Journal of the Korean Physical Society, Vol. 64, No. 1, January 2014, pp. 114∼121
Evaluation of the Dependence of the Exposure Dose on the Attenuation Correction in Brain PET/CT Scans Using 18 F-FDG Eun-Jin Choi Department of Public Health and Medicine, Dongshin University Graduate School, Naju 520-714, Korea
Moon-Taeg Jeong, Seong-Joo Jang, Nam-Gil Choi and Jae-Bok Han Department of Radiological Science, Dongshin University, Naju 520-714, Korea
Nam-Hee Yang Department of Radiological Physics, Dongshin University, Naju 520-714, Korea
Kyung-Rae Dong Department of Radiological Technology, Gwangju Health University, Gwangju 501-701, Korea, and Department of Nuclear Engineering, Chosun University, Gwangju 501-759, Korea
Woon-Kwan Chung∗ Department of Nuclear Engineering, Chosun University, Gwangju 501-759, Korea
Yun-Jong Lee Advanced Radiation Technology Institute, Korea Atomic Energy Research Institute, Jeongeup 580-185, Korea, and Department of Nuclear Engineering, Chosun University, Gwangju 501-759, Korea
Young-Hwan Ryu Department of Radiology, Seoul Medical Center, Seoul 131-865, Korea, and Department of Nuclear Engineering, Chosun University, Gwangju 501-759, Korea
Sung-Hyun Choi Department of Radiology, Kyung Hee University Hospital at Gang-dong, Seoul 134-727, Korea, and Department of Nuclear Engineering, Chosun University, Gwangju 501-759, Korea
Kyeong-Jeong Seong Department of Radiology, Saint-Carollo Hospital, Suncheon 540-719, Korea, and Department of Nuclear Engineering, Chosun University, Gwangju 501-759, Korea (Received 27 June 2013, in final form 8 October 2013) This study examined whether scanning could be performed with minimum dose and minimum exposure to the patient after an attenuation correction. A Hoffman 3D Brain Phantom was used in BIO 40 and D 690 PET/CT scanners, and the CT dose for the equipment was classified as a low dose (minimum dose), medium dose (general dose for scanning) and high dose (dose with use of contrast medium) before obtaining the image at a fixed kilo-voltage-peak (kVp) and milliampere (mA) that were adjusted gradually in 17 - 20 stages. A PET image was then obtained to perform an attenuation correction based on an attenuation map before analyzing the dose difference. Depending on tube current in the range of 33-190 milliampere-second (mAs) when BIO 40 was used, a significant difference in the effective dose was observed between the minimum and the maximum mAs (p < 0.05). According to a Scheffe post-hoc test, the ratio of the minimum to the maximum of the effective dose was increased by approximately 5.26-fold. Depending on the change in the tube current in the range of 10-200 mA when D 690 was used, a significant difference in the effective dose was observed between the minimum and the maximum of mA (p < 0.05). The Scheffe posthoc test revealed a 20.5-fold difference. In conclusion, because effective exposure dose increases with increasing operating current, it is possible to reduce the exposure limit in a brain scan can be -114-
Evaluation of the Dependence of the Exposure Dose on the Attenuation· · · – Eun-Jin Choi et al.
-115-
reduced if the CT dose can be minimized for a transmission scan. PACS numbers: 87.58.-b, 87.58.Fg, 87.58.-Sp, 87.58.Vr Keywords: Attenuation correction, Hoffman 3D Brain Phantom, PET/CT scanner DOI: 10.3938/jkps.64.114
I. INTRODUCTION In recent years, use of positron emission tomography/computed tomography (PET/CT) has become widespread. Since fluorine-18-fluoro deoxy glucose (18 FFDG) has been used to examine pathophysiological changes, PET/CT has become an important diagnostic tool that uses a different diagnostic approach from the existing tool that is used to examine only the morphological change [1]. In addition, as a PET scanner is combined with a CT scanner, which uses X-rays, an anatomical image can be converged with a functional image in terms of hardware. Therefore, PET/CT provides the precise location of the lesion and improves the diagnostic performance of the PET image itself [2,3]. A CT image in PET/CT assists in making a diagnosis and is used as an attenuation map to perform an attenuation correction in a PET emission scan. In existing PET, germanium-68, which emits radiation, such as 511-keV gamma rays that are generated in an annihilation reaction, is used to perform a transmission scan before an attenuation correction is conducted. On the other hand, the method requires significant time to obtain data sufficient for an attenuation correction, which results in a long time being needed for an examination. Nevertheless, CT requires a short time to obtain data sufficient for an attenuation correction. Regarding data, the transmission data are used to correct the emission scan data in PET, which enables a decrease in the scan time [4,5]. Consequently, the attenuation correction time is around 20 – 30 minutes when the GE-68 Phantom in the existing PET is used. The time can be reduced to less than one minute when CT is used. Moreover, the number of photons for a CT image is much higher than that of photons for a transmission image. Hence, a CT image has high resolution with much less noise in an attenuation map. For this reason, it is inevitable to use a segmented attenuation correction (SAC), where a uniform attenuation constant for some tissues is applied, depending on the range of attenuation constant, in order to reduce the noise in the PET image that is obtained with the use of an external source for a transmission scan and to reduce the time that is required for a transmission scan. The measured attenuation correction (MAC) obtained by using a CT image enables the attenuation constant in the pixel unit that is obtained in the attenuation map to be applied by using the Hounsfield Unit (HU), which is a unique value of the CT (Fig. 1). However, the MAC has a weak point in that it increases the patient’s radiation ∗ E-mail:
[email protected]; Fax: +82-62-232-9218
Fig. 1. (Color online) Image reconstruction with an attenuation correction.
exposure [6,7]. In regard to the actual reading of a PET image, the critical issue is to determine if there is any difference in the intake of a PET image obtained using a CT image. On the other hand, CT is used simply as a tool for to obtain an attenuation correction when it is used to simply examine the distribution of a radiopharmaceutical in patients with dementia, epilepsy, Alzheimer’s disease, etc. as in the case of a brain PET/CT scan. Therefore, this study intends to investigate if there is any change in image quality after an attenuation correction is performed for a patient at the minimum dose and the minimum radiation exposure.
II. MATERIALS AND METHODS 1. PET/CT System and Image Acquisition
This study was conducted using the PET/CT systems Biograph Sensation 40 (SIEMENSE, Germany, hereafter BIO 40) and Discovery 690 (General Electric, USA, hereafter D 690). Both systems were reconstructed using an iterative reconstruction algorithm. Water and 18 FFDG (1 mCi, 37 MBq) were poured into and diluted in a Hoffman 3-D Brain Phantom (Fig. 2). An image was obtained in one hour.
-116-
Journal of the Korean Physical Society, Vol. 64, No. 1, January 2014
Fig. 2. (Color online) Hoffman 3-D brain phantom.
Fig. 3. (Color online) Image data acquisition in BIO 40. Fig. 4. (Color online) Image data acquisition in D 690. 2. Image Data Acquisition in BIO 40
As the tube voltage was fixed at 120 kilo-voltage-peak (kV), the milliampere-second (mAs) was increased gradually from 33 mAs to 190 mAs in 10-mAs intervals to obtain Hoffman 3-D Brain Phantom data (Fig. 3). Afterward, the PET image (15 min/1 bed) was obtained to make an attenuation map under each condition before making a correction. In general, the degree of attenuation that takes place when the photon goes through matter can be expressed as Φ = Φ0 [−
S
µ(x, y)dr],
(1)
where Φ and Φ0 represent the numbers of transmitted and incident photons per unit area respectively, dr represents the thickness of the medium, and µ is the attenuation coefficient that represents the degree per unit length to which physical phenomenon, such as the photoelectric effect or the compton scattering effect, takes place when the photon goes through a medium. The equation above can be rearranged to as µ(x, y)dr = loge (
Φ0 ), Φ
(2)
In this case, exp[ µ(x,y)dr] is defined as the attenuation correction factor (ACF) An attenuation correction
for PET data is the process of multiplying emission data by the ACF.
3. Image Data Acquisition in D 690
The tube voltage was fixed to 140 kVp, and the milliampere (mA) was increased gradually from 10 mA to 200 mA in 10-mA intervals to obtain Hoffman 3-D Brain Phantom data (Fig. 4). Afterward, the PET image (10 min/1 bed) was obtained to make an attenuation map for each condition before making a correction.
4. Calculation of CT Effective Dose
The CT effective dose was calculated using the ImPACT CT Patient Dosimetry Calculator program based on a Monte Carlo simulation embedded in the PET/CT workstation (Fig. 5). Dosimetry is the underlying concept generally used to relate the energy imparted by radiation exposure in a volume of tissue to a potential biological effect. This effect can be stochastic in nature, where cancer is induced after a substantial time after exposure, or deterministic, where particular damage and cell death can be predicted after certain levels of radiation
Evaluation of the Dependence of the Exposure Dose on the Attenuation· · · – Eun-Jin Choi et al.
-117-
the radionuclide and the biological half-life determined from the biokinetic behavior of the radionuclide (or radiopharmaceutical). The cumulative activity can be calculated using the following equation: = A0 A
∞
0
f (t)dt,
(4)
where A0 is the initial activity administered at the injection time and f(t) is a time-dependent function. Mostly, f(t) is a combination of one or several exponential functions that, in principle, are specific to each individual and radiopharmaceutical. The S value can be described as follows:
S rT ←rS =
Fig. 5. (Color online) ImPACT CT Patient Dosimetry Calculator program.
exposure are reached. The basic unit that is commonly believed to relate an amount of energy imparted to a biological effect (or the risk of an effect) is the absorbed dose, which is expressed as the mean energy absorbed in as element (J/kg) of tissue. The process of calculating the absorbed dose is quite complicated because it involves several types of coupled photons and chargedparticle interactions in a radiation transport that has a stochastic nature. Therefore, the absorbed dose cannot be measured directly in vivo. As a consequence, all the results are actually estimates with relatively large error bars. The Medical Internal Radiation Dose Committee of the Society of Nuclear Medicine introduced a formalism for calculating the absorbed dose for the medical use of radiopharmaceuticals in the late 1960s [8] and has since written several publications in the area. The formalism is described by the general equation r × Sr ←r , D rT = A S T S
(3)
where D is the mean absorbed dose in a target volume, A is the cumulative activity (e.g., total number of disintegrations over a time interval) in the source volume, and S is a factor that defines the mean absorbed dose in the target volume per unit cumulative activity in the source volume. For internally-administered radionuclides, the cumulative activity depends on the physical half-life of
n · E · ΦrT ←rS , mrT
(5)
where n and E are the number of particles emitted per nuclear transition and the energy of the particles, respectively, and m is the mass of the target volume. The term Φ is called the absorbed fraction and is defined as the fraction of the energy emitted from the source volume that will be absorbed in the target volume. Equation (3) becomes, when broken down into its smallest constituents, D rT =
rS [ArS
·
ni · Φi (rT ← rS )] , mrT
i
(6)
where the absorbed fraction Φ(rT ← rS ) is defined as the fraction of energy emitted from the source rS and absorbed in the target rT . The principles of the MIRDformalism remain valid for magnitudes of the absorbed dose ranging from those at organ levels to those at the cellular levels, but are most commonly used to calculate the mean dose absorbed in whole organs. For practical purposes in dosimetry for risk estimates, the S values are precalculated using a Monte-Carlo calculation on a mathematical phantom [9].
5. Statistical Analysis
The CT effective dose was calculated as the data were entered into the ImPACT CT Patient Dosimetry Calculator program 10 times and depended on the characteristics of the equipment and the CT conditions used in the experiment. ANOVA was performed for statistical analysis. The post-hoc test of Scheffe was conducted when ANOVA showed a significant difference between groups. A p value < 0.05 was considered significant.
-118-
Journal of the Korean Physical Society, Vol. 64, No. 1, January 2014
Table 1. Effective dose measured in BIO 40 by using the ImPACT CT Patient Dosimetry Calculator program.
mAs 33 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190
N 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
BIO 40 Mean ± SD 0.14 ± 0.01a 0.16 ± 0.01b 0.20 ± 0.01c 0.24 ± 0.01d 0.28 ± 0.01e 0.32 ± 0.01f 0.36 ± 0.01g 0.40 ± 0.01h 0.44 ± 0.00h 0.47 ± 0.01i 0.51 ± 0.01j 0.55 ± 0.02k 0.59 ± 0.01l 0.64 ± 0.01m 0.67 ± 0.01n 0.71 ± 0.01o 0.75 ± 0.01o
F
p
3193
0.000∗∗∗
Note: The interaction effect was determined using one-way ANOVA and the post–hoc test of Scheffe. The group of the same alphabet means the same group of the average level. The alphabet order means a difference between groups. The unit is the number of mSv. ∗∗∗ p < 0.001.
Fig. 6. (Color online) Effective dose as a function of the mAs measured in BIO 40 by using the ImPACT CT Patient Dosimetry Calculator program.
Table 2. Effective dose measured in D 690 by using the ImPACT CT Patient Dosimetry Calculator program.
mAs 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
N 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
BIO 690 Mean ± SD 0.04 ± 0.00a 0.08 ± 0.00b 0.12 ± 0.01c 0.16 ± 0.01d 0.21 ± 0.01e 0.25 ± 0.01f 0.29 ± 0.01g 0.33 ± 0.01h 0.37 ± 0.01i 0.41 ± 0.01j 0.46 ± 0.01k 0.49 ± 0.01l 0.53 ± 0.01m 0.59 ± 0.01n 0.62 ± 0.01o 0.66 ± 0.01p 0.70 ± 0.01q 0.75 ± 0.01r 0.79 ± 0.01s 0.82 ± 0.01t
F
p
11269
0.000∗∗∗
Note: The interaction effect was determined using one-way ANOVA and the post-hoc test of Scheffe. The group of the same alphabet means the same group of the average level. The alphabet order means a difference between groups. The unit is the number of mSv. ∗∗∗ p < 0.001.
Fig. 7. (Color online) Effective dose as a function of the mA measured in D 690 by using ImPACT CT Patient Dosimetry Calculator program.
III. RESULTS 1. Total Effective Dose in BIO 40
The ImPACT CT Patient Dosimetry Calculator program was used to fix the tube voltage at 120 kVp in BIO 40. Then, the difference between the minimum and
the maximum mAs was found to be statistically significant when the change in the tube current was 33-190 effective mAs (p < 0.05). A Scheffe post-hoc test showed that the effective dose ratio of the minimum to the maximum dose was increased by approximately 5.26 times (Table 1, Fig. 6).
Evaluation of the Dependence of the Exposure Dose on the Attenuation· · · – Eun-Jin Choi et al.
-119-
Fig. 8. (Color online) Characteristics of the PET attenuation correction.
2. Total Effective Dose in D 690
Fig. 9. (Color online) Image for attenuation correction.
The ImPACT CT Patient Dosimetry Calculator program was used to fix the tube voltage at 140 kVp in D 690. Then, the difference between the minimum and the maximum mA was found to be statistically significant when the change in the tube current was 10 - 200 mA (p < 0.05). The Scheffe post-hoc test revealed a 20.5 fold difference (Table 2, Fig. 7).
IV. DISCUSSION The attenuation correction is a factor with the greatest impact on the quality of a PET image. The attenuation phenomenon in a PET image means that the degree of attenuation on the line of response (LOR) between the two detectors is constant regardless of the location. When two gamma rays are generated at a random point (Fig. 8) and the probability that a gamma ray is detected is multiplied by the probability that the other gamma ray is detected, the product of such probabilities is constant regardless of the location, as shown in Eq. (7): E = E0 × exp(−
l1
0
= E0 × exp(−
µ(x, y)ds) × exp(−
l1 +l2 =L
µ(x, y)ds).
0
l2 0
µ(x, y)ds) (7)
This equation is applied equally to the relationship between a blank scan (B) and a transmission scan (T), as shown in T = B × exp(−
l1 +l2 =L
0
µ(x, y)ds),
(8)
Therefore, when a blank scan (B) anda transmission scan (T) are segmented by each pixel and multiplied by the emission scan data, as shown in the Eq. (9),
E0,γθ =
Bγθ × EBγθ , Tγθ
(9)
it is possible to obtain photon attenuation-corrected emission scan data (Fig. 9). The use ofa CT image for attenuation correction of a PET image has some strong points compared to the
use of an external source for a transmission scan [4,10]. First, much less time is needed to obtain a CT image than a transmission image, which can reduce the scan time. Second, the number of photons for a CT image is much higher than that of photons for a transmission image whereas a CT image has a high resolution with much less noise in an attenuation map. To reduce the noise in a PET image obtained using an external source to conduct a transmission scan and to reduce the time needed for a transmission scan; it is inevitable to use the SAC, where, depending on the range of attenuation constants, uniform attenuation constants for several tissues are applied. On the other hand, in the MAC where a CT image is used, it is possible to apply the attenuation constant obtained in an attenuation map to the pixel unit itself. Unfortunately, the MAC has a weak point in that the patient’s exposure is increased [7,11]. PET/CT consists of stand alone PET and CT systems. CT is performed first because it requires less than several minutes for scanning. The PET image is then obtained. A PET/CT fusion image, which is obtained as the PET image is overlapped with the CT image, is useful for describing the relationship between the biological structure and function. A CT image ensures higher partial volume effects than a PET image and enables an attenuation correction in a more precise manner than a PET image dose. If a quantitative PET image is to be obtained, it is essential to perform an attenuation correction of gamma rays for the human body. In existing PET imaging, external sources, such as 68 Ge and 137 Cs, are used to conduct a transmission scan for an attenuation correction. With the introduction of PET/CT, the attenuation correction is conducted using an attenuation map, where the X-ray attenuation coefficient of the CT is converted to an attenuation coefficient for a 511 keV photon of the PET. Because the γ-rays that are emitted from the human body due to annihilation radiation show an intensity that varies according to the surface and the depth owing to scattering and absorption, it is essential to make a correction of the actually measured data after the measurement in advance of the data on scattering and absorption in the human body to measure the precise value for γ-rays. The collection of data for the purpose is called a transmission scan whereas scanning after injection of a radiopharmaceutical is called emission scan). The absorption correction has strong points. This en-
-120-
Journal of the Korean Physical Society, Vol. 64, No. 1, January 2014
ables a description of the size, shape or location of a lesion more realistically, as well as measurements of the precise value of the radiation concentration in the human body and the tumor by restoring the radiation intensity in the lesion at a specific depth in the human body. In PET, the transmission scan is conducted with an external source that emits γ-rays with the same energy as that for a positron-emitting radioisotope (RI) that has circulated once throughout the human body, as in the case of CT.The scan time is around 2 – 30 minutes on average, which takes up 10 – 50 time. On the other hand, CT requires one minute at most, even though scanning is performed for the whole body. Therefore, when a CT-assisted attenuation correction (CTAC) is used, it is possible to reduce the examination time by as much as the time needed for an existing transmission scan. As the CTAC provides the data in sufficient quantity compared to the case where an external source, such as 137 Cs or68 Ge, is used, the CTAC has a relationship with improved PET image quality. Consequently, an attenuation correction using CT has a strong point of decreasd examination time whereas it has the weak point of increasd patient exposure dose. Therefore, when the tube current, as the scanning condition for the CT, is lowered to approximately 5 mA to reduce the exposure for diagnostic CT, it is possible to secure sufficient data quantity [12–14]. On the other hand, the increase in the exposure dose due to CT can be a cause of medical exposure, which has been the issue recently. In this study, machines manufactured by different companies were used for measurements. Due to differences in the CT structure and the generation method of each machine, it was impossible to conduct experiments under the same conditions. However,the kVp conditions were set up based on the brain protocol that was currently applied to each machine while only the mAs (radiation quantity in mAs) was adjusted for the measurements. Then, an evaluation was conducted for each machine. A Hoffman 3D Brain Phantom was used in the BIO 40 and the D 690 PET/CT scanners, and the CT dose was classified as a low dose (minimum dose), medium dose (general dose for scanning) and high dose (dose with use of contrast medium) before the image was obtained at a fixed kVp and at a mA that was adjusted gradually in 17 – 20 stages. The PET image was then obtained to perform an attenuation correction based on the attenuation map before the effective dose was calculated and compared using the ImPACT CT Patient Dosimetry program. According to the calculation and comparison, the difference between the minimum and the maximum dose value ratio was 5.26 times for BIO 40 in the tube current range of 33 – 190 mAs. In contrast, the difference between the minimum and the maximum of the dose value ratio for D 690 was 20.5 times over the tube current range of 10 – 200 mA The attenuation correction is very important for improving the quality of a brain image. Moreover, because the attenuation correction method is easy to perform us-
ing CT, it is used for brain PET imaging in clinical practice. At the same time, an attenuation correction has the weak point of an increase in the effective exposure dose. To examine the quantitative difference in dose, this study compared the difference between the minimum and the maximum total effective doses in CT, and that difference was used for an actual attenuation correction. The comparison revealed a 5- to 20- fold difference depending on the equipment. Therefore, when a brain examination for dementia, epilepsy, Alzheime’s disease, etc. is conducted to simply check the distribution of a radiopharmaceutical without the need for diagnostic and high-definition CT, the use of low-dose CT in a transmission scan will be helpful for minimizing the patient’s exposure dose.
V. CONCLUSION The attenuation correction is very important for improving the quality of a brain image. Moreover, an attenuation correction is used for brain PET imaging in clinical practice because an attenuation correction is easy to perform when using CT. At the same time, the attenuation correction has a weak point in that it increases the effective exposure dose. Therefore, when a brain examination is conducted to simply check the distribution of a radiopharmaceutical without the need for diagnostic and high-definition CT, the use of low-dose CT in a transmission scan will be helpful for minimizing the patient’s exposure dose.
REFERENCES [1] Y. Nakamoto, M. Osman, C. Cohade, L. T. Marshall, J. M. Links, S. Kohlmyer and R. L. Wahl, J. Nucl. Med. 43, 1137 (2002). [2] R. Bar-Shalom, N. Yefremov, L. Guralnik, D. Gaitini, A. Frenkel, A. Kuten, H. Altman, Z. Keidar and O. Lsrael, J. Nucl. Med. 44, 1200 (2003). [3] G. Antoch, N. Saoudi, H. Kuehl, G. Dahmen, S. P. Mueller, T. Beyer, A. Bockisch, J. F. Debatin and L. S. Freudenberg, J. Clin. Oncol. 22, 4357 (2004). [4] P. E. Kinahan, D. W. Townsend, T. Beyer and D. Sashin, Med. Phys. 25, 2046 (1998). [5] C. Cohade, M. Osman, L. N. Marshall and R. N. Wahl, Eur. J. Nucl. Med. Mol. Imaging 30, 721 (2003). [6] E. Z. Xu, N. A. Mullani, K. L. Gould and W. L. Anderson, J. Nucl. Med. 32, 161 (1991). [7] V. Bettinardi, E. Pagani, M. C. Gilardi, C. Landoni, C. Riddell, G. Rizzo, I. Castiqlioni, D. Belluzzo, G. Luciqnani, S. Schubert and F. Fazio, Eur. J. Nucl. Med. 26, 447 (1999). [8] R. Loevinger and M. Berman, J. Nucl. Med. 9, 7 (1968). [9] L. Michael, S. Sven-Erik and A. K. Michael, Montecarlo Calculations in Nuclear Medicine, 2nd ed. (Taylor and Francis, NY, 2012).
Evaluation of the Dependence of the Exposure Dose on the Attenuation· · · – Eun-Jin Choi et al. [10] F. C. Robiller, K. D. Stumpe, T. Kossmann, D. Weishaupt, E. Bruder and G. K. von Schulthess, Eur. Radiol. 10, 855 (2000). [11] J. S. Kim, J. S. Lee and G. J. Cheon, Nucl. Med. Mol. Imaging 42, 112 (2008). [12] J. Czernin, M. Allen-Auerbach and H. R. Schelbert, J. Nucl. Med. 48, 78S (2007).
-121-
[13] T. Beyer, D. W. Townsend, T. Brun, P. E. Kinahan, M. Charron, R. Roddy, J. Jerin, J. Young, L. Byars and R. Nutt, J. Nucl. Med. 41, 1369 (2000). [14] H. K. Son, T. G. Turkington, Y. Y. Kwon, H. J. Jung and H. J. Kim, Korean J. Med. Phys. 16, 192 (2005).