Università di Pavia - Dipartimento di Fisica INFN - Sezione di Pavia
email: alessia.embriaco@pv.infn.it
The accurate evaluation of the dose distribution is an open issue in Hadrontherapy. MONET (Model of ioN dosE for Therapy) is a fast and accurate code for the evaluation of the 3D dose distribution for protons in water introduced recently by our group [1].
The MONET code
The shape of the lateral profile comes from the combination of two processes: multiple Coulomb scattering and nuclear interactions. The MONET code is based on the Molière theory of multiple Coulomb scattering. To take into account also nuclear interactions, we have added to Molière distribution the Cauchy-Lorentz function to describe the nuclear tail of the distribution [2]. The Cauchy-Lorentz distribution has only two free parameters, which are obtained by a fit to FLUKA simulations.
The next step is the passage from the projected lateral distribution to a 2D radial distribution.
The projected distributions are uncorrelated but not independent. We need to use the Papoulis theorem [3] that allows, in case of cylindrical symmetry, to rebuild the radial distribution starting from the projected one.
For the longitudinal profile, we have implemented a new calculation of the average energy loss that is in good agreement with the simulations. The inclusion of straggling is based on the convolution of energy loss with a Gaussian function. In order to complete the longitudinal profile, also the nuclear contributions are included in the model using a linear parameterization with only two parameters. In conclusion, the MONET code is able to evaluate the total dose profile in a three dimensional
37
mesh with only four free parameters, by calculating at each depth the longitudinal deposited energy and by distributing it laterally on the transverse plane.
Results
We have compared the MONET results with the FLUKA simulations in two cases: a single Gaussian beam and a lateral scan.
We have analysed three energies (100, 150 and 200 MeV). For the single beam, the results are obtained by selecting transverse plane at fixed depths [1]. An example of the comparison between FLUKA and MONET profile is reported in Fig.1. The agreement is good: the relative error, evaluated in a central cross section, is about 3%. We also present the quantile-quantile (QQ) plot to compare the quantiles of the model and MC distribution. We have obtained a similar result for protons of 100 and 200 MeV [1]. The agreement between model and MC simulation is good, the relative error is lower than 10% and also the QQ plot is good.
In order to estimate the accuracy of the model focusing on the tails of the distribution that give rise to the low-dose envelope, we have reproduced a lateral scan as a sum of many pencil beams. In this case, the code calculates the energy deposition for a single beam. Afterwards the dose profile is translated in order to obtain a lateral scan and the total energy deposition will be given by a sum of all single beams. We have considered the field size factor for our model and for MC simulations: the ratio between the total dose deposition inside of concentric square fields of increasing size and the same quantity in a square reference Dref of size 10 cm.
The results of FSF show a good correspondence [1]. The maximum difference is about 5%, of the
Figure 2: Energy 150 MeV at depth z=11 cm.
same order of accuracy of the single beam tests. These results indicate a good description of the low dose contribution.
Another advantage of the MONET code is the fast calculation time. Indicatively, for each depth the calculation time is about 2 seconds for the single beam and 4 seconds for the full lateral scan.
38
These times are competitive compared to the simulation time (about 1-10 hours). Conclusions
The MONET code accounts for all the physical interactions of protons with water and is based on well known and validated theories. The advantages of our model are the physical foundation, the small number of parameters, the fast calculation time and the accuracy.
Recently, MONET has been extended to the case of Helium beams and preliminary results have been recently published [4]. A possible development is the creation of a dose database of clinical interest and an online fast dose evaluation tool.
References
1. A. Embriaco, V. E. Bellinzona, A. Fontana and A. Rotondi (2017), An accurate model for the
computation of the dose of protons in water. Phys. Med. 38, 66–75.
2. E.V. Bellinzona, M. Ciocca, A. Embriaco, A. Ferrari, A. Fontana, A. Mairani, K. Parodi, A. Rotondi, P. Sala and T. Tessonnier (2016), A model for the accurate computation of the lateral scattering of protons in water. Phys. Med. Biol. 61(4), N102.
3. A. Papoulis (1968), Joint densities with circular symmetry. IEEE Trans. on Inf. Theo. 14, 164-165
4. A. Embriaco, V. E. Bellinzona, A. Fontana and A. Rotondi (2017), On the lateral dose profile of
4He beams in water. Phys. Med., http://dx.doi.org/10.1016/j.ejmp.2017.07.007
CHARACTERIZATION OF A CdZnTe DETECTOR PROTOTYPE FOR BORON IMAGING BY SPECT
S. Fatemi1,2, C.H. Gong5, S. Bortolussi1,2, I. Postuma2, N. Protti2, G. Benassi3, N. Zambelli3, M.
Bettelli4, A. Zappettini4, X.B. Tang5 and S. Altieri1,2
1University of Pavia, Department of Physics, Via A. Bassi 6, IT-27100 Pavia, Italy
2National Institute of Nuclear Physics INFN, Unit of Pavia, Via A. Bassi 6, IT-27100 Pavia, Italy
3due2lab, Viale Mariotti 1, 43121 Parma, Italy
4IMEM-CNR, Parco Area delle Scienze 37/A, IT-43100 Parma, Italy
5Nanjing University of Aereonautics and Astronautics, Nanjing, China
email: setareh.fatemi@pv.infn.it
Introduction
Boron Neutron Capture Therapy (BNCT) is a binary radiation therapy which is able to selectively destroy malignant cells while sparing the normal tissue. BNCT consists of the administration of a
10B-enriched drug to the patient and the subsequent irradiation of the tumour specifically targeted by the boronated drug with thermal neutrons.10B has a high thermal neutron capture reaction cross section (3840 barns) and produces an alpha particle and a 7Li nucleus which in the 94% of the cases is in an excited state and emits a 478 keV gamma ray. Both particles have a high Linear Energy Transfer (LET) thus having a small range comparable to the cell diameter and as such ensuring the selectiveness of the therapy [1]. Therefore BNCT treatment effectiveness strongly depends on the radiation dose deposited locally by the 10B(n,α)7Li reaction in the tumour; however, the local and real time measurement of this quantity during the neutron irradiation is a big challenge, not yet solved by the BNCT researchers’ community. The deposited dose evaluation needs a correct knowledge of the
39
Both quantities are presently not measured directly and in real time, exacerbating the uncertainties in BNCT dosimetry evaluation. In order to solve this problem it is possible to exploit the 478 keV dis-excitation photons of the 7Li and as such avoid the difficulty of measuring the Boron concentration and the neutron flux as separated quantities [2]. Therefore our project aims to develop a Single Emission Computed Tomography (SPECT) imaging system based on a CdZnTe (CZT) semiconductor detector.
Materials and Methods
Room-temperature semiconductor detectors, such as CZT, have favourable physical characteristics for medical applications; moreover the material is present on the market at affordable prices and generally shows very good energy resolution [3].
We are studying a 5x5x20 mm3 CZT prototype. First we tested its response to standard calibration gamma sources to understand the energy resolution and efficiency and then we measured the detector’s performances in presence of a neutron field; such measurements were conducted in the thermal column of the Triga Mark II reactor of Pavia University using two different configurations. The detector was positioned in the thermal column behind a Boral shield that has a high quantity of
10B. The Boral shield can be opened or closed. In the first case a thermal neutron flux of 105-104 n/cm2/s reaches directly the CZT prototype giving way to a capture reaction on the cadmium present in the detector itself and emitting a 558 keV gamma. When the Boral shield is closed the thermal neutron flux is attenuated enough not to have the cadmium capture reaction.
After the prototype characterization we simulated the response of a CZT element of 20x20x20 mm3
that can be considered as one of the base elements to be employed for a BNCT-SPECT system; we simulated the performances of such detector using Geant4 Monte Carlo code.
The detector has 80 pixels and was placed at 21 cm from a point-like source without any collimation system, moreover the detector was rotated to obtain 36 projections over 360 degrees.
The point like source was first positioned in air and then inside a tissue-equivalent small animal phantom to simulate the attenuation and diffusion of the 478 keV gammas from the source. The simulation was made in an ideal condition. Even when the point-like source is embedded in a tissue equivalent phantom, no background contribution was simulated, despite the main noise in a BNCT scenario is due to the boron in the healthy tissue. Furthermore we used the Filtered Back Projection reconstruction method to evaluate its imaging capabilities.
Results
The energy resolution of the 5x5x20 mm3 CZT prototype is 2.8% at 356 keV and 2.9% at 511 keV; we can expect a similar energy resolution for the 478 keV gamma emitted during BNCT treatment. Moreover the efficiency of the prototype detector is 10.7% at 356 keV and 4.3% at 511 keV, thus we calculated the efficiency for the 478 keV gamma to be 6.1%; this value can be improved by a better read out electronic and with a bigger sized detector.
The measurements done in the thermal column of the Triga Mark II reactor of Pavia University showed us that in both Boral configurations the 478 keV peak is clearly visible. When the shield is opened the 558 keV peak due to the cadmium is clearly distinguishable from the 478 keV peak due to the good energy resolution of the CZT detector. To suppress the 558 keV gamma we shielded the detector with lithium carbonate enriched in 6Li which is a good thermal neutron absorber thus avoiding the neutrons to reach the detector.
The simulations carried out on the SPECT system base element of 20x20x20 mm3 show us the capabilities of the detector for an imaging system. When simulating only one 478 keV point-like source without any background gammas we are able to determine the spatial resolution of the detector to be 8 mm. The spatial resolution obtained is for a collimator-less system but we have to take into account that the simulations were made without any background gammas that could interfere with the correct image reconstruction; to avoid this problem a collimator system should be studied. Moreover when simulating two point-like sources of 478 keV gamma orthogonally positioned with
40
respect to the CZT detector and along one phantom diameter, we are able to see that when their relative distance is below 8 mm the two sources are superimposed in the reconstructed image and as such are not distinguishable, while when their relative distance is longer than 8 mm they are correctly distinguishable. This confirms that the spatial resolution of our CZT detector in this ideal simulation is 8 mm and it could be further improved introducing a collimator.
Conclusions
The energy resolution and efficiency of a 5x5x20mm3 prototype CZT detector were tested and found to be appropriate for a BNCT-SPECT system even more since there is the possibility to further improve them.
The spatial resolution of a 20x20x20mm3 CZT detector was simulated without a gamma background and without collimation and found to be 8 mm; therefore in our future work we aim to simulate a system with an appropriate collimator to improve the spatial resolution and see how the system responds to the presence of 478 keV background photons due to the presence of small quantities of
10B in the healthy tissue surrounding the tumour. Moreover we plan to study various reconstruction algorithms to improve the imaging capabilities of the BNCT-SPECT system.
References
[1] Sauerwein, W. A., Wittig, A., Moss, R., & Nakagawa, Y. (Eds.). (2012). Neutron capture
therapy: principles and applications. Springer Science & Business Media.
[2] Kobayashi, T., Sakurai, Y., & Ishikawa, M. (2000). A noninvasive dose estimation system for
clinical BNCT based on PG‐ SPECT—Conceptual study and fundamental experiments using HPGe and CdTe semiconductor detectors. Medical physics, 27(9), 2124-2132.
[3] Del Sordo, S., Abbene, L., Caroli, E., Mancini, A. M., Zappettini, A., & Ubertini, P. (2009).
Progress in the development of CdTe and CdZnTe semiconductor radiation detectors for astrophysical and medical applications. Sensors, 9(5), 3491-3526.
41
BEAM RANGE MONITORING IN PARTICLE THERAPY WITH CHARGED