monte carlo calculation of dose to spinal structures from intrathecally administered yttrium-90...
TRANSCRIPT
MONTE CARLO CALCULATION OF DOSE TO SPINAL STRUCTURES FROM INTRATHECALLY MONTE CARLO CALCULATION OF DOSE TO SPINAL STRUCTURES FROM INTRATHECALLY ADMINISTERED YTTRIUM-90ADMINISTERED YTTRIUM-90
Michael Hall1 MSc, George Mardirossian1 Ph.D, Joseph Montebello1 MD, Nina Mayr1 MD and William Yuh1 MD.1 Oklahoma University Health Sciences Center, Oklahoma City, Oklahoma
MONTE CARLO CALCULATION OF DOSE TO SPINAL STRUCTURES FROM INTRATHECALLY MONTE CARLO CALCULATION OF DOSE TO SPINAL STRUCTURES FROM INTRATHECALLY ADMINISTERED YTTRIUM-90ADMINISTERED YTTRIUM-90
Michael Hall1 MSc, George Mardirossian1 Ph.D, Joseph Montebello1 MD, Nina Mayr1 MD and William Yuh1 MD.1 Oklahoma University Health Sciences Center, Oklahoma City, Oklahoma
Objective: The treatment of cerebrospinal fluid (CSF)
malignancies by intrathecal administration of Iodine-131 radiolabeled monoclonal antibodies has been previously studied. The assumption that more healthy tissue will be spared when a pure beta-emitter is employed has lead to the proposal of Yittrium-90 as a preferable radioimmunoconjugate to 131I. It is believed that the shorter range of beta particles will offer better localization of the dose to metastases found within the subarachnoid space. However, such assumptions have not been accurately verified and validated. The purpose of this study is to quantitatively evaluate the dose distribution to the CSF space and its surrounding spinal structures following intrathecal administration of 90Y based upon Monte Carlo modeling of the 90Y source coupled with a 3-D digital phantom of a region of the spinal cord.
Method: A 3-D digital phantom of a section of the T-spine
was constructed from the Visible Human Project series of images1. The original color photographs of the female cadaver were used to produce an 8-bit gray-scale series of images with dimensions 0.66 x 0.66 x 0.66 mm3. This series of images was further processed so that only certain anatomical features of interest were represented. These features of interest include the spinal cord, central canal, subarachnoid space, pia mater, arachnoid, dura mater, vertebral bone marrow, and intervertebral disc.
Monte Carlo N-Particle (MCNP) version 4C2 was used to model the 90Y radiation distribution. The size of the cells modeled with MCNP was made to coincide with the voxel size of the 3-D phantom. A set of images of just the CSF compartment was convolved with the radiation distribution obtained from the MCNP calculations to determine the dose distribution within the subarachnoid space and surrounding tissues. To achieve this, a fast Fourier transform (FFT) was utilized to transform the two sets of data into frequency space where they were multiplied together. The inverse Fourier transform was performed on their product in frequency space to obtain the convolution3.
The cumulated activity of 90Y in the CSF compartment was calculated based on previously-published 131I pharmacokinetic biodistribution data to obtain the overall dose due to the presence of 90Y in the CSF4.
ResultsResultsResultsResults
Table 1. Summary of Non-Zero MCNP Tallies indicating reliability of data obtained.
Range of RDistance from
Source Cell (mm)% of Total
Deposited EnergyNumber of cells
Quality of the Tally
< 0.10 0 to 0.692 96.680% 4417 Generally reliable
0.1 to 0.2 0.653 to 0.814 2.490% 2615 Questionable
0.2 to 0.5 0.753 to 0.974 72.00% 3478 Factor of a few
0.5 to 1.0 0.837 to 1.12 0.11% 3586 Not meaningful
Conclusion: Confirms that 90Y would be a suitable radionuclide in the treatment of intrathecal
neoplasms and Offers an improvement over 131I in such an application. Favorable dose distribution to critical tissues such as bone and bone marrow. dose distribution of 90Y radionuclide would be confined largely within the vertebral
foramen.
Figure 5. Binary Phantom Slice before (A) and After (C) Convolution With Energy Deposition Kernel (B).
Radionuclide Kernel: MCNP4C6 is used to simulate the tracks of
betas in tissue when the betas have energies sampled from the 90Y source probability distribution file PDF7. The convolution technique developed to calculate dose required that the MCNP cells have the same dimensions as the phantom voxels: 0.66 x 0.66 x 0.66 mm3 (or 2.87 x 10-4 cm3). The published range of 2.28 MeV electrons was found to be approximately 1 cm for water and adipose tissue. With the origin defined at the center of the source, a matrix of 31x31x31 would provide sufficient distance for the maximum energy emission of 2.28MeV to be stopped within the tissue. The energy deposition was computed in each voxel lattice (MeV/dis) for 1 millions histories. This choice led to a run time of 46.7 hours, or approximately 2 days, on PC with a 2.99 GHz processor and the Windows XP operating system.
3D convolution: A convolution macro was written in IDL3 to perform the
discrete convolution of the 3-D binary phantom with the 3-D kernel matrix. The digital images of VisWoman were convolved with the radiation kernel distribution obtained from MCNP4C calculations to determine the dose distribution.
Figure 3. (A) Transverse and (B) Sagittal slice of the Anatomical Phantom Used for Dosimetry .
Figure 1. (A)- A sample VisWoman image. (B)- A region of spine including T8 and parts of T7 and T9 was chosen as a representative section for the phantom (rectangle).
Pharmacokinetics : In the Papanastassiou studies4, forty patients
with a variety of malignancies administered a single injection of 131I-labeled mAb’s via an Ommaya reservoir. The activity was sampled from the ventricles and bloodstream as a function of time to determine absorption and elimination constants in these compartments. They reported that the radioimmunoconjugate enters the bloodstream quickly, with peak blood levels of radioimmunoconjugate occurring 24 to 56 hours post-injection. Using the Papanastassiou value of Tb=12.55 hr for biological half-life and incorporating the physical Tp = 64.1 hr for 90Y radioimmunoconjugate, one obtains Teff =10.50 hr. Consequently, it is assumed that for patients injected with 90Y radioimmunoconjugate intrathecally, the measured elimination constant would be 0.066 hr-1 would be used. Recalculating the cumulated activity in the subarachnoid space, a value of 15.163 mCi•hr per mCi injected activity, which is the value of cumulated activity used in the dosimetry calculations presented in this project.
Table 3: Dose Comparisons of 131I and 90Y. The WB and BM doses are calculated using MIRD.
References: 1- Peitgen, H.-O. Bremen, Germany, Berghorn,Wilhelm; Biel Matthias; Eberhardt,T.;
Habermalz,E.; Juergens, H.; Lang, M., Netsch, T. G., Prause, U. Scheil, T. Schindewolf.The Complete High-Resolution Male and Female Anatomical Datasets from the Visible Human Project (TM), CD-ROM, ISBN 0387142479, Copyright Year 1998.
2- Briemeister JF, Editor: MCNP—A general Monte Carlo N-particle transport code, version 4C. Los Alamos National Laboratory Report, LA-13709-M, 2000.
3- Research Systems, Inc.: IDL version 5.6, 2002. 4- Papanastassiou V, Pizer BL, Chandler CL, Zananiri TF, Kemshead JT, Hopkins KI:
Pharmacokinetics and dose estimates following intrathecal administration of 131I-monoclonal antibodies for the treatment of central nervous system malignancies. Int J Radiat Oncol Biol Phys 31: 541-552, 1995.
5- Rasband W: ImageJ 1.31v. National Institude of Health, USA Public Domain.6- Regents of the University of California and Los Alamos National Laboratory: Monte Carlo
N-Particle Code version 4C, 2001. 7- Prestwich WV, Nunes J, Kwok CS: Beta dose point kernels for radionuclides of potential
use in radioimmunotherapy. J Nuc Med 30: 1036-1046, 1989.
Visible Human Project: The Visible Human
Project was undertaken by the National Library of Medicine and the result is an extraordinary collection of transverse cross-sectional images of the human anatomy. High-resolution anatomical images (radiographic, CT, MRI, and color photographs) of human cadavers have been collected for a male (38 year-old, 186 cm in height, 90 kg in weight) and a female (59 year-old, 167 cm in height, 72 kg in weight), both of whom were considered representative of the population. The female images have higher resolution and were used in this work. The VisWoman images consist of approximately 5600 slices with a slice thickness of 0.33 mm. Images are also available in resolutions of 832 x 608 (0.66 x 0.66 mm2), 416 x 304 (1.32 x 1.32 mm2) and 208 x 152 (1.32 x 1.32 mm2) on the same CD-ROM. a single 8-bit slice containing voxels with dimensions 0.66 x 0.66 x 0.66 mm3.
To create the dosimetry phantom, a total of 55 transverse slices were processed and each image was cropped to 55 x 55 pixels per slice, yielding a 3-D matrix of cubic voxels with a total dimension of 36.3 x 36.3 x 36.3 mm3. The process started with 8-bit images consisting of 256 gray-scale and resulted in slices consisting of nine gray-scale values: 0 to 8. Because each of these nine values was assigned to a separate anatomical structure, a system of identifying the different anatomical structures was created. Consequently, nine different anatomical structures were modeled.
0
20
40
60
80
100
120
140
0 5 10 15 20 25 30 35 40 45 50
Pixel
Gra
y S
cale
Valu
e
Spinal Cord
CSFCSF
Vertebra
Vertebra
0
30
60
90
120
150
180
210
0 5 10 15 20 25 30 35 40 45 50
Pixel
Gra
y S
cale
Valu
e
Spinal Cord
CSFCSF
Ligament (Interver-tebral Disc)
Vertebra
Figure 2. Plot profiles of gray scale values as a function of position along the yellow rectangles. The plot profiles illustrate the variation of gray scale values according to tissue type. Thresholds are set for the intervertebral disc ligament and the CSF compartment. Using ImageJ, it was determined that almost all pixels with a gray scale value above 140 corresponded to ligament and almost all pixels with a gray scale value below 74 corresponded to the CSF compartment. Any tissue with a gray scale between these values corresponds to spinal cord, bone, fat, or meninges.
0.9
90
0.9
24
0.8
58
0.7
92
0.7
26
0.6
60
0.5
94
0.5
28
0.4
62
0.3
96
0.3
30
0.2
64
0.1
98
0.1
32
0.0
66
0.0
00
0.0
66
0.1
32
0.1
98
0.2
64
0.3
30
0.3
96
0.4
62
0.5
28
0.5
94
0.6
60
0.7
26
0.7
92
0.8
58
0.9
24
0.9
90
0.00%
0.65%
1.30%
1.95%
2.60%
3.25%
3.90%
4.55%
5.20%
5.85%
6.50%
Figure 4. Energy deposition normalized to total energy deposited in the kernel
A B C
Table 2. 90Y Dose to Spinal Structures.
Spinal Structure
Dose (cGy/mCi)
Dose (cGy/MBq)
Normalized Dose
CSF 142.1 3.84 1.00
Pia Mater 94.3 2.55 0.66
Arachnoid 72.7 1.96 0.51
Central Canal 62.7 1.69 0.44
Spinal Cord 58.7 1.59 0.41
Dura Mater 46.7 1.26 0.33
Epidural Fat 24.8 0.67 0.17
Bone Surface 18.7 0.51 0.13
Bone Average 4.77 0.13 0.03
Ligament 1.0 0.03 0.01
RadionuclideInitial Activity
(MBq)
Dose (Gy)
CSF Pia WB BM131I 1020 20.0 3.67 0.14 0.9590Y 521 20.0 13.3 0.13 1.21131I 5556 109 20.0 0.78 5.1790Y 784 30.1 20.0 0.20 1.82
Discussion: Consider the specific case of treating the CSF with 20 Gy
after one injection. In the 131I scenario, the amount of initial activity required to administer this dose to the CSF is 1,020 MBq, which, in turn, would deliver 3.67, 0.14, and 0.95 Gy to the pia mater, whole body, and bone marrow respectively. In the case of 90Y, the amount of initial activity required is 521 MBq, which, in turn, would deliver 13.3, 0.13, and 1.21 Gy to the pia mater, whole body, and bone marrow respectively. This is about one-half of the initial activity required compared to 131I, and would result in almost four times the dose to the pia mater. Furthermore, the whole body would receive less dose overall and only slightly greater dose would be delivered to the bone marrow. Alternatively, consider the case of treating the pia mater with 20 Gy after one injection. In the 131I scenario, the amount of initial activity required to administer this dose to the pia would be 5,556 MBq, which, in turn, would deliver 109, 0.78, and 5.17 Gy to the CSF, whole body, and bone marrow respectively. In the case of 90Y, the amount of initial activity required is 784 MBq, which, in turn, would deliver 30.1, 0.20, and 1.82 Gy to the CSF mater, whole body, and bone marrow respectively. In this case, treating the pia mater with 90Y could be accomplished with about one-seventh of the initial activity required if 131I were used, but the CSF would receive significantly less dose than if 131I were used. Administering 90Y to treat the pia mater would also result in less dose to the whole body and bone marrow.
Radionuclide CSF Dose
(cGy/MBq ) Pia Mater Dose
(cGy/MBq ) Whole Body Dose
(cGy/MBq) Bone Marrow Dose
(cGy/MBq) 90Y 3.84 2.55 0.025 0.232 131I 1.96 0.36 0.014 0.093
Specific cases: •Treat CSF with 20 Gy
or •Treat Pia with 20 Gy
Figure 6. Transverse (A) and Frontal (B) Slice of Dose Distribution.
A B
A B