesses of 1 and 10 cm. The introduction of wax
gypsum at a SSD of 80 cm for a field size of 30 30 cm2
could cause the relative surface dose to rise from the 40%
level, for the open beam, to 70%.
The RSI enhancement due to the shadow tray relative to
the open beam is on the order of 4% for the 5 5 cm2 case at
80 cm SSD and on the order of 1% at 100 cm SSD. For the
30 30 cm2 field size the enhancement is 15% and 3% re-
spectively.
Figure 9 bottom indicates that at 100 cm SSD the sur-
face dose would be less than 50% for field sizes less than
30 30 cm2. At 80 cm SSD the surface dose of 50% is ex-
ceeded for the 30 30 cm2 field size at all energies. From
Eq. 6 the calculated field at 6 MV and 80 cm SSD, that
would result in a RSI of 50%, relative to the maximum dose,
was calculated as 14.8 14.8 cm2, while measurements not
shown in this study indicated it to be 15.4 15.4 cm2. Here
was taken from Table V as 2.36 and RSI5 was read from
Fig. 8 top as 27% for a 1 cm waxgypsum thickness. It is
also observed from Fig. 8 that the RSI is not critically de-
pendent on the thickness of the waxgypsum material except
at 15 MV for the 30 30 cm2 field case where the RSI dif-
ference is on the order of 8% at 80 cm SSD. From Fig. 8
top the open beam RSI differs by 3% between 80 and 100
cm SSD, while for the large field case bottom the differ-
ence is 8%.
The influence of SSD on the RSI is clearly evident from
Fig. 9 at any constant beam energy. At shorter SSDs the RSI
is enhanced. The effect on field size on the RSI is also evi-
dent; for the 30 30 cm2 field size case, the RSI is larger at
any constant beam energy when compared to the small field
case. This enhancement can be ascribed to relatively more
scattered electrons from the compensator that can reach the
point of measurement on the phantom surface. Based on
these observations the scatter angles of the electrons exiting
the compensator are assumed to be relatively large.
The KleinNishina equation K38 describes the relative
probability for a photon to scatter through an angle . For all
megavoltage beam energies the relative probability is highest
1254 F. C. P. du Plessis and C. A. Willemse: Radiological properties of a waxgypsum compensator material
for small scatter angles. For a 15 MV beam K 0.1 for
20, and for an effective beam energy of 4 MeV it will be
35. Thus 90% of photon scatter in a 15 MV beam will be for
35. At 6 and 8 MV with effective beam energies of 2
and 2.5 MeV this photon scatter angle will be within 45. At
these scatter angles and effective photon beam energies the
electrons will have maximum energies of 1.11.4 MeV for 6
and 8 MV and 2.3 MeV for 15 MV as can be calculated
from: h = h / 1 + 1 cos where and = h / 0.511, h molar masses of 268646 g/mole for paraffin wax is on the
is the original energy of the photon, and h is the energy of
the scattered photon, with the electron energy given by E
= h h. It can be shown that the electron scatter angle is
given by
1 2
= sin1
h sin , 1 2 = 0.511
0.511 E + 0.511
For 35 the electrons will scatter between 22 and 90 at
15 MV in the compensator. At 6 and 8 MV 45 and the
electrons would scatter between 27 and 90 in the compen-
sator. A few percent of the electrons are produced from pho-
ton scatter larger than 35 at 15 MV that would scatter in the
forward direction in the compensator. From these numbers,
large angle scatter is predominant and would explain why the
RSI values increase closer to the compensator and for larger
field sizes, these calculations ignore electron multiple scatter
in the compensator that would further enlarge the average
scatter angle for electrons exiting the compensator.
The influence of the compensator thickness is such that
the RSI values between the 1 and 10 cm thicknesses are
virtually invariant for each beam energy and SSD, but devi-
ate for the 15 MV case at 80 cm SSD. For 6 and 8 MV
photon beams electrons produced further upstream in the
compensator material are more efficiently absorbed in the
compensator, compared to 15 MV since the maximum elec-
tron energy in the compensator is 2.3 MeV at 15 MV for
photons scattered within 35 and an effective beam energy of
4 MeV is assumed. The linear CSDA range for the electrons
in waxgypsum is 0.7 and 1.1 cm at 6 and 8 MV and for 15
MV it is 1.6 cm based on the above energy of the electrons
produced in the compensator. From these values most of the
electrons must be produced within the bottom 1 cm slab to
exit the compensator for 6 and 8 MV. At 15 MV additional
material will cause an increase in the liberated electrons,
since its linear CSDA range is 1.6 cm.
At 6 and 8 MV electrons reaching the detector are pro-
duced within the bottom 1 cm slab. Thus additional thick-
nesses would not increase the amount of scattered electrons
that reach the detector. At 15 MV additional thickness would
enhance the number of electrons reaching the detector.
The mass-scattering power T / for waxgypsum was
calculated along a formulation of Rossi as described by
others.39 It was found that for maximum electron energies of
1.1, 1.4, and 2.3 MeV T / was 0.43, 0.29 and
0.12 rad2 cm2 g1. Thus in 6 and 8 MV photon beams, elec-
trons will spread out at much larger angles due to multiple
scatter from their original Compton scattered directions than
at 15 MV. An empirical fit to T / versus electron energy
Medical Physics, Vol. 32, No. 5, May 2005
1254
yielded the equation: T / = 0.505E1.773, where E represent
the energy of the scattered electrons between 1 and 15 MV.
The accuracy over this energy range was within 6%. It
should be kept in mind that the calculated values of T / for
waxgypsum depends on the molar mass M for paraffin
wax which may have a considerable variation due to varia-
tions between 18 and 45 in the number of carbon atoms in
each molecule. The variation in T / for waxgypsum for
order of 4% and is invariant with electron energy.
IV. CONCLUSIONS
In this study the radiological properties of a waxgypsum
.
compensator material were investigated. The main focal ar-
eas were the determination of EACs at 80 and 100 cm SSD,
its variation with field size and off-axis position. It was
found that the EAC value could be parametrized in terms of
its measured narrow beam EAC as a function of field size, S,
and off-axis location, x. The accuracy of the parametrization
over field size was within 3.6% for fields less than 20
20 cm2 at both SSDs. The EAC could also be predicted
accurately at dmax for a 10 10 cm2 field as a function of
beam energy, E, and SSD, z. It was found that the addition of
gypsum to wax reduced the thickness required by 26% for a
10 10 cm2 field at 8 MV.
The RSI was also studied. It was found that the RSI could
be expressed in terms of the RSI for a 5 5 cm2 field with an
equation similar to that used by Arora et al.4 for all beam
energies studied. It was found that the waxgypsum materi-
als had larger values for parameter compared to a pure
gypsum compensator described by Arora et al. This indicated
that the wax component leads to a larger increase in the RSI
as a function of field size as compared to the pure gypsum
compensator material. RSI values at 5 5 cm2 and 30
30 cm2 indicated that the RSI is less than 50% at a SSD of
100 cm. At 80 cm SSD the RSI of 50% will be exceeded if
the field size is larger than 15 15 cm2. For larger fields this
could have implications on the surface dose that a patient
receives if such a compensator material is used as beam in-
tensity modifiers in IMRT based treatments, e.g., the 30
30 cm2 field results in a RSI of up to 80% compared to the
maximum build-up dose at 80 cm SSD for beam energies in
the 815 MV range. It was also found that the RSI is virtu-
ally the same for 1 and 10 cm thickness of waxgypsum
material, the exception being for the 15 MV case at 80 cm
SSD.
ACKNOWLEDGMENT
We would like to thank Omer A. Ali for the careful trans-
mission and surface ionization measurements.
a
Current address: Medical Physics Department, Faculty of Medicine, Uni-
versity of the Free State, P.O. Box 339, Bloemfontein, 9300 South Africa;
electronic mail: Freek.duplessis@fccc.edu
1
F. Ellis, E. J. Hall, and R. Oliver, A compensator for variations in tissue
thickness for high energy beams., Br. J. Radiol. 32, 421425 1959.
2
T. Xu, P. M. Shikhaliev, M. Al-Ghazi, and S. Molloi, Reshapable physi-
1255 F. C. P. du Plessis and C. A. Willemse: Radiological properties of a waxgypsum compensator material
cal modulator for intensity modulated radiation therapy, Med. Phys. 29,
22222229 2002.
3 22
R. J. Boge, R. W. Edland, and D. C. Matthes, Tissue compensators for
megavoltage radiotherapy fabricated from hollowed Styrofoam filled with
wax, Radiology 111, 193198 1974.
4
V. R. Arora and K. J. Weeks, Characterization of gypsum attenuators for
radiotherapy dose modification, Med. Phys. 21, 7781 1994.
5 24
K. P. Mandal, D. H. Baxter, and P. Ray, Thin lead sheets as tissue
compensators for larger field irradiation, Int. J. Radiat. Oncol., Biol.,
Phys. 6, 513517 1980.
6 25
E. J. Hall and R. Oliver, The use of standard isodose distributions with
high energy radiation beamsThe accuracy of a compensator technique
in correcting for body contours, Br. J. Radiol. 34, 4352 1961.
7 26
J. van de Geijn, The construction of individualized intensity modifying
filters in cobalt 60 teletherapy, Br. J. Radiol. 38, 865870 1965.
8 27
J. P. C. van Santvoort, D. Binnekamp, B. J. M. Heijmen, and P. C. Lev-
endag, Granulate steel as compensator material, Radiother. Oncol. 34,
7880 1995.
9
K. J. Weeks, B. A. Fraass, and K. M. Hutchins, Gypsum mixtures for
compensator construction, Med. Phys. 15, 410414 1988.
10
M. S. Thomsen, and N. Uls, Attenuation of 420 MV x rays by a new
compensator material of cement, Med. Phys. 29, 24272432 2002.
11
F. M. Khan, V. C. Moore, and D. J. Burns, The construction of compen-
sators for Cobalt teletherapy, Radiology 96, 187192 1970.
12
W. Sewchand, N. Bautro, and R. M. Scott, Basic data on tissue-
equivalent components for 4-MV X-rays, Int. J. Radiat. Oncol., Biol.,
Phys. 6, 327332 1980.
13
P. C. Shragge and M. S. Patterson, Improved method for the design of
tissue compensators, Med. Phys. 8, 885891 1981.
14
W. Ansbacher, D. M. Robinson, and J. W. Scrimger, Missing tissue
compensators: Evaluation and optimization of a commercial system,
Med. Phys. 19, 12671272 1992.
15
A. L. Boyer, Compensating filters for high energy x rays, Med. Phys.
9, 429433 1982.
16
P. S. Basran, W. Ansbacher, G. C. Field, and B. R. Murray, Evaluation
of optimized compensators on a 3D planning system, Med. Phys. 25,
18371844 1998.
17
F. R. Bagne, N. Samsami, S. W. Hoke, and D. G. Bronn, A study of
effective attenuation coefficient for calculating tissue compensator thick-
ness, Med. Phys. 17, 117121 1990.
18
P. M. K. Leung, J. van Dyk, and J. Robins, A method for large irregular
field compensation, Br. J. Radiol. 47, 805810 1974.
19
B. J. Gerbi, Compensator filter design using radiographic stereo shift
information, Med. Phys. 12, 646648 1985.
20
G. G. Beck, W. J. McGonnagle, and C. A. Sullivan, Use of a Styrofoam
block cutter to make tissue-equivalent compensators, Radiology 100,
694696 1971.
21
D. B. M. Watkins, A proposed method for making reduced wax com-
Medical Physics, Vol. 32, No. 5, May 2005
1255
pensators for use with high-energy radiation beams, Br. J. Radiol. 48,
760762 1975.
J. F. Laursen, H. C. Andersen, and H. P. Hansen, 3D thin lead sheet
compensating system, Med. Phys. 9, 741745 1982.
23
J. W. Andrew, J. E. Aldrich, M. E. Hale, and J. A. Perry, A video-based
patient contour acquisition system for the design of radiotherapy compen-
sators, Med. Phys. 16, 425430 1989.
J. W. Andrew and J. E. Aldrich, A microcomputer-based system for
radiotherapy beam compensator design and contour plotting, Med. Phys.
9, 279283 1982.
W.-C. Lam, K.-S. Lam, D.-J. Lee, and M. J. ONeill, Tissue compensator
fabrication using a simple photographic technique, Radiology 149, 322
323 1983.
A. L. Boyer and M. Goiten, Simulator mounted Moir topography cam-
era for constructing compensator filters, Med. Phys. 7, 1925 1980.
J. E. Aldrich and J. W. Andrew, Self-adhesive lead for compensator
production and radiation shielding, Med. Phys. 19, 361362 1992.
28
E. Papiez and G. Froese, The calculation of transmission through a pho-
ton beam attenuator using sector integration, Med. Phys. 17, 281286
1990.
29
P.-H. Huang, L. M. Chin, and B. E. Bjrngard, Scattered photons pro-
duced by beam-modifying filters, Med. Phys. 13, 5763 1986.
30
E. E. El-Khatib, E. B. Podgorsak, and C. Pla, The effect of lead attenu-
ators on dose in homogeneous phantoms, Med. Phys. 13, 928935
1986.
31
M. E. Castellanos and J. C. Rosenwald, Evaluation of the scatter field for
high-energy photon beam attenuators, Phys. Med. Biol. 43, 277290
1998.
32
B. E. Bjrngard and H. Shackford, Attenuation in high-energy x-ray
beams, Med. Phys. 21, 10691073 1994.
33
W. F. Hanson, L. W. Berkley, and M. Peterson, Off-axis beam quality
change in linear accelerator x-ray beam, Med. Phys. 7, 145149 1980.
34
B. J. Gerbi and F. M. Khan, Measurement of dose in the buildup region
using fixed-separation plane-parallel ionization chambers, Med. Phys.
17, 1726 1990.
35
F. M. Khan, V. C. Moore, and S. H. Levitt, Effect of various atomic
number absorbers on skin dose for 10-MeV x rays, Radiology 109,
209212 1973.
36
J. H. Hubbell, H. A. Gimm and I. verb, Pair, triplet, and total atomic
cross sections and mass attenuation coefficients for 1 MeV-100 GeV
photons in elements Z = 1 to 100, J. Phys. Chem. 4, 10231148 1980.
37
R. Mohan, C. Chui, and L. Lidofsky, Energy and angular distributions of
photons from medical accelerators, Med. Phys. 12, 592597 1985.
38
M. K. Islam and J. Van Dyk, Effects of scatter generated by beam-
modifying absorbers in megavoltage photon beams, Med. Phys. 22,
20752081 1995.
39
X. A. Li and D. W. O. Rogers, Electron mass scattering powers: Monte
Carlo and analytical calculations, Med. Phys. 22, 531541 1995.