Prompt gamma and neutron detection in BNCT utilizing a CdTe detector

Applied Radiation and Isotopes ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso

Prompt gamma and neutron detection in BNCT utilizing a CdTe detector Alexander Winkler a,b,n, Hanna Koivunoro b, Vappu Reijonen b, Iiro Auterinen c, Sauli Savolainen a,b a

Department of Physics, University of Helsinki, P.O. Box 64, FI-00014 Helsinki, Finland HUS Medical Imaging Center, Radiology, University of Helsinki and Helsinki University Hospital Finland, POB 340, FI-00029 HUS, Finland c VTT Technical Research Centre of Finland, Espoo, POB 1000, FI-02044 VTT, Finland b

H I G H L I G H T S







We tested a modern CdTe spectrometer for simultaneous gamma and neutron detection. Prompt gamma (PG) signals from both sources were able to be distinguished. PG from 10B(n,a) and 113Cd(n,g) were also seen in a phantom used for BNCT. The results open the possibility to determine 10B dose in real time for BNCT.

art ic l e i nf o

a b s t r a c t

Article history: Received 30 January 2015 Received in revised form 21 May 2015 Accepted 25 July 2015

In this work, a novel sensor technology based on CdTe detectors was tested for prompt gamma and neutron detection using boronated targets in (epi)thermal neutron beam at FiR1 research reactor in Espoo, Finland. Dedicated neutron filter structures were omitted to enable simultaneous measurement of both gamma and neutron radiation at low reactor power (2.5 kW). Spectra were collected and analyzed in four different setups in order to study the feasibility of the detector to measure 478 keV prompt gamma photons released from the neutron capture reaction of boron-10. The detector proved to have the required sensitivity to detect and separate the signals from both boron neutron and cadmium neutron capture reactions, which makes it a promising candidate for monitoring the spatial and temporal development of in vivo boron distribution in boron neutron capture therapy. & 2015 Elsevier Ltd. All rights reserved.

Keywords: BNCT CdTe Prompt gamma Neutron detection Neutron dosimetry Spectrometry Radiotherapy

1. Introduction Boron neutron capture therapy (BNCT) (Menéndez et al., 2009; Kankaanranta et al., 2007, 2011; Barth et al., 2012) is based on high-LET radiation released from the neutron capture reaction 10 B(n,α)7Lin, in which a prompt gamma (PG) photon of 478 keV is emitted with a probability of 94%. These photons leave the patients body with small attenuation (Minsky et al., 2009). Previously, the estimation of the 10B concentration in the tumor and healthy tissues during the treatment has been based on average tumor concentrations for several patients analyzed from tissues at n Corresponding author at: Department of Physics, University of Helsinki, P.O. Box 64, FI-00014 Helsinki, Finland. E-mail address: alexander.winkler@helsinki.fi (A. Winkler).

a certain time after the boron administration. A procedure that includes an uncertainty for the individual patient and time of measurement. A more accurate determination of the boron concentration could be achieved by detecting the 478 keV PGs during patient treatment (Minsky et al., 2009; Munck af Rosenschöld et al., 2001; Verbakel et al., 2003; Verbakel and Stecher-Rasmussen, 1997; Kobayashi et al., 2000). However, the technical realization of a capable detector system requires further development to identify the 478 keV PG from the photon spectrum, emitted from patient during neutron irradiation (Kobayashi et al., 2000). In 2003 a study on 478 keV PG detection using a CdTe/CdZnTe detector was carried out in Finland, with the result that about 3% energy resolution and appropriate neutron shielding and collimation techniques are needed (Morozov et al., 2006). Since then further work has been done by Murata et al. from the University of

http://dx.doi.org/10.1016/j.apradiso.2015.07.040 0969-8043/& 2015 Elsevier Ltd. All rights reserved.

Please cite this article as: Winkler, A., et al., Prompt gamma and neutron detection in BNCT utilizing a CdTe detector. Appl. Radiat. Isotopes (2015), http://dx.doi.org/10.1016/j.apradiso.2015.07.040i

A. Winkler et al. / Applied Radiation and Isotopes ∎ (∎∎∎∎) ∎∎∎–∎∎∎

2

TOPVIEW

FRONTVIEW

beam aperture

B4C

beam aperture

SPM 2

SPM 1

B4C

SPM 2

SPM 1

PC

PC

Fig. 1. Schematic drawing of the B4C setup.

Osaka, pointing out an improvement in energy resolution (Murata et al., 2014). In the present work, we evaluate the performance of the latest CdTe detector technology capable of high resolution spectrometry. Additionally, we evaluate the ability of the same device to detect and distinguish neutron and γ signals, by comparing the PG photons from both cadmium and boron neutron capture reactions. Previously, the strong background caused by the cadmium neutron capture reaction, 113Cd(n,γ)114Cd, has been considered a drawback, since the 559 keV PG peak from that reaction occurs close to the 478 keV peak of the 10B neutron capture reaction. We ultimately aim to calculate a tomographic image of the 10B concentration of the tumor and surrounding healthy tissue, similar to the approach published by Minsky et al. (2011). However, doing so by involving signals from both boron neutron and cadmium neutron reactions.

2. Experiment 2.1. Detectors Two CdTe based spectrometers were studied, the first (SPM1) is a recently manufactured and customized Amptek X-123 CdTe detector with a 3  3  1 mm3 crystal and state of the art readout electronics. The second spectrometer (SPM2) is a ten year old custom built device utilizing a single 6  6  1 mm3 CdTe crystal together with the readout electronics of the same age. Both devices were protected with a 6 mm lead housing that did not cover the detector crystals. 2.2. Experimental setup The epithermal neutron beam of the FiR 1 reactor was used at a power of 250 kW for clinical BNCT (Joensuu et al., 2003). Our experiments, however, were performed at lower power levels (0.25, 1.25 and 2.5 kW) to avoid radiation damage to the detector electronics. In turn excessive device shielding in front of the detector crystals was omitted and a 6 mm layer of lead was used only to protect the electronics from the expected secondary radiation. The detectors were placed perpendicular to the 14 cm diameter circular neutron beam port at 11.5 cm (SPM1) and 18 cm (SPM2) distance from the beam center. The applied beam consists mainly of epithermal neutrons ( 0.5 eV < E < 10 keV ) with about 7% thermal ( <0.5 eV ) and fast ( >10 keV ) neutrons. Thermalized neutrons from the phantom are mainly detected in this study (Auterinen et al., 2004).

The expected data quality strongly depends on the ability of the detector system to distinguish signals from the 10B(n,α)7Lin and the 113Cd(n,γ)114Cd reactions, as well as both signals from the background. The background is consisting of secondary photons emitted from neutron capture reactions in beam shaping assembly, surrounding structures and the phantom. The associated PG peak of the 10B(n,α) reaction is 477.59 (53) keV (http://www-pub. iaea.org/MTCD/publications/PDF/Pub1263_web.pdf) and of the 113 Cd(n,γ) reaction 558.46 keV (http://www.nndc.bnl.gov/capgam/ byn/page128.html). The irradiation was performed in four setups: (1) Free beam. (2) A bottle (90 mm diameter) filled to 11.5 cm height with B4C powder (774 g Borcarbid MG 55,29, grain size 40–60 μm, Schuchardt, München) placed in front of the neutron beam. (3) A cylindrical PMMA phantom with a radius of 10 cm and a length of 24 cm placed in front of the neutron beam. (4) Two boronated plastic disks (diameter 30 mm and 19 mm length, 3-wt% 10B) placed inside the PMMA phantom at the depth of 0–38 mm. In this work, each setup will be referred to either by number or by the names Free Beam, B4C, PMMA and Neutrostop, respectively. The experimental setups are shown in Fig. 1 for the setup 2 and Figs. 2 and 3 for the setups 3 and 4. The setups 1 and 2 are essentially identical, with the only difference of the added B4C bottle in the beam line of setup 2. In setups 3 and 4, no in-line arrangement of the SPM1 with the center of the disks (dashed line, Fig. 2) was possible, due to the size of the lead shielding. An irradiation time of 900 s was used for each measurement. The Free Beam measurement was performed to detect the baseline spectrum, which includes the scattering from the surrounding structures that are present in all setups (excluding the PMMA phantom). The aim of the second setup was to record a clearly identifiable signal from the boron neutron capture reaction by placing a large amount of B4C directly in the neutron beam. The setups 3 and 4 simulated the treatment situation more closely. The setup 3 is a reference irradiation condition with no 10B present within the phantom, while in the setup 4 boronated PMMA disks were placed in the phantom to simulate a patient's tumor. The detectors were calibrated using the common isotopes (22Na, 133Ba, 137Cs and 152Eu). Background spectra were recorded before and after the measurements. The background spectra recorded after the experiments were also used to examine possible activation of the detector structures as well as a change in the background radiation in the treatment room. The data was recorded with the energy range of 1–800 keV, which also is the area of sensitivity of the detectors. All spectra were reduced by the background spectrum before

Please cite this article as: Winkler, A., et al., Prompt gamma and neutron detection in BNCT utilizing a CdTe detector. Appl. Radiat. Isotopes (2015), http://dx.doi.org/10.1016/j.apradiso.2015.07.040i

A. Winkler et al. / Applied Radiation and Isotopes ∎ (∎∎∎∎) ∎∎∎–∎∎∎

TOP VIEW

3

0.25 kW 1.25 kW 2.5 kW

14000

55

9k

eV

12000

10000

counts

beam aperture

PMMA

Co

8000

m

n pto

ed

ge

6000

4

4000

eV 5k 53 eV 1k 51 V ke 78 57

6k

65

2000

0

SPM 2

eV

100

200

300

400

500

600

1k

eV

700

800

keV Fig. 4. Recorded spectra of different reactor powers in the PMMA setup of the SPM1 detector. The most important peaks are marked with their energies.

PC

3000 0.25 kW 1.25 kW 2.5 kW

Fig. 2. Schematic drawing of the PMMA and Neutrostop setup. The two PMMA disks are replaced by boronated PMMA disks for the setup 4.

2500

counts

2000

47

8k

eV 56

0k

eV

1500

1000

500

Fig. 3. Photograph of the same setup as presented in Fig. 2.

0 0

comparison and a linear baseline correction as well as the spectrum smoothing was performed. As a consequence background and Compton spectra are cut from these processed spectra. In order to identify the γ peaks of interest, Gaussian distributions were fitted to the processed spectra (curve fitting tool, http:// www.mathworks.com). The recorded peaks however are not Gaussian shaped, but pose shoulders on the low energetic side. This is a consequence of the hole-tailing effect in CdTe together with the used setting for pile-up rejection and pulse-rise-time discrimination used for our experiments. The resulting fits therefore require a combination of multiple Gaussian peaks.

3. Results 3.1. Calibration and linearity The neutron flux of the reactor increases linearly with its power the recorded spectra scale up similarly, as shown in Fig. 4. The data, however, indicates that only the reactor power of 2.5 kW provides a reasonable amount of signal. Therefore, all further discussion in this work will refer to the reactor power of 2.5 kW. The observed linear response of the detectors suggests that the system is in principle scalable and thus with proper collimation and shielding, measuring PG spectra at higher, possibly even full BNCT reactor powers could be possible.

100

200

300

400

500

600

700

800

KeV Fig. 5. Recorded spectra of increasing reactor power taken with the SPM2. The spectrum is dominated by noise and a down shift of the peaks occurs with increasing fluxes.

3.2. Comparison old vs. new spectrometer Fig. 5 presents the spectrum at the Free Beam setup as recorded with the SPM2. The resolution and sensitivity is clearly worse compared to the newer device (SPM1, Figs. 4, 6, 7). In addition, pile-up and undershoot of the readout electronics degrade the spectrum with increasing reactor power. As a consequence decreasing energy resolution, down shift and higher noise can be observed. Only the lowest reactor power provides a useful spectrum in which the peaks from the boron neutron and cadmium neutron reactions are recorded with about 120 and 90 counts. Their FWHM energy resolution is 15 keV and 14 keV, respectively. 3.3. Free beam and B4C setup The peaks marked in Fig. 4 are Compton edge, 477.60 keV 10B(n,

α) PG peak, 511.00 keV annihilation peak, 535.26 keV CdK,α single escape peak (5% yield), 558.46 keV 113Cd(n,γ) PG peak (main emission, 100%), 576.08 keV 113Cd(n,γ) PG peak (minor emission, 6%), 651.26 keV 113Cd(n,γ) PG peak (minor emission, 19%). Fig. 6 presents the spectra of the Free Beam and the B4C setups. In order

Please cite this article as: Winkler, A., et al., Prompt gamma and neutron detection in BNCT utilizing a CdTe detector. Appl. Radiat. Isotopes (2015), http://dx.doi.org/10.1016/j.apradiso.2015.07.040i

A. Winkler et al. / Applied Radiation and Isotopes ∎ (∎∎∎∎) ∎∎∎–∎∎∎

4

11000 10000

3

47

8

10

Free Beam B4C Area of Difference

V ke

PMMA NeutroStop Area of Difference

9000

55

9k

eV

2.5

8000

55

9k

eV 2

6000

counts

counts

7000

5000

1.5

4000

1

3000 2000

1 51

1000

ke

V

53

5k

eV 57

6k

eV

0.5

47

0

460

480

500

520

540

560

8k

eV 5

0

580

keV

460

480

k 11

eV

500

53 520

to improve the analyzing process both the spectra have been smoothed and their difference is marked with a shaded area. The first point of interest is the peaks from the 113Cd(n,γ) reaction (dotted line), which are found at 559.13 keV and 559.64 keV with fitting errors of 2.7% and 2.8% for setups 1 and 2, respectively. The 113 Cd(n,γ) peak is prominently present in both spectra and at the expected energy. The B4C spectrum of setup 2 contains in addition the 10B(n,α) peak (solid line), which is found at 476.97 keV (error 2.8%). Table 1 lists the peak details for each setup. Fig. 6 shows that the peaks from both boron and cadminum neutron capture reactions can clearly be differentiated from each other, which indicates that the SPM1 can be used to detect simultaneously the needed data from γ and neutron radiation. The 113Cd(n,γ) peak of the B4C spectrum is slightly smaller than the same peak in the Free Beam setup. The difference is small and the influence of noise cannot be ruled out. However, decreased count numbers that only occur in the intermediate area of the boron and cadmium neutron PG peaks is speaking against noise related reasons. 3.4. PMMA vs. Neutrostop setup For the setups 3 and 4, no in-line arrangement of the detector and the boronated disks (inside the phantom) was possible due to the size of the lead shielding around the SPM1, see Fig. 2. This was especially problematic for the setup 4, where our trials showed that the peak from the boron neutron reaction decreases along the axis of the phantom and with increasing distance from the beam aperture. Nevertheless, a small difference in the measured spectra of the setups 3 and 4 can be observed at the location of the 10B(n, α) peak. The fitted spectra with highlighted differences are shown in Fig. 7. The PG peak from 113Cd(n,γ) is five times larger than the same peak in the setups 1 and 2. It can clearly be identified and

cts

1 2 3 4

436 10 547 876 1780

10

B (n, α ) peak

ΔE

10

B (n, α ) peak [keV]

4.7 70.12 10.7 70.30 3.1 70.04 0.9 70.02

cts

113

4956 4437 22 148 19 944

Cd (n, γ ) peak

57 540

560

6k

eV

580

Fig. 7. Smoothed spectra of the PMMA and Neutrostop setups. Both 10 B(n,α) peaks can be identified.

113

Cd(n,γ) and

the centroids are found at 559.33 keV and 559.43 keV for the PMMA and Neutrostop setup, respectively. A peak at the location of the 10B(n,α) reaction was found at 477.95 keV (setup 3) and 478.07 keV (setup 4) with fitting errors of 1.3% (PMMA) and 2.1% (Neutrostop), respectively. Peak details are listed in Table 1 for comparison.

4. Discussion The PG peaks originating from both 10B(n,α) and 113Cd(n,γ) reaction were detectable and distinguishable with both spectrometers. As expected, the SPM2 was not able to record spectra with the adequate resolution and at higher reactor power levels. The SPM1 on the other hand performed exceptionally in comparison. Especially the signal originating from the 113Cd(n,γ) reaction was clearly present and pronounced in every spectrum acquired with the SPM1. As expected the detected neutron signal is larger for the setups (3 and 4) involving the phantom due to the higher number of thermal neutrons that are scattered towards the detector. For the B4C setup, a large amount of 10B was placed in the neutron beam to produce a strong PG peak from the 10B(n,α) reaction, which can clearly be identified in Fig. 6. The PG peak from the 113Cd (n,γ) reaction can be distinguished from the PG peak of 10B(n,α) reaction without any overlap of the peaks. This verifies our initial assumption that CdTe based detectors of the present technological level can be used simultaneously as γ and neutron detectors. In Section 3.3 we noted that the PG peak from 113Cd(n,γ) in the B4C setup is smaller than that in the Free Beam arrangement. This is due to the large amount of neutrons converted in the B4C powder, leaving a smaller amount to be scattered and to propagate towards the detector. The additional count difference between the 10 B(n,α) and the 113Cd(n,γ) peak that is located between 490 and

Table 1 Comparison of peak count numbers (cts), their energy resolutions ( ΔE ) and the peak count differences (Δ cts) for the Setup

eV

keV

10

Fig. 6. Smoothed spectra of the Free Beam and B4C setup. The peak from the B neutron reaction can clearly be distinguished from the peaks of the 113Cd reaction in each spectra. The shaded area visualizes the difference between both setups.

5k

ΔE

113

Cd (n, γ ) peak [keV]

8.2 7 0.22 9.2 7 0.26 13.17 0.17 10.5 7 0.20

10

B(n,α) and

Δcts

10

113

Cd(n,γ) peak of all setups.

B (n , α ) peak

Δ cts

113

þ 10 111

 519

þ 904

 2204

Cd (n, γ ) peak

Please cite this article as: Winkler, A., et al., Prompt gamma and neutron detection in BNCT utilizing a CdTe detector. Appl. Radiat. Isotopes (2015), http://dx.doi.org/10.1016/j.apradiso.2015.07.040i

A. Winkler et al. / Applied Radiation and Isotopes ∎ (∎∎∎∎) ∎∎∎–∎∎∎

550 keV can be accounted towards the same reason, see Fig. 6. Furthermore, the background spectra taken before and after the measurements showed negligible differences only, which implies that a change of noise sources can largely be ruled out. Nevertheless, the count difference of the 113Cd(n,γ) peak is too small to be entirely certain. Figs. 6 and 7 possess a small peak at the location of the 10B(n,α) PG peak, even though no boron was added in these setups. We have identified two possible sources for this signal. First, boron is a standard component used for the production of printed circuit boards (PCB), which are present inside the detector device. This source however is rather unlikely to be the major cause, as the amount of boron used in PCBs is only a fraction compared to the amount used in the boronated disks of the setup 4. The second source is the six known γ emissions from the 113Cd(n,γ) reaction that occur between 470 and 490 keV (http://www.nndc.bnl.gov/ capgam/byn/page128.html). The strongest is at 477.60 keV, which is almost exactly at the same location as the 10B(n,α) peak (477.59 (5) keV). Even though the intensities of these additional 113Cd(n,γ) emissions are ≤1% compared to the main emission at 559 keV, they still add up to be significant ( E50%) compared to the counts of the 10B(n,α) peak. Nonetheless, these emissions are present in all spectra similarly and can therefore be accounted to the background. The PG peaks from 113Cd(n,γ) reaction in the setups 3 and 4 match each other well in shape, yet a small count difference exists between them. In the case of the Neutrostop setup, the PG peak from the 113Cd(n,γ) reaction is smaller by E2200 counts compared to the same peak of the PMMA setup. Otherwise, the two spectra are very similar, except for the 10B(n,α) PG peak area and minor cadmium neutron reactions peaks. The resulting PG peak from the boron neutron reaction is increased by E900 counts and can be identified when comparing the spectra from the PMMA and Neutrostop setups to each other (Fig. 7). The minor 113 Cd(n,γ) reaction peaks are all reduced in their count rates in setup 4. The number of counts of the 10B(n,α) PG peak is only about twice the background, which currently is too low for an effective usage in BNCT. An alignment of the detector with the center of the boronated disks would increase this PG peak height, as well as additional detectors would do. A supplementary collimator can furthermore reduce background and will result in a more clear spectrum. These options will be investigated in further studies. Nevertheless, the reduced counts of the 113Cd(n,γ) peak as well as the increased counts of the 10B(n,α) peak between setups 3 and 4 demonstrate that the employed detector is in principle capable of providing the necessary information to calculate the ratios between both signals and from that the boron dose in the tumor and the surrounding healthy tissues. All results were obtained at a reactor power of 1% from the level applied in BNCT treatments, which demonstrates the exceptional sensitivity of the detector to both γ and neutron radiation. This circumstance could be utilized to determine the optimal time to start the patient irradiation in BNCT. With a SPECT-like system based on the sensitive detector studied in this work, the patient could be imaged with a low neutron flux (e.g. 1–5% reactor power in the case of FiR 1) before the actual treatment irradiation in order to monitor the spatial and temporal development of 10B concentrations of the tumor vs. healthy tissues. Once the optimum concentration has been reached, the irradiation could be started with full reactor power. This approach could allow a timely and direct 10B concentration determination with only a small additional dose to the patient. The details of this approach, e.g. an estimation of the required and obtainable count numbers, as well as a discussion of the unforeseen size of the annihilation peak are content of a follow up study. The achieved results have become accessible recently only as

5

previous detector generations did not possess the energy resolution and electronic stability required to perform this task as indicated by the data obtained with the spectrometer SPM2.

5. Conclusion In this work we have studied the capability of a modern CdTe detector to measure and distinguish γ and neutron radiation. According to our measurements, the improved energy resolution of the CdTe detector makes the detector type now feasible for detecting and quantifying the 478 keV PG peak from the 10B(n,α) reaction within other peaks present in the spectrum. If a thermal neutron shielding is not present in the measurement setup, the CdTe detectors can also be used to measure the 559 keV PGs from the 113Cd(n,γ) reaction, allowing the detector to simultaneously measure and separate signals from both the 10B(n,α) and 113Cd(n,γ) reactions. To simulate a BNCT treatment setup, we measured PG signals from boronated disks within a PMMA phantom. An increased 478 keV PG signal could be distinguished, compared to the reference setup (same PMMA phantom without the boronated disks). All experiments were carried out at 1% reactor power compared to the patient treatments levels. Such an arrangement could be used to determine the optimal time to start the patient treatment as real time monitoring of the development of the boron distribution in the tumor and healthy tissues would be possible.

Acknowledgments We acknowledge VTT for supporting the use of the FiR1 reactor. We are grateful to Pertti Tikkanen from the accelerator lab of the University of Helsinki for input for the spectra interpretation.

Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.apradiso.2015.07. 040.

References Auterinen, I., et al., 2004. Measurement of free beam neutron spectra at eight BNCT facilities worldwide. Appl. Radiat. Isot. 61, 1021–1026 http://dx.doi.org/10.1016/ j.apradiso.2004.05.035. Barth, Rolf F., et al., 2012. Current status of boron neutron capture therapy of high grade gliomas and recurrent head and neck cancer. Radiat. Oncol. 7 (7), 146. http://dx.doi.org/10.1186/1748-717X-7-146. Joensuu, H., et al., 2003. 2003 Boron neutron capture therapy of brain tumors: clinical trials at the Finnish facility using boronophenylalanine. J. Neurooncol. 62 (1–2), 123–134. Kankaanranta, Leena, et al., 2007. Boron neutron capture therapy in the treatment of locally recurred head and neck cancer. Int. J. Radiat. Oncol. Biol. Phys. 69 (October 1 (2)), 475–482. Kankaanranta, Leena, et al., 2011. L-boronophenylalanine-mediated boron neutron capture therapy for glioblastoma or anaplastic astrocytoma progressing after external beam radiation therapy: a phase I study. Int. J. Radiat.: Oncol. Biol. Phys. 80 (2), 369–376 http://dx.doi.org/10.1016/j.ijrobp.2010.02.031. Kobayashi, Tooru, et al., 2000. A noninvasive dose estimation system for clinical BNCT based on PG-SPECT conceptual study and fundamental experiments using HPGe and CdTe semiconductor detectors. Med. Phys. 27 (9) http://dx.doi.org/10. 1118/1.1288243. Matlab and curve fitting toolbox – Mathworks 〈http://www.mathworks.com〉. Menéndez, P.R., et al., 2009. BNCT for skin melanoma in extremities: updated Argentine clinical results. Appl. Radiat. Isot. 67 (July (7–8 Suppl)), S50–S53 http:// dx.doi.org/10.1016/j.apradiso.2009.03.020. Minsky, Daniel M., et al., 2009. Experimental feasibility studies on a SPECT

Please cite this article as: Winkler, A., et al., Prompt gamma and neutron detection in BNCT utilizing a CdTe detector. Appl. Radiat. Isotopes (2015), http://dx.doi.org/10.1016/j.apradiso.2015.07.040i

6

A. Winkler et al. / Applied Radiation and Isotopes ∎ (∎∎∎∎) ∎∎∎–∎∎∎

tomograph for BNCT dosimetry. Appl. Radiat. Isot. 67, 179–182. http://dx.doi. org/10.1016/j.apradiso.2009.03.044. Minsky, Daniel M., et al., 2011. First tomographic image of neutron capture rate in a BNCT facility. Appl. Radiat. Isot. 69, 1858–1861. http://dx.doi.org/10.1016/j. apradiso.2011.01.030. Morozov, Boris, et al., 2006. The 3D tomographic image reconstruction software for prompt-gamma measurement of the Boron Neutron Capture Therapy. In: Nakagawa, Y., Kobayashi, T., Fukuda, H. (Eds.), Advances in Neutron Capture Therapy 2006. Proceedings of ICNCT-12. 12th International Congress on Neutron Capture Therapy, Kagawa, Japan, pp. 409–412. Munck af Rosenschöld, P.M., et al., 2001. Toward clinical application of prompt gamma spectroscopy for in vivo monitoring of boron uptake in boron neutron capture therapy. Med. Phys. 28 (5) http://dx.doi.org/10.1118/1.1367281. Murata, Isao, et al., 2014. Characterization measurement of a thick CdTe detector for BNCT-SPECT Detection efficiency and energy resolution. Appl. Radiat. Isot. 88,

129–133 http://dx.doi.org/10.1016/j.apradiso.2014.01.023. B – Thermal Neutron Capture Prompt Gamma Rays, IAEA – International Atomic Energy Agency 〈http://www-pub.iaea.org/MTCD/publications/PDF/Pub1263_ web.pdf〉. 113 Cd – Thermal Neutron Capture Gammas National Nuclear Data Center; Brookhaven National Laboratory 〈http://www.nndc.bnl.gov/capgam/byn/page128. html〉. Verbakel, W.F.A.R., Stecher-Rasmussen, F., 1997. A gamma-ray telescope for on-line measurements of low boron concentrations in a head phantom for BNCT. Nucl. Instrum. Methods Phys. Res. A 394 (1–2), 162–172. http://dx.doi.org/10.1016/ S0168-9002(97)00667-0. Verbakel, W.F.A.R., et al., 2003. Concentrations in brain during boron neutron capture therapy: in vivo measurements from the phase I Trial Eortc 11961 using a gamma-ray telescope. Int. J. Radiat. Oncol. Biol. Phys. 55 (3), 743–756. http: //dx.doi.org/10.1016/S0360-3016(02)04392-4. 10

Please cite this article as: Winkler, A., et al., Prompt gamma and neutron detection in BNCT utilizing a CdTe detector. Appl. Radiat. Isotopes (2015), http://dx.doi.org/10.1016/j.apradiso.2015.07.040i