Download The validation of the treatment planning system BDTPS - INFN-LNL

The validation of the treatment planning system BDTPS through an “invivo” comparison
G. G. Daquino1, S. Caria2, N. Cerullo2,3, T. Aihara4, J. Hiratsuka4, H. Kumada5, , G. Lomonaco2, L. Muzi 6, A.Sainato7,
1. Fondazione CUOA, University of Padua, Italy
2. DIMNP, University of Pisa, Italy
3. DIPTEM, University of Genova, Italy
4. Kawasaki Medical School, Japan
5 Tokai Research and Development Center, JAERI, Japan
6. Dept. of Science and Engineering, School of Medicine, Oregon Health & Science University, USA
7. Dept. of Oncology and Radiotherapy, AOUP, Pisa, Italy
Introduction.
Boron Neutron Capture Therapy (BNCT) is a therapeutic technique, still under development, which exploits the interaction
between thermal neutrons and B10 isotope administered to the patient using drugs. The reaction B10(n,α) Li7 produces two
high LET and short range fragments causing tumor cell damage.
BNCT may be useful for treatment of locally advanced tumors that, in addition to the primary tumor mass, show
microscopic infiltration into healthy tissue depths
Correct boron level measurements both in tumor and in normal tissue are fundamental for therapeutic success. The method
used in the past consists of an analysis of tissue samples taken during surgery [1][2][3]. However, the need to move
towards non invasive techniques led to development of innovative methods. Nowadays, Positron Emission Tomography
(PET) is able to provide boron concentration estimations through in vivo information [4][5].
The idea to couple the Treatment Planning System (TPS) to the information on the real boron distribution in the patient is
the main added value of the new methodology set-up at DIMNP of University of Pisa, in collaboration with the JRC of
Petten (NL). The methodology has been implemented in the new TPS, called BDTPS (Boron Distribution Treatment
Planning System) [6], which takes into account the actual boron distribution in the human volume of interest, while the
standard TPS assumes a uniform boron distribution, far from the reality.
After an “in-phantom” validation at JRC (Petten) [7], an “in-vivo” validation of BDTPS was possible thanks to a Japanese
researchers’ team collaboration.
A Japanese 74 years old woman affected by parotid cancer was the medical case proposed for BDTPS implementation.
Both CT and PET stacks, together with other information about the case, have been provided by the Tokai Research and
Development Center of JAERI.
Actually, BDTPS was implemented both in standard and in PET-based approach, in order to make possible an exhaustive
comparative analysis between the two approaches. The first case, named Back, was necessary to derive the same results of
Japanese standard system (JCDS) [8] used for the application.
Materials and Methods.
A 74 years old Japanese woman case affected by a stage IV parotid cancer is the study reference. The objective is to
compare the standard approach (adopted with JAEA Computational Dosimetry System - JCDS) and the PET-based
procedure. Both runs have been conducted through the Boron Distribution Treatment Planning System (BDTPS), designed
as a complete TPS containing all the main characteristics of JCDS (e.g. pre-processing based on CT images of the patient,
fiducial markers, organs at risk, Monte Carlo modeling, post processing, etc.). In BDTPS, the boron dose is evaluated on the
basis of the real macro-distribution of the boron compound by modules fully integrated into its internal structure.
A. Back.
Hereafter, the approach used with JCDS will be called “Back”. This method consists of five phases:
• 3D reconstruction on CT stack
• T/N evaluation on PET stack
• Assigning the 3D model with two 10B concentrations according to T/N ratio
• Doses and fluences evaluations
• Irradiation time and plan definition
The CT scan of the patient, 60 Kg, was used for the head 3D model reconstruction.
Later on, the patient has been PET-scanned after an infusion of 220 MBq of F-BPA for a period of 40 minutes in order to
assess the T/N ratio. Thus, two 10B concentrations are assigned, one to the tumor and the other to the healthy tissue
respecting this ratio. In particular:
• T/N =5
• [10B]tumor= 146.5 ppm
• [10B]tissue= 29.3 ppm
Afterwards, 500 mg of BPA for 60 kg in 3 hours have been infused in the patient and the radiation therapy performed. As
neutron source, the Japan Research Reactor N. 4 (JRR-4) was used; it is a 3.5 MWth swimming pool style reactor,
moderated and cooled with light water, that uses enriched uranium ETR-type as fuel element.
B. PET-based approach (BDTPS)
BDTPS needs the construction of three models of the human zone of interest:
• 3D Model, which takes into account the real geometry through CT images.
• Monte Carlo Model (MC), to simulate irradiation calculation.
• Boron Model, to take into account the heterogeneous boron distribution in the various tissues.
During the MC model construction, the Boron model is formed, hereafter referred as B Model. In fact, each universe in the
MC model defines a region, which in turn defines a material with its own average boron concentration and density.
Moreover, being MC model built directly from 3D model, this raises the need to have the CT stack perfectly aligned with
the PET stack. This is not a requirement in JCDS standard methodology, where the PET scans are used only to assess the
T/N ratio. An ad hoc CT-PET alignment work was carried out through the use of the software Amide 0.9.1.
Results
The 3D model is composed of 49 CT images, and the tumor extends from CT slice 9 to CT slice 20. We refer to this range
because of the presence of high 10B concentrations that allows us to better appreciate the differences between the two
methods. In particular we consider the slice 16 as best representative of the dynamics encountered in the central tumor
range. The two approaches comparison shows how the boron dose in the center of the tumor assumes smaller values in the
PET-based approach. This is even more evident through the “pixel by pixel” difference between the boron dose rate
mappings in the PET-based and the Back methods. In particular, this behavior is highlighted from the slice 14 till the slice
17.
The results of the comparison show that there is an overestimation of the boron dose rate in the tumor center of about 40%
in the Back case when compared with PET-based case. The situation is made even clearer if we look at the slice 16, and in
particular only to the tumor mass. In fact, considering the area of the rectangle bounding the tumor and evaluating the dose
values within the box, a regular surface distribution comes out from the Back approach due to the homogeneous
concentration of 10B . The PET-based approach, instead, shows a central depression as a consequence of the heterogeneous
distribution resulting from the activity values recorded in the PET images.
Making the difference (PET)−(Back), the formation of a concave area is representative of the boron dose rate (Gy/h)
overestimation in the Back approach.
Because the boron dose rate is one of the parameters from which mainly the treatment planning depends on, according to
BDTPS results one can argue an underestimation of the irradiation time in JCDS protocol. This hypothesis is supported by
the patient follow-up. In particular, this document shows a recurrence raised about eighteen months later the radiation
therapy, placed just where the overestimation is shown by BDTPS.
Conclusions.
Previous studies[9][10][11] evidenced the possibility to improve the TPS efficiency, introducing the boron localization in
Monte Carlo code. In particular, the use of PET images, obtained through a positron-emitter agent (F18) linked to the boron
drug BPA [12], enhances the precision and realism of the results of the treatment planning system applied to Neutron
Capture Therapy (NCT) [13] . In fact, the knowledge of the boron distribution in the tissues is a key factor for the precise
dose evaluation.
The comparison between the first results obtained with PET-based analysis and those from JCDS Japanese method
evidenced the importance of considering PET images and boron distribution at the level of the treatment planning in neutron
capture therapies[14].
Moreover, an underestimation of boron dose around the tumor ROI is detected using the PET-based approach. Possible
contributions could derive also from the larger, and therefore incorrect, definition of region contour by the radiologist. This
implies that the PET data integration in suitable treatment planning allows not only a valid description of the macroscopic
boron distribution in various tissues, but gives support in the definition of region contours.
The follow-up of the patient confirms how BDTPS led to results very close to the reality.
References.
1.
Haritz D, Gabel D, Huiskamp R. Clinical phase-I study of Na2B12H11SH (BSH) in patients with malignant glioma
as precondition for boron neutron capture therapy (BNCT). Int. J. Radiat. Oncol. Biol. Phys. 1994; 28:1175-81.
2. Haselberger K, Radner H, Pendl G. Boron neutron capture therapy: boron biodistribution and pharmacokinetics
of Na2B12H11SH in patients with glioblastoma. Cancer Research 1994; 54:6318-20.
3. Horn V, Pharm D, Slansky J, Janku I, Strouf O, Sourek K, Tovarys F. Disposition and tissue distribution of boron
after infusion of borocaptate sodium in patients with malignant brain tumors. Int. J. Radiat. Oncol. Biol. Phys.
1998.
4. Kabalka GW, Smith GT, Dyke JP, Reid WS, Desmond Longford CP, Roberts TG, Reddy NK, Buonocore E,
Hubner KF. Evaluation of fluorine-18-BPA-fructose for boron neutron capture treatment planning. J. Nucl. Med.
1997
5. Mazzini M.,Cionini L., Menichetti L., Sauerwein Wolfang a.g., Salvadori P., Positron Emission Tomography: a
crucial role for BNCT ?,Atto di convegno internazionale, Firenze,2008.
6. G. Daquino, “A PET-based Treatment Planning System to evaluate the Influence of the Heterogeneous Boron
Distribution in Boron Neutron Capture Therapy (BNCT), Phd thesis, 2002
7. G. Daquino, N. Cerullo, M. Mazzini, R. Moss, L. Muzi, “Experimental and computational validation of BDTPS
using a heterogeneous boron phantom”, Applied Radiation and Isotope , 2004
8. H. Kumada, K. Y., “Development of JCDS, a computational dosimetry system at JAEA for boron neutron capture
therapy”, Journal of Physics: conference series 74, 2007.
9. N. Cerullo, G. Daquino, "Use of Monte Carlo code in neutrons behavior simulation in BNCT (Boron Neutron
Capture Therapy)" Technologies for the New Century, ANS, Nashville (TN), 19-23 April 1998
10. N. Cerullo G. G. Daquino, R. Moss, L. Muzi – "Evaluation of the Heterogeneous-Boron-Distribution Influence on
the computed dose in BNCT Treatment Planning Systems" – in “Research and Development in Neutron Capture
Therapy”, W Sauerwein, R. Moss, A. Wittig eds., pag. 541-545, Monduzzi Editore, Bologna (Italy) – 2002
11. G. W. Kabalka, T. L. Nichols, G. T. Smith, L. F. Miller, M. K. Khan, P. M. Busse, “The use of positron emission
tomography to develop boron neutron capture therapy treatment plans for metastatic malignant melanoma”,
Journal of Neuro-Oncology, 2003; 62:187-195
12. K. Nagata, Y. Sakurai, M. Suzuki, Y. Kinashi, S. Masunaga, N. Morita, T. Aihara, J. Hiratsuka, Y. Imahori, A.
Maruhashi, K. Ono, "Estimation of intratumour boron concentration of boron neutron capture therapy for head
and neck tumours using 18FBPA-PET image" Advances in Neutron Capture Therapy, 2006G. G. Daquino,
"PET-based approach to treatment planning systems: an improvement toward successful Boron
Neutron Capture Therapy (BNCT)", European Publication Office, Luxembourg, 2003, ISBN 92-894-5507-1G.
G. Daquino. N. Cerullo,T. Aihara, J. Hiratsuka, H. Kumada, G. Lomonaco, S.Caria, L. Muzi, R. Moss, A.
Sainato, O. Sorace, D. Bufalino, “An In-vivo Comparative Study Between Standard and PET-Based Approach to
Assess the Therapeutic Dose in Neutron Capture Therapies” NSS IEEE 2008, pp 1-8, Dresden (Germany), 2008 .