Differences in reported backscatter factors, BSF, for low energy X-rays. A literature study
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Institutionen för medicin och vård
Avdelningen för radiofysik
Hälsouniversitetet
Differences in reported backscatter factors,
BSF, for low energy X-rays. A literature study
Carl A Carlsson
Department of Medicine and Care
Radio Physics
Faculty of Health SciencesSeries: Report / Institutionen för radiologi, Universitetet i Linköping; 66 ISSN: 1102-1799 ISRN: LIU-RAD-R-066 Publishing year: 1991 © The Author(s)
-1-
1991-01-30 ISSN 0348-7679
Differences in reported backscatter factors,
BSF, for low energy X-rays. A literature study.
Carl A Carlsson
Dept of Radiation Physics
University of Linköping
Report ULi-RAD-R-066 SSI P 587.90-2-
Differences in reported backscatter factors, nSF,
for low energy X-rays. A literature study.
Table of contents
l. Introduction .... 3
2. Application of nSF in a code of practice • ••• 4
3. The concept backscatter factor, nSF • ••• 4
4. The concept hal f value layer, HVL • ..• 5
5. Determination of HVL • ••• 6
5.1 Calculations •••• 6
5.2 Measurements .... 7
5.3 Common errors in HVL-measurements · .•. 8
6. HVL as a measure of radiation quaiity .•. 10
7. Review of literature about backscatter factors •.. 12
7.1 BJR Suppl No 17 (1983) ... 12
7.2 IAEA No 277 (1987) ... 12
7.3 Grosswendt (1984) ... 12
7.4 Klevenhagen (1989) ... 13
7.5 Chan and Doi (1981) · .• 14
7.6 Harrison et al (1990) · .• 15
7.7 Grosswendt (1990) · .• 15
7.8 Knight and Nahum (1990) ... 16
8. Discussion · .. 16
9. Conclusions ... 19
10. Acknowledgement ... 20
11. References · •. 21-3-
1. Introduction
This work was ini tia ted by the large discrepancy in published values of the
backscatter factor (BSF) as function of the half value layer (HVL) that
exists in a new code of practice (IAEA, 1987) compared to the earlier,
general ly used code (BJR, 1983). Change to the IAEA-code changes reported
values of absorbed dose with up to 10%. These deviations apply for X-rays
with HVL < 8 mm Al. Fig 1 shows BSF as function of field diameter with the
HVL as a parameter.
BSF
HVL, mm Al
1.4
------- === 4.0
//
/
/
/
/
/
/
1.3 /
/
/
/
/
/
/
/
/
I
1.2 / 1.0
I
I
- - - - --------
-- --- ---
/
/
~
/ /-
//
/ //
I 0.4
/
1.1 /
/
------=-'--=-=---=-=---~,
/
, -----
I
/1 ----::.-=:::- ------
/
I , 0.1
/'
------------------------------
1.0 '---- L-- L-- " ' -_ _- - - '
O 5 10 15 20
Field diameter, cm
Fig 1: Backscatter factor as a function of field diameter and half value
layer. From BJR (1983) with values from IAEA (1987) (Klevenhagen et al 1991)
inserted (dotted lines).-4-
The aim of this work is to try to explain the discrepancy between the two
sets of data in Fig 1. To reach this end, the concepts of BSF and HVL are
first scrutinized and the experimental and calculational methods repor ted in
the literature for their determination then critically reviewed.
2. Application of BSF in a code of practice
In radiation therapy with X-rays, 10-300 kV, tabula ted values of central
axis depth doses and BSF are mostly used instead of alaborious dose
planning. Relative absorbed doses are given by BJR (1983) as
[D(z)/D(O») • 100 as function of HVL and field diameter for different source
to surface distances (8SD). The absorbed dose in tissue at depth z, D(z), is
determined from a measurement of Ke,a i r at the en trance surface in the
absence of scattering material (patient, phantom) and is given by
D(z) = ~ Kc,alr
. (1)
D(O)
where (~en)t~ssue
p aIr
is a weighted average of the ratio of the mass energy
absorption coefficients for tissue and air (Carlsson and Alm Carlsson,
1990) .
3. The concept backscatter factor, BSF
The backscatter factor can be defined as the quotient between the absorbed
dose in the surface layer of a phantom to the absorbed dose in the same
material in the absence of scattering material (the phantom), that is
~
f ('fE + 'fE ) ~dE
p K
O 'P ,s c'D + s + s
BSF "'----=---..,,------- = (2)
ro lJ en K D
c,p P
f 'f dE
O E,p p-5-
Here is the energy fluence differentiated by photon energy E, indices p
~E
Ilen
and s stand for primary and secondary photons, respectively, and p is the
mass energy absorption coefficient for tissue (water).
The last step in Eq (2) is valid under charged particle equilibrium as then
collision kerma is equal to absorbed dose (Alm Carlsson 1985). Grosswendt
(1984) has expanded the definition to
Il
~ )(~) dE
E,s," p K
BSF" Il (3)
K ro
J ~ (~) dE
O E, P p K
where" stands for phantom material and K for detector material. Corrections
for determination of BSF = BSF~ from measured BSF~ are given. These
corrections are small for nearly tissue equivalent materials as air and LiF.
The determinations of BSF reviewed here have been made either
experimentally (Klevenhagen 1989; Harrison et al 1990) or by Monte Carlo
calculations (Chan and Doi 1981; Grosswendt 1984, 1990, Knight and Nahum
1990). In both experiments and calculations the detector should ideally be a
point detector. The finite volume necessary to get sufficient energy
impartation is preferably chosen as a thin plane detector as the dose
gradient varies faster with depth than orthogonally to the normally incident
central ray.
4. The concept half value layer, HVL
BJR (1983) defines HVL as "The thickness of a specified material which
attenuates the beam of radiation so that the exposure rate or absorbed dose
rate at a point is reduced by half. The contribution of all scattered
radiation, other than that which might be present in the beam initially, is
deemed to be excluded". For calculation of HVL it is defined implicitly in
Eq (4)
ro Ilen HVL
J (-p- )air ~E e- Il ' dE
O
(4)
OJ }len
J (--p--)air ~E dE
O-6-
where ~ is the linear attenuation coefficient for the attenuator material.
The other symbols are defined in Eq (2).
Comment 1. In all publications cited here, HVL is defined for the quantities
exposure X, air kerma K.
aIr
or air collision kerma Kc,alr
.• These three HVL:s
are practically identical.
Comment 2. The radiation must be monodirectional. Otherwise the thickness of
the attenuator is not well-defined. It is then necessary to modify the BJR-
definition above and also exclude "scattered radiation present in the \Jeam
ini tially".
5. Determination of HVL
S.l Calculations
With a known photon energy spectrum, RVL is calculated from Eq (4) by using
different attenuation thicknesses, d, instead of RVL so that the quotients
are both above and below 1/2. Interpolation gives HVL just as in measuring
HVL.
In early calculations of HVL it was common to regard the coherently
scattered radiation as primary radiation. The coherently scattered photons
were assumed to be strongly forward directed and to be detected as primary
photons in measurements of HVL. In old tabulations of attenuation
coefficients (Rubbeli 1969; Storm and Israel 1970) the total mass
attenuation coefficient was tabula ted both with and without the contribution
from coherent scattering. Calculations excluding the coherently scattered
photons from the primary ones agreed better with measurements using narrow
beam geometry (Carlsson 1963). Using modern cross sections for coherent
scattering in liquid water derived from Morin (1982), Persliden and Alm
Carlsson (1986) calculated the angular distribution of scattered photons
behind water slabs and found a minimum of scattering at angles smaller than
2_S o • There is reas on to suppose that this effect is more evident in
aluminium, supporting the idea that is possible to measure HVL with
detection of a negligible amount of incoherently and coherently scattered
photons.-7-
5.2 Measurements
To fulfil the requiremen ts of the definition of HVL, the measuremen ts mus t
be made in narrow beam (good) geometry to avoid deteetion of seattered
radiation from the attenuator (HVL-material). This eondition is best
, o
approaehed by a eollimation so that the opening angle, a, is less than ~ 4
and with the attenuator foils as elose to the souree as possible (Fig 2a).
ej ,/
,,
- fT
s ,,
II
M
A
°1 °2
Fig 2: Geometries used in HVL-measurements. S = radiation souree, M =
monitor, D = diaphragm (eollimator), A = attenuator, a = opening
angle of the beam, I = ionization ehamber (deteetor).
a) Good geometry, minimal opening angle, minimal distanee souree to
attenuator.
b) Geometry propos ed by Trout et al (1960). Three positions of the
attenuator - eollimator assembly is shown.
e) Geometry propos ed by IAEA (1987).-8- Trout et al (1960) made a thorough investigation of the performance of HVL- measurements and applied accurate corrections for the detected scattered radiation. A common practice to place the attenuator half-way between the source and detector originates presumably from this work. This, in turn, is based on the idea that the same area of the attenuator should be irradiated, independently of its position between source and detector (Fig 2b). With this presumption, the recommended position is correct. The presumption itself is, however, fundamentally wrong. The IAEA (1987) protocol has abandoned the idea of eons tant irradiated attenuator area (Fig 2c). The geometry is as in Fig 2a, but the introduction of a monitor, M in Fig 2c, means that the monitor may detect scattered radiation from the attenuator. In this case, the attenuator must be placed at a sufficient distance from both detector, I, and the monitor, M (Fig 2c). To get the same good result as with a measurement in the geometry in Fig 2a, the opening angle in Fig 2c has to be even smaller than in Fig 2a. Comment. If the monitor M in Fig 2c is in permanent use the discussion above is valid. On the other hand, if a monitor is introduced only for the HVL- measurement, it causes an extra filtration of the beam. The output and photon energy spectrum are changed as is the HVL. In this case, the monitor, M, should be placed outside the beam that reaches the detector, I (see Fig 2a). 5.3. Common er rors in HVL-measurements Fig 3 shows results from HVL-measurements in different geometries: (a) correct as in Fig 2a, (b) as in Fig 2b with scattered radiation detected by the ionization chamber, I, (c) as in Fig 2c with scattered radiation detected by the monitor M.
-9-
Ke, air (d)
Ke, air (O)
0.5
bJ
aj
ej
0.2
0.1 L - - - ' - - - - - - - ' - - - - - - ' - - - - 4 : - -
O 1 2 3
Attenuator thickness, d, mm
Fig 3: Results from HVL-measurements in different geometries.
a) good geometry (Fig 2a)
b) the ionization chamber, I, has detected scattered radiation
(Fig 2b)
c) the monitor, M, has detected scattered radiation (Fig 2c)
In geometry b), HVL is overestimated while in geometry c) it is
underestimated. The attenuator should ideally be perpendicular to all of the
beam. In a divergent beam, the outer rays pass longer distances through the
attenuator. They are thus more attenuated which should result in an
underestimate of the HVL. This effect is, however, mostly negligible, and
counteracts the effect of detection of scattered radiation by the ionization
chamber. Detection of scattered radiation is generally the more severe
problem in HVL-measurements.
Influences of ionization chambers, purity of the attenuator material and
measurements of their thickness on the accuracy of HVL-measurements are
discussed in a recent paper by Wagner et al 1990.-10-
6. HVL as a measure of radiation quality
The same value of HVL can be obtained with different energy spectra. Two
extremes are: a beam obtained with a thin window X-ray tube and no added
filter and a monoenergetic beam with the same HVL.
Fig 4 shows the energy distributions of the differential fluence, PE , and
the air collision kerma, Kc,alr,
. E' for a beam generated with 100 kV, 0.8 mm
added aluminium filter (Seelentag et al 1979). The monoenergy (26.4 keV)
that has the same HVL (1.61 mm Al) as the spectrum is inserted.
dKc,air ,and dCP __ , relative units
dE dE
1.0 ,-----.------..r,-r-,------.---.r----,,-----.----r-r-,---.----,--,---,
r ,
I l C 66
r l
I l
I \
I \
0.8 I
I
\
l
I l
I l
, I \
I \
0.6 I \
,,
I \
\
\
, I I
\
0.4 .,
nll!
I
l
\
\
\
'
,
I
III II
I'
\\ 'II"
\ ~\
, , I
0.2 \ \
...
I l
'I ~
...... l II
- ",II.
\
---
o L--=..L--L----LL.----l---L_L.----l---L_L-.:-y-_..1..---'------'-_-'--.-1
O 20 40 60 80 100 120 140
Photon energy, E, keY
Fig 4: The relative energy distributions of the quantities fluence (solid
line) and air collision kerma (broken line) for an X-ray beam 100 kV,
0.8 mm Al added filtration . (From Seelentag et al 1979). The
monoenergy (26.4 keV) with the same HVL as the spectrum is inserted.-11-
Fig 5 shows the result of applying Eq 4 to the spectrum in Fig 4 as weIl as
to the monoenergetic beam with the same HVL.
0.5 , "
"'"
"',',,
- HVL~",
0.2 ", ",
, "
"o
- - HVL -·I-----HVL2·---~---+-1
Attenuator thickness, d, mm
Fig 5: Attenuation curves obtained with the beams presented in Fig 4. Solid
line: 26.4 keV monoenergy, dashed line: 100 kV energy spectrum with
the same HVL. Second HVL, HVL 2 , is indicated for the spectral beam.
From Figs 4 and 5 one can conclude that the low energy photons are decisive
for the HVL-value of the beam while the high energy photons determine the
long distance penetration. The influence of the spectral shape on the values
of BSF will be discussed later.-12- 7. Review of literature about backscatter factors 7.1. BJR Suppl No 17 (1983) This dosimetry protocol (code of practice) gives BSF and percentage central axis depth dose data for different HVL, SSD and field areas. The tables in BJR (1983) are, for the HVL-ranges discussed here, identical with those in its earlier editions (BJR Suppl Nos 11 and 10 1972 and 1961) and contain material from BJR Suppl No 5 (1953). The tabula ted values in BJR (1983) are in many cases weighted averages of results from several investigators. The values in the later editions (1983, 1972, 1961) are basedon the use of smaller (thinner) ionization chambers than in the earlier one (1953), a fact that motivated the revision. The revised BSF-values generally exceed the previous ones. BJR (1983) states a scatter of ± 15% of their data in the HVL-range 0.01 to 0.7 mm Al. This scatter is explained by variations in measuring technique, phantom material and energy spectra. The BJR code gives correlated values of depth doses and BSF. 7.2. IAEA No 277 (1987) In this dosimetry protocol two methods for absolute determination of absorbed dose in the range 10-100 kV are mentioned. a) Use of a plane-parallel ionization chamber directly on the phantom and b) Measurement of air kerma free in air. In the latter case the re is a need for knowledge of the BSF •. BSF-values are tabulated in the code but there are no depth dos e values given since "the main dosimetric task in this photon energy range is the determination of the absorbed dose at the surface qf the phantom ". 7.3. Grosswendt (1984) The BSF-values tabula ted by IAEA (1987) are derived from this work. Grosswendt has calculated BSF with a Monte Carlo method. The calculations are weIl explained and seem reliable. Perhaps the use of a collision density estimator could increase the efficiency of the calculations (Persliden and Alm Carlsson, 1986). Grosswendt used eight published energy spectra for his calculations. The spectra were chosen so that no overlapping HVL:s occurred. These spectra from Seelentag et al (1979) are presented later in Grosswendt
-13- (1990). The paper (Grosswendt 1984) is more methodological than quantitative. The results are presented in small graphs showing general trends only. 7.4. Klevenhagen (1989) Klevenhagen has presented careful experiments with a thin, plane-parallel ionization chamber. He uses increased added filtration with increased tube potential. As a result no HVL-s overlap. The results from Klevenhagen agree weIl with those of Grosswendt (1984). Significant differences appear at the smallest field diameter and the highest HVL (1-4 mm Al). This is explained as an effect of the lateral extension of the ionization chamber being less close to a point on the central ray than the sphere simulated in the calculations. Other attempts to analyze the reason for small differences are not equally successful. For example, Klevenhagen cites Trout et al(1960) and states that Grosswendt has not accounted for the variation of beam quaiity (HVL) with changing field size. The measured HVL is said to decrease with decreasing field diameter. As has been discussed above, this effect is a systematic error in the HVL-measurement when using large field diameters (see Figs 2 and 3). There is, however, a variation of HVL over the field area due to attenuation of the bremsstrahlung in the anode. At the cathode side of the field, the HVL is lower and the photon fluence higher than on the central ray while the opposite is true (higher HVL and lower photon fluence) at the anode side (heel effect). HVL-measurements in different directions from a diagnostic X-ray tube have been repor ted by Carlsson (1963) and energy spectra in different directions by Svahn (1977). The heel effect is less pronounced using X-ray tubes with larger anode angles than those used in diagnostic units (10-20 0 ). The tube used for Grosswendt's calculations has an anode angle of 40 0 • Klevenhagen also tentatively states that his results are 1 - 1.5% higher than those of Grosswendt due to differences in their SSD:s used (Grosswendt 100 cm, Klevenhagen 15 and 25 cm). Ile claims that the more. divergent beam should give a higher BSF due to alarger phantom volume being involved in scattering interactions. On the contrary, it can be argued that the BSF should be lower with a shorter SSD as the same number of scatter interactions are then distributed over alarger volume. That is, the scattered photons have lo~ger distances to the point at zero depth on the
-14- central ray. Another possible explanation of these small differences could be that Klevenhagen in general produces his HVL:s with lower filtration and higher tube potential than Grosswendt. This will be further discussed below. Another criticism is that Klevenhagen discusses the energy distribution of scattered photons by means of effective energies. The effective energy, hveff , defined from ~(hveff) = ~~~ is a doubtful concept. The energy spectra of primary and scattered photons need be fully appreciated (Alm Carlsson and Carlsson, 1984; Carlsson and Alm Carlsson, 1990). 7.5. Chan and Doi (1981) Chan and Doi repor t Monte Carlo calculations of BSF for mammography units with both a Mo-anode (0.03 mm added Mo-filter) and a W-anode (1 mm added AI- filter). They used a geometry with all photons impinging perpendicularly on the phantom at the same point (pencil beam) and register all energy impartations throughout,a laterally infinite phantom. This approach is equivalent to calculate depth doses in the phantom irradiated with a broad beam of perpendicularly incident photons. The backscatter was received by extrapolating the depth doses to zero depth. They report BSF for four phantom materials, water, fat, 50% water + 50% fat, and Lucite. Their results for water agree weIl with those of Klevenhagen for a 20 cm field diameter. Chan and Doi repor t an interesting simulation of how a LiF-dosemeter (dimensions 3.2 x 3.2 x 0.89 mm 2 ) measures BSF. Tltey calculate the average absorbed dose in the dosemeter with and without the phantom present. The resulting BSF is significantly lower than that calculated without a dosemeter (up to 7% with a water phantom). These calculated "BSF:s" agree weIl with reported measurements using thermoluminescent LiF of the given dimensions. Their calculations of the average absorbed dose in the dosemeter are not presented in detail. It is somewhat astonishing that, in their calculations, use of a phantom equivalent dosemeter of the same dimensions as the LiF-dosemeter resuIts in approxilllately the same (only slightly . sIllaller) underestilIlate of the JlSF even if LiF has 2.5 to 3.5 higher density than the different phantom materials.
-15-
7.6. Harrison et al (1990)
In order to avoid the large perturbation caused by the displacement of
phantom material by the relatively large air volume in an ionization chamber
(Liden 1961), Harrison et al performed BSF-measurements with
thermoluminescence (TL) dosemeters. They used chips of Li ZB40 with
7
dimensions 3.Z x 3.Z x 0.9 mm 3 which, they assumed, would not significantly
perturb the radiation field. Their precision in measured BSF is low. The
given standard deviation is about 4%. This is caused by alow-precision
TL-technique. TL-measurements with a precision of 0.3% has earlier been
repor ted (Carlsson, Mårtensson and Alm Carlsson 1968; Alm Carlsson 1973). It
should also have been advantageous to use much thinner TL-dosemeters to
avoid the perturbation effect repor ted by Chan and Doi (1981). The scatter
of their results is in general large. For one of the beams used, they repor t
measured and calculated HVL:s of 0.93 and 1.77 mm Al, respectively. In
general, their BSF-measurements agree better with IAEA (1987) than with BJR
(1983).
7.7. Grosswendt (1990)
Here, Grosswendt presents BSF-values for a great number of radiation
qualities, SSD:s and field diameters. Contrary to the author's earlier work
(Grosswendt 1984) this work is quantitative. The BSF-values are
systematicallY presented in a table, that is a valuable resource for
everyone with questions about BSF. Besides an increased number of energy
spectra, three monoenergies (ZOO, 400 and 661) are introduced. The BSF-
values are restricted to water as both detector and phantom. Coll~sion
kerma, Kc ' is used instead of kerma, K, in the definition of BSF. This is to
prefer as Kc is identical to absorbed dos e in charged particle equilibrium
and the absorbed dos e is the important quantity. A further advantage with
this definition is that it uses mass energy absorption coefficients instead
of mass energy transfer coefficients (Eq Z instead of Eq 3). The former are
~he ones given in modern tabulations. The Monte Carlo technique has als o
been changed. Now the absorbed dose on the central ray at zero depth is
. calculated byestimating the spectral fluence of photons traversing a small
area on the surface. Earlier (Grosswendt 1984), spheres of various radii
. were used and values extrapolated to zero radius. Eq Z in Grosswendt 1990-16-
1
dE dE (A I cosa 1-
d4i = dn ) (Sa)
could more correctly be written
dEdeda.L (Sb)
where the area dal is perpendicular to the photon's direction of motion a.
If da is an area element on the surface traversed by dN photons then da~ =
dalcosal. In the estimat~ of ~: ' finite values of the differentials must be
used and an integration over all angles a be performed.
A curious detail is that the aU thor has here accepted the cd ticism by
Klevenhagen that the HVL.should increase when the field area increases (see
the above review, Section 7.4). Grosswendt uses, however, correct HVL-values
calculated for narrow beam geometry (Eq 4).
7.8. Knight and Nahum (1990)
These authors have with Monte Carlo technique calculated values of BSF for
monoenergetic photons and derived BSF for spectra by weighting. Their
results, repor ted in a graph in this abstract, agree weIl with those of
Grosswendt 1990.
8. Discussion
In summary, among modern and old literature about backscatter factors
Grosswendt's (1984, 1990) results seem the most reliable. The experiments by
Klevenhagen (1989) are also very convincing at least with large field
diameters. Other calculations and experiments verify the results of these
two authors. All authors use the HVL for describing radiation quaiity and
are aware of its limitations. Nevertheless, they use a series of spectra
that is produced by increasing added filtration and tube potential
simultaneously. No one has used different spectra with nearly the same HVL.
In Fig 6, results of Grosswendt (1990) and Klevenhagen (1989) are compared.
Fig 6a shows the added filtrations, Fig 6b the tube potentials and Fig 6c
the backscatter factors. All these quantities are presented aS function of
HVL. From Fig 6 a) and b) it is evident that for the same HVL Klevenhagen-17-
uses less added filter material and instead higher tube potential than
Gro.sswendt (at least for HVL > 1 mm Al). Fig 6 c shows that the BSF:s from
Klevenhagen are larger than those from Grosswendt at the large field
diameter (20 cm) and vice versa at smaller field diameters (5 and 2 cm).
Addod Al filt or, mm
a
4
3
2
. .0- 0'-'
o 2 4 5
Tuba potontlal, kV b
140
120
100
80
60
40- [7
20 ,!
Ol-_ _~_ _~ ~ ·-18-
An explanation of Fig 6c could be: because Klevenhagen's beams have both
lower and higher energies than those of Grosswendt with the same IIVL, they
produce seattered photons of higher energies and the BSF:s saturate at
larger field diameters. If, on the other hand, the field area is small, the
high energy scattered p~otons have a large probability for escaping th~
phantom outside the entranee field. The low energy photons «20 keY)
eontribute little to the BSF (Persliden and Alm Carlsson 1984). Maybe,
Klevenhagen's BSF-values for small field diameters are in fact eorrect and
not too low, as he himself suspected because of the geometry of his
ionization chamber (see Seetion 7.4).
The differences in the BSF-values presented in Fig 6 for equal HVL are
examples on the weakness of HVL as a measure of radiation quaiity. The
spectral distribution of air collision kerma, Kc,a i r, E. from a known fluenee
speetrum ~E is given by (ef Eq 4)
Ilen
Kc,alr,
• E = (-p-)air • E • ~
E
(6)
Ilen
The product (---).p all'
. E contains two terms, one deereasing the other
increasing with inereasing photon energy, E. As seen in Fig 7, the product
deereases very fast with increasing energy at energies up to 30 keY, has a
broad minimum around 70 keY and increases slowly above 120 keY. As a result
Kc,a i r, E is more dominated by low energy photons than is ~E' as illustrated
in Fig 4. Crudely speaking, the low energy photons of a diagnostic X-ray
sp~ctrum determine the HVL and the high energy photons the penetration power
and energy eontent of the beam (Carlsson and Alm Carlsson, 1990).
Going back to Fig 1, the differences between BJR (1983) and IAEA (1987)
could partly be explained as caused by using lower filtration and higher
tube potentials for the BJR BSF-values than for those of rAEA. The former
could be better than suggested by all the modern investigators. It is, of
course, time to exchange these old BSF-values as they are compilations of
many investigators and as relatively large ionization chambers have been
used. BJR give both BSF and pereentage depth doses; IAEA means that depth
doses are of little interest with these spectra. This is in accordance with
the practice at many hospitals. A new code of practice, Klevenhagen et al
(1991) presents BSF derived from Grosswendt (1990), supported by data from-19-
both Knight and Nahum (1990) and Klevenhagen (1989). These data are, in
fact, used instead of those from IAEA (1987) for producing Fig 1.
2
Ilen air' E -1J1,-. keV
lp '''9
10
9
8
7
6
5
4
3
2
1
oL_-=:::::::::=====::;:::======-
O 50 keV 100 150 E,
Fig 7: The product of the mass energy absorption coefficient, ~en/p, and
photon energy, E, as function of photon energy.
9. Conclusions
The new backscatter factors given by Grosswendt (1990) and Klevenhagen
(1989) seem reliable taking into account the energy spectra they are derived
from. With knowledge of filter, tube potential, HVL, anode material and
anode angle one can try to find the actual energy spectrum from the spectral
catalogue (Seelentag et al 1979). If possible, one can, for a given HVL,
change filter and tube potential to the values given by Grosswendt (1990)
and use his BSF-values with confidence.-20- The use of BSF from Grosswendt or IAEA together with depth doses from BJR JI1~Y, however, be problema tic. It should be interesting to get BSF-values for beams with very broad spectra, e.g., from X-ray tubes with Be-windows and no added filter and also for the other extreme, monoenergetic radiation. The problem of the perturbation of a TL-dosemeter on the determination of BSF (Chan and Doi 1981) should benefit from further investigations. Experimental determinations of BSF suffer of insufficient knowledge of the energy distribution of the scattered radiation. Publication of such results from Monte Carlo calculations should be valuable. 10. Acknowledgement This work was initiated by Dr Jonas Karlberg and has been supported by grants from the Swedish National Institute of Radiation Protection, SSI.
-21-
11.· References
Alm Carlsson G (1973). Dosimetry at interfaces. Acta Radiol Suppl 332,
Stockholm.
Alm Carlsson G (1985). Theoretical basis for dosimetry. In The dosimetry of
ionizing radiation Vol I, Eds Kase, Bjärngard, Attix, Academic
Press, Orlando, pp 1-75.
Alm Carlsson G and Carlsson C A (1984). Effective energy in diagnostic
radiology. A critical review. Phys Med Biol 29, 953-958.
BJR (1953). Central axis depth dos e data. Brit J Radiol Suppl No 5.
BJR (1961). Depth dose tables for use in radiotherapy. Brit J Radiol Suppl
No 10.
BJR (1972). Central axis depth dos e data for use in radiotherapy. Eds
Cohen, Jones and Greene. Brit J Radiol Suppl No 11.
BJR (1983). Central axis depth dose data for use in radiotherapy. Brit J
Radiol Suppl No 17.
Carlsson C (1963). Determination of integral absorbed dose from exposure
measurements. Acta Radiol Ther, Phys Biol !' 433-458.
Carlsson C A and Alm Carlsson G (1990). Dosimetry in diagnostic radiology
and computerized tomography. In The dosimetry of ionizing
radiation. Vol III. Eds Kase, Bjärngard and Attix. Academic
Press, San Diego, pp 163-257.
Carlsson C A, Mårtensson B K A and Alm Carlsson G (1968). High precision
dosimetry using thermoluminescent LiF. Proc 2nd Int Gonf Lum
Dosimetry. USAEG Rept CONF-680920.
Chan H-P and Doi K (1981). Monte Carlo simulation studies of backscatter
factors in mammography. Radiology 139, 195-199.-22-
Grosswendt B (1984). Backscatter factors for X-rays generated at voltages
between 10 and 100 kV. Phys Med Biol 29, 579-591.
Grosswendt B (1990). Dependence of the photon backscatter factor for water
on source-to-phantom distance and irradiation field size. ,Phys
Med Biol 35, 1233-1245.
Harrison R M, Walker C and Aukett R J (1990). Measurement of backscatter
factors for low energy radiotherapy (0.1 - 2.0 mm Al HVL) using
thermoluminescence dosimetry. Phys Med Biol 35, 1247-1254.
Hubbell J H (1969). Photon cross sections, attenuation coefficients and
energy absorption coefficients from 10 keY to 100 GeV. Report
NSRDS-NBS 29. National Bureau of Standards, Washington.
IARA (1987). Absorbed dose determination in photon and electron beams. An
international code of practice. Techn Rep Ser No 277.
International atomic energy agency, Vienna.
Klevenhagen S C (1989). Rxperimentally determined backscatter factors for X-
rays generated at voltages between 16 and 140 kV. Phys Med Biol
34, 1871-1882.
Klevenhagen S C, Aukett R J, Burns J R, Harrison R M, Knight R T, Nahum A R
and Rosser K R (1991). Report of the Institute of physical
sciences in medicine working party on low- and medium energy X-
ray dosimetry. Submitted to Phys Med Biol.
Knight R T and Nahum A E (1990). Improved backscatter fac\ors for low energy
X-rays. Abstract.
Liden K V H (1961). Errors introduced by finite size of ion chambers in
depth dos e measurements, pp 161-172 in Selected topies in
radiation dosimetry (Proc lnt Con f Vienna), International
atomic energy agency, Vienna.
Morin LRM (1982). Molecular form factors and photon coherent scattering
cross sections of water. J Phys Chem Ref Data ~, 1091-1098.-23-
Persliden J and Alm carlsson G (1984). Energy imparted to water slabs by
photons in the energy range 5-300 keY. Calculations using a
Monte Carlo photon transport model. Phys Med Biol 29, 1075-
1088.
Persliden J and Alm Carlsson G (1986). Calculation of the small-angle
distribution of scattered photons in diagnostic radiology using
a Monte Carlo collision density estimator. Med Phys 13, 19-24.
Seelentag V V, Panzer V, Drexler G, Platz L, Santner F (1979). A catalogue
of spectra for the calibration of dosemeters. GSF Bericht 560.
Gesellschaft fUr Strahlen - und Umweltforschung MBH, MUnchen.
Storm E and Israel H I (1970). Photon cross sections from 1 keY to 100 MeV
for elements Z = l to Z = 100. Nuclear Data Tables A7, 565-681.
Academic Press, New York and London.
Svahn G (1977). Diagnostic X-ray spectra. A study of the effects of high
tension ripple, large x-ray tube currents, extra-focal
radiation and anode angulation with Ge (Li)-spectrometry.
Thesis, Radiation Physics Dept, University of Lund, 'X-221 85
Lund, Sweden.
Trout D E, Kelley J P and Lucas A C (1960). Determination of half-value
layer. Radiology 84, 729-740.
Wagner L K, Archer B R and Cerra F (1990). On the mesurement of half-value
layer in film-screen mammography. Med Phys 17, 989-997.You can also read
