Calculo Ontecarlo Metodologia

Monte-Carlo Method Python Library for dose distribution Calculation in
Brachytherapy
R.D. Randriantsizafy
Madagascar Institut National des Sciences et Techniques Nuclaires (Madagascar INSTN), BP
4279, Antananarivo 101, Madagascar
R. Raboanary
Institut of @stro and High-Energy Physics Madagascar (@HEP-MAD), University of Antananarivo,
Madagascar
M.J. Ramanandraibe
Madagascar Institut National des Sciences et Techniques Nuclaires (Madagascar INSTN), BP
4279, Antananarivo 101, Madagascar
The Cs-137 Brachytherapy treatment is performed in Madagasacr since 2005. Time treatment
calculation for prescribed dose is made manually. Monte-Carlo Method Python library written at
Madagascar INSTN is experimentally used to calculate the dose distribution on the tumour and
around it. The first validation of the code was done by comparing the library curves with the
Nucletron company curves. To reduce the duration of the calculation, a Grid of PC's is set up with
listner patch run on each PC. The library will be used to modelize the dose distribution in the CT
scan patient picture for individual and better accuracy time calculation for a prescribed dose.
1. INTRODUCTION
Brachytherapy is the short distance treatment of cancer using encapsulated small sources.
In Madagascar, oncology department uses intracavitary treatment brachytherapy type with
Cs-137 27mCi Source. It is a Low Dose Rate (LDR) treatment.
For good practice, the dose at a point of treatment (tumour) have to be known with a good
accuracy, and the dose distribution in the body too. Doses can be determined using three
ways: The Treatment Planning System (TPS), Manual Calculation and Standard 2D Dose
distribution curves. Madagascar Oncology Department uses Dose curves and Manual
calculation to determine the time treatment corresponding to the prescribed dose.
The objectif of this work is the conception of Monte-Carlo Python Library which follows the
path of a photon from Cs-137 source in an homogeneous water equivalent body.
2. METHOD
For
the Cs-137
brachytherapy
treatment,
using Nucletron
Selectron
LDR
remote
afterloaders unit, the photons come from around ten Cs-137 sources of 2mm size positioned in
fletcher applicators. 1.5cm to 3cm ovoids are used to maintain the applicators inside vagina
and rectum. For this model we consider one Cs-137 source as a point source. The calibration
of Cs-137 sources was performed with the applicators, so the activity of the source is an
apparent activity and it is not necessary to take into account the presence of the aluminium
thickness of the applicators. The calibration is based on Air Kerma measurement at 1m.
From the source, the photons are considered as interacting directly with the body which is
water equivalent material. As the photon energy from Cs-137 is 662 keV and Z=10 for the
water, the curve of proportion interaction effect (see Figure 1) shows that the Compton effect
is dominant in front of Photoelectric effect and Pair production effect.
Figure 1: Interactions proportion
Figure 2: Compton interaction
After travelling a random free path (see Figure 2) the incident photon (h) interacts with on
free electron. The interaction produces one Compton electron and one scattered photon (h').
The random free path d of the incident photon between two interaction is determined with
the following formula.
d=
1
1
(1)
where is a random number and the linear attenuation coefficient depending on the
incident photon energy and the target material. We consider that the Compton electron
energy and the difference between incident photo energy and scattered photon energy are
absorbed by a voxel of 1cm3 at the interaction point. The energy of the scattered photon is
given by
h
h'=
1
h
1cos
m0
(2)
The scattered photon which is the new incident photon after interaction will continue in the
same way until it leaves the volume of 20cmx20cmx20cm or until its energy is below 100keV.
The probability distribution function (PDF) used by the method is based on Klein Nishina
cross section

d
2 '
=
d 2m0
'
2
sin
'
(3)
from which a database of a cumulative probability distribution function (CPDF) for each
diffusion angle (0 to 2 radian) and for each photon energy (From 100 to 662 keV) is precalculated (see Figure 4). The incident photon from the source are fixed as a packet of 1000
photons.
Figure 4: The model
Table I: Margin specifications

CPU
Pentium
Random Memory
256 MBytes
Number of Photons
5900000
Voxel size
Duration
1 cm
1000 sec
733MHz
3. RESULTS
Figure 5: Dose proportion around the source

We validate the model by comparing the Monte Carlo distribution dose curve with the
Nucletron curves furnished by the Nucletron Company.
Figure 6: Dose proportion around the source

The values on the Nucletron curve are placed on the Monte Carlo (MC) curve. The MC curve
fit the Nucletron curve, that means that the Monte Carlo library can be used as basis engine
to calculate the dose distribution in a body. The calibration of the source is done according to
the TG 43 recommendation (4)
K air d air =
Aapp AKR
d2
(4)
The dose is determined from the air kerma at 1 m of the source which is used to calculate the
apparent activity of the Cs-137 source and the fluence. Then the fluence is introduced in the
model.
4. DISCUSSION
The calculation is done for one source. For one treatment, around ten pellet sources will be
taken into account, that means more time for the calculation. To solve this state a python
patch is installed on machines listens the PC called calculation server which share the
calculation packet to the PC member of the grid (Fig 7).
5. OUTLOOK
The library will be used in a GUI Tkinter for clinical application. The dose distribution
visualisation will be performed with VTK library. The introduction of the code in the hospital
will be done in 2 steps:
First step: Code generate 3 D dose curves which will be stored in the computer. Befor a
treatment, the prescribed curve will be placed in patient radiography picture to see the dose
for each organ.
Second step: When the CT Scan of the Radiotherapy service will be installed, the model will
be used to draw the 3D Dose curve for one patient. The CT scan picture and the Source
positions in the applicator are the inputs of the model.
6. CONCLUSION
The comparison of the Monte Carlo curve and the Nucletron Curve (see Figure 6), for
validating the model, shows that the results of the model fot the values of the manufacturer
(Nucletron). The energy is considered as been absorbed by the voxel of 1 cm3. But for more
accuracy, the volume of the voxel has to be decreased, that means more photons and more
computation time. For clinical use, the calculation has to be performed with more powerful
CPU (more than 3GHz) and if it is not enough (calculation too long), through a grid of
computers.
The advantage of Monte Carlo method is that it has the capability to take into account the
real volume model of the patient and shape an individual 3D dose curve for each patient, wich
is not not possible with the manufacturer standard curve.
Ackowledgments
We would like to express our sincere thanks to the International Atomic Energy Agency
(IAEA) for all technical and scientific supports to the dosimetry department through the
provision of dosimetry standard, training and expertise. We are very gratefull to High Energy
Physics Madagascar (HEPMAD) for allowing us to show the first step of this work and for the
fruitfull discussion we have during the HepMad 2007 conference. Many thanks to all
Madagascar INSTN Staff especially the Dosimetry and Radiation Protection Department
member.
References
[1] Introduction to Monte-Carlo Methods for Transport and Diffusion Equations. B.Lapeyre,
E.Pardoux and R.Sentis. Oxford University Press 2003- ISBN 0-19 852592 3
[2] Radiation Oncology Physics: A Handbook for Teachers and Students. IAEA-COMP/CCPM,
EFOMP, ESTRO, IOMP, PAHO, WHO. IAEA 2005. ISBN 92-0-107304-6

Calculo Ontecarlo Metodologia

Uploaded by

Copyright:

Available Formats

Calculo Ontecarlo Metodologia

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Calculo Ontecarlo Metodologia

Uploaded by

Copyright:

Available Formats

Monte-Carlo Method Python Library for dose distribution Calculation in

Figure 1: Interactions proportion

Figure 2: Compton interaction

Figure 4: The model

Table I: Margin specifications

Figure 5: Dose proportion around the source

Figure 6: Dose proportion around the source

You might also like