Hatano Y., Katsumura Y., Mozumder A. (Eds.) Charged Particle and Photon Interactions with Matter - Recent Advances, Applications, and Interfaces

Подождите немного. Документ загружается.

630 Charged Particle and Photon Interactions with Matter

and JF (Sato et al., 2009) have a ner voxel size of 0.98 × 0.98 × 1 mm

. The performance of

CTscanners greatly improved in a short period and made it possible to take CT pictures at a high

resolution

with less exposure, resulting in the development of the high-resolution phantoms.

Both

the male phantoms Otoko and JM have body sizes close to the Asian Reference Man

(ARM) dened by Tanaka (Tanaka and Kawamura, 1996). The body size of JF is also close to the

Asian Reference Man, Female (ARMF), but Onago has larger body size than ARMF. We compared

the body thicknesses and widths of the developed voxel phantoms with the reference values for the

thorax, the abdomen, and the buttocks; they are all within 2σ deviations around the reference val-

ues, and it was conrmed that the developed voxel phantoms do not much deviate from the average

body

shapes of the Japanese.

radiation protection, dose calculations are usually performed using an assumed upright posi-

tion of human bodies, while voxel phantoms have been constructed from CT data taken in lying

position. The structures of a human body and its organs are considered to change slightly according

to the posture; therefore, to investigate the effect of posture on organ doses, the voxel phantom JM2

was constructed from CT data taken in an upright position for the same person as JM (Sato et al.,

2007b). According to comparison of doses between JM and JM2, it was concluded that the effect

of body posture is not signicant from a viewpoint of radiation protection, though apparent dose

difference

is observed in some cases (Sato and Endo, 2008; Sato et al., 2008).

23.3.3 icrp reference phantoMS

In Publication 103, ICRP (2007) concretely dened the reference phantoms that should be used

in dosimetry. Currently ICRP has not dened any specic reference phantoms even though the

committee has published data on the reference anatomy (ICRP, 1975, 2003). Therefore, various

reference phantoms had been used to obtain fundamental dosimetric quantities; some are her-

maphrodite and some are sex specic; some have arms and some do not. Now ICRP specied the

reference voxel phantoms of the adult reference male and female developed by Zankl et al. (ICRP,

2010). Therefore these voxel phantoms are required usage to obtain basic dosimetric data from

now on. Since the phantoms were originally constructed from CT data for individuals, they had

anatomies different from the reference male and female. Thus, the ICRP reference phantoms were

adjusted by image processing to have body sizes and organ masses approximately the same as the

reference Caucasian male andfemale. Namely, the reference body heights are 176 and 163cm and

the

reference body weights are 73 and 60

for male and female, respectively (ICRP, 2003).

23.4 dose CalCulation For radiation proteCtion purposes

23.4.1 Monte carlo radiation tranSport calculation

In order to obtain organ doses and related quantities using computational phantoms, radiation inter-

actions inside phantoms are simulated using Monte Carlo transport calculation. For example, when a

photon enters the phantom, it produces secondary electrons and deposits energy mainly through sec-

ondary electrons. In Monte Carlo simulation, all the trajectories of primary and secondary radiations

are followed. Then an organ dose is calculated as the sum of energy deposited by all these radiations

inside the organ divided by the mass. Radiation transport calculations can be carried out for differ-

ent types of radiations by selecting an appropriate simulation code. Here, some explanations will

be given about photon–electron transport calculation with reference to EGS4 (Nelson etal., 1985).

Interaction processes considered in EGS4 are photoelectric absorption, Compton scattering, pair

production, coherent scattering for photons, energy loss by collisions and Bremsstrahlung, Moliere

multiple scattering, Moller and Bhabha scatterings for electrons and positrons. When a pair pro-

duction takes place, a positron–electron pair is produced and the positron emits two annihilation

gamma gays of 0.511MeV after it stops moving. Basically radiation interactions are probabilistic

ComputationalHuman Phantoms and Their Applications toRadiation Dosimetry 631

events submitted to specic distributions; therefore, in simulation, properties of radiation interac-

tions and emitted secondary particles are determined randomly using random numbers and specic

distributions of physical quantities. Then, energy deposited, to calculate the organ dose in an organ,

the sum of a large number of radiation incidence.

Monte

Carlo code systems combined with the voxel phantoms for calculating doses and related

quantities have been constructed (Funabiki et al., 2000, 2001; Saito et al., 2001; Kinase et al., 2003,

2007) as user codes of EGS4. In the whole code systems, we added the function to calculate radia-

tion

transport in voxel geometry that EGS4 originally did not contain.

23.4.2 external doSe calculation

In this section, some typical results for external exposures are shown and discussed. In radiation

protection for external exposures, usually monoenergetic irradiations of broad-parallel beams in

idealized geometries are assumed. The geometries generally assumed are anterior posterior (AP),

posterior anterior (PA), lateral (LAT, which represents both left lateral [LLAT] and right lateral

[RLAT]),

rotational (ROT), isotropic (ISO), which are schematically shown in Figure 23.5.

23.4.2.1

photon

In Figure 23.6, examples of effective dose curves for external photon exposures are indicated (ICRP,

1996). Effective doses and organ doses are often normalized to air kerma that is a kind of absorbed

dose in air assuming that secondary electrons entirely lose their energies at the origin points. From

the gure, some typical features of effective doses are observed. Effective doses decrease rapidly

below several tens of keV because of the self-shielding effect of the human body. Low-energy

photons are absorbed by tissues at the body surface with a high probability; therefore, it is dif-

cult for these to reach important organs existing in the middle of the body. On the other hand, in a

high-energy region, photons can enter the body deeply and deposit energy rather homogeneously.

ISO

L AT ROT

Figure 23.5 Idealized irradiation geometries of external exposure assumed in radiation protection.

632 Charged Particle and Photon Interactions with Matter

Thus,the effective dose curves are rather at in the MeV region. The curve has a peak just below

100 keV, resulting from a build-up effect; contributions from scattered photons and secondary elec-

trons

are maximum in this energy region.

Further,

it is clear that effective doses are different depending on geometry. Among the six

geometries considered here, the effective dose is the highest in AP irradiation. This is because

dominant organs having high radiation sensitivities exist mostly at the front side of the body and

the AP beam directly irradiates these organs, while the doses are small in lateral geometries where

photons need to travel long distances in the body to reach the dominant organs. The effective dose

curves

for PA, ROT, and ISO exist between those for AP and LAT.

Variations

in organ doses for external photons were investigated using the developed voxel phan-

toms at JAEA (Sato et al., 2008). Organ doses were calculated for six typical external irradiation

conditions of AP, PA, LLAT, RLAT, ISO, and ROT with 25 kinds of monoenergetic photons rang-

ing from 0.01 to 10MeV. In order to clarify the difference in organ doses between the Japanese and

the Caucasian phantoms, analyses were made using the calculated data (Sato et al., 2008; Saito

etal., 2009). The ratio of an organ dose in a Japanese phantom to that in the Caucasian phantom of

the corresponding gender was calculated for the lungs, stomach, bladder wall, liver, esophagus, thy-

roids, skin, brain, colon, testes, breast, and ovaries. Organ doses for Otoko and JM were normalized

to that for Rex (Schlattl et al., 2007), and organ doses for Onago and JF were normalized to that for

Regina

having body characteristies close to the references (Schlattl et al., 2007).

Figure

23.7 shows the maximum and minimum values of the ratios for AP, ROT, and LLAT

geometries in the energy range of 50keV to 3MeV. This energy range was selected because the

typical exposure in the environment is due to photons within this range. The data indicate that

organ dose differences between the Japanese and the Caucasian phantoms vary within a factor

of 2 except for lateral geometries. On an average, organ doses were found to be slightly larger for

Japanese, which is considered to be due to differences in body size; and relatively large differences

are observed in lateral geometries. However, the difference in organ doses averaged over diverse

organs

and conditions was found to be very small.

Figure

23.8 shows energy dependencies of thyroid doses for Onago, JF, and Regina in LLAT

geometry. The positions of the thyroids tend to change greatly according to individuals, and this

could vary the dose. The positions are clearly different even among Japanese phantoms. The thy-

roids of Regina are expected to locate at deeper positions where the shielding effect by the shoulder

is signicant in lateral irradiation geometries, though the detailed anatomical data have not been

released on Regina. It is not yet clear if the large difference is due to racial difference or due to

individual difference, and further analysis is necessary for understanding the systematic difference.

1.5

1.0

0.5

0.01 0.1

RLAT

ROT

Photon energy (MeV)

Effective dose/air kerma (Sv Gy

–1

)

Figure 23.6 Variation in effective doses for external photons.

ComputationalHuman Phantoms and Their Applications toRadiation Dosimetry 633

In total, concerning individual organs, the dose difference among different phantoms comes

up to a several factor. The difference in effective doses between the Japanese and the Caucasian

phantoms was within 10% for almost all the cases. It must be noted that, in case of partial irradia-

tion by narrow low-energy beams, organ doses change drastically by the incident position and angle

(Ohnishi et al., 2004). In such cases, an appropriate individual phantom is considered necessary to

obtain realistic organ doses and effective doses.

23.4.2.2 e

lectron

Features

of energy deposition by incidences of electrons or charged particles are greatly different

from those of photons. Electrons have nite penetration ranges and it makes a large difference

whether they reach target organs or not. Saito et al. (2008) calculated external electron doses

for Otoko and Onago in AP, PA, and ISO geometries and compared the data with a MIRD-type

phantom by Ferrari et al. (1997). Here, organ doses were normalized to incident electron uence.

5.0

1.0

0.5

AP, Max

ROT, Max ROT, Min

LLAT, Max LLAT, Min

AP, Min

–1

(a) (b)

–1

Photon energy(MeV) Photon energy(MeV)

Dose ratio(JF or Onago/Regina)

5.0

1.0

0.5

Dose ratio(JM or Otoko/Rex)

Figure 23.7 Range of the organ dose ratios of the Japanese phantoms to the Caucasian phantoms. (a) Male

and

(b) female.

1.4

1.2

1.0

0.8

0.6

0.4

0.2

–2

–1

Onago

Regina

Photon energy(MeV)

Organ dose per air kerma (Gy Gy

–1

)

Figure 23.8 Comparison of thyroid doses among Onago, JF, and Regina for external photon exposure

inLLAT

geometry.

634 Charged Particle and Photon Interactions with Matter

Two examples of the dose coefcients for the three phantoms are shown in Figure 23.9. The liver

doses in AP geometry show a good agreement among the three phantoms; while the kidney doses

in PA geometry show apparent difference between the MIRD phantom and the voxel phantoms. In

external electron exposure, organ depth is a dominant factor to determine the energy dependency of

the dose coefcients. The gure indicates that the effective depth is quite similar for three phantoms

in the case of liver, but obviously different in the case of kidney.

Generally, the energy dependency curve of organ dose for electrons consists of three regions with dif-

ferent tendencies: (1) At low energy, the organ dose is very small; (2) at middle energy, the curve steeply

increases with energy; and (3) at high energy, the dose is nearly constant or slightly increases with energy.

In region (1), the organ dose is close to zero, since electrons do not reach the target organ because

their penetration ranges are too short. However, even in this energy region an extremely small

amount of energy is delivered to the target organ. This is considered mainly due to Bremsstrahlung

radiation emitted by electron transport. In region (2), the dose increases rapidly with energy,

because electron penetration becomes great enough that some of the electrons can reach the organ

and deposit energy directly. The electron energy at which this rapid increase starts varies accord-

ing to the depth of the organ from the body surface. In region (3), the dose saturates and exhibits a

slower increase. In this region the electron penetration range exceeds the order of body thickness,

and electrons deposit their energy rather homogeneously in the body. The slow increase of dose is

considered

mainly due to buildup of Bremsstrahlung radiation.

Figure

23.10 shows organ doses for Otoko and Onago normalized to those for the MIRD phan-

tom. Large differences between different phantoms are observed around 10MeV where the electron

range is several cm; dominant organs exist mostly at depths of several cm. The maximum dose

difference observed in the study was a factor of 50 in kidney dose in PA geometry. A factor of 50

does not have any concrete meaning, but the electron dose for an individual organ sometimes shows

a large discrepancy between different phantoms. The maximum difference observed in effective

doses was a factor of two, which coincides with data presented by Kramer (Kramer, 2005).

23.4.3 internal doSe calculation

23.4.3.1 absorption Fraction and specic absorption Fraction

Figure 23.11 gives a scheme concerning internal exposure. Radionuclides accumulated in a source

organ emit gamma rays, beta rays, and in some cases alpha rays. The emitted radiations deposit

their energy inside the source organ itself and also in other target organs. Then, the fraction of

Liver(AP) Kidneys (PA)

Otoko

Onago

MIRD

Electron energy (MeV) Electron energy (MeV)

Equivalent dose per fluence(fSv· m

)

Equivalent dose per fluence(fSv· m

)

Figure 23.9 Examples of organ doses for external electron exposure calculated using Otoko, Onago, and

MIRD

phantoms. (From Saito, K. et al., Jpn. J. Health Phys., 43, 122, 2008. With permission.)

ComputationalHuman Phantoms and Their Applications toRadiation Dosimetry 635

energy per emitted energy deposited in the target organ is designated as absorption fraction (AF),

and the AF divided by the mass of the target organ is the specic absorption fraction (SAF). The

AF and SAF are basic quantities for the evaluation of internal doses for monoenergetic radiations

prepared using computational phantoms. Further, by integrating

AFs or SAFs considering types and energy distributions of emitted

radiations, S-values are obtained to convert radionuclide concentra-

tion

to organ doses or effective doses.

The

SAFs obtained from calculations using MIRD phantoms

have been currently used. However, positional relation among

organs, which is an important factor to determine SAFs, could

change signicantly according to the phantom; MIRD-type phan-

toms may not reect positional relations among organs in human

bodies accurately. Thus, SAF calculations using sophisticated

phantoms

are desired to accurately evaluate internal doses.

Sato

et al. (2005) calculated self-absorption fractions (self-AFs)

using Otoko, JM,anda MIRDphantom.In self-AFs, the target organ is

takenas the same as thesource organ. Examplesofself-AFs are shown

for six organs in Figure 23.12. It was found that self-AFs aresimilar

among different phantoms for large organs having simple identical

shapes, such as brain and kidneys. In case of small organs, however,

the sizes and shapes affect the self-AFs signicantly. When the shapes

of small organs are identical, the portion of energy absorbed by the

organ itself is largely affected by the size. The difference in spleen

doses could be explained by the sizes, while in pancreas, thyroid, and

urinary bladder wall, the self-AFs are not explained by the size, and

this indicated that shape is an important factor for these cases.

0.1

0.01

Effective dose

ROT

0.1

0.01

Relative dose

0.1

Relative dose

0.1

Relative dose

Lungs(PA)

Small int.(AP) Bone marrow(AP)

Stomach(AP) Kidneys (PA)

Electron energy (MeV) Electron energy(MeV) Electron energy (MeV)

Otoko/MIRD

Onago/MIRD

Otoko/MIRD

Onago/MIRD

Otoko/MIRD

Onago/MIRD

Otoko/MIRD

Onago/MIRD

Otoko/MIRD

Onago/MIRD

Figure 23.10 Organ doses and effective doses for Otoko and Onago normalized to those for a MIRD phan-

tom in external electron exposure. (From Saito, K. et al., Jpn. J. Health Phys., 43, 122, 2008. With permission.)

AF:

Fraction of energy

absorbed in the target organ

per energy emitted from the

source organ

Target

Source

Figure 23.11 Schematic rep-

resentation oftheabsorption frac-

tion

(AF) in internal exposure.

636 Charged Particle and Photon Interactions with Matter

Sato et al. (2005) examined the relation of SAFs to organ distance using Otoko, Onago, JF, and

JM. In Figure 23.13, SAFs for several different combinations of source and target organs in the four

Japanese voxel phantoms are plotted together as functions of distance between the centers of grav-

ity. The distance between the centers of gravity may not be the most suitable parameter to analyze

SAFs in every case; nevertheless, it was conrmed that SAFs would be roughly expressed as a func-

tion of this distance.

23.4.3.2 s-value

Kinase and Saito (2007) and Kinase et al. (2004) evaluated S-values, which are the mean absorbed

doses per unit cumulated activity to the target organ, from uniformly distributed radioactivity

within the source organ, for several beta-ray emitters. S-values have been used for dose estimates

in radiological protection and also in medical diagnostic and treatment procedures. In particular,

S-values for positron emitters in the brain, heart, and urinary bladder play an important role

for accurate dose evaluation of patients with administered radiopharmaceutical for clinical PET

imaging. S-values for positron emitters within the urinary bladder are currently derived on the

simple assumption that the dose at the surface of the content is approximately half the dose within

their volume.

Figure 23.14 shows the ratios of the S-values derived from the currently used simple assump-

tion by ICRP to those by Monte Carlo calculations using Otoko and Onago. The S-values for the

urinary bladder wall for several beta-ray emitters in the bladder contents have been evaluated. The

ratios increase as beta-ray energy decreases. The discrepancies between Otoko and Onago may

be explained by the different ratios of the bladder wall mass to the bladder contents mass. It was

–1

Otoko

Voxel MIRD

Otoko

Voxel MIRD

Otoko

Voxel MIRD

Otoko

Voxel MIRD

Otoko

Voxel MIRD

Otoko

Voxel MIRD

–2

–1

–2

–1

–2

–1

–2

–3

–1

–2

–1

–2

–1

Energy (MeV)Energy (MeV)

–2

–1

–2

–1

–2

–1

–2

(a)

(b)

(c)

(d) (e) (f)

–1

–2

–1

Energy (MeV)

Self-SAF (–)Self-SAF (–)

Figure 23.12 Self-AFs for six different organs calculated using Otoko, JM, and MIRD phantoms. (a)Brain,

(b)

kidneys, (c) spleen, (d) pancreas, (e) thyroid, and (f) urinary bladder wall.

ComputationalHuman Phantoms and Their Applications toRadiation Dosimetry 637

conrmed that the S-values from the simple assumption currently used by ICRP are conservative.

However, the difference is quite large from the realistic value, and realistic dose evaluation would

desirable in the future.

23.4.4 environMental expoSure

Usually in radiation protection dosimetry for external exposure, it is assumed that reference adults

are irradiated by monoenergetic radiations in simplied geometries. When we consider exposure

in the environment, the situations are different. Radionuclides are distributed in the environment

–1

–2

–4

–5

–6

–3

SAF(per kg)

Distance between gravity centers(cm)

Kidneys

Liver

Lower colon

Thyroids Stomach

Liver Lower colon

Liver

Brain

Thyroids

Lower colon

Figure 23.13 SAFs as a function of distance between the gravity centers. Data calculated using Otoko,

Onago,

JM, and JF are shown together.

100

Otoko

Onago

137

147

204

Beta-ray emitter

S-value ratio

(the simple assumption/EGS-UCSAF)

Figure 23.14 Ratios of the S-values for positron emitters derived from the currently used method to those

by Monte Carlo calculations with voxel models. (From Kinase, S. et al., J. Nucl. Sci. Technol. Suppl., 4, 136,

2004.

With permission.)

638 Charged Particle and Photon Interactions with Matter

with their specic distributions depending on the origin; thus, the photons incident to a human body

have angular and energy distributions even for the case where sources emit monoenergetic photons.

Further, people covering a wide range of age exist in the environment, and this is different from

exposure

in nuclear facilities.

Saito

et al. simulated exposure of adults and children in the environment assuming three typical

source distributions: (a) uniformly distributed volume source in the air; (b) uniformly distributed

plane source in the ground; and (c) uniformly distributed volume source in the ground (Saito etal.,

1990, 1991, 1998; Petoussi et al., 1991; Zankl et al., 1997). Sources (a) and (b) are modeled on

anthropogenic radionuclides released from nuclear facilities. Soon after the release of radionuclides

into the environment, they stay in the air; source (a) simulates this situation. If some successive

release of radionuclides arises, they deposit on the ground; source (b) simulates this situation.

Source (c) models natural radionuclides of

238

U series,

232

Th series, and

K in the soil. A human

body was assumed to stand vertically on the ground, and transport of photons in the environment

and

in the body were both accurately simulated.

Figure

23.15 gives angular distributions of air kerma at 1m above the ground as cosines to a vec-

tor normal to the ground. Each source has its specic angular distribution. In case of volume sources,

that is sources (a) and (c), gamma rays enter the body in isotropic directions from the half space where

source exists, and small amounts of scattered gamma rays enter the body from the other half space.

In plane source on the ground, the dominant part of gamma rays enters the body from the horizontal

directions. This is because the amount of source per unit solid angle increases greatly at horizontal

directions.

In Figure 23.16 are indicated effective doses normalized to air kerma in sources (a) and (b)

for adults, a 7-year-old child, and an 8-week-old baby. It was conrmed that effective doses are

higher for infants, and the maximum dose difference between the adult and the baby is a factor

of two above 50 keV where exposure in the environment is generally signicant. Further, the

maximum dose difference for a single organ was found to be a factor of 3 (Saito et al., 1991).

In source (c), effective doses for the baby were found, some 70% higher that those for adult.

Further, the variation in effective doses due to several reasons such as bias of source distribution,

human body posture, and relational change to the source have been also investigated (Saito etal.,

1998).

–1

–2

–1

–2

–1

–2

–1

1 MeV 662 keV

232

Th series

Angle of incidence (cos)

Relative air kerma

Angle of incidence(cos) Angle of incidence (cos)

1 –1 0 1 –1 0 1

(a) (b) (c)

Figure 23.15 Angular distributions of environmental gamma rays at 1m above the ground for three

typical environmental source distributions. (a) Volume source in air, (b) plane source in ground, and

(c)

volume source in ground.

ComputationalHuman Phantoms and Their Applications toRadiation Dosimetry 639

23.5 appliCation to other Fields

23.5.1 application to radiation therapy

Radiation therapy is an important application of dose calculation technique using voxel phantoms

with which accurate individual doses can be obtained. It is recommended that the dose accuracy

be within 5% in radiation therapy (ICRU, 1976), and this means that the net error in dose calcula-

tion should be within a few percent. This dose accuracy would be difcult to attain by conventional

analytical methods in some cases where the elemental composition and density change drastically

according to position in the patient body, since the disequilibrium of electrons could lead to signi-

cant

errors in the dose calculation.

Some

commercial dose planning systems have been released that utilize Monte Carlo calcula-

tions potentially able to perform accurate simulations. However, Monte Carlo calculations need

enormous computational time, and also users need some basic knowledge to properly perform

the calculations. These requirements have prevented the Monte Carlo systems from wide spread

use in Japan.

Saito et al. (2004, 2006) developed a system for providing accurate dose distributions through

networks to multiple medical facilities based on voxel phantoms and Monte Carlo calculations. The

basic concepts of the system design are to provide reliable accurate doses for any conditions so that

the calculated doses can be used as standard values and to construct an environment where users

1.2

1.0

0.8

0.6

0.4

0.2

0.10.01

(a)

(b)

1 10

1.2

1.0

0.8

0.6

0.4

0.2

0.1 10.01

Baby (8 weeks)

Child (7 years)

Adult

E/air kerma (Sv Gy

–1

)E/air kerma (Sv Gy

–1

)

Source energy (MeV)

Figure 23.16 Effective doses for humans at different ages due to environmental gamma rays. (a) Volume

source

in air and (b) plane source in ground.