Radiation Protection Dosimetry (2006), Vol. 118, No. 1, pp.

131136 Advance Access publication 12 January 2006

doi:10.1093/rpd/nci338

TECHNICAL NOTE

PHYSICAL PHANTOM OF TYPICAL KOREAN MALE FOR RADIATION PROTECTION PURPOSE

J. I. Kim1, H. Choi1, B. I. Lee1, Y. K. Lim1, C. S. Kim1, J. K. Lee2 and C. Lee2, 1 Radiation Health Research Institute of Korea Hydro and Nuclear Power, 388-1 Ssangmoon, Dobong, Seoul, Korea 2 Hanyang University, 17 Hangdang, Seongdong, Seoul, Korea Received January 1 2005, amended May 11 2005, accepted May 16 2005
Dose distribution within a human body can be measured using physical anthropomorphic phantoms. In an effort to establish reference Korean physical model, the rst Korean physical phantom of average Korean adult male was constructed using computed tomography (CT) images of a healthy volunteer. The body dimension of the subject was close to that of average Korean male. The source images were obtained using fusion positron emission tomography machine at Radiation Health Research Institute in Korea, and ported into rapid prototyping process. The physical phantom was composed of three tissueequivalent materials: epoxy resin, urethane foam and polyurethane representing bone, lungs and soft tissues, respectively. The densities of the tissue-equivalent materials were close to those recommended by the International Commission on Radiation Units and measurements. To facilitate dose mapping, the phantom was sliced into 2 cm sections. Hole grids for thermoluminescence (TL) dosemeter chips were drilled. To verify the appropriateness of the physical phantom, organ doses of selected organs were measured for reference photon beam, and compared with those computed by tomographic model constructed from the same CT images. Absorbed doses converted from TL relative response showed good agreement within 7% with those calculated.

INTRODUCTION A structure that contains tissue substitutes and that used to simulate radiation interactions in the human body is dened for a phantom(1). Absorbed dose distribution within a human body can be calculated by using computational human models or measured using physical phantoms. Various kinds of physical human phantoms have been fabricated for several purposes. The representative physical phantoms for research in radiation protection, diagnosis and treatment are RANDO(2) and ATOM (CIRS Inc., Norfolk, VA). Pediatric phantoms based on computed tomography (CT) images have been developed at University of Florida(3). The height and weight of them agree with those of the reference man, which was recommended by the International Commission on Radiological Protection (ICRP)(4,5). However, the reference man is basically a Caucasian having Western Europe or North American habitat and custom. The characteristics and parameters of the ICRP Reference man may not be directly applicable to other populations since there must be racial differences in custom, dietary habits and climatic conditions. The project to formulate reference Asian has been performed as the project of the International Atomic Energy Agency (IAEA), and included in revised ICRP reference man(5,6). The scope of the study included the information of appearance, total diet and food-stuff of Chinese,

Indian, Japanese, Korean and other south east Asian people, and provides important data for setting up the reference Asian model. Efforts to establish the Korean reference man system also took place in Korea. As part of our efforts, the rst Korean specic physical phantom was constructed using whole-body CT images of a male volunteer. This article describes the whole construction process of the phantom. Absorbed dose was measured by using dosemeter and compared with those from tomographic model based on the same CT images using Monte Carlo computation to verify the appropriateness of the phantom.

MATERIALS AND METHODS Source images A healthy volunteer, whose body dimension is close to that of average Korean male, was recruited to be a model for phantom construction. The body dimension of average Korean male was obtained from the data of the Fourth Survey of National Physique Standard by Korean Agency for Technology and Standard (KATS). The survey has been periodically carried out since 1979. In the fourth survey, the body size data of 13,062 various aged Koreans were obtained (6578 males and 6484 females). Table 1 tabulated the body dimension of the representative volunteer and that of average Korean male formulated by KATS. CT image of the volunteer was scanned using fusion positron emission tomography

Corresponding author: cslee@itrs.hanyang.ac.kr

The Author 2006. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

J. I. KIM ET AL. Table 1. Body dimension of the male volunteer used in phantom construction and that of average Korean male. Body dimension Average Korean male Mean Body height (cm) Body weight (kg) Head circumference (cm) Neck circumference (cm) Chest circumference (cm) 170.9 68.1 56.7 36.3 90.9 SD 5.0 8.2 1.5 1.9 5.9 172 68 54.1 37.2 88.3 Figure 1. STL-formatted 3-D views of lungs (left), bone structure with lungs (middle) and skin-covered (right) rendered by V-works. Volunteer

(PET) machine (Siemens Somatom Emotion Duo system). Fusion PET scanning service for cancer nding has been performed at Radiation Health Research Institute of Korea since 2003. It was approved that no cancer was found in the resulting fusion PET images of the volunteer by radiologists. Total time required only for CT image acquisition was $20 min. Of transversal whole-body CT images 1788 slices were obtained at 1 mm interval, and 68 slices of them were overlapped. CT images from head top to mid-thigh were adopted for physical phantom construction. The CT data in Digital Imaging and Communications in Medicine (DICOM) format were converted into Stereo Lithography (STL) format to be ported into manufacturing tool. Phantom construction Most of physical phantoms are fabricated by pouring soft tissue material(7) into outermost cast, in which bone and lungs are positioned. In case of bone material, natural human skeletons were used in RANDO, and bone-equivalent material was used in ATOM. In this study, the Rapid Prototyping and Manufacturing (RP&M) technique was adopted to construct physical phantom based on CT images of real human subject. The term rapid prototyping refers to a class of technologies that can automatically construct physical models from Computer Aided Design data(8). The RP&M has been usually utilized to develop a master pattern of complex components directly from tomography by stacking material layer upon layer. The RP&M technique was well appropriate for manufacturing complex structure like human body. CT images in DICOM format were converted into STL format using V-works (CyberMed, Korea). V-works performs volume rendering, surface rendering and image segmentation using graphical user interface. Image segmentation was automatically carried out for three tissues according to CT numbers: bone, lungs and soft tissues. The resulting physical phantom also was composed of three tissue

materials. Figure 1 shows the STL-formatted 3-D views of bone, lungs and skin. Since the biggest model size manageable by the RP&M machine, SLA3500 (3D systems, CA), was 350 350 400 mm3, the body data were separated into several parts using SolidView (Solid Concepts, CA). Separated STL les of 3-D part of body were ported into SLA3500. SLA3500 constructed 3-D parts of body from liquid photosensitive polymers that solidify when exposed to ultraviolet light. Body parts were built upon a platform situated just below the surface in a vat of liquid epoxy resin, SL5510 (Vantico, 3D systems, CA). SL5510 was directly used for composing human bone since its density (1.23 g cm3) was so close to average density of human bone ($1.3 g cm3). Urethane foam (0.28 g cm3) was used to construct lungs. Lungs could not be directly constructed by SLA3500 since liquid photosensitive material with density of human lungs ($0.29 g cm3) was not available. Cast for lung model was built using SL5510 and SLA3500, and lung-equivalent material, urethane foam (0.28 g cm3), was injected into the cast. Outer cast was removed after hardening for 8 h. Polyurethane (1.07 g cm3) was used for constructing soft tissue. Cast for outermost skin contour was constructed using SLA3500, and then the lungs and bone structure were positioned in the skin cast. Polyurethane was poured into the skin cast, and hardened at 80 C for 24 h. The male gentile region was simplied as square shaped box covering testes and penis of the subject as the same case with other adult male physical phantoms, RANDO and ATOM. Table 2 tabulated the characteristics of the tissue-equivalent materials used in phantom, and atomic composition. Those of RANDO and ATOM were included for comparison. Figure 2 shows the views of assembling bone structure including lung models, and sectional outermost cast containing bone and lung structures. To facilitate

132

TYPICAL KOREAN MALE PHYSICAL PHANTOM Table 2. Characteristics of the tissue-equivalent materials used in the composition of physical phantoms. Soft tissues This ATOM RANDO Ref.a Density (g cm3) 1.07 Elemental composition H 6.45 N 5.56 C 63.20 O
a b

Bone

Lungs Ref. 0.29 10.3 2.9 10.2 75.7

This ATOM RANDOb Ref. This ATOM RANDO 1.23 9.53 3.83 64.20 14.50 1.6 4.86 0.86 71.9 1.4 0.28 0.21 8.32 3.15 83.82 47.9 16.50 18.1 0.32 5.7 2.0 74.0

1.055 8.47 1.65 82.03

1.00 8.9 3.1 66.8 21.1

1.04 10.5 2.5 22.7 63.5

7.3 6.29 3.1 5.42 25.5 60.90

21.10

ICRP 23 (1975) Natural human skeleton

Figure 2. Views of assembling bone structure including lung models (left), and sectional outermost cast containing bone and lung structures (right).

dose measurement, the phantom was sliced into 2 cm sections as other commercial physical phantoms by using Computer Numeric Control (CNC) cutting machine. Even though 2 mm thick material was lost during CNC cutting process, margin for the lost thickness was considered in the STL generating step. Sectioned slices were coated with urethane paint. Hole grids (2153) which were 7 mm in diameter for thermoluminescence (TL) chips were drilled with the interval of 2 cm. Verication of phantom To verify the completeness and appropriateness of the phantom, two methods were used: nondestructive CT examination and comparison of measured dose with simulated dose. First, CT images of the resulting phantom were scanned to examine internal vacancy. CT images (20 coronal and 40 sagittal) were obtained from combined physical phantom, and examined. Secondly, relative response of TL dosemeter (TLD) was measured in standard photon beams using disk type

LiF:Mg,Cu,P TLDs of 4.5 mm in diameter and 1 mm in thickness. Each hole was indexed with the identication number of organ to which the hole belonged. 184 TLDs were plugged into hole grids as close as possible to the center of organs. TLDcontaining phantom was exposed to 137Cs gamma irradiation dose of 0.01 Gy at Korea Atomic Energy Research Institute. Relative response of 184 TLD was readout, and the values of TLDs belonging to the same organ were averaged to be representative response for the organ. To convert the relative response to absorbed dose, the tomographic model that was constructed from the same CT image set with those used for the physical phantom was adopted for Monte Carlo dose calculation. Radiosensitive organs and tissues (22) are segmented in the model, which was ported into Monte Carlo transport code, MCNPX 2.4(9). MCNPX 2.4 is general purpose Monte Carlo radiation transport code that tracks 34 types of particles at all energies. It consists of 300 150 343 voxels of a size of 2 mm 2 mm 5 mm. Broad parallel photon beams with the energy of 662 keV in anteroposterior geometry were simulated. The energy deposition tally, f6, was used with the unit conversion option to obtain the mean organ absorbed dose, and described as Z Z Z r dV f6 a H E F~, E, td Edt r , 1 rg V t E V where ra is atom density, rg is gram density and H(E) is heating response summed over nuclides in a material. The original unit of the energy deposition tally is MeV g1. An asterisk changes the unit to jerks g1 (1 MeV 1.60219 1022 jerks). Histories (108) were applied to reduce relative error <10%. Relative error is dened to be one estimated SD of the mean divided by the estimated mean. Results with errors <10% are generally reliable in Monte Carlo calculation(9). Relative response of TLD

133

J. I. KIM ET AL.

Figure 3. Outer view of typical Korean physical phantom.

indexed as soft tissue was selected as calibration value since it was uniformly distributed through the whole-body. Dose conversion factor, which converts relative response of TLD into absorbed dose, was calculated for soft tissue. The following for different types of elemental tissue compositions were considered for organ dose calculation: soft tissue (1.05 g cm3), lung (0.26 g cm3), skin (1.09 g cm3) and bone (1.3 g cm3). Then the TL response values were compared with the organ doses calculated by Monte Carlo simulation. RESULTS AND DISSCUSSION Typical Korean male physical phantom The physical phantom of typical Korean male was constructed using whole-body CT images and rapid prototyping technique. The phantom was sectioned into 43 slices with the thickness of 2 cm. Arms and legs were absent. Figure 3 shows the outer appearance of the completed Korean male phantom. Verication of phantom First, CT examination was performed to evaluate internal vacancy. Although small vacancy caused by the shrinkage of soft tissue material (polyurethane) was found, that would be negligible for radiation dose measurement. To overcome the vacancies, another manufacturing method except cast-based construction should be applied. Selected views of transversal CT images of manufactured physical

Figure 4. Chest tomographic images of the volunteer (upper) and the phantom scanned with PET/CT device (lower).

phantom and original human subject were shown in Figure 4. Second, relative response of TLD was measured using 184 disk type LiF:Mg,Cu,P TLDs, situated at the representative position in organs and tissues. At the same time, absorbed doses of selective organs and tissues were calculated using MCNPX 2.4 and tomographic model. The relative response of TLD for 662 keV 137Cs source and absorbed dose per single photon calculated from tomographic models were tabulated in Table 3. Dose conversion coefcient (Gy per relative response) was calculated as 2.96 1032 for soft tissue. Organ absorbed doses for other organs in physical phantom could be calculated using this dose conversion coefcient, and compared with those from calculation. Ratio of TLD dose to MCNP dose was depicted in Figure 5. Two absorbed doses from TLD and Monte Carlo code were in agreement within 7%. Adrenals, gall

134

TYPICAL KOREAN MALE PHYSICAL PHANTOM Table 3. Relative response of TLD and absorbed dose calculated by Monte Carlo method. Relative response of TLD (a.u.a) Adrenals Bladder Brain Colon Eye ball Fat Gall bladder Gonad Heart Kidney Liver Lungs Small bowel Soft tissue Stomach Thyroid
a

glands. Just one TLD chip was not enough to represent mean adrenal absorbed dose. CONCLUSION In an effort to establish reference Korean physical model, the rst physical phantom of typical Korean male was constructed using CT images of a healthy volunteer. The body dimension was close to that of average Korean male. Whole-body CT images were obtained using fusion PET machine, and processed to construct physical phantom and tomographic model for verication study. The resulting Korean physical phantom was composed of three different tissue-equivalent materials: epoxy resin, urethane foam and polyurethane representing bone, lungs and soft tissues, respectively. To facilitate dose mapping, the resulting phantom was sliced into 2 cm sections, and 2153 holes for TLDs were drilled. Non-destructive examination using CT showed up the small vacancy, which may be neglected for radiation protection measurement. Dose measured using TLD was compared with that computed by tomographic model, and agreement <7% was shown for all selective organs and tissues. The Korean physical phantom constructed in this study was the rst one, which will be followed by additional typical Korean physical phantoms. Rapid prototyping and manufacturing techniques applied in this study will be valuable fundamentals for further study. The phantom will be applied for several goals including the measurement of radiation protection quantities and dose distribution in radiation treatment and diagnosis, and verication of Monte Carlo calculation. ACKNOWLEDGEMENTS We wish to acknowledge the support by the Ministry of Science and Technology of Korea. We also would like to give special thanks to Health Physics Department of Korea Atomic Energy Research Institute (KAERI) for partial support to this work. REFERENCES

Absorbed dose calculated per photon (Gy) 2.13 2.53 1.98 2.48 2.93 2.33 2.47 2.81 2.42 1.91 2.29 2.40 2.57 2.31 2.43 2.91 10 1028 1028 1028 1028 1028 1028 1028 1028 1028 1028 1028 1028 1028 1028 1028
28

6793.6 8678.6 7115.3 8323.2 9992.7 8016.4 7882.6 9585.8 7589.9 6377.9 7746.1 7596.4 8599.7 7808.4 7940.0 9986.7

Arbitrary unit

Figure 5. Ratio of absorbed doses measured from TLD to those from Monte Carlo calculation for 16 organs and tissues.

bladder, heart, lungs and stomach in the Korean physical phantom received $5% lower absorbed dose than those of tomographic model. Brain dose of the phantom was 6% higher than that of the calculation model. These discrepancies may be caused by limited number of TLD chips placed in selective organ region. In case of adrenals as an example, only one TLD chip was situated in right adrenal gland although there are right and left

1. International Commission on Radiation Units and Measurements. Phantoms and computational models in therapy, diagnosis and protection. ICRU Publication 48 (Oxford: Pergamon Press) (1992). 2. Saylor, W. L. and Adams, B. L. The patient equivalence of the RANDO phantom for cobalt gamma rays. Radiology 92, 165 (1969). 3. Jones A. K., Hintenlang D. E. and Bolch W. E. Tissueequivalent materials for construction of tomographic dosimetry phantoms in pediatric radiology. Med. Phys. 30, 20722081 (2003).

135

J. I. KIM ET AL. 4. International Commission on Radiological Protection. Reference man: anatomical, physiological and metabolic characteristics. ICRP Publication 23 (Oxford: Pergamon Press) (1975). 5. International Commission on Radiological Protection. Basic anatomical and physiological data for use in radiological protection: Reference values. ICRP Publication 89 (Oxford: Pergamon press) (2003). 6. International Atomic Energy Agency. Compilation of anatomical, physiological metabolic characteristics for a reference Asian man Volume 1: data summary and conclusions. IAEA TECDOC Series No. 1005, Vol. 1 (Vienna: IAEA) (1998). 7. International Commission on Radiation Units and Measurements. Tissue substitutes in radiation dosimetry and measurement. ICRU Report 44 (Oxford: Pergamon Press) (1984). 8. Ashley, S. Rapid prototyping is coming of age. Mech. Eng. 117(7), 6268 (1995). 9. Waters, L. S., Ed. MCNPX users manual, version 2.4.0. Los Alamos National Laboratory LA-CP-02-408 (Los Alamos, NM) (2002).

136