ARTICLE IN PRESS

Applied Radiation and Isotopes 61 (2004) 739–743

Calculations of cellular microdosimetry parameters

for alpha particles and electrons
C.J. Tunga,*, C.S. Liua,b, J.P. Wangc, S.L. Changa
a
Department of Nuclear Science, National Tsing Hua University, Hsinchu 300, Taiwan
b
Department of Nuclear Medicine, Taipei Veterans General Hospital, Taipei 112, Taiwan
c
National Synchrotron Radiation Research Center, Hsinchu Science Park, Hsinchu 300, Taiwan

Abstract

The cellular microdosimetry parameters including the cellular S-value and the single-event specific energy
distribution for alpha particles and electrons are important in radiation dosimetry and biology. These parameters
may be used to determine the relative biological effectiveness of radiations in the boron neutron capture therapy. In the
present work, such parameters were calculated for different source to target region combinations, i.e. cell surface,
cytoplasm, nucleus and cell. Calculations were made using a semi-analytical model that simulated the emission of alpha
particles or electrons by the Monte Carlo method and calculated the energy imparted to the target volume by the
analytical method. Delta particle equilibrium and partial delta particle equilibrium were applied to alpha particles and
electrons, respectively. Range–energy relations were employed to determine the incident and emerging energies of the
primary particles. For electrons, the fraction in the energy loss resulting from the generation of bremsstrahlung and
high-energy secondary electrons was estimated. The energy loss straggling of electrons entering and leaving a target
volume was also estimated. Calculated cellular S-values were compared to corresponding data of the MIRD
Committee. Calculated single-event specific energy distributions were also compared to results calculated using the
Penelope code.
r 2004 Elsevier Ltd. All rights reserved.

Keywords: Microdosimetry; Cellular S-value; Specific energy

1. Introduction source and target regions in the cell. The radiation

The dosimetry in boron neutron capture therapy
(BNCT) involves both the high and low linear energy
transfer (LET) particles. The relative biological effec-
tiveness (RBE) of these particles is determined by the
microdosimetric distribution of the energy imparted to
the matter in cellular volumes. Stochastic quantities,
including the lineal energy and the specific energy
(ICRU, 1998), and nonstochastic parameters, such as
the cellular S-value (Goddu et al., 1997), are used in
cellular microdosimetry. Their values depend on the
*Corresponding author. Tel.: +886-3-5727300; fax: +886-3-
5718649.
E-mail address: cjtung@mx.nthu.edu.tw (C.J. Tung).
source is usually assumed to be uniformly distributed in
one of the regions of the cell: throughout the cell (C),
cytoplasm (Cy), cell surface (CS), or cell nucleus (N).
For instance, studies of the use of lipiodol as a drug
carrier to transport boron to hepatoma cells (Chou et al.,
1995) showed that these cells were able to take up
lipiodol actively and selectively by endocytosis with
prolonged cytoplasm retention for as long as the
cells lived. The biological target is generally assumed
as the cell nucleus or the entire cell. In the case of source
to target region combination N’Cy, the microdosi-
metric parameters are contributed by crossers and
stoppers. Whereas, these parameters are contributed
by starters and insiders for source to target combination
N’N.

0969-8043/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.apradiso.2004.05.003
 ARTICLE IN PRESS
740 C.J. Tung et al. / Applied Radiation and Isotopes 61 (2004) 739–743
The cellular microdosimetric parameters may be
calculated using event-by-event Monte Carlo simula-
tions (Turner et al., 1988; Wilson et al., 1988; Wilson
and Nikjoo, 1999). However, these simulations are
computationally expensive, and cross sections needed
to transport the simulations are quite limited. Therefore,
some analytical and mixed simulation models were
developed. For instance, the Medical Internal Radiation
Dose (MIRD) Committee of the Society of Nuclear
Medicine developed an analytical model (Goddu et al.,
1997) that involved the continuous slowing down
approximation (CSDA) to calculate the cellular S-values
for alpha particles and electrons. The PENELOPE code
(Stewart et al., 2002), on the other hand, applied a mixed
algorithm, i.e. simulations of hard electron collisions by
an event-by event way and soft electron collisions by the
multiple scattering theory, in microdosimetric calcula-
tions. This paper presents a semi-analytical model that
simulates the emission of alpha particles or electrons by
the Monte Carlo method and calculates the energy
imparted to the target volume by the analytical method.
The present model used the delta particle equilibrium for
alpha particles and the partial delta particle equilibrium
for electrons to calculate the energy imparted to a
cellular target volume. Range–energy relations were
employed to determine the incident and emerging
energies of the primary particles. For electrons, the
fraction in the energy loss resulting from the generation
of bremsstrahlung and high-energy secondary electrons
was estimated using data of the restricted collisional
stopping power and the radiative stopping power. The
energy loss straggling of electrons entering and leaving a
target volume was evaluated by applying a Gaussian
distribution with the mean square energy loss data. This
model was used to calculate the microdosimetric
quantities such as the cellular S-values and the single-
event specific energy distribution for alpha particles and
electrons. The application of this model to cellular
microdosimetry of BNCT is feasible.
2. Theory
Fig. 1 illustrates the model of a cell consisting of two
concentric spheres making up the N (dark inner sphere),
Cy (concentric shell), and CS (outer surface). Assuming
a primary particle (arrow) emits from the CS (the source
region) and penetrates through the N (the target region),
the energy imparted to the target volume (ICRU, 1998),
e, may be written as
e ¼ Ein;p þ Ein;s 2Eout;p 2Eout;s 2Eout;b ;
where Ein;p is the energy of primary particle incident on
the volume, Ein;s is the energy of secondary particles
(dots) entering the volume, Eout;p is the energy of
primary particle emerging from the volume, Eout;s is
Eout,s
Ein,p
Eout,p
Ein,s
N
Eout,b
CS Cy
Fig. 1. A schematic diagram for a radiation particle emitted
from the cell surface (CS), travelled through the cytoplasm (Cy)
and penetrated into the nucleus (N). The energy of primary
particle (arrow) incident on the nucleus, Ein;p ; the energy of
secondary particles (dots) entering the nucleus, Ein;s ; the energy
of primary particle emerging from the nucleus, Eout;p ; the energy
of secondary particles leaving the nucleus, Eout;s ; and the energy
of bremsstrahlung escaped from the nucleus, Eout;b ; are all
labeled.
the energy of secondary particles leaving the volume,
and Eout;b is the energy of bremsstrahlung escaped from
the volume.
In the case of alpha particles, the bremsstrahlung may
be disregarded and a delta particle equilibrium (Carls-
son, 1985) may be adopted by assuming
Ein;s ¼ Eout;s : ð2Þ
Thus the energy imparted to the volume is given by
e ¼ Ein;p 2Eout;p : ð3Þ
In the case of electrons, the partial delta particle
equilibrium (Carlsson, 1985) may be adopted, i.e.
Ein;soD ¼ Eout;soD ; ð4Þ
where Ein;soD and Eout;soD are, respectively, the incom-
ing and outgoing secondary electron energies below a
cutoff value D: Therefore, the energy imparted to the
volume is given by
e ¼ Ein;p 2Eout;p þ Ein;s>D 2Eout;s>D 2Eout;b : ð5Þ
The radiation source is assumed to be uniformly
distributed in the source region. The particle emitting
point and the particle path are determined by random
ð1Þ
numbers in the Monte Carlo simulations. If the path is a
crosser, both incident and emerging energies of the
primary particle, Ein;p and Eout;p ; are calculated using
range–energy relations, RðEÞ (Watt, 1996; Berger et al.,
2000; ICRU, 1993; ICRU, 1989). Let din and dout be,
 ARTICLE IN PRESS
C.J. Tung et al. / Applied Radiation and Isotopes 61 (2004) 739–743 741

Table 1
Cellular S-values for alpha particles (RC=10 mm and RN=5 mm)

Alpha S(C’C) (Gy/Bq s) S(C’CS) (Gy/Bq s) S(N’N) (Gy/Bq s) S(N’Cy) (Gy/Bq s) S(N’CS) (Gy/Bq s)
energy (MeV)
MIRD Present MIRD Present MIRD Present MIRD Present MIRD Present

1 — 3.14E02 — 1.68E02 — 2.01E01 — 1.48E02 — 2.57E04

1.47 — 4.17E02 — 2.31E02 — 2.31E01 — 2.97E02 — 4.54E03
1.78 — 4.56E02 — 2.63E02 — 2.26E01 — 3.97E02 — 1.11E02
2 — 4.76E02 — 2.82E02 — 2.15E01 — 4.50E02 — 1.68E02
3 4.40E02 4.41E02 3.02E02 3.02E02 1.57E01 1.57E01 4.06E02 4.07E02 2.51E02 2.51E02
4 3.34E02 3.36E02 2.33E02 2.33E02 1.25E01 1.25E01 2.99E02 2.99E02 1.74E02 1.73E02
5 2.74E02 2.75E02 1.87E02 1.87E02 1.05E01 1.05E01 2.44E02 2.44E02 1.37E02 1.37E02

respectively, the distances between the particle source
and its intersecting points to the volume. The CSDA
incident and emerging energies are given by
RðEin;p Þ ¼ RðE0 Þ2din ð6Þ
and
RðEout;p Þ ¼ RðE0 Þ2dout ; ð7Þ
where E0 is the emitted energy of the primary particle. If
the path is a starter or stopper, Ein;p ¼ E0 or Eout;p ¼ 0:
If the path is an insider, Ein;p ¼ E0 and Eout;p ¼ 0: In the
case of electrons, the exclusion of bremsstrahlung and
high-energy secondary electrons from the energy im-
parted, i.e. Ein;s>D  Eout;s>D  Eout;b in Eq. (5), is esti-
mated using data on the restricted collisional stopping
power and the radiative stopping power (Watt, 1996;
Berger et al., 2000; ICRU, 1993; ICRU, 1989). Due to
the randomness of inelastic interactions and energy
losses, the incident and emerging energies of particles
may deviate from their average values. To the first-order
approximation, this energy loss straggling (Tung et al.,
1986; Tung and Kwei, 1991) is a Gaussian distribution
with its standard deviation determined by the mean
square energy loss. To consider the energy loss strag-
gling, the indent and emerging energies are characterized
by Ein;p 7DEin;p and Eout;p 7DEout;p ; respectively. For an
alpha particle of crossers, for instance, its contribution
to the cellular S-value of source region rh and target
region rk, i.e. Sðrk ’rh Þ; is given by
Ein;p  Eout;p
Sðrk ’rh Þ ¼ ; ð8Þ
mk
where mk is the mass of the target region. The standard
deviation of the cellular S-value is given by
 1=2
2 2
Ein;p þ Eout;p
DS ðrk ’rh Þ ¼ : ð9Þ
mk
Similar expressions for electrons are apparent. Similarly,
equations for the lineal energy, y, and the single-event
specific energy, z, can be derived.
3. Results and discussion
Table 1 lists the cellular S-values for alpha particles
calculated in the present work and compares them with
data of the MIRD (Goddu et al., 1997). Tabulations are
for different source to target regions, i.e. cell to cell
(C’C), cell surface to cell (C’CS), nucleus to nucleus
(N’N), cytoplasm to nucleus (N’Cy), and cell surface
to nucleus (N’CS). The radii of the cell and cell
nucleus are 10 and 5 mm, respectively. Because of the
short range of alpha particles, cellular S-values are large
for (N’N) and small for (N’CS) at low energies. In
all cases the difference between present work and MIRD
is negligibly small. To apply the cellular S-values to
BNCT, results for alpha particles of 1.47 (94%) and
1.78 MeV (6%) (Hsu et al., 2003) are included in the
table. It is noted that the units, Gy Bq1 s1, are for
radionuclides. In BNCT, Bq s should be interpreted as
nuclear reaction rather than nuclear decay.
Table 2 gives a similar comparison of the cellular S-
values for electrons calculated in the present work and
by the MIRD. Here the radii of the cell and cell nucleus
are also 10 and 5 mm, respectively. For electron energies
of 1 and 10 keV, cellular S-values for (N’CS) are zero
since electron ranges are smaller than the shortest
source–target distance. Again, the difference between
present work and MIRD is negligibly small. However,
the standard deviation associated with the cellular S-
value was not evaluated by the MIRD. This deviation,
resulting from the energy loss straggling, was calculated
in the present work and included in Table 2 (values in
parentheses). The energy loss straggling made significant
contribution to the cellular S-values for electrons at,
especially, high energies. It is noted that the S-values for
electrons are two to three orders of magnitude smaller
than the S-values for alpha particles. This indicates that
alpha particles are more effective than electrons in
imparting energies to the target volume of micrometer
size and thus resulting higher RBEs than electrons.
Fig. 2 shows a plot of the single-event distribution,
f1 ðzÞ; on the specific energy, z, for a 100 keV electron
source located on the cell surface of a 10 mm-diameter
 ARTICLE IN PRESS
742 C.J. Tung et al. / Applied Radiation and Isotopes 61 (2004) 739–743

Table 2
Cellular S-values for electrons (RC=10 mm and RN=5 mm)

Electron S(C’C) (Gy/Bq s) S(C’CS) (Gy/Bq s) S(N’N) (Gy/Bq s) S(N’Cy) (Gy/Bq s) S(N’CS) (Gy/Bq s)
energy
(keV)
MIRD Presenta MIRD Presenta MIRD Presenta MIRD Presenta MIRD Presenta

1 3.82E05 3.81E05 1.91E05 1.91E05 3.04E04 3.04E04 2.55E07 2.43E07 0.00E+00 0.00E+00
(7.73E7) (4.38E7) (8.69E6) (3.31E6) (0.00E+0)
10 3.35E04 3.34E04 1.76E04 1.75E04 2.30E03 2.29E03 1.10E04 1.09E04 0.00E+00 0.00E+00
(2.84E5) (1.65E5) (3.17E4) (1.24E4) (0.00E+0)
100 1.20E04 1.20E04 8.01E05 8.04E05 4.75E04 4.78E04 1.05E04 1.05E04 5.71E05 5.66E05
(1.54E4) (1.26E4) (8.71E4) (5.70E4) (5.03E4)
1000 4.96E05 4.98E05 3.31E05 3.32E05 1.98E04 1.99E04 4.35E05 4.35E05 2.34E05 2.32E05
(1.54E4) (1.26E4) (8.73E4) (5.72E4) (5.05E4)
a
Values in parentheses are the standard deviations of the cellular S-values resulting from the energy loss straggling.

1000 distribution of the present approach may be comparable

to that of the purely analog Monte Carlo simulations.
The difference between curves D and A is due to the
800
C effects of elastic scatterings. The single, large-angle
scattering (Chen et al., 1995) causes electrons that move
f1(z) (Gy-1)

600
A D
400
B
A
200 C
D even broader and shifting to small specific energies.
0 approach to the calculations of the single-event dis-
10-5 10-4 10-3
z (Gy)
Fig. 2. A plot of the single-event distribution, f1 ðzÞ; as a
function of specific energy, z, for 100 keV electrons emitted
from the CS (cell diameter=10 mm) and imparted energy to the
N (nucleus diameter=7.5 mm). Curves B,C and D are results
calculated in the present work using, respectively, Eq. (3), (5)
and the energy loss straggling. The corresponding data
calculated using the Penelope code are also plotted (curve A)
for comparison.
cell. The cell nucleus of 7.5 mm-diameter is the target site.
Curves B,C and D are results calculated in the present
work using, Eq. (3), (5) and the energy loss straggling.
Curve A is the corresponding data calculated using the
Penelope code (Stewart et al., 2002). A comparison
between curves B and C shows that the exclusion of
bremsstrahlung and high-energy secondary electrons
from the energy imparted makes the single-event
distribution shifting toward small specific energies. A
comparison between curves C and D indicates that the
inclusion of the energy loss straggling makes the single-
event distribution broader. Further, a comparison
between curves D and A reveals that the single-event
B away from the target site changing their directions and
passing through the site. The multiple, small-angle
scatterings (Chen et al., 1992) cause electrons to travel
tortuous paths rather than straight paths inside the
target. These effects make the single-event distribution
Despite of these effects, the application of the present
10-2 tribution, frequency-mean specific energy per event and
hit probability in cellular microdosimetry is feasible.
4. Conclusions
A semi-analytic approach was developed to calculate
the microdosimetric quantities such as the cellular S-
values and the single-event specific energy distribution
for alpha particles and electrons. This approach applied
the delta particle equilibrium for alpha particles and the
partial delta particle equilibrium for electrons to
calculate the energy imparted to a cellular target volume.
Range–energy relations were used to determine the
incident and emerging energies of the primary particles.
The fraction in energy loss resulting from the generation
of bremsstrahlung and high-energy secondary electrons
was estimated using data of the restricted collisional
stopping power and the radiative stopping power. The
energy loss straggling of electrons entering and leaving a
target volume was evaluated by applying a Gaussian
distribution with the mean square energy loss data.
Cellular S-values for alpha particles calculated in the
present work are in excellent agreement with those by
the MIRD. Cellular S-values for electrons of the present
 ARTICLE IN PRESS
C.J. Tung et al. / Applied Radiation and Isotopes 61 (2004) 739–743
work are also in very good agreement with data of the
MIRD. In addition, the present work evaluated the
uncertainties associated with the cellular S-values that
were not considered by the MIRD. The energy loss
straggling, responsible for such uncertainties, is also
important in the determination of the single-event
specific energy distribution for electrons. A comparison
of the single-event distribution calculated in the present
work and by the Monte Carlo Penelope code shows that
the present approach is feasible to calculate the single-
event distribution. These cellular microdosimetric para-
meters can be used to determine the RBE of radiations
in the BNCT (Hsu et al., 2003; Amols and Reinstein,
1994).
Acknowledgements
This research was supported by the National Science
Council of the Republic of China.
References
Amols, H.I., Reinstein, L.E., 1994. Stochastic considerations in
BNCT: microdosimetry and cell survival. In: Zamenhof, R.,
Solares, G., Harling, O. (Eds.), Topics in Dosimetry &
Treatment Planning for Neutron Capture Therapy. Ad-
vanced Medical Publishing, Madison, Wisconsin.
Berger, M.J., Coursey, J.S., Zucker, M.A., 2000. Stopping power
and range tables for electrons, protons, and helium ions.
http://physics.nist.gov/physrefdata/star/text/contents.html.
Carlsson, G.A., 1985. Theory basis for dosimetry, In: Kase,
K.R., Bjarngard, B.E., Attix, F.H. (Eds.), The Dosimetry
of Ionizing Radiation, Vol. I. Academic Press, Inc.,
Orlando, FL.
Chen, Y.F., Kwei, C.M., Tung, C.J., 1992. Electron inelastic
mean free paths versus attenuation lengths in solids. J. Phys.
D: Appl. Phys. 25, 262–268.
Chen, Y.F., Kwei, C.M., Su, P., 1995. Angular distribution of
electrons elastically backscattered from non-crystalline solid
surfaces. J. Phys. D: Appl. Phys. 28, 2163–2169.
743
Chou, F.I., Fang, K.C., Chung, C., Lui, W.Y., Chi, C.W., Liu,
R.S., Chan, W.K., 1995. Lipiodol uptake and retention by
human hepatoma cells. Nucl. Med. Biol. 22, 379–386.
Goddu, S.M., Howell, R.W., Bouchet, L.G., Bolch, W.E.,
Rao, D.V., 1997. MIRD cellular S values. Society of
Nuclear Medicine, 1850 Samuel Morse Drive, Reston, VA
20190-5316.
Hsu, F.Y., Tung, C.J., Watt, D.E., 2003. Microdosimetric
spectra of the THOR neutron beam for boron neutron
capture therapy. Radiat. Prot. Dosin. 104, 121–126.
ICRU, 1989. Stopping powers for electrons and positions.
ICRU Report No. 37. International Commission on
Radiation Units and Measurements, 7910 Woodmont
Avenue, Washington, DC, 20814.
ICRU, 1993. Stopping power and ranges for protons and alpha
particles. ICRU Report No. 49. International Commission
on Radiation Units and Measurements, 7910 Woodmont
Avenue, Washington, DC, 20814.
ICRU, 1998. Fundamental quantities and units for ionizing
radiation. ICRU Report No. 60. International Commission
on Radiation Units and Measurements, 7910 Woodmont
Avenue, Washington, DC, 20814.
Stewart, R.D., Wilson, W.E., McDonald, J.C., Strom, D.J.,
2002. Microdosimetric properties of ionizing electrons in
water: a test of the PENELOPE code system. Phys. Med.
Biol. 47, 79–88.
Tung, C.J., Kwei, C.M., 1991. Penetration of electrons in
condensed media II. Proc. Natl. Sci. Counc. ROC(A) 15,
371–381.
Tung, C.J., Lou, J.J., Lin, S.W., 1986. Pathlength distributions
of low energy electrons in water. Nucl. Instrum. Methods.
16, 83–87.
Turner, J.E., Hamm, R.N., Wright, H.A., Ritchie, R.H.,
Magee, J.L., Chatterjee, A., Bolch, W.E., 1988. Studies to
link the basic radiation physics and chemistry of liquid
water. Radiat. Phys. Chem. 32, 503–510.
Watt, D.E., 1996. Quantities for dosimetry of ionizing
radiations in liquid water. Taylor & Francis Ltd., London.
Wilson, W.E., Metting, N.F., Paretzke, H.G., 1988. Micro-
dosimetric aspects of 0.3 to 20 MeV proton tracks I:
crossers. Radiat. Res. 115, 389–402.
Wilson, W.E., Nikjoo, H., 1999. A monte carlo code for
positive ion track simulation. Radiat. Environ. Biophys. 38,
97–104.